Skip to main content

Alert for All

Final Report Summary - A4A (Alert for All)

Executive Summary:
Alert4All (A4A) focuses on improving the effectiveness of alert and communication to the population during crises within the context of the European Union. To achieve this goal, A4A developed an extensive and interdisciplinary alerting framework that integrates the key enablers to achieve significant improvements in terms of alerts penetration, cost-benefit ratio and intended vs. actual impact of alerting strategies. To this end, the project investigates five major research areas: authority operations, human behaviour in crises, the emerging role of new media, information management and communications technologies. Concretely, the Alert4All project conceived, designed, developed and demonstrated a proof-of-concept of an integrated, modular and scalable multi-hazard public alert system in close cooperation with end users. Alert4All encloses five integrated components to manage effectively the dissemination of alert messages to the population at risk during crises:
• The Information Management Portal (IMP) implements user management and map-centred incident management. The IMP fosters cooperation between authorities by enabling easy and efficient (but controlled) exchange of relevant information to public alert related to an incident, the Alerting Common Operational Picture (A-COP). Furthermore, the IMP acts as the A4A unique interface to the user, providing access to the rest of functionalities in an integrated manner.
• The Global Alerting Gateway (GAG) is an alert message dispatcher that connects with several communications systems to disseminate alerts through a variety of consumer devices, including terrestrial and satellite TV broadcast, mobile satellite services, mobile sirens activated by satellite and navigation devices. The corresponding receiver applications have been implemented as well. Furthermore, the GAG implements a complete communication framework developed to increase the effectiveness of public alert. This consist of (i) the guided use of best practices when formulating alert messages; (ii) the application of alerting libraries to ease formulation and avoid jargon, errors and ambiguities; (iii) the implementation of a communication protocol at transport level that on the one hand provides support for conveying best practices and on the other hand allows cost-effective transmission of alert messages by reducing significantly the required capacity in the transmission channel. By combining these components, Alert4All enables harmonised procedures to implement alerting in any European country and at the same time it enables customised decoding at the receiver side, allowing multi-language and multi-modal delivery of alerts, so that people with special needs are concretely addressed.
• The Screening of New Media (SNM) tool provides real-time screening of posts in new media and post processing of this data to elaborate statistics about the perception of the population about a crisis. Many recent research articles have confirmed that citizens are increasingly using various social media services during emergency situations to communicate and express opinions. Hence, the SNM offers a new dimension at the stakeholder side for the support of decision processes and enables adapting strategies in still running crisis situations.
• The Alert Simulation Module (ASM) is intended to analyse the reaction of citizens to specific alerting plans and in specific alerting scenarios based on agent-based simulations. The ASM shall be mainly used as a decisional support tool for developing efficient alerting strategies towards the population and better preparing decisions on the deployment of channels to alert the population.
• The Training Module (TM) comprises training materials such as tutorials ad quizzes with self-assessment, videos and access to the full Alert4All system in “Training Mode”. The “Training Mode” can be accessed by registered users to learn and train the system at any time in its full functionality (except the actual dissemination of alert messages to the population).
Furthermore, the relevant institutional framework for an A4A-based pan-European public alert system, as well as the cost-benefit analysis and service roll-out have been investigated.
As a result of the intensive dissemination work, the project team established several cooperation frameworks. Two of them enabled extending the project scope with further alert channels (building security systems) and by integrating with an operational public alert system in Germany. Finally, the project team is active in relevant standardisation bodies to contribute with project outcomes.
Project Context and Objectives:
Public alert systems have the potential to reduce injuries and losses to life and property from existing hazards. This potential is maximised if (i) the alert messages are issued with awareness of the potential consequences and (ii) it can be ensured that the citizens at risk actually receive, notice and understand the alert message.
This is indeed a challenge when Europe as a whole is addressed. On the one hand, the migration movements within Europe are significant, creating a challenging socio-demographic landscape when it comes to communication. On the other hand, each Member State follows a completely independent (and often different) strategy to alert the population, uses different technologies or approaches and follows different governance structures, often lacking of cooperation frameworks in disaster relief across countries. On top of all this, hazards do not understand about borders.

In this context, the Alert4All project has two main objectives:
• In the near term, it aims at improving effectiveness of alerting systems in the European Union by enabling cooperative management of diverse communications means.
• In the long term (10 years), it aims at laying the foundations for a pan-European alerting system.

The Alert4All vision on a future European alert and communication concept overcomes the challenges listed above:
• It allows cooperation between authorities beyond border limits
• It harmonises the strategies to formulate alert and communication messages in European countries
• It is aware of the impact of alert and communication messages in the population and applies this awareness to improve strategies and communication
• It is inclusive, addressing people with special needs
• It can use any communication technology to reach timely the maximum number of citizens at risk

Keeping the Alert4All vision and top level objectives in mind, the project addresses an extensive and interdisciplinary alerting framework that integrates the key enablers to achieve significant improvements in terms of alerts’ penetration, cost-benefit ratio and intended vs. actual impact of alert strategies. To this end, the project investigated five major research areas: authorities operations, human behaviour in crises, the emerging role of new media, information management and applicable communications technologies.

To accomplish the top-level goals and implement the Alert4All vision, the following scientific and technological objectives were targeted:
• To prepare the foundations for a pan-European alerting operating concept providing effective and coordinated crisis management to the population by (i) proposing optimised operational concepts to improve effectiveness of procedures for alerting the population in crises, which exploits the knowledge, tools, information and communication technologies provided by Alert4All; (ii) preparing training materials for authorities and responders to adapt to the new operational concepts that encompass the knowledge, tools, information and communication technologies provided by Alert4All;
• To investigate the impacts of alerting operational plans in the population by providing alert penetration and behaviour models;
• To develop a ’what if?’ tool based on the models mentioned above to support decision making processes in crises management, in particular for alerting and communication to the population;
• To investigate the role of mass and new media during / after a crisis situation and its impact on the effectiveness of institutional alert and communication towards citizens;
• To design tools for screening new media, allowing authorities to understand the perception of citizens about the crisis and adapt their procedures consequently;
• To propose solutions to use new media as an additional channel to disseminate alerts.
• To design a portal for secure and controlled information sharing among authorities and first responders across different levels (regional, national, international) in order to (i) enable cooperation between authorities in operational procedures and responses and (ii) allow authorities and responders to have a common picture of the crises.
• To design an alert message dispatcher, the Global Alerting Gateway, that allows authorities and first responders to disseminate alert messages over a variety of commercial communications systems and media by two major design components. First, designing a modular architecture for integration of several commercial communications systems and already well-established dedicated alert systems; second, designing a universal alerting protocol that provides two major features. First it allows reliable and secure transport of alerting messages over different communications channels (e.g. terrestrial and satellite broadcast, mobile networks, satellite navigation signals) to personal devices. Second, it minimises the required transmission capacity (as the CAP protocol consumes resources that will not be available for alerting purposes in commercial systems).
• To analyse business models and service sustainability for the Alert4All concept.
• To show the Alert4All benefits in critical scenarios in terms of operational feasibility, resilience, reliability and effectiveness of successful alert management.

The result is an integrated, modular and scalable multi-hazard public alert system that encloses the key enablers to manage the dissemination of alert messages to the population at risk during crises, with the following top-level features:
• Endorsing cooperation and information exchange among involved authorities by means of common interfaces.
• Providing support for decision making by means of alert impact awareness tools: (i) an alert impact ‘what if’ simulation tool that can be applied for planning and preparedness purposes and (ii) real-time screening of new media posts, providing knowledge on the citizens’ perception about the crisis situation.
• Fostering cost-effective alert message dissemination over a variety of communications systems, including satellite links, by means of novel communication protocols that promote the use of harmonised best practices.

Project Results:
Alert systems have the potential to reduce injuries and losses of life and property from existing hazards. This potential is maximised if (i) the alert messages are issued with awareness on the potential consequences and (ii) it can be ensured that the citizens at risk actually receive, notice and understand the alert message.

This is indeed a challenge when Europe as a whole is addressed. On the one hand, the migration movements within Europe are significant, creating a challenging socio-demographic landscape when it comes to communication. On the other hand, each Member State follows a completely independent (and often different) strategy to alert the population, uses different technologies or approaches and follows different governance structures, often lacking of cooperation frameworks in disaster relief across countries. On top of all this, hazards do not understand about borders.

The Alert4All vision on a future European alert and communication concept overcomes these challenges.
• It allows cooperation between authorities beyond border limits
• It harmonises the strategies to formulate alert and communication messages in European countries
• It is aware of the impact of alert and communication messages in the population and applies this awareness to improve strategies and communication
• It is inclusive, addressing people with special needs
• It can use any communication technology to reach timely the maximum number of citizens at risk

With this vision in mind, the Alert4All project aims at improving the effectiveness of alert and communication to the population in crisis with focus on a pan-European perspective. This goal has been addressed in the project by means of an extensive and interdisciplinary alerting framework that integrates the key enablers to achieve significant improvements in terms of alert penetration, cost-benefit ratio and intended vs. actual impact of alerting strategies. To this end, the project investigated five major research areas: authorities operations, human behaviour in crises, the emerging role of new media, information management and applicable communications technologies.

The Alert4All project conceived, designed and developed a proof-of-concept of an integrated, modular and scalable multi-hazard public alert system. Alert4All integrates the key enablers to manage the dissemination of alert messages to the population at risk during crises, enclosing the following features:
• Endorsing cooperation and information exchange among involved authorities by means of common interfaces.
• Providing support for decision making by means of alert impact awareness tools: (i) an alert impact ‘what if’ simulation tool that can be applied for planning and preparedness purposes and (ii) real-time screening of new media posts, providing knowledge of the citizens’ perception about the crisis situation.
• Fostering cost-effective alert message dissemination over a variety of communications systems, including satellite links, by means of novel communication protocols that promote the use of harmonised best practices.

To this end, the project started with the definition of a target scenario and a state-of-the-art analysis to identify gaps and room for improvement in the area of public alert at European level, scaling down to national and regional level. Based on this pre-study, user and system requirements were identified to derive the Alert4All system architecture as well as to design and implement its components. In the process, periodic consultation with the Advisory Board of End Users (AB) was carried out to tailor Alert4All to end user needs. The project concluded with the integration and demonstration of the overall concept in a public event in the presence of relevant stakeholders from the politics, industry, end users and research communities.

4.1.1 State-of-the-art analysis Scenario and Requirements

One of the key benefits in early stages of the project was the identification of the difference in interpretation of the word ‘scenario’ by technologists and practitioners; to the former, a scenario is an instance within a ‘use-case’, which itself defines a goal-oriented set of actions and interactions between the users of the system and the system itself. To the practitioner, a ‘scenario’ is a realistic series of actual or hypothetical (but, nonetheless feasible) events in the real world in which they operate. The user community will contend that in any project which aims to deliver a proof-of-concept solution, the starting point should be a realistic scenario in the operational sense: from this, key pinch-points can be identified and these can be used to produce use-cases, from which user (and later, system) requirements can be derived. Additionally, it was felt that using an operationally realistic scenario, as opposed to a multitude of small-scale, discrete scenarios would, in addition to eliminating duplication of effort, provide a higher level of credibility to the project and its development to those who may become future users of the Alert4All system.

In addition to the development of the pre-Alert4All scenario, another valuable tool has been the development and employment of the ‘post-Alert4All’ scenario, which has greatly assisted the consortium members in achieving a common understanding of how the world might change with the adoption of the A4A system. Also of interest was an analysis of the currently available methods used to communicate with the population in times of crisis, using real incidents dating as far back in time as 1931. Current practices

An examination of the Command and Control (C&C), incident management and organisational governance practices in Europe reveals that, despite the fact that terminologies and structures may differ significantly, there is nonetheless a general acceptance and recognition of the three tiers of Command and Control: strategic, tactical and operational. No tabulated presentation however can illustrate accurately the complexities that lie behind major incident operations: the research was able to demonstrate just how diverse the various governmental and organisational structures can be. Strategy should be decided at the top – but at the top of what? Most incidents will have a political presence, influence or perhaps command. To what extent political presence influences the command and control structure varies by country.

Four case studies were undertaken in selected regions of four different European countries to understand the use of technologies nowadays at operational level for public alert purposes. These case studies helped defining the approach to target an effective public alert system that should be usable in any European country. The main conclusion was that any system which is developed cannot be designed for deployment and use by specific authorities using a defined hierarchy with discrete responsibilities. It must be designed with the maximum of flexibility in the system, allowing authorities and agencies to take the system and adapt it to their particular requirements. In practice, this is not as complex as it first appears: the adaption process is not really a technical one relating to the system; rather, it is a managerial one relating to the C&C processes and lines of communication relevant to a particular area. This is relatively easily achieved but will require those authorities to be prepared to think laterally and adapt their requirements to what the system is capable of delivering, rather than to simply use the system as an addition to what they already do and continue to operate as before.

Furthermore, Alert4All performed a critical analysis of the state-of-the-art communication technologies used for public alerting purposes. The focus in this analysis is not anymore the operational dimension, but the technology capabilities compared against relevant attributes for public alert. This analysis has addressed four main areas: (i) a review of the communication technologies proposed by the research community; (ii) a collection of the current or emerging communication technologies used for alerting systems; (iii) accessibility for people with special needs; (iv) a survey of existing and proposed initiatives for alerting the population in European countries. As an outcome of this analysis, the synergies between the different communication channels have been identified and an overview of the currently deployed systems in Europe has been obtained. Additionally, a model for evaluating the effectiveness of alerting systems has been developed. According to the previous analysis, the model has been used to evaluate the effectiveness of a significant number of public alert systems which are used in several European countries with the purpose of identifying gaps. The model evaluates the capability of the different communication systems used for alerting purposes to fulfil a series of requirements. The results of this analysis show that none of the communications technologies fulfil satisfactorily all requirements. This translates into the conclusion that complementarities between different technologies for the purpose of alerting need to be exploited to achieve truly effective public alerting. Furthermore, the use of mobile satellite-based technologies becomes important to ensure resilience.

The analysis of the alerting technologies was completed with the state-of-the-art in information management. This took into account reception, merging and association of information, together with management of geographical information and the selection and scheduling of alerting systems. Best practices collection

A collection of best practices in warning the population has been carried out in the project by reviewing relevant scientific literature. This compilation was carried out with three main purposes: (i) to serve as basis for warning officials in general and the Alert4All users in particular, (ii) to compare current to best practices and identify misalignments and (iii) to embed best practices in the design and implementation of the Alert4All proof-of-concept..

The collection of best practices has allowed designing the system in an efficient manner, applying the identified critical recommendations into its design and implementation. The result has been an alert system that guides the user to apply best practices when formulating alert messages. Additionally, this collection of best practices has helped addressing satisfactorily autonomy aspects (as ethical issue), so that users are aware about how to best use Alert4All to prevent from false alert.

Furthermore, the best practices collection has been included in the developed A4A training platform that shall be used as training material by the end-users of the system. In the analysis of the comparison of the current practices in Europe vs. best practices it was found that despite most of the countries are somewhat aligned with the identified best practices, several major gaps exist that need to be addressed. Therefore it is of major importance that the end-users of the system have access and make use of this collection to enhance the warning effectiveness in their countries. This collection of best practices has been included in the “Encyclopedia of Natural Hazards” (Taylor & Francis Group) that will be published in January 2015. Human behaviour

A knowledge base of human behaviour in alarming situations was developed in Alert4All, including those factors that might have an impact on the behaviour in the crisis situations, focusing in the particular situation of Europe. Two different methodical approaches were combined. On the one hand, a literature review was carried through with regard to human behaviour in times of crisis and when being alerted or warned before/during a crisis. On the other hand, a series of interviews with practitioners from the field of operational crisis management was conducted; opinions, estimations and experiences from practitioners were gathered regarding human behaviour in crisis and after warnings. In both tracks the methodical approach of collecting and describing impacting factors based on heuristics was used to make them comparable. In the final step the described factors were then grouped and classified, bringing out a consolidated picture of human behaviour during crisis warnings and the impacting factors thereon.

Furthermore assumptions and statements regarding the concrete values of those factors’ impact were added. The obtained result covers all aspects of human behaviour related to alerts and is a compromise between representativeness within the European Union and the feasibility within the project boundaries. However, substantial aspects have been identified that add significant complexity such as the close inter-dependencies between the named factors as well as the complex dependencies of preconditions like regional circumstances, crisis type and severity of the situation. This research served as a basis for the definition of a model for alert message dissemination and human reaction developed in the project. Requirements

Consultation with the AB together with the pre-study comprising the Alert4All scenario definition, the analysis of use cases derived from the scenario, the state-of-the-art research in current practices and technologies as well as best practices and human behaviour has been applied to derive user and system requirements. These requirements were treated as ‘active’ statements throughout the project and were subject to review and revision, in close consultation with the AB in the context of four workshops.

4.1.2 The Alert4All Architecture
Based on the user and system requirements, the A4A team defined the overall A4A architecture, following a modular approach. Modules that address each of the research areas proposed in the project were identified and described in terms of their scope and functionalities, together with the required interfaces to provide an integrated system from the user (alerting system operator) perspective.

The adopted modular approach has several advantages. On the one hand, it allows addressing the different research areas that the A4A project is intended to fulfil (authorities’ operations, human behaviour, role of new media, information management and communication technologies) in an independent manner, but still integrating them and establishing inter-dependencies (where applicable) through the defined interfaces. On the other hand, this modular approach enables several degrees of freedom in the A4A design, implementation and deployment, as well as flexibility and scalability. This flexibility will allow a tailored deployment of the system to the organisational, governance and administrative needs of the corresponding authority.

The term scalability is twofold. One side of the coin is the capability to deploy the A4A modules incrementally, so that modules can be added when the needs require and capacities allow them. The other side of the coin is the capability to deploy A4A from local through regional to national scope and allow cross-border cooperation. In this regard, several operational architectures were discussed, concluding with a partially distributed operational architecture. This operational architecture is based on the principle that an A4A instance is deployed at a specific administrative level, letting all lower administrative levels access this instance to alert within their jurisdictional reach. A4A instances at the same administrative level are interconnected to efficiently exchange relevant information in a controlled manner and therefore enable cooperation; the information to be shared or exchanged by practitioners through this interface has been called Alerting Common Operational Picture (A-COP) in the project and it refers to all context information related to alerting the population within a specific incident, e.g. sent alert messages and further data collected by the system that shall be used for decision support purposes. With this architecture, authorities have full governance of the A4A system, but are able to share relevant information related to a crisis in an effective manner, enabling (even cross-border) cooperation.

4.1.3 The Alert4All Components

Five modules were identified to constitute the Alert4All system. Cooperation is enabled by the Information Management Portal (IMP), acting as the Alert4All interface to the user, enabling the A-COP exchange and user as well as incident management. Features for decision support are provided through two different modules: the Screening of New Media (SNM) tool and the Alert Simulation Module (ASM). On the one hand, the ASM provides alert impact ‘what-if’ simulations that can be applied mostly for planning and preparedness purposes; on the other hand, the SNM provides real-time screening of posts in new media and post processing of this data to elaborate statistics about the perception of the population about a crisis. Cost effective alert message dissemination promoting the use of harmonised best practices is provided by the Global Alerting Gateway (the GAG), connecting to a variety of communications technologies, and receiver software applications to deliver the alert message. Finally, the Training Module (TM) provides training facilities for practitioners to make proper use of the A4A system. The scope and design of each A4A component (IMP, ASM, SNM, GAG and TM) are described below. The Information Management Portal (IMP)

The IMP is the A4A module that allows user and incident management. At the same time, the IMP provides access through a user friendly graphical user interface (GUI) to the core capabilities offered by the rest of A4A modules (TM, SNM, ASM and GAG).

Based on recommendations by the AB, the IMP GUI has been developed as a map-centred geographical information support system with features for quick identification of the incident area as well as the geographical areas to be alerted. Through the IMP, the civil protection user can manage incidents (create, link, check history, etc.), alerting plans and exchange (import/export) the Alerting COP (A-COP) with other civil protection agencies that shall be involved in a specific crisis situation, enabling this way efficient cross-border cooperation. In particular, the IMP is capable of easily collecting all context information generated in the A4A system related to a specific incident and creating a data structure, the A-COP, that can be exported to other A4A instances (at other civil protection agencies). By importing the A-COP, the receiving A4A instance incorporates the incident originally created by a different A4A instance, including all information (history) into its map-oriented GUI. Through the support of Alerting Plans, the civil protection practitioners can pre-configure the A4A system to cope with expected incidents (e.g. in specific disaster-prone areas such as around chemical industry, close to volcanos, tsunami-risk areas, etc.). This feature is a preparedness asset that serves to save time at hazard onset, as basic information is already contained in the system, as well as guidelines to implement public alert in such circumstances.

Making use of open standards to ensure interoperability with other disaster management systems, the IMP has been designed and implemented to support TSO (Tactical Situation Object), a data format used to support different actors with different emergency management systems. In terms of governance and administrative aspects, the IMP provides support for different political, geographical and jurisdictional issues, while cooperation (even cross-border) is enabled by hierarchical access complemented by the A-COP exchange between A4A instances.

Finally, the IMP implements the GUI to access the rest of A4A modules. For this purpose, a SOA (Service Oriented Architecture) approach has been applied using web services to integrate all the subsystems within the IMP. This approach allowed independent implementations and databases for the different modules behind the IMP and the required flexibility and modularity to enable incremental deployment and being “future-proof”. This distributed implementation is however transparent to the user, who accesses one unique integrated system through the GUI. The Screening of New Media (SNM) tool

An important objective for command and control personnel during crises is to find out how people perceive the crisis and how they react to the communicated alert messages. The Alert4All Screening of New Media (SNM) tool allows for better situational awareness of people’s opinions and alertness through collecting and harvesting large amounts of user generated content from the web, and applying tailor-made affect analysis algorithms on the acquired user generated content.

An assumption made in the Alert4All project is that a representative selection of the population will be using social media such as Twitter and blogs to communicate their personal experiences during a crisis. This is an assumption that seems to be reasonable, as many recent research articles have confirmed that citizens are increasingly using various social media services during emergency situations in order to communicate, search for information, ask for help, etc. Examples of crises where an extensive use of social media for such purposes have been reported are, e.g. the Fukushima disaster, the terrorist attacks in Norway, and the Sandy hurricane. The SNM tool is capable of collecting social media posts in real-time from a large number of sources including Twitter, Facebook, blogs, Flickr, YouTube, and comments on web sites (e.g. news sites).

The SNM system architecture comprises two main subsystems: the data acquisition module and the data analysis module, each fulfilling their specific tasks in a quite independent manner and communicating with each other by means of a well-defined data communication channel, implemented on top of a database. Although both modules implement quite independent processes, there is also an obvious dependency between the two in that the data analysis process works on data delivered by the data acquisition process, following a relationship of producer-consumer between them: the data acquisition process generates pieces of content that are consumed by the data analysis process right after they are generated.

During the lifetime of an incident, alert messages are sent to the population in crisis through the A4A system in which, e.g. protective actions are recommended. Both for the incident itself and for each sent alert message, the data acquisition module then launches the corresponding search processes in order to obtain related user generated posts. As social media posts are continuously acquired and added to the database by the data acquisition process, the data analysis module applies tailor-made affect analysis algorithms and classifies the posts as belonging to one of the different considered emotions, such as fear, anger, or happiness. The used algorithms are based on natural language processing and machine learning, as well as on various lexical resources. In this way, automatic processing of a large number of social media posts within a limited amount of time is carried out.

Once the data has been acquired and classified, the user of the system can analyse the crisis by interacting with visualisations showing how the number of posts containing various emotions varies over time. By using different filtering selections it is also possible to adjust the analysis so that only posts from a certain time period, source, or containing certain keywords are included in the shown information. This gives the user a good idea of how the citizens in general are reacting to the crisis or to particular alert messages, but it is also possible to drill down into the underlying data set to see what people are afraid of, what they are upset about, etc., on a more detailed level.

When the user has become aware of how people have reacted to a communicated alert message or the crisis itself, a decision on whether to eventually disseminate a new clarifying alert message can be taken. If such a message is sent out, the user can again monitor the social media to get an idea of how the message was received, and so forth until the crisis is over or the SNM process has been cancelled.

Taking input from crisis responders into account has been important for the tool design to make sure that the included functionalities are both useful and usable in a real crisis situation. As part of this iterative design process, a series of advisory board workshops has been undertaken in order to understand and support the relevant working procedures. Much effort has also been devoted to privacy considerations. The Alert Simulation Module (ASM)

The Alert Simulation Module (ASM) is intended to analyse the reaction of citizens to specific alerting plans and in specific alerting scenarios based on simulations. The ASM shall be used as a decisional support tool for developing efficient alerting strategies towards the population. This functionality fills a gap, since no existing tools are able to provide information regarding the dissemination of warning and alerting messages within the population, including local and situational characteristics. Currently, simulations are used in various cases as decision support tools within crisis management. Nevertheless, those approaches were designed to support very specific use-cases such as the dispersion of toxic smoke clouds or crowd evacuation. The ASM aims at simulating the impact of disseminating warning messages to the population. It was implemented with an agent based simulation model that contains the following main components (following the definitions of Macal and North of an agent based simulation model):
• A set of agents, including their attributes and behaviour;
• A set of agent relationships and methods of interaction, including an underlying topology of connectedness, which defines how and with whom agents interact;
• The agent’s environment, with which the agents can interact, in addition to the other agents.

In this context a model regarding the dissemination of messages and human reactions was designed, referring to the general model “Diffusion of Emergency Warning: Comparing Empirical and Simulation Results” by Rogers and Sorensen. The model calculates diffusion of emergency warning as a logistic function, where alternative warning systems are characterised in terms of the parameters of the model. The developed model combines both individual processing of warning and alert messages and dissemination through a network of individuals. Therefore different theories and concepts of communication theory, human behaviour, warning message dissemination, crisis perception and information diffusion were analysed and own research regarding human behaviour in crisis situations was conducted. The complex model includes specifications regarding the agents, their attributes, the semantic relationships and the specific environment.

The ASM system architecture corresponds to a modular system of several very distinctive subsystems, each fulfilling their very specific task:
• ASM Information Processing (generally also called ASM) and ASM Web Services, as main switch/communicator between all other elements;
• CAST Simulation Business Logic, which operates the simulation model/logic;
• ASM Database, which includes previously saved runs as well as geographical base data;
• Additionally, the IMP from the Alert4All project is used for all end-user interactions.

The design of the tool was based on the user-centred design approach, starting with a requirements’ analysis and a use case definition with the help of crisis management experts. During the following development process different further evaluation steps with experts helped to design a tool matching the needs of crisis management, including early stage and paper based usability tests as well as user-guided walkthroughs using a full working prototype of the tool.

Within the development process of the ASM two main scientific developments have been achieved:
• A knowledge base on Human reactions to alerts was developed, using results from conducted expert interviews and literature research. The gained information was used to design the simulation model and its components. Finally, the impact factors used by the tool were again evaluated by crisis management experts.
• The development process of the tool shows the feasibility of the user-centred design approach in the context of a crisis management software solution. The gained knowledge in form of lessons learned will be helpful for comparable future developments of interfaces for crisis management software tools. The Communications System

The understanding of the alert message and trust by the recipients can be influenced by the alert message content, delivery mode and style. Furthermore, the variety of cultures and languages combined with significant migration and tourism movements make it even more complex to effectively inform the (whole of the) population at risk for a given incident in a European context. The content of alert messages is managed in different ways across European countries: it can be decided on demand, in some cases templates are used or pre-defined messages are applied for cases in which they exist and are applicable. Furthermore, pre-defined messages for a specific area are not necessarily applicable to other areas, and templates used by different authorities do not necessarily match.

According to the collection of best practices carried out in the project, recipients are more likely to trust the alert and implement the requested protective actions when the alert message includes sufficient information about the hazard, i.e. type, location, time, intensity and certainty, as well as guidance on protective actions and issuing source. Specific guidelines related to the alert message style are also identified, e.g. most important information should be provided first, so that the use of headlines is recommended. It is also important to avoid the use of ambiguous and complex words, expert jargon or complex sentences and misspellings in the alert message that could cause a complete different meaning in the message than the desired one. Best practice-compliant and harmonised procedures to formulate alert messages all over Europe (and even beyond) could significantly help addressing these issues, improving the effectiveness of the received alert messages, simplifying cross-border alerting at the recipient dimension.

With this challenge in mind, the Alert4All project developed a complete communications paradigm to formulate and disseminate alert messages that encloses three major components:
• Guided use of best practices when formulating alert messages;
• Application of alert libraries to ease formulation and avoid jargon, errors and ambiguities;
• A communication protocol at transport level that on the one hand provides support for conveying best practices and on the other hand allows cost-effective transmission of alert messages by reducing significantly the required capacity in the transmission channel.

By combining these components, Alert4All enables harmonised procedures to implement alerting in any European country and at the same time it enables customised decoding at the receiver side, allowing multi-language and multi-modal delivery of alerts, so that people with special needs are concretely addressed.

To this end, and based on the identified best practices, an abstraction was made to identify the information components required (and their order) to formulate alert messages with the maximum desired impact. The alert message must contain information about the ‘hazard type’, the ‘location at risk’, the ‘time’, the expected ‘intensity’, ‘certainty’ of the given information, ‘protective action’ to be implemented by the affected citizens and ‘source’ of the message. While it is unrealistic to think of a pre-defined alert message for each possible situation, it is realistic to envisage a limited dictionary, what we have called “Alerting Libraries”, for each information item. Hence, by combining values contained in the “Alerting Libraries” for each information item that are relevant to a specific crisis situation, a very large spectrum of situations can be addressed. With this principle, the paradigm applied in Alert4All to formulate alerts was the creation of exemplary libraries (for a proof-of-concept) and a software module that formulates a “human-readable” sentence out of the selection of keywords from the library based on basic syntax rules of a specific language. Furthermore, the values contained in the libraries can be encoded. This allows transmitting only an encoded list of values (instead of the full text message) to the receiver. By hosting the syntax rules of the desired language at the receiver side and a “decoder” software (alert message processing at client-based application), the receiver can “re-construct” the alert message in the desired language out of the received code and even present the alert message in different modes (e.g. text and/or speech). Transmitting an encoded version of the alert message implies a significant reduction of the required transmission capacity.

To support the encoded transmission, a flexible protocol was designed, the A4A protocol, capable of transporting CAP (Common Alerting Protocol) compatible alert messages with very low capacity requirements by making use of the “Alerting Libraries” paradigm. The protocol is based on a core primary header complemented by a secondary and extension headers that are appended only if required (when additional features are required or free text not contained in the libraries needs to be added). The primary header contains the essential information to process the alert message. It contains the “alert” part of a CAP message plus some other fields that are moved here for the sake of the information organization. The secondary header contains the alert message text, with the free text body expressed in different languages as well as additional resources (multimedia, documents, or URI), if necessary. In case the alerting libraries are used and no free text is added to the message, it is not necessary to include the free-text part, since the alert message can be recovered from the encoded part, even in multiple languages, by properly decoding the message. A set of possible (optional) extensions is foreseen. It is derived from the optional fields of the CAP message, such as the “area” part, plus some additions due to the protocol requirements, such as retransmission or coding to increase reliability of the transmission or authentication and encryption features.

The complete paradigm has been implemented in the A4A communications system, which is composed of the following elements:
• The Global Alerting Gateway (GAG): an alert message dispatcher that connects to several communications technologies to disseminate alert messages;
• The Alerting Channel Access Points (ACAPs): points of access to the networks corresponding to the different communications technologies that shall further disseminate alert messages;
• The receiver applications: software support for the delivery and presentation of alert messages in consumer devices of different technologies.

Assuming that the user will access the GAG through the GUI hosted in the IMP, the complete A4A communications path encloses the IMP, the GAG, the ACAPs and the Receiver applications. In particular, the GAG interfaces the IMP and the ACAPs by means of web services.

The Global Alerting Gateway (GAG)
The GAG is the core component of the Alert4All communications system. It is an alert message dispatcher that practitioners can use to formulate alert messages and disseminate them easily through several communications means. Making use of the best practices in alerting and the “Alerting Libraries” concept, the GAG GUI has been conceived to guide the user systematically when creating alert messages. The user can easily select keywords that are relevant to the current incident to indicate the ‘hazard type’, ‘location at risk’, ‘time’, expected ‘intensity’, ‘certainty’ and ‘protective action’ to be implemented by the affected citizens (the message source is inserted by default); the user can also select (or pre-configure) the channels through which the alert message shall be disseminated.

The information inserted by the user through the selection of keywords and configuration options is processed and encoded in the GAG and forwarded to the ACAPs of the selected communication technologies to disseminate the alerts. Furthermore the GAG exchanges metadata and signalling information with the ACAP to gather information related to availability and status of alerts. For backward compatibility with CAP-compliant systems that do not yet support A4A, the GAG can provide also a plain XML CAP output towards the ACAPs.

Based on the “Alerting Libraries” concept, the communications system can implement multi-language automatically. However, it was also required by the AB to have the option to insert free text for cases in which the provided keywords of alert formulation method does not match the situation. The GAG provides the option to insert free text in the alert message; however, when the user adopts this option the GAG cannot implement the automatic multi-language feature for the full alert message. Hence, in case the user inserts free text, the GAG will warn the user that the multi-language feature will not be available for this message and will advise the user to insert the translation of the alert message into relevant languages by hand. This is also recognised at the receiver devices side.

The ACAPs and Receiver devices
Solutions to disseminate alert messages over several technologies were designed and implemented during the project with the aim to provide innovative and effective alert delivery means. In particular, the project team focused on solutions for different flavours of fixed satellite and terrestrial TV broadcast, mobile satellite systems and navigation systems as alert delivery means. The necessary features in the access points to such technologies as well as in the receivers (where applicable) were developed to allow delivery of alert messages in a multi-modal, multi-language and at the same harmonised way.

The solutions for TV broadcast systems were developed with two different approaches. The first solution makes use of a widely extended standard for smart TVs (HbbTV), delivering the alert in a web-like environment directly in the TV. The second solution intends to make full use of the A4A protocol features by an add-on to state-of-the-art digital TV receivers that are DVB-compliant. Mobile satellite systems in S-band (not available yet in European market) were demonstrated by integrating a prototype S-band receiver that acts as gateway with handheld and other devices, such as smartphones, tablets and mobile sirens. Additionally, the delivery of alert messages through navigation signal was emulated in the testbed and its impact was investigated through simulations.

In addition, the project team established two major cooperation frameworks with Euralarm and Mecom to extend the scope of the initially planned communications system in Alert4All. Mecom is a news agency that additionally provides the German Modular Warning System (MoWaS) that has been rolled out in Germany at national, regional and locality level during the project lifetime. Euralarm is an association representing the European electronic fire and security industry.

On the one hand, Alert4All was extended by integrating with the MoWaS system. This integration enabled that a MoWaS user is capable of disseminating alert messages through the novel alert channels provided by Alert4All. On the other hand, Alert4All was extended by making use of building security systems (such as mass notification systems, voice evacuation and advanced fire alarms) for public alert purposes, in cooperation with the Euralarm project PEARS (Public Emergency Alert & Response System in Buildings). This was an excellent example of cooperation between research and industry.

TV Broadcast: Alert delivery over HbbTV (terrestrial and satellite)
Hybrid Broadcast Broadband TV, or HbbTV, is a major new pan-European initiative aimed at harmonising the broadcast and broadband delivery of entertainment to the end consumer through connected TVs and set-top boxes. HbbTV combines broadcasting and internet services in order to create brand new services and usage experiences for TV, including catch-up TV, video on demand (VoD), interactive advertising, personalisation, voting, games and social networking, as well as programme-related services such as digital teletext, Electronic Programme Guides and second screen applications. In a nutshell, HbbTV applications are based on XHTML, and the receiver uses a browser to display the HbbTV page, where the user can find information and navigate to other webpages in the Internet (whenever the TV set is connected to internet). However the broadband connection is not mandatory and applications can make use of the broadcast part only. HbbTV has become a de-facto standard and almost all of the new COTS (commercial off-the-shelf) TVs and set-top boxes implement it, for both terrestrial and satellite broadcast (DVB-T/T2/S/S2) and all network operators (ADSL, cable, DVB-C, fibre and satellite).

For these reasons, HbbTV has been chosen for implementation of the alerting channel on TV broadcast. Leveraging on the interoperability of this standard, a single application and a unique ACAP have been implemented in Alert4All and then installed and operated on a DVB-T demonstration platform, as well as on a DVB-S operational platform (a demonstration channel inserted in the Fransat bouquet by Eutelsat).

When the GAG sends a request for sending an alert, the CAP-message generated by the GAG and that is part of such request will be searched for the specific XML-objects that are needed to create the alert-message in the HbbTV application. The plaintext of the XML-objects is transmitted within multiple Stream Events towards a multiplexer. The multiplexer will then include them, together with the static HbbTV-application, inside the DSM-CC (Digital Storage Media Command And Control) carousel. The carousel is part of the DVB transport stream (TS) that also includes the video and audio (V/A).

In the solution being developed by the A4A project, an A4A-HbbTV-Application (i.e. the HbbTV-XHMTL) is transmitted constantly within the DSM-CC carousel in the transport stream. This application is static and is automatically running when a service has been chosen. The application is invisible as long as there are no new or updated alerts. It is only if an alert is sent with a Stream Event that the application will be shown on the television screen. When using HbbTV as an alerting-system with terrestrial or satellite transmission it is important that the system works on both broadcast and broadband communication channels independently so that the risk of unavailability of the service in case of crisis is minimized. Therefore the implementation for the A4A HbbTV alerting channel has been structured on 2 levels: (i) all the essential information is delivered within the transport stream, i.e. the transport container for video, audio and additional data via Broadcast; (ii) complementary information, e.g. detailed maps, or accessory instructions which are not mandatory to handling the crisis can be retrieved via broadband when such a connection is available.

Enhanced DVB-T/T2 Receiver
DVB-T2 is the second generation of the digital terrestrial standard. It was standardised in 2009 and in the same year first regular services were on air. The standard comes up with many improvements to make the signal even more robust, deliver more data rate or a mixture of both, e.g. new and more interleavers. In the Alert4All project, an end-to-end solution to deliver alert messages in DVB technologies has been developed supporting the A4A protocol and multi-language, multi-modal delivery based on the Alerting Libraries principle described above. The developed solution is compatible with any DVB technology (DVB-T/T2, -S/S2, or –C), since it is applied at MPEG-TS layer, compliant to all DVB technologies.

At the transmission side, a software module has been developed that gathers alert messages delivered by the GAG to the broadcaster side and injects them into the DSM-CC data carrousel. At the receiver side, a software package has been integrated with a DVB-T/T2 USB receiver in a regular computer, upgrading the capabilities of a COTS DVB-T/T2 receiver to enable the delivery of A4A alert messages. This software package is composed of two modules. The first gathers A4A messages from the DSM-CC carrousel, parses and decodes them; the second formulates a “human readable” message out of the decoded A4A message, based on basic syntax rules, and delivers the alert message in a suitable language as overlay to the programme that is being watched. The message can be reproduced in text and/or speech and is repeated periodically until the TV owner closes the message. Through the user interface created in the project, the user can configure the preferred language and mode in which alert messages should be delivered, as well as the location (e.g. postal code), so that alert messages are only displayed if they are meant for the area in which the TV is located and filtered out otherwise. Hence, this solution makes full use of the A4A protocol benefits.

Alert delivery over satellite mobile services (DVB-SH) to mobile phones and sirens
The DVB-SH (Digital Video Broadcasting- Satellite to Handeld) is an ETSI standard, derived from the DVB-H standard, to broadcast video, audio and data to mobile devices. DVB-SH tries to exploit the S band frequencies, where there are opportunities for Mobile Satellite Services (MSS), following the attribution of harmonized licenses for S-band exploitation in Europe by the European Commission. This standard, created specifically for distributing content via satellite in mobility situation, makes a major innovation in the telecommunications world over satellite: it enables the addition of a network of terrestrial repeaters, called CGC (Complementary Ground Component) to complement the satellite coverage.

The choice of the satellite, and of the DVB-SH, as the transmission system component on the downlink component is easily justified by the desire on the one hand to avoid congestion or interruption of the terrestrial networks which can become harmful in case of a major crisis, and, on the other hand, to quickly connect a warning system or a single isolated but important subscriber. In terms of cost, due to its broadcasting capabilities, the DVB-SH network allows to transport the alert only once to reach the whole set of connected terminals reducing strongly the capacity price.

At the time of the Alert4All R&D activities, no DVB-SH handheld terminal with satellite reception was available in the market. However it is still relevant to demonstrate the benefits of such a solution, since it represents a robust and reliable alternative to GSM, 3G or 4G when the terrestrial networks are affected by the emergency situation. The implemented solution in the project consists on the use of the so-called “Green-Box”, a small DVB-SH device that acts as IP gateway between the satellite and the COTS smartphones and tablets, emulating a satellite-enabled handheld. Differently from the DVB solution for fixed systems (DVB-T/T2, -S/S2 and C), this solution needs to be IP-compatible so as to allow the interconnection with the handheld devices through their WLAN interface.

At the receiver side, a software module has been installed in smartphones and tablets (as an app) with the same functionality as the software module developed for the DVB-T/T2 receiver, i.e. parsing and decoding of the A4A message based on the Alerting Libraries, as well as compilation of a “human readable” alert message and delivery in text and/or speech in different languages. The major difference is the graphical user interface that has been implemented for the operative systems of different vendors (Android and iOS) and including further information such as a map to pinpoint the area at risk. Additionally, the application archives alert messages so that the user can always go back to previous messages to analyse or review the evolution of the disaster situation.

The same satellite technology used as gateway to handhelds has been used for remotely triggered sirens. The existing Satellite Alert System prototype of Eutelsat, developed in previous R&D projects, has been interfaced with the GAG as an additional example of alerting channel in the Alert4All system.

With this systems, alerts are broadcasted on a large area and received by all the terminals, but geo-fencing is used by each terminal to determine whether the terminal is concerned by the alert and therefore triggering the siren, or not.

Alert over Global Navigation Satellite Services (GNSS)
Global Navigation Satellite Services (GNSS) is a large container of all technologies providing positioning and location-based services by means of satellite technologies. In particular, the satellite-based augmented system (SBAS) is an interesting technology, as it complements functionalities already provided by regular positioning systems (e.g. GPS, GLONASS) by improving the position precision and extending the overall coverage. To this regard, the European version of this technology, the European Global Navigation Overlay System (EGNOS) has been taken as reference in terms of GNSS technology for the potentials of its employment in alerting applications. In more detail, the availability of such a technology (and in general of GNSS services) is quite relevant for alerting applications because of the widespread use of navigation devices in everyday life which concur to improve the penetration of alerting messages in case of crisis management situations.

The integration of GNSS in the A4A system was achieved by means of emulation systems, responsible for reproducing a real EGNOS device and a receiver displaying the received message so as to mimic the arrival of an alert message on navigation device.

The main achievement of this integration has been the study of the implications in multiplexing alerting messages with the EGNOS intrinsic messages, especially in terms of the traffic shaping and scheduling policies (priority assignment) to be configured in order not to penalise the EGNOS service and simultaneously let the alert message to be timely and correctly received. The Analysis has been carried out two-fold: 1) real alert messages where fed into the EGNOS system through the GAG and a dedicated ACAP, 2) assessment performance was also carried out through MATLAB ® simulations.

The overall lesson learnt during this investigation and integration into the A4A alerting system is that the harsh operative conditions of EGNOS mandate the use of a very efficient transport protocol for alert messages. Further to this, proper assignment of priority is instrumental to achieve good performance in terms of message delivery delay. Finally, it was also observed that in case of security requirements, whereby longer messages should be transmitted, the alerting message service is anyhow penalised even at the highest priority, thus requiring the EGNOS system to be provisioned with higher capacity to better accommodate this service. Training material

Training material and an educational platform that integrates all the materials have been developed to train the end-users of the system. This was done with the aim to enable the end-users of the system to get familiar with the system during non-crisis time, to allow them alerting the population in an effective manner.

The development of the training material has been based on a training methodology which was developed in the A4A project. In the definition of the methodology two main types of user profiles were identified, one with wide knowledge in the use of IT systems and the other with low knowledge or ability to operate such systems. To address this reality and to allow an effective training of both end-user types, the materials have been developed with a basic level and a more advanced level.

Training is a key requirement in the operation of emergency management IT systems. It can be even considered the most valuable component of such a system for first responders. Without a proper training the end-user will not be familiar with the system during a crisis situation, and as a consequence will not be able to effectively alert the population. In an increasingly operationally demanding and budget-restricted environment, it is strictly necessary that a balance between effectiveness of the training and time coexists. In order to answer to this requirement practical tutorials have been developed in text as well as in video to train the users. The last will allow just-in-time training for end-users. Furthermore, the educational platform allows the end-users to assess their learning progress in the completion of the tutorials by answering to self-assessment quizzes after finishing each tutorial. In addition the end-user has the possibility to continue the quizzes in case he or she needs to interrupt the training due to an emergency which is something that is likely to happen in a daily workday of a first responder.

4.1.4 Proof-of-concept and final demonstration

The Alert4All components described above were successfully integrated by means of RESTful web services into a large proof-of-concept testbed accessible by users through the graphical user interface hosted by the IMP. The integrated testbed and its most remarkable functionalities and capabilities were successfully presented in a live demonstration that took place in the premises of the German Aerospace Center (DLR) in Oberpfaffenhofen (near Munich) on October 15, 2013 in the presence of several stakeholders from research, industry, civil protection and politics. The demonstration day programme was divided in three major blocks:
• An operational demonstration based on a fictive scenario developed for this purpose;
• A guided tour into the system features, were attendees to the event could enjoy thorough explanations about the R&D work and further small scale demonstrations;
• A panel session in which the next steps to achieve a real pan-European public alert system were discussed with selected speakers and the audience.

During the operational demonstration, a fictive complex but realistic crisis scenario in an imaginary location was described along with the operational use of the Alert4All system in the context of this fictive scenario. During the operational demonstration, the use of the Alert4All system to formulate and deliver multi-language and multi-modal alert messages in a coherent manner through a variety of terrestrial and satellite communication channels was shown live. It was shown that it is technically feasible to deploy a public alert system that is user friendly for the practitioners and at the same time can reliably deliver alert messages in the language selected by the recipient and presented in a suitable mode for the recipient, addressing also people with special needs and regardless of their location.

The delivery of alert messages using several technologies was demonstrated. Major focus was given to a variety of solutions developed using terrestrial and satellite based broadcast technologies, including different TV solutions, smartphones and tablets connected over satellite, an emulated satellite navigation receiver and a siren activated over satellite. Furthermore, also systems that have been originally conceived for building security were integrated with Alert4All through the cooperation with the Euralarm project PEARS. In this context, it was demonstrated that mass notification and evacuation systems that are normally installed in office buildings, malls, etc., can be also successfully applied for public alert.

Moreover, the use in an operational context of the Alert4All screening of new media features to observe how the population understand the situation and to make decisions was extensively demonstrated. The impact that the crisis situation has on the population is inferred from the information that people generates on social media (Twitter, Facebook, Flickr, YouTube and collaborative web sites) and presented to the user in terms of sentiments, namely how people feel about the crisis and the communicated alert messages. Thus, during the operational demonstration, it was shown how the visualisation of the screening of new media results presented to the user during the management of a crisis can be used for estimating people’s reactions to specific communicated alert messages and how this knowledge can help for taking decisions such as whether to send an update message. During the final demonstration, the SNM tool was evaluated using a fake dataset, which had been constructed especially for the live demonstration in order to make it possible to use the tool to its full extent without having to worry about privacy concerns. In general, the SNM tool obtained much attention and positive reviews, and the overall concept was not criticised at all. In part, this is hopefully due to the tool being well-developed, but probably it is also due to the fact that it takes time to fully grasp a new concept and elaborate on it in relation to one’s own working procedures. At large, the final demonstration visitors meant that the idea of using social media input for informing crisis management decisions must be considered novel related to their experiences and current working procedures.

Finally, the simulation of alerting effects in the population was shown to the audience. It allows reflecting if alert messages and more detailed crisis information have been transmitted to the relevant people and not only to technical devices. For example, during a concert most persons will not watch their mobile devices, and in the night radio and TV messages are only recognised by very few persons. Moreover, the crisis managers can share their assumptions if messages may have been understood, considering multi-language barriers, illiteracy or handicaps. And in the end, the population has to believe the messages and follow instructions, which may differ for the specific kind of hazard (e.g. if you can recognise it yourself) and for the trust in governmental and media institutions that send and transmit the messages. This alerting effect simulation is considered by the involved crisis managers to be very useful both in preparedness planning and for investment decisions about additional channels.

4.1.5 The institutional framework and business aspects

First the Institutional Framework needed to bring the A4A Service into operation was defined and the concept of the European Warning and Informing Agency (EWIA) introduced. The EWIA entity would act as a federation of member states, with each member having a national A4A operator that could either provide a nation-wide service (centralised) or a regional/local service (de-centralised), depending on the level of the existing civil protection structure. The EWIA would co-ordinate and federate all these individual National, Regional or local level A4A entities. An analysis of existing alerting systems as identified in early stages of the project was undertaken and the characteristics that should be included in the structure of the proposed EWIA entity and any operational constraints were identified. The Alerting-Common Operational Picture developed and defined within the project would be used as the co-ordinating channel between these individual national/regional/local entities and the EWIA.

It was proposed that the EWIA would enter into a number of service level agreements with alerting channel providers to ensure the effective and efficient delivery of the warning and informing messages to citizens. It was concluded that the creation of this pan-European entity would achieve significant economies of scale by aggregating demand and providing common systems.

The activities of the EWIA would be overseen by an advisory European Warning and Informing Supervisory Authority (EWISA), which in turn would be guided by a Council formed by representatives of each participating nation. EWIA and EWISA would be Intergovernmental Organisations whilst the individual national/regional/local operating entities would be formed of limited liability companies subject to European and national commercial law, contractual interpretation and the arrangements of the nation in which each is sited.

The Cost-Benefit Analysis and Sustainability Modelling for the A4A Service operators, for both the centralised and de-centralised deployments, were then considered. First we developed a brand new methodology to evaluate the benefits of the system due both to the citizens and the emergency personnel in a crisis. We began by closely analysing five disaster types to assess the impact that the A4A System would have on a citizen’s alert status in each crisis type. These disasters were Wildfire, Heat Wave, Flood, Storm and Earthquake. For each we carried out a population survey to obtain grass roots data which, in conjunction with data analysis techniques, enabled us to estimate the penetration of the A4A system benefits through the population. Using this information, we evaluated the size of the benefits of damage mitigated and lives saved due to the system over a forecast period 2014-2023 and carried out a comprehensive sensitivity analysis on the model’s critical assumptions. We then considered the Secondary Benefits due to emergency responders, such as improvements in situational awareness and effectiveness of response. Combining the Primary and Secondary benefits we estimated that the damage that could be mitigated across Europe over the forecast period would be €1,276m€440m and the number of lives that could be saved at 263(+183,-56).

We then looked at the costs of the system and estimated the Capital Expenditure (CAPEX) and the Operational Expenditure (OPEX) for both the centralised and de-centralised deployments on a European Scale. CAPEX incorporated elements such as development of the proof of concept into a deployable product, the purchase of the hardware needed to run the system and its installation. Our budget for the OPEX included maintenance and on-going support, terrestrial and satellite bandwidth for message dissemination and premise facilities. We estimated that the costs for the European deployment of the A4A system, over the ten year forecast period, would be €150m and €246m for the centralised and decentralised models respectively.

We then carried out a Net Present Value analysis on the costs and the benefits, using a 10% discount rate, over the forecast period. The data showed that the A4A system would provide an excess of benefits over costs of some €633m for the centralised deployment and €537m for the decentralised deployment. The analysis unequivocally demonstrated that the A4A System would be sustainable over the period 2014-2023.

We concluded the report with a brief discussion on the effectiveness of the system at saving lives and concluded that in a centralised deployment it would cost approximately €500k to save one life, while in the de-centralised deployment it would costs approximately €850k. We suggested that this implies the A4A program would be highly cost-effective at saving lives.

4.1.6 The Advisory Board

The success of the project has in no small part been due to the involvement of the Advisory Board of end users. The project has three end-user organisations (BAPCO, BBK and DRK) within the consortium. It could be argued that they may be biased towards the project and its outcomes as a result of their being part of its development. A better solution therefore is to involve an independent AB, whose members’ interests can be represented by the end-user organisations, but who can be brought in at key points in the project development to define and produce key needs, such as scenarios and user requirements and can test and validate key project elements and solutions. The membership of the AB is made up of people active in the fields of emergency service (fire police, ambulance and medical), press and media in various areas of Europe.

In the Alert4All project, the AB members assisted in the development of the scenarios and user requirements; took part in workshops to assist the design and development of key system components and tested and validated the results. They also took part in the final demonstration, providing feedback on the content and noting the progress made since (and as a result of) their last meeting with the consortium.

The inclusion of the AB is considered an unqualified success by both its members and the consortium. Members are themselves interested in promoting the findings of the Alert4All project amongst their contacts and organisations, which adds significantly to the dissemination opportunities and will further the cause of public alerting across Europe beyond the scope of the actual project.

4.1.7 The Ethics Board

At the onset of the project, an ethics advisory board (EAB) was brought together for the purpose of identifying, analysing and monitoring critical ethical issues. The EAB work has been continuously documented in an ethics register, which contains a list of the identified critical ethical issues along with their analyses and recommended responses. The ethics register has been used throughout the project to discuss and highlight relevant issues, and for providing guidelines to the respective consortium partners regarding how to handle or avoid the identified issues.

The issues in the register include topics such as privacy, data security, social inclusion, dual use, autonomy, and human dignity and human rights. The content and the status of the ethics register have, however, varied as a result of the EAB discussions. Some of the issues could be closed at an early stage by showing to be out of scope with regard to the research project.

Further issues were closed during the course of the project by being adequately tackled by the concerned consortium partners. Three major examples can be mentioned here: social inclusion, human dignity and human rights and privacy. The issue on social inclusion was addressed by including special requirements for persons with special needs in the design of simulation tools and the alert communication paradigm developed in the project, allowing multi-modal and multi-language alert delivery. The issue on human dignity and human rights concerned the activities related to the cost-benefit analysis planned in the project. While there is a pragmatic need to calculate the cost of saving human lives, which includes the need to calculate the potential economic damage caused by the loss of human lives, it was required to be careful not to put a “price tag” on human life. The value of human life goes beyond economic commodification as it represents a deeper value without which potentially saved money becomes meaningless. To address this issue, the initially proposed single-objective analysis (minimising costs) was replaced by a multi-objective analysis (saving lives, minimising costs). Finally, the data privacy issue could be closed during the final demonstration where the screening of new media tool was demonstrated using solely fabricated data.

An important part of the Alert4All approach to ethics management, and also a prerequisite for the EAB work, was that it remains the responsibility of the consortium partners to raise possible ethical issues and to make the EAB aware of these issues. Therefore, apart from its internal working procedures, the EAB has been made an integrated part in the project by, e.g. being present in telephone conferences, participating in meetings, and holding presentations where issues and recommendations have been discussed.

Potential Impact:
Alert4All (A4A) has conceived, developed and implemented a proof-of-concept for a cooperative and integrative platform for public alert in crisis situations and has demonstrated that it can significantly improve the effectiveness of alert communications, if applied in any European country, but especially if it is applied in several (or all) European countries. In the communications domain, this is achieved by significantly increasing the penetration, resilience, reliability and impact of alert messages by using diverse communications means, including terrestrial and satellite-based solutions, harmonising best practices, addressing people with special needs by multi-modal delivery of alert messages and providing suitable solutions for multi-language alerting. In the area of operations, A4A contributes to significantly enhance the operational picture; it enables improvements in planning efficiency and the means to develop and test strategies to alert citizens in crisis situations through simulations of the alert impact; furthermore, alert strategies can be improved almost in real time by the capability of gathering the opinions of citizens about the current crisis situation and presenting those in the form of statistical sentiment trends.

Alert4All brought experts together from a number of disciplines combining operational civil protection experience, with scientific, technological, human science, and industrial knowledge. This has clearly re-enforced the European capabilities in the field of public alert with the associated benefits to the safety of European citizens.

The acceptance at the end user side of the concepts proposed in Alert4All and their impact in future alert systems is ensured by the active participation of end users in the project either as full partners (the German Civil Protection, BBK, the German Red Cross (DRK) and the British APCO) or as members of the Advisory Board (AB) of end users. Indeed, the Advisory Board has followed the project throughout its entire lifetime, providing user requirements and steering the design and development phases. The result has been a system completely tailored to the user needs, from both the operational and the user-friendliness perspectives. The participation of the Advisory Board has been of significant benefit for both its members and the consortium. Members are themselves interested in promoting the findings of the Alert4All project amongst their contacts and organisations. Furthermore, they continue supporting A4A contributions to standards of relevant scope.

In general terms, the potential societal benefit of Alert4All has been estimated by means of interviews to different population groups in different scenarios during the project. The results prove that the benefit of deploying the Alert4All system is measurable in terms of cost-benefit ratio, the number of affected citizens reached by alerts in a timely manner, trust of citizens on alerts and reduction of losses in terms of lives, injured people and goods. In particular, as mentioned in the previous section, the benefit analysis implemented during the project has provided estimations on the damage that could be mitigated by the deployment of Alert4All across Europe over the forecast period 2014-2023 in term of economic value at €1,276m(+/-€440)m and of number of lives that could be saved at 263(+183/-56). In terms of costs, the estimation including Capital Expenditure and Operational Expenditure depend on the deployment fashion, i.e. centralised or decentralised. The estimated costs over the forecast period are €150m and €246m for the first and second deployment option, respectively. By applying a Net Present Value analysis on the costs and the benefits, using a 10% discount rate, over the forecast period, the data showed that the A4A system would provide an excess of benefits over costs of some €633m for the centralised deployment and €537m for the decentralised deployment in Europe over the forecast period.

Improved planning and response through better situational awareness
Information share between authorities and first responders through the A4A Information Management Portal (IMP) and making use of the Alerting Common Operational Picture (A-COP) concept will allow the convergence of stakeholders’ picture of the situation and even cross-border cooperation. This contributes significantly to the efficiency of communications towards citizens as information flows are optimised; in many cases, rapid and coordinated communication towards the population is fundamental to minimise the impact of crises at all levels. A converging picture of the situation contributes to avoid the overuse of alert messages and the dissemination of multiple uncoordinated messages from different sources that might be contradictory and confuse the citizens. The avoidance of overuse and the coordination of alerts are of paramount importance to increase the trust of citizens in the alert and to ensure proper understanding and reaction of the citizen to the alert.

Furthermore, Alert4All goes far beyond the benefits of having a converging picture of the situation due to its awareness about the role and impact of new media in crisis situations. Stakeholders need to extend their picture of the situation by being aware of processes related to the communications capabilities of citizens during and after the crisis situation, which have dramatically changed in the last decade and also have a major impact on the social behaviour and understanding of the situation. Alert4All provides suitable means to incorporate this factor in the crisis management. The Alert4All Screening of New Media (SNM) tool allows monitoring the exchange of information among citizens during (even after) a crisis to complete the picture and gain awareness of the perception of the affected citizens about the current situation. This is a major innovation that provides a new dimension at the stakeholder side for the support of decision processes and enables adapting strategies in still running crisis situations (not only in the aftermath). As a matter of fact, the handling of new media is a relevant issue for Civil Protection and the results in this area aroused considerable interest to Civil Protection bodies.

More effective communications planning using decision support tools.
Through the Alert Simulation Module (ASM) integrated in Alert4All, alert strategies can be tested in concrete societal environments to support the decision making processes, especially in the planning phase, and to execute alert and communications plans. The ASM applies a model for calculating the diffusion of alert messages, combining the individual process of warning propagation through communications systems and the dissemination through the network of individuals. Hence, effective alert and communication strategies can be selected for each specific crisis, while strategies showing low performance can be discarded. Furthermore, decisions on the deployment of communication channels can be better prepared.

Improved penetration and enforcement of alerts through diverse communications means
Alert4All has proposed a complete communications paradigm for public alert that encloses the key enablers to achieve effective communication. On the one hand, the graphical user interface has been conceived to guide the stakeholders in the use of best practices when formulating and managing alert messages. On the other side, the application of the “Alerting Libraries” concept results in harmonised alert message formulation at any stakeholder, independently of their scope, background or nationality. Hence, a European citizen would be alerted with the same style and wording in any country making use of Alert4All. This is a major enabler to foster that recipients of alert messages understand and trust alert messages. Moreover, the use of the “Alerting Libraries” concept – supported by a communication protocol (the A4A protocol) and client-based applications at the receiver side – enables multi-modal and multi-language delivery of alert messages, so that impaired people can better understand the alert message, as well as tourists and immigrants. In addition, the A4A protocol allows cost-efficient transmission of alert messages by minimising the required capacity in communication networks.

On top of this, Alert4All supports increased penetration and re-enforcement of alert messages to the population by using a communications strategy involving diverse communications bearers. It provides a flexible and modular communications architecture that allows stakeholders disseminate alert messages through different mass market communications technologies. This way, citizens can receive alert messages in a diverse range of domestic or personal communication devices (e.g. digital TV, mobile phone, vehicular terminal, navigation terminal, etc.). The diversity concept has several advantages: first, it maximises the probability that a citizen is consumer of any of the targeted communications technologies; second, if a citizen is consumer of more than one of the technologies, it maximises the probability that he/she is using the corresponding devices at the time the alert is sent; third, the centralised coordination for the dissemination of alerts through different technologies enforces that alerts received by the citizen through different technologies are coherent with each other; fourth, the incorporation of satellite-based technologies as well as navigation devices provide resilience for crises in which terrestrial communication infrastructures are disrupted; fifth, the multi-channel approach allows exploiting complementarities in the warning (wake-up effect), information capabilities and other attributes of different means for alerting the population.

Contribution to European capabilities in Civil Protection and alerting
All the issues discussed above, together with the compilation of best practices enclosed in the development of the Alert4All system, the training materials and the development of the service roll out provides a significant contribution to the enhancement and harmonisation of alert and communications towards the citizens in crisis situations in European countries. In addition, the modularity and scalability of the Alert4All concept also allow its deployment element by element, where required, and at different levels, namely local, regional, national and European, both bottom-up and top-down.

Cooperation with other activities
The project team has also setup several cooperation frameworks with different initiatives during the project lifetime:

• Cooperation with the European GNSS Evolution Program – MRS Land Users Testbed (MLU TB): this cooperation involves the European Space Agency as European GNSS Evolution Program coordinator and funding entity for the MLU TB project, as well as the company Logica Ltd., coordinator of the MLU TB project. This cooperation was of high interest to A4A as the MLU TB project was implementing a testbed for disseminating alert messages (among other services) over GNSS. The cooperation was implemented at requirements and specification level. It was not possible to carry out a common demonstration given that ESA could not guarantee the access to the satellite capacity during the period for which the A4A final demonstration was planned.

• Cooperation with EC FP7 SEC – Opti-Alert project: Opti-Alert is a parallel project under the same funding programme and topic as A4A. Hence, cooperation between both projects was of high interest to share and harmonise results among different projects under the same programme. During the project lifetime, Alert4All and Opti-Alert held a joint workshop in Berlin to exchange results and a further workshop open to participants of the IDRC conference 2012.

• Cooperation with EC FP7 SEC – OSMOSIS Support Action: this cooperation was purely cross-dissemination cooperation that was concluded at early stages of the project since the time overlap between both projects was very short (OSMOSIS concluded early 2012).
• Cooperation with Mecom: Mecom is the company providing alert dissemination services to the German Federal Office of Civil Protection. This cooperation was set up with the aim of targeting a possible integration of the German alerting system (MoWaS) with the A4A proof-of-concept. Within this cooperation, it was possible to integrate Alert4All and MoWaS and carry out a demonstration showing that MoWaS users can make use of the alerting channels provided by Alert4All through integration.

• Cooperation with Euralarm: Euralarm is the association of European manufacturers and installers of fire and security systems. The cooperation was setup with high interest from both parties to complement the A4A proof-of-concept testbed with connection to fire alarm and/or evacuation systems within buildings. During the second phase of the project, A4A was integrated with several indoor security systems, such as voice evacuation and mass notification systems through cooperation with the Euralarm project PEARS. This achievement demonstrated that systems originally designed and deployed for indoor security purposes can be also applied for public alerting purposes.

The cooperation frameworks with ongoing activities have been kept for the project duration to exploit synergies and some of them (Euralarm, mecom and Opti-Alert) have been highly beneficial for the impact of the project and the involvement of industry stakeholders and raising the attention of further stakeholders.

Dissemination activities
The dissemination activities in the A4A project focus on making available the foreground generated during the project to the relevant audience; on the one hand to improve the concept with expert feedback and on the other hand to increase the project impact. The identified relevant audience splits in three groups: (i) the general public (as beneficiaries of an alerting system), (ii) the scientific community in the research areas tackled by the project and (iii) the potential A4A system end users, i.e. civil protection authorities and responders in European countries.

To reach these relevant communities, a strategy has been followed that applies a variety of dissemination channels, using the most suitable channels for each target group. It shall be noted that, even if the use of new media had been planned to reach the general public, the limitations faced by the special clause on limited dissemination in terms of timing and the high effort vs. impact ratio drove the team to discard the use of new media to reach the general public.

The project has implemented dissemination activities in three major areas:
• General high level dissemination activities, including the project website, newsletters, a video, handouts and a flyer.

• Focused dissemination to end users, by setting up an advisory board of end users that participates to project workshops, interviewing end users and participating to events that are relevant to the end users, such as conferences and congresses. The organised workshops within the project have been applied not only to disseminate the results to end users, but to involve them in the conception and design phases of the project ideas and system. In the last project phase, Alert4All participated in a number of events and exhibitions (Euralarm General Assembly, EENA workshop, BAPCO exhibition, 4th Civil Protection Forum), where small scale demonstrations could be carried out. The project team could present the project to Mrs. Kristalina Georgieva, the European Commissioner for International Cooperation, Humanitarian Aid and Crisis Response, in two occasions: the 7th European Civil Protection Congress in 2011 and the 4th Civil Protection Forum in 2013.

In addition, the project coordinator was invited as speaker in several relevant fora, such as the panel on “Alerting the Population” in the Security Research Event “Research meets Industry” organised by the European Commission in 2012, the conference on “Civil Alert in Europe 2020 - an Outlook” organised by the German European Security Association (GESA) in 2012 and the workshop on security research organised by the Commission on Civil Protection of the Federal Ministry of Interior in 2012.

Furthermore, the project team organised a final demonstration with guests from politics, industry, research and civil protection stakeholders, where the most remarkable features of the developed system were successfully demonstrated live. Further focused small scale demonstrations have been carried out during December towards end users in Sweden and in January (after the official completion of the project) towards German-speaking broadcasters. Furthermore, a small scale demonstration is planned on request from Euralarm to be held during the Euralarm General Assembly in May 2014.

• Focused dissemination to the scientific community by participating and contributing to relevant conferences and journals to the research areas of the project. A total of 17 scientific papers and four presentations in conference proceedings have been published, 5 poster contributions have been provided to conferences, 3 further scientific papers are accepted for publication in the proceedings of upcoming conferences in 2014, two journal papers are accepted for publication in 2014, a contribution to one book chapter will be published in 2015 and further three journal paper submissions are planned after the project official conclusion.

Exploitation plans
The Alert4All team is active in different parallel fronts to foster the deployment of a public alert service based on Alert4All first at local/regional or national level and to grow European. The open fronts are in the areas of standardisation, stakeholder involvement and further dissemination, manufacturing, further projects and service roll out.

Many developments carried out in Alert4All would require standardisation processes to foster the applicability at European level: e.g. the best practices guided composition of alerts, transport protocol, application of alerting libraries for multi-language / multi-modal delivery, A-COP and alerting plans templates. Hence, part of the project team has become active in ETSI-EMTEL, ETSI-SES-SatEC and ITU to propose technical reports and specifications based on the outcomes of the project. Furthermore, the team is investigating ways to join ISO TC 223 to propose contributions that allow the use of alerting libraries, e.g. definition of taxonomy, syntax and logic rules to be applied to a message composition engine in different languages. The intention is to create a number of complementary documents in different (the relevant) standardisation bodies for each type of content, so that the technology and the processes can be defined in a coherent manner in the relevant framework.
Concretely, the team is following the proposal from the Chairman of ETSI-EMTEL to create a number of Work Items related to the outcomes of Alert4All: the complete framework, the use of each communication technology for alerting purposes (e.g. HbbTV, navigation signal, direct broadcast, etc.). Furthermore, the current ETSI-SES-SatEC STF activity on the technical specification of an encapsulation method for alert messages is being followed by DLR with the intention to include contents from the Alert4All protocol in this new specification. Furthermore, through cooperation with Euralarm, the progress in ISO TC 223 is being followed.

Stakeholder Involvement and Further Dissemination
During the Alert4All project and through the final demonstration, the A4A team has gained a large number of contacts to civil protection agencies. This is being exploited currently to establish discussion on the needs and interest of the different countries in the concept after their impressions during the project and the final demonstration. To complete the picture, the different Alert4All partners have agreed to create a subtitled version of the A4A video in their own languages for distributing the video also to non-English speakers.

The team shall now establish concrete discussions with each of the civil protection agencies to understand how Alert4All could be deployed in their regions/countries in parallel. In particular, referent people at the Portuguese, Spanish, British, German, Netherlands, Swiss, Swedish and Norwegian civil protection authorities are being contacted (and will be further contacted during the first half of 2014) to carry out such discussions. This shall help understanding how the local/regional/national deployment could work in different countries and explore how to grow European by interconnecting them. The involvement of end users shall also support the standardisation process (for example in ISO, where they are directly represented in many cases). Hence, carrying out the standardisation and stakeholder involvement in parallel is positive. Furthermore, if the technology development (prototyping and implementation) is also carried out in parallel, the three activities can culminate together in pilot phases in selected localities or regions. Concretely, one A4A partner is proposing to provide consultancy services to end-users for the adaptation of the system to their working manner/needs for a successful system implementation and deployment.

Furthermore, other stakeholders need to be involved as well to foster prototyping, implementation and service deployment. On the one hand, manufacturers need to be involved to implement the required SW updates to receivers or applications for consume. On the other hand, service providers and especially broadcasters need to be involved and agree on the direct injection of alert messages in the TV content. For this purpose, the A4A team plans to carry out further dissemination activities after the official closing of the project, including small scale demonstrations. For this purpose, the A4A team has agreed to keep the system up and running for 6 months after the project closing and accessible for the project partners for dissemination and research purposes free of charge. Two concrete examples of further dissemination activities are already in the agenda:

• A colloquium focused on German-speaking broadcasters (German, Austrian and Swiss) organised by IRT with the participation of DLR. The invitation is published in the link below:

• A presentation and small scale demonstration in the Euralarm General Assembly in May 2014 (the agenda is not yet public).

The implementation of different components of the A4A system for consumers requires the involvement of manufacturers. DLR is carrying out first discussions with Eutelsat to propose ideas for the implementation of the DVB-T2 enhanced receiver with A4A decoder in external plug-in devices, settop boxes and other developments (e.g. the recently presented SmartLNB) to deliver alert messages over terrestrial and satellite TV broadcast to future but also state-of-the-art TVs.

Additionally, the involvement of Euralarm during the final phase of Alert4All eases the involvement of indoor building security manufacturers to implement the A4A protocol and interfaces to deliver alerts in indoor security systems. The discussion after the small scale presentation during the Euralarm General Assembly will provide more insights on the steps to follow.

Further Projects
Partners from the A4A Consortium will make further use of the gained foreground in further already acquired projects. In particular, the EU-FP7 projects PHAROS (Space Programme) and DRIVER (Security Programme) will benefit from developments carried out in Alert4All. Furthermore, in the PHAROS project, it is planned to further develop the Global Alerting Gateway to be ready for operational use and further investigate solutions for alert message delivery through navigation signal.

Service Roll-Out
The service roll-out is dependent on the consumer technology readiness (implementation of receiver applications in mass market consumer devices), the willingness of civil protection agencies to pay for the system and service and the creation of a service provider (or adoption of the service by an existing service provider). One A4A partner is investigating the option to become the service provider in Spain. In parallel, potential service providers in other countries shall be contacted to prepare the terrain. The contacts to these players must start already in the next half year to understand their interest, apply consultancy for costs-benefits and establish a business plan that would allow going back to the civil protection agencies for service offering and negotiation. Furthermore, service level agreements need to be set up with intermediate providers (e.g. satellite broadcasters, terrestrial broadcasters, operator of EGNOS, mobile networks operators, etc.) to ensure the alert delivery and establish the conditions.

List of Websites:

Contact details:
Mrs. Cristina Párraga Niebla
Alert4All Project Manager

DLR – German Aerospace Center
Member of the Helmholtz Association
Institute of Communications and Navigation

82234 Weßling

Telephone: +49 8153 28-2824
Telefax: +49 8153 28-2844