Periodic Reporting for period 2 - OPTIMUM (Multi-source Big Data Fusion Driven Proactivity for Intelligent Mobility)
Reporting period: 2016-11-01 to 2018-08-31
OPTIMUM developed an information platform supported by big-data analytics allowing greater integration of the modal networks and promotion of collective traveling options. Through a mobile app, OPTIMUM enabled the provision of multimodal trip alternatives and the integration of proactive information in real time to improve access levels of public transport and promote sustainable means of mobility and proactive decision making plus sustainable transportation behaviors.
OPTIMUM’s smart sensing system was able to cope with a huge amount of heterogeneous data in real-time and the app was trialed in 3 European cities, Birmingham, Vienna and Ljubljana.
From the field trials, OPTIMUM achieved its original targets for improvements in travel efficiency of daily journeys when using public transport; shift from individual transport towards sustainable modes of transport; improvement of private car users willing to use public transport; improvement of well-being by increasing use of walking/cycling of travelers with high environmental awareness and increase of accepting new incentives to change modes.
In the domain of freight transport, OPTIMUM ventured to address the underutilization of tolled motorways for the transport of goods, evidenced that transport operators choose non-toll-based routes to cut down on operational costs. These routes often involve the use of former national networks causing high maintenance costs and congestion since these networks were not projected for such high levels of traffic especially from heavy vehicles. OPTIMUM developed a dynamic charging model trialed in Portugal based on prevailing conditions related to traffic and travel times. The developed model after evaluation of the trials’ results revealed reduction in travel times and fuel consumption for the fleet of the participating vehicles and demonstrated an increase in revenues for the toll operator.
The developed architecture guided the modelling and implementation of all R&D activities structured into 4 layers. For Observe layer, a working data infrastructure constituting a single point of access for all data was built. The initial data architecture was enhanced and fine-tuned to integrate all necessary data operators for automated retrieval of the ITS-sensor and multimodal information from distributed data sources allowing for the harmonisation and enrichment of the data sources.
Travel demand and forecasting models were developed and fine-tuned in terms of predictive accuracy and execution efficiency as part of the Orient layer while a complex event processing engine was designed, developed, integrated in the information flow, validated and fine-tuned.
For Decide layer, a dynamic charging model was developed and fine-tuned calculating the discount for any given road section containing a specific toll. The multimodal routing algorithm was also enhanced and fine-tuned to integrate the feedback received, increase accuracy and integrate additional querying parameters.
The (Pro-)Act layer involved development of the OPTIMUM user model and a range of information personalisation and persuasive recommendation services. These services were also enhanced and fine-tuned to integrate the feedback received, increase accuracy and deliver more personalised recommendations.
All components were integrated together as part of OPTIMUM applications used as part of the two rounds of the pilots during the project lifecycle.
The overall dissemination strategy was applied and the targets were reached engaging the specified audience. OPTIMUM engaged key stakeholders such as decision makers, technology experts, researchers, policy makers and potential users and will further liaise with already existing clusters on the field even after the project end.
The exploitation methodology was successfully defined and the consortium identified the business models that could be employed for the commercialisation of OPTIMUM.
• Big-data architecture for traffic forecasting Development of a social mining component for identification of situations of interest on transport networks using public input
• Optimisation technique for deriving missing information as part of a region-wide microsimulation model
• Big-data architecture based on open-source tools, Development of a data harmonization service and user interface responsible for mapping transportation related data into OPTIMUM reference model
• Contribution for a web interface responsible for presenting toll prices and best routing alternatives taking into account the cost of the trip
• Advanced Hybrid Choice Model estimating the probability of an individual to choose a specific multimodal alternative depending on its socio-economic characteristics, attitudes, perceptions and multimodal attributes
• OPTIMUM framework for sustainable mobility supporting transportation decisions via personalised behavioral change interventions through data driven user profiles, analysis of user feedback on past persuasive interactions and contextual information
• Route recommendation service operating on lists of alternative routes for traveling provided by routing engine and returns a personalized ranked list of recommended routes
• Dynamic charging model for freight vehicles employing optimal toll price taking into consideration time and congestion level on the road in various time-of-day intervals
• Mobility patterns detection and prediction service with integrated real time automatic travel mode detection
• Crediting model for the multimodal case comprised of two choice models: a reward choice model exploring the individuals’ preferences on different types of rewards and a mixed discrete choice model estimating the probability of an individual to choose a multimodal alternative in the presence of the aforementioned types of rewards
• Motorhome AI Communication Hardware (MACH) and software for Car2X pilot. MACH is installed into a motorhome and allows monitoring via graphical user interface
• Novel approach for Complex Event Processing (CEP), a complex data processing flow (instead of pattern matching) aligned with the need for an intensive stream processing as required in the transportation domain. Additionally, we introduce the concept of the Traffic Processing Network (TPN) as a specialization of the Event Processing Network (EPN) defining a new process representing a meta processing layer for the use cases, reflecting the complex (data-intensive) and dynamic nature of relevant scenarios