Periodic Reporting for period 2 - TARGET (Training Augmented Reality Generalised Environment Toolkit)
Período documentado: 2016-11-01 hasta 2018-10-31
Security Critical Agents (SCAs) also known as first responders require a range of competencies, knowledge and practical experience to cope with a complex civil protection and terrorist events. These challenges can range from managing large scale cyber-attacks at the strategic level in a command centre through to tactical fire arms training undertaken by officers on the ground. The diversity, complexity and relative rarity of situations creates unprecedented challenges for effective SCA training. In addition, due to the increasing visibility of emergency services owing to social media, societal pressures are imposing more expectations from technology, more accountability from emergency services, and less tolerance for errors. While there is a demand for SCA training, it can be costly, complex, time consuming and potentially dangerous. In addition the frequency of training can often be limited due to the inability to easily customize or re-run scenarios or due to regulatory restrictions. Furthermore, assessment techniques often side step key issues such as situational awareness or team performance. Taken together while existing approaches have been successful in many respects the need to increase the amount and diversity of training options due to the current security climate means that cost-effective and more efficient complementary approaches need to be found. In order to address these challenges TARGET defined five objectives: a generic platform and supporting content which can be connected to third party technologies, a component based approach to allow for the modification and creation of new scenarios, the development of six demonstration training scenarios, assessment with real end users and the creation of long-term sustainable impacts.
TARGET undertook an extensive requirements capture exercise which involved observing current training practices within the end-user partner organisations. This allowed for the identification of possible scenarios, from which aspects such as training objectives, roles of the participants, use of objects, stress factors and locations were more clearly understood. It also allowed for the identification of which hardware and software technologies and components would be the most suitable. Throughout, the project the technical and end-user partners developed and revised the TARGET platform and scenarios. This also led to the identification of when the use of serious games or gamification may or may not be appropriate within such training scenarios. Six scenarios were proposed and developed to a TRL level of 4-7, rather than a TRL level of 9 for a commercial product.
A major challenge within TARGET was the integration wide variety of technologies. These ranged from specifically developed components such as Mixed Reality (MR), assessment tools and geospatial environments through to legacy police command and control systems. An example is collecting data in real time from interactions within augmented reality (AR) and displaying training metrics. A message oriented middleware approach was used which in future allows for the connection of new devices and services. TARGET provides tools which let trainers develop new or customize existing scenarios. A run-time environment is also provided so that the trainer can change aspects of the scenario in real time, they can also view assessment data.
TARGET developed two drone approaches for capturing models of real world outdoor locations. The first an unmanned photogrammetry survey drone was developed and used to capture a variety of locations which were used within two of training scenarios these include a Berlin Airport and a motorway in Germany. A second iteration which used a swarm drone approach was developed and successfully tested. MR formed a key part of TARGET and the scenarios supported AR, virtual reality and also the use of tracked objects within AR scenes. As the project was security related, an overall MR platform was developed which could be easily installed and run on local machines or remote servers. Positioning of real world objects indoors is also supported and these can be tracked and behaviors attached which impact on what is displayed within the AR HMDs.
A major challenge within TARGET was the integration wide variety of technologies. These ranged from specifically developed components such as Mixed Reality (MR), assessment tools and geospatial environments through to legacy police command and control systems. An example is collecting data in real time from interactions within augmented reality (AR) and displaying training metrics. A message oriented middleware approach was used which in future allows for the connection of new devices and services. TARGET provides tools which let trainers develop new or customize existing scenarios. A run-time environment is also provided so that the trainer can change aspects of the scenario in real time, they can also view assessment data.
TARGET developed two drone approaches for capturing models of real world outdoor locations. The first an unmanned photogrammetry survey drone was developed and used to capture a variety of locations which were used within two of training scenarios these include a Berlin Airport and a motorway in Germany. A second iteration which used a swarm drone approach was developed and successfully tested. MR formed a key part of TARGET and the scenarios supported AR, virtual reality and also the use of tracked objects within AR scenes. As the project was security related, an overall MR platform was developed which could be easily installed and run on local machines or remote servers. Positioning of real world objects indoors is also supported and these can be tracked and behaviors attached which impact on what is displayed within the AR HMDs.
A key part of TARGET was the development and testing of novel and often immature technologies with a TRL level of 4-7 for use within SCA training. The project was not aiming for a commercial level product with high levels of graphical realism but instead to explore the potential future use of such technologies for SCA training. This led to some mismatch between end user expectations and the final outputs of the project. One key limitation being the field of view. This resulted in the project identifying that current AR technologies are not suitable for some of the training scenarios in TARGET, however as they mature many of the limitations encountered may be overcome. The project has identified one very strong use case which has attracted significant interest and further development funding.
From a platform perspective TARGET has moved beyond the state-of-the-art used in training scenarios as it connects both legacy tools and new innovative technologies together so that organisations can train using the components which best reflect their needs. As it stands TARGET connects MR technologies, a base platform, tracked real world objects, advanced drone imaging technologies and geospatial systems. When combined with the scenario editing tools and the advanced assessment engine, they provide a strong basis for future research and work in SCA training. Furthermore, the combination of these components allows scenarios which can be used by different command levels.
TARGET is flexible and allows organisations to assemble new scenarios based on a library of components. While the editing tools require some improvement, they allow for (compared to current solutions) ways to create and modify scenarios on site, therefore making it lower cost and easier to run scenarios. A key finding in the project is that first responder organisations are often reluctant to adopt serious games of gamification approaches within sensitive training scenarios. These were for a variety of reasons including the negative perception by the trainees and members of the public to “game-like” training approaches. Also reducing training scenarios to points, levels and badges was viewed by some to trivialize scenario or be mismatched to overall learning objectives. TARGET never intended to replace existing training techniques, but instead aimed to complement them and has the potential to provide a new phase in the training cycle and at relatively low-cost.
While the currently available AR hardware may not be suitable for all use scenarios at this time, other aspects of the project have illustrated high levels of immediate promise. For example, the advanced 3D photogrammetry drone systems developed in TARGET have won a major award, are now going through a commercialisation phase. While work on the 2D geospatial tools for use within the command post environments continues to attract interest. The TARGET systems were demonstrated at varying TRL (4-7) levels within end-user organisations and a thorough assessment of the strengths and weaknesses was undertaken. A start-up has also been set-up to exploit other results from TARGET. Finally, two scenarios will be possibly be used for training.
From a platform perspective TARGET has moved beyond the state-of-the-art used in training scenarios as it connects both legacy tools and new innovative technologies together so that organisations can train using the components which best reflect their needs. As it stands TARGET connects MR technologies, a base platform, tracked real world objects, advanced drone imaging technologies and geospatial systems. When combined with the scenario editing tools and the advanced assessment engine, they provide a strong basis for future research and work in SCA training. Furthermore, the combination of these components allows scenarios which can be used by different command levels.
TARGET is flexible and allows organisations to assemble new scenarios based on a library of components. While the editing tools require some improvement, they allow for (compared to current solutions) ways to create and modify scenarios on site, therefore making it lower cost and easier to run scenarios. A key finding in the project is that first responder organisations are often reluctant to adopt serious games of gamification approaches within sensitive training scenarios. These were for a variety of reasons including the negative perception by the trainees and members of the public to “game-like” training approaches. Also reducing training scenarios to points, levels and badges was viewed by some to trivialize scenario or be mismatched to overall learning objectives. TARGET never intended to replace existing training techniques, but instead aimed to complement them and has the potential to provide a new phase in the training cycle and at relatively low-cost.
While the currently available AR hardware may not be suitable for all use scenarios at this time, other aspects of the project have illustrated high levels of immediate promise. For example, the advanced 3D photogrammetry drone systems developed in TARGET have won a major award, are now going through a commercialisation phase. While work on the 2D geospatial tools for use within the command post environments continues to attract interest. The TARGET systems were demonstrated at varying TRL (4-7) levels within end-user organisations and a thorough assessment of the strengths and weaknesses was undertaken. A start-up has also been set-up to exploit other results from TARGET. Finally, two scenarios will be possibly be used for training.