Periodic Reporting for period 1 - HEIDI (Holistic and adaptivE Interface Design for human-technology Interactions)
Periodo di rendicontazione: 2022-09-01 al 2024-02-29
The HEIDI project aims at enabling safe, efficient and acceptable interactions between vulnerable road users and vehicles, including vehicles with different levels of automation and users with reduced (mental and/or physical) abilities due to age or other temporary conditions. The holistic approach of HEIDI considers the users on both sides of the vehicle (internal and external) as part of the same CCAM system, therefore providing cooperative solutions to a large variety of potentially critical situations. The HEIDI solutions use data collected from inside and outside the vehicle to detect the state of the respective interaction agent and to recognize the type of interactive situation in advance. The optimal solution, which reduces the risk of danger and optimizes the comfort and efficiency on both sides, is calculated for the respective agents. Internal and external interfaces communicate the proposed solution (e.g. yield/pass) considering the evolution of the situation and the intentions and responses of all the interaction partners. For this, interfaces inspired by fluid interaction concept are designed and tested in co-simulation environments to enable foresight-safety, i.e. the human-like ability to prevent dangerous situations before a hazard occurs.
The overall objectives of the HEIDI project are: 1. Develop and demonstrate fluid, cooperative HMI solutions; 2. Develop technical innovation modules for mutual awareness between road users and drivers; 3. Develop suitable validation methods for assessing fluid, cooperative HMI solutions; 4. Recommendations for regulation, standardisation, and development of adaptive internal and external HMIs.
The project specifically targets the development of user-centred solutions to enable safe and efficient interactions between road users with different needs. Therefore, the integration of Social Sciences and Humanities has occurred at several stages throughout the project. Considerations for different user categories and respective needs (for both drivers and pedestrians) have guided the requirement specifications of several use cases in WP1, HMI solutions in WP2-4, research methodologies in WP5. Similarly, the iterative development and testing of the internal and external (virtual) prototypes in WP2 and WP3 have throughout adopted user-centred design principles. Also, the cooperative logics developed in WP4 is based on human factors aspects that optimize the benefits for different user categories in terms of safety and efficiency. Specific tasks in WP5 are dedicated to elaborating and implementing ethical research methodologies that are being specifically developed to test cooperative fluid interfaces. In line with this, an ethics advisor with strong Humanities background have been appointed and ethics guidelines developed to ensure the adherence of the research work to the highest ethical standards.
The overall objectives of the HEIDI project are: 1. Develop and demonstrate fluid, cooperative HMI solutions; 2. Develop technical innovation modules for mutual awareness between road users and drivers; 3. Develop suitable validation methods for assessing fluid, cooperative HMI solutions; 4. Recommendations for regulation, standardisation, and development of adaptive internal and external HMIs.
The project specifically targets the development of user-centred solutions to enable safe and efficient interactions between road users with different needs. Therefore, the integration of Social Sciences and Humanities has occurred at several stages throughout the project. Considerations for different user categories and respective needs (for both drivers and pedestrians) have guided the requirement specifications of several use cases in WP1, HMI solutions in WP2-4, research methodologies in WP5. Similarly, the iterative development and testing of the internal and external (virtual) prototypes in WP2 and WP3 have throughout adopted user-centred design principles. Also, the cooperative logics developed in WP4 is based on human factors aspects that optimize the benefits for different user categories in terms of safety and efficiency. Specific tasks in WP5 are dedicated to elaborating and implementing ethical research methodologies that are being specifically developed to test cooperative fluid interfaces. In line with this, an ethics advisor with strong Humanities background have been appointed and ethics guidelines developed to ensure the adherence of the research work to the highest ethical standards.
WP1 Requirements, User Needs and Use Cases (progress: 70%) has produced a systematic description and classification of interaction use cases between pedestrian(s) and vehicles, including the specific needs of different pedestrians (regular, older, child, with reduced mobility) and drivers (regular, distracted, older) that must be considered in the design of interfaces and warning functions.
WP2 iHMI and Internal Sensing (progress: 100%) has iteratively designed and tested an internal human-machine interface (iHMI) to inform, warn and alert the driver of crossing pedestrians. The results of the exploratory studies in simulation environment provide solid evidence that the current implementation is effectively improving the perceived safety, acceptance and usability of such a system.
WP3 eHMI and External Sensing (progress 100%) has advanced our understanding of how to effectively recognize pedestrians and communicate with them. The limited visibility and intelligibility of external interface (eHMI) and the timings of the action recommendations sequence in a fast-evolving uncertain situation, as the one where a pedestrian is about to cross in front of an incoming (automated) vehicle, have been considered.
WP4 Cooperative HMI & Situation Understanding (progress: 80%) has produced the concept for the cooperative HMI and the interface formats to allow data exchange across components. Different modules are being developed and integrated in co-simulation setup to build the situation understanding, create joint behavior recommendations and track the situation evolution.
WP5 Validation Methods & Standardisation Recommendations (55%) has established the ethical framework and protocols to conduct research studies for the assessment of adaptive interfaces in co-simulation environments. Moreover, validation methods that encompass objective, subjective and qualitative data are being developed in preparation for evaluation studies of the internal, external and cooperative interfaces.
WP6 Prototyping (progress: 50%) The hardware required to provide the sensor data, the output units (HMIs) and the necessary infrastructure in the vehicle has been designed, while the implementation is ongoing. The requirements of real time transmission and application-specific data structure have been incorporated in the Data Distribution Service (DDS) as an essential component of the Osmotic Layer software which is being integrated in the prototype vehicle.
WP2 iHMI and Internal Sensing (progress: 100%) has iteratively designed and tested an internal human-machine interface (iHMI) to inform, warn and alert the driver of crossing pedestrians. The results of the exploratory studies in simulation environment provide solid evidence that the current implementation is effectively improving the perceived safety, acceptance and usability of such a system.
WP3 eHMI and External Sensing (progress 100%) has advanced our understanding of how to effectively recognize pedestrians and communicate with them. The limited visibility and intelligibility of external interface (eHMI) and the timings of the action recommendations sequence in a fast-evolving uncertain situation, as the one where a pedestrian is about to cross in front of an incoming (automated) vehicle, have been considered.
WP4 Cooperative HMI & Situation Understanding (progress: 80%) has produced the concept for the cooperative HMI and the interface formats to allow data exchange across components. Different modules are being developed and integrated in co-simulation setup to build the situation understanding, create joint behavior recommendations and track the situation evolution.
WP5 Validation Methods & Standardisation Recommendations (55%) has established the ethical framework and protocols to conduct research studies for the assessment of adaptive interfaces in co-simulation environments. Moreover, validation methods that encompass objective, subjective and qualitative data are being developed in preparation for evaluation studies of the internal, external and cooperative interfaces.
WP6 Prototyping (progress: 50%) The hardware required to provide the sensor data, the output units (HMIs) and the necessary infrastructure in the vehicle has been designed, while the implementation is ongoing. The requirements of real time transmission and application-specific data structure have been incorporated in the Data Distribution Service (DDS) as an essential component of the Osmotic Layer software which is being integrated in the prototype vehicle.
The HEIDI project goes beyond the state-of-the-art as it adopts, develops and demonstrates: Fluid design approach for "human-like" adaptive interfaces; Osmotic Layer for wireless data communication within and between vehicles and infrastructure; Internal HMI (for drivers), external HMI (for pedestrians), and cooperative HMI (for optimized joint behavior), working across all automation levels and adapted to different user needs and abilities; Sensing systems & algorithms to identify different types of pedestrians and drivers and predict their intentions; Validation methods for investigating complex real-time interactions between VRUs and vehicles in co-simulation systems. The outcomes of this work will serve to provide recommendations necessary for the regulation and standardisation of the HEIDI system and similar.