HIDDEN · HYBRID INTELLIGENCE FOR ADVANCED COLLECTIVE PERCEPTION AND DECISION MAKING IN COMPLEX URBAN ENVIRONMENTS

Urban mobility poses various challenges to CCAM systems. Timely and reliable detection of occluded objects and especially vulnerable road users (VRUs) is a major one. To tackle this challenge, HIDDEN is developing advanced collective awareness (CA) and decision-making algorithms, with or without road infrastructure support. The CA system is focusing on the detecting occluded objects, including VRUs, while predicting their short-term trajectories using advanced behavioural models. In parallel, a novel driver gaze tracking and status monitoring system, is developed. The output of the CA and the driver gaze tracking is utilised by the real-time decision-making algorithms, which are designed to be explainable and aligned with human driving styles and ethical principles. HIDDEN deploys Hybrid Intelligence techniques across the whole chain, from perception till decision, to benefit from the combination of human with machine intelligence. During this process AI-related ethical and legal aspects are carefully considered via the development of a dedicated framework. HIDDEN developments will be evaluated both in real-world and in simulation. Real-world tests will be conducted with the AVs owned by the consortium (8 in total) in the respective testing facilities or on public roads, while for virtual tests novel co-simulation environments will be deployed. HIDDEN consortium has pre-selected four main use cases that drive the technical developments of the project. HIDDEN is developing CCAM systems which are not just technologically advanced but also deeply aligned with human driving styles, ethical principles and regulations, setting a new benchmark for the future of AVs technology.To increase HIDDEN’s impact, the consortium intends to reach out to CCAM stakeholders, in EU and beyond, engage in a continuous discussion with EU type approval authorities and UNECE working groups and promote mature results to standardisation.