Intention recognition for real-time automotive 3D situation awareness (IRRA)
About the project
Background and needs
- New services for vehicles and for traffic regulation often involve data-driven prediction of drivers, vehicles and traffic systems.
- Since data is scant, or irrelevant, knowledge of road users' (not necessarily directly observable) intentions is an important factor in being able to make predictions (e.g., of traffic flows after an accident).
- Intentional recognition (IR) uses structure in observations to derive the intention of drivers, vehicles and traffic systems
- Existing IR algorithms are limited to directly observable intentions and assume access to large amounts of training data. These algorithms need to be significantly improved to be able to be applicable when data is scant or when intentions cannot be observed directly.
Method and implementation
- Selection of use cases based on important industrial and social application areas for IR: drivers, vehicles, and systems
- Improvement of existing algorithms for IR using state-of-the-art in machine learning, computer vision, multi-agent systems, automatic reasoning, localization and new sensor technology.
- Evaluation of results is performed through proof-of-concept implementation, vehicle-based tests, and through scientific publication.
- Knowledge transfer is ensured through an industrial doctoral student, workshops, and publications.
Knowledge transfer from academy to industry and society:
- New PhD.
- Dissemination of scientific results.
- Increased competitiveness for the Swedish automotive industry through new types of services based on IR with application in driver assistance, autonomous vehicles and road safety.
- Methods to improve traffic flow in connection with disruptions.
- Academic profiling in a growing AI area.
The project is a collaboration between Volvo Car Corporation (coordinator), the Swedish Transport Administration (Trafikverket), the University of Skövde, RISE SICS and Smart Eye. The project is funded by Swedish Governmental Agency for Innovation Systems (Vinnova) within the Strategic vehicle research and innovation (FFI) program and the total grant is 12 million SEK during 2019-01–2022-12 of which the University of Skövde receives 2.3 million.
See also Vinnova's description of the project.