Case study KoSi – Cooperative autonomous driving with safety guarantees
Ensuring safety in autonomous driving remains a key challenge, particularly in mixed traffic. In the KoSi project (Cooperative Autonomous Driving with Safety Guarantees), fortiss has developed a cooperative motion planning system using AI to improve behavioural safety and decision-making. This enabled successful demonstrations of coordinated automated vehicle platoons and paved the way for safer autonomous mobility.
Solution
fortiss developed a novel approach to cooperative motion planning for highly automated and autonomous vehicles in mixed traffic. This included direct and indirect communication mechanisms with other road users, supported by algorithms for behaviour and trajectory prediction. In addition, the team integrated compliance with traffic regulations into the planning process and validated the approaches through simulations and road tests with prototype vehicles. The project culminated in a live demonstration featuring the fortiss research vehicle fortuna and the EDGAR vehicle from the Technical University of Munich (TUM), which drove cooperatively in a platoon.
Result
- Development of AI-based cooperative motion planning methods for autonomous vehicles in mixed traffic
- Integration of behaviour prediction and compliance with traffic rules into trajectory planning
- Validation of the approaches through simulations and real-world vehicle demonstrations
- Successful demonstration of cooperative platooning journeys with fortuna (fortiss) and EDGAR (TUM)
- Progress in safety guarantees for autonomous driving under dynamic urban conditions
- Strengthening readiness for future safe and cooperative autonomous mobility systems
Outcome
KoSi demonstrated that the integration of cooperative decision-making, AI-based perception and behaviour prediction significantly enhances the safety of autonomous vehicles in mixed traffic. fortiss also proved that real-time interaction with other road users is essential for practical and trustworthy automated driving, thereby laying the foundation for future scalable and safe mobility solutions.
Project partner
