Di Liu

Project title: Self-organizing platoons of cooperative vehicles in heterogeneous traffic conditions

Host Institution: Technical University of Munich (TUM)

Host Supervisor: Prof. Matthias Althoff

Co-host Institution:École Polytechnique Fédérale de Lausanne (EPFL)

Co-host Supervisor: Prof. Alexandre Alahi

Summary project: Problems related to increasing traffic and limited road infrastructure are emerging in several countries. Cooperative automated vehicles promised to solve such problems by envisioning vehicles autonomously forming platoons, adaptively merging/splitting according to traffic conditions,
and safely interacting with human drivers. Unfortunately, such promises are largely unattended. A crucial bottleneck in cooperative vehicles technology lies in the control algorithms, which are still unable to operate autonomously and provably safe in actual traffic conditions.
The SELF-PLATOON project aims to enhance cooperative vehicles technology with self-organizing functionalities 1) Self-organized mixed traffic: SELF-PLATOON will study how adaptive and learning-based control can form platoons in heterogeneous traffic with human drivers. Tools from socialaware artificial intelligence will predict the behaviour of human drivers and promote safe interaction between human-driven and automated vehicles.
2) Self-organized communication: SELF-PLATOON will co-design control and
communication, to guarantee safety even when the communication degrades or fails. Optimal control of hybrid/switched systems (switched-based dynamic programming) will make communication switch and dynamically evolve depending on traffic conditions.
3) Self-organized platoon manoeuvres: in addition to vehicle-following functionality, SELFPLATOON will use formal methods based on reachability analysis to prove the correctness of each driving action (e.g. merging/splitting functionalities with safety specifications).