Travail de fin d'études et stage[BR]- Travail de fin d'études : Preliminary Design and Development of a Behavior Planner for Cooperative Highway Merging in Self-Driving Vehicles[BR]- Stage d'insertion professionnelle : FEV.io

Abstract

Advanced Driver Assistance Systems and Automated Driving Systems improve mobility as a whole by increasing security, enhancing comfort and enabling new transportation services. In this context, the EU co-funded the Hi-Drive project, which aims at extending the capabilities of self-driving vehicles by decreasing the amount of situations where the driving tasks have to be performed by the driver. One of the enabling technologies is the Vehicle to Vehicle communication, which allows vehicle cooperation in complex situations such as intersections, merging or traffic jams. To avoid the use of several communication protocols, one of the project's goal is to demonstrate the convenience of standardized cooperation methods. The objective of this master's thesis is then to design and develop a behavioral planning software which defines the most appropriate maneuver for the vehicle to perform when merging onto the highway in the presence of other vehicles. It takes advantage of vehicle cooperation in order to generate safe and smooth trajectories. After a short overview of the state-of-the-art of behavioral planning algorithms and vehicle cooperation as well as the definition of relevant concepts in automated driving systems, this master's thesis describes the design and implementation of a Decision Tree, the selection of the most appropriate maneuver and the computation of trajectories. The development process includes high level requirement analysis based on test scenarios defined by Hi-Drive, software architectural design, software implementation as well as unit testings of the Behavior Planner. Integration of the software in the vehicle and validation processes are however not covered. In the end, this work contributes to move towards a higher level of standardization for vehicle cooperation processes in automated driving systems by proving the interoperability of cooperation protocols. It also proposes a logic-based algorithm for behavioral planning in cooperative highway merging situations. Based on the tests that have been performed, the software generates consistent behaviors and trajectories with regard to the test cases and functional requirements.