Êtes-vous un étudiant de l'EPFL à la recherche d'un projet de semestre?
Travaillez avec nous sur des projets en science des données et en visualisation, et déployez votre projet sous forme d'application sur GraphSearch.
This paper presents a real-time motion planning scheme for urban autonomous driving that will be deployed as a basis for cooperative maneuvers defined in the European project AutoNet2030. We use a path-velocity decomposition approach to separate the motion planning problem into a path planning problem and a velocity planning problem. The path planner first generates a collision-free piecewise linear path and then uses quintic Bezier curves to smooth the path with C-2 continuity. A derive-free optimization technique Subplex is used to further smooth the curvature of the path in a besteffort basis. The velocity planner generates an optimal velocity profile along the reference path using Model Predictive Control (MPC), taking into account user preferences and cooperative maneuver requirements. Simulation results are presented to validate the approach, with special focus on the flexibility, cooperative-awareness and efficiency of the algorithms.