Velocity planning approach for autonomous vehicles
Abhiram Rahatgaonkar · Chalmers Publication Library (Chalmers University of Technology) · 2016
Motion planning being one of the core functionalities of autonomous driving system, has been studied and investigated with great interest by the developers across industries and academia.Collision avoidance is one of the major challenges within trajectory planning problem (TPP).This master thesis is inspired from the 'concept of decomposition of trajectory planning problem in to path planning and velocity planning' which results in great reduction in the computational complexity of the TPP.The work in this project is focused on the velocity planning with the known path so that the collision with the moving vehicles is avoided.Concept of space time scheduling is used along with the A* (A star) search algorithm in a particular fashion to achieve the time optimal speed profile.The approach developed in this work is deterministic but has possibility to work with probabilistic motion model of the vehicles in the traffic.It also has great potential to work with the different road geometries and different vehicle maneuvers such as merging, yielding, intersection, etc.Hence, it is very promising approach for velocity planning with possibilities to work with variety of scenarios.