Application of Chance-Constrained and Sample-Based Path Search for High Level Behavioural Planning: A Case Study of Autonomous Highway Lane Change Scenario
Kamran Turkoglu, Jhanani Selvakumar · AIAA Scitech 2020 Forum · 2020
In this paper, we address the problem of high-level path-planning, for the application of an autonomous vehicle on a highway, using a chance-constraint based rapidly exploring random tree (CC-RRT*) methodology. Random exploration tree search quickly explores the free space available to the vehicle to generate feasible and/or optimal paths from one point to another. The inclusion of chance constraints allows us to effectively handle the uncertainty that arises from the behavior of other agents on the highway, as well as the noise in the measurements and process noise associated with our vehicle model. The novel contribution of this work is the adaptation of the constraints of high way driving (such as kinodynamic model, collision avoidance, and routing) to the framework of chance constraints and point-mass sample-based tree search. A succinct description of our problem formulation and solution method are presented along with some numerical simulations to illustrate the efficacy and performance of the proposed algorithm for the application of highway driving.