A novel approach for obstacle avoidance maneuverer controls to a 49ton autonomous drive heavy-duty truck in GCV

Abstract

This paper presents a novel obstacle avoidance maneuver for a 49-ton heavy duty truck in a narrow campus environment, which is surrounded by a narrow two-way street. In the beginning of the paper, the autonomous drive system implemented in the GCV (Geely Commercial Vehicle) project are introduced. The proposed planning algorithm combine a state machine and lateral sample method to achieve a balance between real-time functionality and controllability. To simplify the path planning process, a global path from routing module is used as a reference path for vehicle to follow if there’s no obstacle exists. Planning algorithm calculates a feasible local path based on the reference path and avoids obstacles on the reference path. Our test shows that our method is easier to calibration, quicker and safer comparing to an A star search method, which is widely used in autonomous robot domain.