Driving With Single Deformable Polyline Tentacle and Velocity Obstacles

Abstract

Intelligent transportation systems are emerging as the approach to dramatically improve efficiency while at the same time leading to the goal of zero accidents. Autonomous driving is now possible but still greatly limited to low speed driving. This is mainly due to computational load in detecting the road and objects, path planning and limited controller cycles. The focus of this work is to develop faster trajectory planning scheme given the sensed environment map (occupancy grid). This work utilizes interaction of single projected trajectory from the non-holonomic vehicle kinematic model with the objects in the road. Furthermore, the planned trajectory is made sensitive to car speed. Simulation results with non-linear vehicle kinematic model shows that the proposed combination of single trajectory and velocity objects gives faster satisfactory trajectory with safe obstacle avoidance while following prescribed way points.

Key words: Deformable poly-line tentacle, Velocity objects, trajectory planning and non-holonomic constraints.