Hybrid Collision Avoidance for Autonomous Surface Vehicles

Abstract

Autonomous vehicles and collision avoidance (COLAV) systems are advancing rapidly. However, the majority of the COLAV methods developed are not designed for vehicles with second-order nonholonomic constraints, such as autonomous surface vehicles (ASVs). This paper proposes the hybrid dynamic window (HDW) algorithm, which in addition to acting as a reactive COLAV method, functions as a trajectory tracker in a hybrid COLAV architecture. The algorithm serves as an interface to any deliberate COLAV method which generates time parameterized trajectories, enabling vehicles to avoid local minima and track optimal trajectories towards the goal. Furthermore, a new distance function is developed for the dynamic window (DW) algorithm, improving the algorithm trajectory planning when operating close to obstacles. A case study is done for an ASV model using the HDW algorithm and the rapidly-exploring random tree (RRT) algorithm, which together form a hybrid COLAV system. The performance is evaluated through simulations using a defined set of performance metrics.