MPC-Based mid-level collision avoidance for ASVs using nonlinear programming

Abstract

In this paper, we present a mid-level collision avoidance algorithm for autonomous surface vehicles (ASVs) based on model predictive control (MPC) using nonlinear programming. The algorithm enables avoidance of both static and moving obstacles, and following of a desired nominal trajectory if there is no danger of collision. We compare two alternative objective functions, where one is a quadratic function and the other is a nonlinear function designed to produce maneuvers observable for other vessels in compliance with rule 8 of the International Regulations for Preventing Collisions at Sea (COLREGS). The algorithm is implemented in the CASADI framework and uses the IPOPT solver. The performance of the algorithm is evaluated through simulations which show promising results. Furthermore, the algorithm is considered computationally feasible to run in real time. This algorithm serves as a base algorithm for further development in order to ensure full COLREGS compliance.