Real-Time Control Allocation for Autonomous Surface Vehicle Using Constrained Quadratic Programming

Abstract

Control system is very important for each autonomous surface vehicle (ASV), which involves the problem of maintaining the vehicle’s position and heading using feedback controller and achieving the desired forces through thrust allocation. In this paper, we present a practical thrust allocator for under-actuated and fully-actuated vehicles, which can be represented as a quadratic programming (QP) problem with linear constraints. Such an optimization method allows us to consider common propulsion system, including tunnel thruster, azimuth thruster, and fixed propeller with rudder. These linear constraints enable us to explicitly account for the rate, amplitude and azimuth constraints of each propeller on the vessel. The proposed methods have been illustrated by simulated and experimental maneuvers for different thruster layout of a vehicle.