Guidance and control of an unmanned surface vehicle exhibiting sternward motion

We present algorithms for planning dynamically feasible backward paths and controlling unmanned surface vehicles along such paths using a nonlinear backstepping trajectory tracking controller. These algorithms have been developed for Virginia Tech's riverine unmanned surface vehicle (USV). The USV is a rigid hull inflatable boat with a servo-actuated outboard motor. The backward path planning algorithm uses the receding horizon approach and solves an optimal control problem by converting it into a nonlinear programming (NLP) problem. The control law is a modification of the forward nonlinear backstepping control law, augmented to compensate for the bistable sternward behavior exhibited by the riverine USV. Specifically, the bistable sternward behavior is modeled as two linear, stable left and right turnrate regions separated by an unstable zero turnrate region. Experimental results indicate that this backstepping controller adequately compensates for model and environmental disturbances.