A predictive motion planner for guidance of autonomous UAV systems

Abstract

This paper investigates Unmanned Aerial Vehicle (UAV) systems motion planning for ground attack missions involving enemy defenses. The UAV dynamics are modeled as a unicycle, linearized using dynamic extension and expanded over a finite prediction horizon as a piece-wise affine function. The motion planning problem is then formulated as a constrained, convex minimization in the form of Linear Quadratic Model Predictive Control (LQMPC). Avoidance of enemy defenses is achieved using linear inequality constraints. The design is tested in a simulated ground attack mission involving a layered enemy defense system using MATLAB. Preliminary results demonstrate the feasibility of using LQMPC to guide a UAV in ground attack missions involving complex enemy defenses.