Fixed-Wing UAV Path Planning and Collision Avoidance using Nonlinear Model Predictive Control and Sensor-based Cloud Detection

This article proposes a Nonlinear Model Predictive Control (NMPC) method for path planning and obstacle avoidance of a fixed-wing Unmanned Aerial Vehicle (UAV) coupled to an online monocular electro-optical cloud detection. A Nonlinear Programming (NLP) problem is solved at each discrete-time iteration of the NMPC using an inhouse Sequential Quadratic Programming (SQP) method. The proposed NMPC method enables real-time path planning while avoiding clouds, whose distance and extent are determined using electro-optical sensors and computer vision methods. Cloud detection and avoidance is intended to allow UAVs to operate under visual flight rules (VFR) to reduce the risk of collision with other air traffic. The mission task here consists of airborne reconnaissance of targets that can be defined by the user and updated online. The user interacts with the simulation by setting reconnaissance targets for the UAV at anytime during the flight.