Collision Avoidance Algorithm for a Fixed-Wing UAV Utilizing Geometric Estimation Considering Multiple Dynamic Obstacles

Navigating through environments replete with obstacles while adhering to dynamic constraints poses a formidable challenge for fixed-wing Unmanned Aerial Vehicles (UAVs), as they remain vulnerable to collision threats. Moreover, in scenarios involving multiple dynamic obstacles, a quick and reactive maneuver is needed. To address this challenge, this paper introduces a collision avoidance algorithm tailored for fixed-wing UAVs operating within dynamic constraints amidst a multi-obstacle environment. Leveraging a geometric approach with low computational load, this algorithm offers a pragmatic solution for ensuring UAV safety and maneuverability in complex airspace scenarios. The algorithm uses a discrete Kalman filter to predict obstacle trajectories and generates a two-dimensional plane containing the minimum of the obstacle trajectory. A target point for collision avoidance is computed according to the resulting two-dimensional plane considering the movement trajectory of the obstacle and dynamic constraints of the unmanned aerial vehicle. Moreover, the collision times are predicted for scenarios with multiple obstacles. Numerical simulations are performed to demonstrate the effectiveness of the proposed algorithm.