An Obstacle avoidance technique of small UAV (Unmanned Aerial Vehicle) using PSO (Particle Swarm Optimization)

Abstract

We discuss a new obstacle avoidance technique for a UAV (Unmanned Aerial Vehicle) using PSO (Particle Swarm Optimization) called optimization imitated animal foraging. In this method, the avoidance orbit is generated by the response by step change of the target position in the left and right direction and the response by the step change of the target speed in the forward direction. The magnitude of the change of the two target values is determined by the PSO based on the unique evaluation function. The evaluation function is a function consisting of collision risk based on the closest distance to the obstacle, energy based on magnitude of step change of the target value to be applied, and avoidance time based on speed change in the forward direction. Here, the calculation of the closest approach distance is obtained by performing a simulation based on the mathematical model expressing the flight trajectory of the obstacle and the own vehicle each time the evaluation function calculation of the repetitive calculation in the optimization calculation of the PSO is performed. We confirmed that this technique is effective by the numerical simulation.

Keywords

• pso
• uav
• drone
• obstacle avoidance
• optimization
• simulation