Modelling of time-to collision for unmanned aerial vehicle using particles swarm optimization

Abstract

A method for the development of Time-to-Collision (TTC) mathematical model for outdoor Unmanned Aerial Vehicle (UAV) using Particles Swarm Optimization (PSO), are presented. TTC is the time required for a UAV either to collide with any static obstacle or completely stop without applying any braking control system when the throttle is fully released. This model provides predictions of time before UAV will collide with the obstacle in the same path based on their parameter, for instance, current speed and payload. However, this paper focus on the methodology of the implementation of PSO to develop the TTC model for 5 different set of payloads. This work utilizes a quadcopter as our testbed system that equipped with a Global Positioning System (GPS) receiver unit, a flight controller with data recording capability and ground control station for real-time monitoring. The recorded onboard flight mission data for 5 different set of payloads has been analyzed to develop a mathematical model of TTC through the PSO approach. The horizontal ground speed, throttle magnitudes and flight time stamp are extracted from the on-board quadcopter flight mission. PSO algorithm is used to find the optimal linear TTC model function, while the mean square error is used to evaluate the best fitness of the solution. The results of the TTC mathematical model for each payload are described.