Implementation of the Speed Approach for UAV Collision Avoidance in Dynamic Environment

Abstract

Location of many UAV in a common aviation area increases collision probability. Since 2015 the number of papers related to the avoidance of such collisions has increased significantly. The most part of methods and algorithms require increasing computational capacity of on-board computer systems that is essential for detecting and tracking dynamic threats and implementing maneuvers of collision avoidance. Many researchers suggest the usage of automatic dependent surveillance system (ADS-B) or analogues to decrease computational costs for detecting and tracking dynamic threats. Computational costs can be also decreased via replacing numerical algorithms by analytical ones. In the paper a new method of selection and implementation of collision avoidance maneuvers is suggested. UAV trajectories are approximated by two or three trajectory points obtained from ADS-B system. Then, using analytical expressions, two cutoff values of the critical speed range, at which a collision is possible, are calculated. As calculation expressions for meeting points and cutoff values of the critical speed are represented in the analytical form, even if an on-board computer system has limited computational capacity, the time for calculation will be far less than the time of receiving data from ADS-B. For this reason, calculations can be updated at each cycle of new data receiving, and the trajectory approximation can be bounded by straight lines. Such an approach allows developing compact algorithm of collision avoidance, even for a significant number of UAVs (more than several dozens). Research adequacy is proved by modeling using a software system developed specifically for this purpose.