Evolutionary path planning for unmanned aerial vehicles cooperation

Abstract

We suggest an evolutionary based off-line/on-line path planner for cooperating Unmanned Aerial Vehicles (UAVs) that takes into account the environment characteristics and the flight envelope and mission constraints of the cooperating UAVs. The scenario under consideration is the following: a number of UAVs are launched from the same or different known initial locations. The main issue is to produce 3-D trajectories that ensure a collision free operation with respect to mission constraints. The path planner produces curved routes that are represented by 3-D B-Spline curves. Two types of planner are discussed: The off-line planner generates collision free paths in environments with known characteristics and flight restrictions. The on-line planner, which is based on the off-line one, generates collision free paths in unknown static environments, by using acquired information from the UAV’s on-board sensors. This information is exchanged between the cooperating UAVs in order to maximize the knowledge of the environment. Both off-line and on-line path planning problems are formulated as optimization problems, with a Differential Evolution algorithm to serve as the optimizer.