Path generation for rendezvous of dissimilar UAVs using Particle Swarm Optimization of Dubin’s curve sets
Vincent Jouffroy, Xavier Bovier-Lapierre, Omar Kassim Ariff, Thomas Richer · 2016
The one dominant trend of development of small scale UAVs is toward multi-platform or multi-agent systems, where each agent is a single flying platform. Numerous intelligent algorithms for formation flight from literature focus on the positioning and relative motion of individual agents. The purpose of formation flight is numerous. It can be to refuel or repower each other, as well as to perform a mission by pooling resources. Prior to adopting close proximity, however, the agents have to be guided to rendezvous conditions such that these intelligent algorithms may be implemented. The guidance and navigation for optimal approach of multiple agents have to be examined from a point of view of arbitrary initial conditions, which are relative position and velocity vectors. This paper proposes a method to generate flyable trajectories for two agents in three dimensional Cartesian space. It is able to do this in real time by applying particle swarm optimisation procedure to a candidate set of Dubin’s curves. By simulating a series of critical test cases, it is shown that the curve generated is the optimal trajectory for rendezvous for the given initial and terminal conditions.