Spacecraft Swarm Orbital Formation Optimisation Using Evolutionary Techniques

Robot swarms have already demonstrated their ability to accomplish complex missions collectively while solely relying on individual behaviours. The inherent resilience and scalability properties of such systems make them highly attractive for space applications, and thus sees a growing interest, including from the NASA and ESA. In this article we propose the Orbital Formation Algorithm (OFA) to address formations of autonomous spacecraft with the aim of surveying asteroids. The objective is to spread evenly the swarm members around the orbiting body to maximise its coverage, while minimising propellant consumption. We introduce a set of parameters for the swarm formation problem and optimise them using a genetic algorithm to obtain general optimal solutions for each case study, comprising swarms of 2, 3, 5, and 10 satellites. Moreover, we propose an alternative method to obtain optimal parameters individually for each satellite using a light-weight simulated annealing algorithm to be run onboard. Results from simulations show that the OFA performs well on the 400 scenarios analysed and that the onboard optimisation approach is more accurate than the general solutions, although it uses computing resources from each satellite.