Use of Genetic Algorithms for Collision Avoidance Manoeuvre in space RVD missions

The following thesis aims at assessing Collision Avoidance Manoeuvrer (CAM) strategies during the rendezvous and docking phase between two satellites of different sizes. This study focuses on analysing the final phase of the SROC mission, particularly the last 30 meters of the rendezvous and docking manoeuvrer with the Spacerider vehicle, where the SROC satellite must be able to escape from the target trajectory and recover the initial position of the docking in less than 24 hours. The analysis identifies a set of manoeuvres that, based on the initial distance, meets both requirements while minimizing the manoeuvrer cost and avoiding the Spacerider’s Keep Out Zone during the desired period. For optimal solution search, genetic algorithms available in Matlab software were employed, allowing for random exploration within the system’s domain. Through the use of populations, generations, genes, and mutations, the algorithm refines the search, providing a starting point for further optimization processes. Although the random nature of the solutions does not guarantee finding the global minimum, with proper adjustments, a minimum can be obtained that satisfies the conditions set by the genetic algorithm’s fitness function. A large set of possible manoeuvres is assessed through a simplified model. Once an initial analysis was completed, the manoeuvrer that ensured the greatest minimum distance between the chaser and the target was selected. Subsequently, the search was refined to optimize this solution, and a robustness analysis was performed using specific Monte Carlo simulations on the obtained solutions to validate the results. This procedure was also repeated with a more accurate propagation model, yielding results that better align with a real-world model. The result of the thesis provide a look up table of the CAM that can be safely performed during the docking phase of the SROC mission.