The Use of Genetic Algorithms for Novel Geostationary Satellite Collocation Trajectories

As a direct research assistant to PhD Rahul Rughani at the USC Space Engineering Research Center, I have been working to adapt his original RPO free trajectory algorithms for LEO swarms to the GEO stationary collocated satellites.

In MATLAB I wrote and tested functions that introduced specific Geostationary orbital perturbations into the motion pieces of the original code, and have been adapting the code to run parallel for loops. With these new adaptations, testing on the USC supercomputer for swarms of satellites greater than 12 members will be conducted and results published in summer 2020.

The use of genetic algorithms allows convergence to an optimal set or family of trajectories for each member of the swarm regardless of the size of the swarm or mission functions required, should a solution exist. These trajectories will be such that each spacecraft can perform their required individual actions while minimizing the fuel required for maneuvering and also avoiding conjunctions, to a prescribed probability of collision, for a given amount of time. Genetic algorithms have been used previously for optimization of low thrust orbit transfers, drone delivery networks, and control of self-driving cars, among other applications.