The increasing number of satellites for specific space missions is hindering efficient control from the ground stations. In addition, the risk of failing or deteriorating elements in a multi-satellite system (MSS) leads to the uncertainty in its operating environment which challenges traditional centralized planning algorithms. Centralized algorithms depend on one coordinator to overview entire system, which is a single poitn of failure, so if it malfunctional, the MSS may go down. To overcome this problem, a distributed onboard mission planning algorithm based on the Hybrid Dynamic Mutation Genetic Algorithm (HDMGA) is proposed for such MSS. The satellites are considered as autonomous agents which can communicate and negotiate with other satellites. The mission objective is decomposed through the team negotiation procedures, and the local goals are dispatched to the participant satellite. The local planning procedure on each satellite is modeled as local search problems, while the planning problem for entire team is formulated as strongly-coupled distributed optimization problem. The verification shows that the proposed distributed approach can achieve the same results as the centralized approach. The simulation results illustrate that the proposed approach shows superior performance on both computation time and success rate, as compared with other two state-of-the-art distributed mission planning approaches.
- Distributed planning approach
- Genetic algorithm
- Multi-satellite system