A Maximum Independent Set Method for Scheduling Earth-Observing Satellite Constellations

Operating Earth-observing satellites requires efficient planning methods that coordinate activities of multiple spacecraft. The satellite task-planning problem entails selecting actions best satisfy mission objectives for autonomous execution. Task scheduling is often performed by human operators assisted heuristic or rule-based tools. This approach does not efficiently scale to assets as heuristics frequently fail properly vehicles over long horizons. Additionally, the becomes more difficult solve large constellations complexity scales exponentially in number requested observations and linearly It expected new commercial optical- radar-imaging will require automated meet stated responsiveness throughput objectives. paper introduces a solving satellite-scheduling generating an infeasibility-based graph representation finding maximal independent set vertices graph. tested on scenario up 10,000 imaging locations SkySat constellation optical well simulated 24 satellites. Performance compared with contemporary graph-traversal mixed-integer linear programming approaches. Empirical results demonstrate improvements solution time along scheduled collections beyond baseline methods. For problems, maximum able find feasible schedule 8% 75% less time.