How to Be a Team Player: Multi-Agent Autonomy for Space Exploration

Teams of autonomous robots and spacecraft can enable new space exploration missions, enhance the state-of-the-art (e.g. shortening the data collection period), or change the mission risk posture by introducing redundancy. Present-day monolithic systems (e.g. single spacecraft or rover) could be replaced by a team of interconnected and coordinating assets, a multi-agent system. These teams can increase the science return by cooperatively exploring an area of interest (e.g. rovers and helicopters) or making distributed measurements from different vantage points (e.g. smallsats). Multi-agent autonomy reduces the reliance on ground in the loop decision making and coordination among various assets and instills a level of intelligence in teams to autonomously and cooperatively achieve the mission goal. This talk presents examples of the multi-agent autonomy work currently under development at JPL including an autonomy architecture for multi-robot collaborative autonomy and dives a bit deeper into communication aware orbit optimization for collaborative exploration of a small body using a team of smallsats.