IMM-MPC framework saves 59.8% of satellite fleets from lethal faults

Operating a fleet of remote robotic systems (like satellites) with intermittent communications requires careful scheduling of limited contact opportunities. Operators face a key ambiguity: when an asset fails to check in, it's impossible to distinguish between a lethal hardware fault and a benign communications failure. A team from Stanford (Schreiber, Eddy, Kochenderfer) presents Interacting Multiple Model Model Predictive Control (IMM-MPC), a receding-horizon framework that maintains a probabilistic belief over discrete fault modes with time-inhomogeneous dynamics. The framework optimizes a two-term objective coupling acquisition value with information gain, allowing the scheduler to prioritize correctly even when different failure modes produce identical observations.

Tested on satellite launch and early orbit communications scheduling, IMM-MPC dramatically outperforms existing methods: it recovers 59.8% of spacecraft experiencing lethal faults, compared to just 9.0% for binary-MPC and 2.0% for a bipartite graph-based formulation solved through matching. These results hold across 200 randomized trials, while maintaining identical acquisition of healthy satellites and near-identical solve times. The paper also characterizes when observationally aliased fault modes can be disambiguated through scheduled actions and when aliasing is permanently unresolvable, providing a theoretical foundation for future scheduling in space missions and other robotic fleet applications.