New AI Lets Satellites Outsmart Each Other in Space
The same math could someday stop self-driving cars from crashing into each other.
Researchers developed an algorithm called WOLF for nonconvex, game-theoretic motion planning problems subject to disturbances over long time horizons — posed as a partially-decoupled generalized Nash equilibrium problem where each agent's dynamics depend only on its own state and control.
WOLF applies receding-horizon model predictive control to an open-loop differential games solver based on sequential convexification. Unlike robust formulations that fix the uncertainty description offline, the robustness tube here is itself a dynamic state, co-optimized with the trajectory, with its thickness directly setting the tightening of the shared coupling constraints. The authors derive a sufficient condition under which a nominal trajectory satisfying constraints tightened against all agents' error bounds remains feasible for every admissible disturbance realization.
They demonstrate the method on two adversarial on-orbit games with coupled translational-attitude dynamics: a stealthy co-orbital jamming game under an active detection-probability bound, and a sun-blocking game in which an adversary disables an evader by decreasing solar power.
- WOLF helps several machines plan their moves at once, even when they're competing and conditions keep changing
- Researchers tested it on satellites, including one that blocks a rival's sunlight to drain its solar power
- It's simulation only so far — no real satellites, cars or drones were used
Why It Matters
This kind of planning is what self-driving cars and delivery drones need before they can safely share our roads and skies.