Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
This study introduces privileged critic training to enable sensor-free fault adaptation in thruster-actuated robots, addressing the challenge of continuous and partially observable thruster failures. By allowing the PPO value function to access the true degradation state during training while the actor relies on standard observations, the method shapes a robust policy that effectively compensates for failures in real-time deployment. The proposed RAFT policy, evaluated on a floating-platform robot, achieved a 70.2% success rate under multiple thruster failures, significantly closing the gap to an oracle policy that has full state visibility.
Sensor-free fault adaptation for thruster-actuated robots can achieve 70.2% success even under multiple concurrent failures, leveraging privileged critic training.
Fault-tolerant navigation for thruster-actuated robots requires online adaptation to failures that are neither binary nor fully observable: thrusters may degrade continuously, fail dead, or jam stuck-open. Classical fault detection pipelines require dedicated sensors unavailable at deployment; oracle controllers that observe the true failure state are equally impractical. We show that privileged critic training is sufficient for sensor-free fault adaptation: giving the PPO value function access to the true degradation state dgt during training, while the actor receives only standard task observations, shapes a policy that compensates for failures at deployment without any dedicated fault sensing. We propose RAFT (Recurrent Asymmetric Fault Tolerant), a policy with recurrent memory trained with a privileged asymmetric critic. Evaluated on a floating-platform robot (8 thrusters, 1 reaction wheel) under up to four simultaneous thruster failures, RAFT achieves 70.2% success at four concurrent failures, closing 84% of the gap from a failure-naive baseline (4.8%) to an oracle policy that sees the full degradation state at deployment (82.4%). All code, checkpoints, and data are open-source.