Failure Forecasting Boosts Robustness of Sim2Real Rhythmic Insertion Policies
This work addresses the problem of maintaining precision and consistency in robotic assembly tasks, which is incremental as it builds on existing sim-to-real and failure prediction methods.
The paper tackles the challenge of Rhythmic Insertion Tasks (RIT) for robots, such as screwing nuts into bolts, by proposing a sim-to-real framework that integrates a reinforcement learning policy with a failure forecasting module, achieving high success rates and robust performance over repetitive tasks.
This paper addresses the challenges of Rhythmic Insertion Tasks (RIT), where a robot must repeatedly perform high-precision insertions, such as screwing a nut into a bolt with a wrench. The inherent difficulty of RIT lies in achieving millimeter-level accuracy and maintaining consistent performance over multiple repetitions, particularly when factors like nut rotation and friction introduce additional complexity. We propose a sim-to-real framework that integrates a reinforcement learning-based insertion policy with a failure forecasting module. By representing the wrench's pose in the nut's coordinate frame rather than the robot's frame, our approach significantly enhances sim-to-real transferability. The insertion policy, trained in simulation, leverages real-time 6D pose tracking to execute precise alignment, insertion, and rotation maneuvers. Simultaneously, a neural network predicts potential execution failures, triggering a simple recovery mechanism that lifts the wrench and retries the insertion. Extensive experiments in both simulated and real-world environments demonstrate that our method not only achieves a high one-time success rate but also robustly maintains performance over long-horizon repetitive tasks.