Failure Forecasting Boosts Robustness of Sim2Real Rhythmic Insertion Policies

TL;DR

This paper introduces a sim-to-real framework for high-precision robotic insertion tasks that combines reinforcement learning, failure prediction, and pose representation improvements to enhance robustness and transferability.

Contribution

It presents a novel sim-to-real approach integrating failure forecasting and pose representation in the nut's frame, improving insertion accuracy and robustness in repetitive tasks.

Findings

High success rate in simulated and real-world tests

Enhanced sim-to-real transferability through pose representation

Robust long-term performance in repetitive insertions

Abstract

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.