Long-Horizon Prediction and Uncertainty Propagation with Residual Point Contact Learners
This work addresses the challenge of accurate contact simulation for robotic tasks, but it appears incremental as it builds on existing simulator frameworks with residual corrections.
The paper tackles the problem of discrepancy between simulator predictions and real-world observations in robotics by proposing a self-supervised approach to learn residual models for rigid-body simulators, which improves predictive performance and uncertainty propagation, as evidenced by empirical evaluation on planar dice rolls.
The ability to simulate and predict the outcome of contacts is paramount to the successful execution of many robotic tasks. Simulators are powerful tools for the design of robots and their behaviors, yet the discrepancy between their predictions and observed data limit their usability. In this paper, we propose a self-supervised approach to learning residual models for rigid-body simulators that exploits corrections of contact models to refine predictive performance and propagate uncertainty. We empirically evaluate the framework by predicting the outcomes of planar dice rolls and compare it's performance to state-of-the-art techniques.