Particle-Based Conformal Prediction for Contact-Aware Uncertainty Calibration in Stratified Configuration Spaces
2026-08-10 • Robotics
RoboticsMachine Learning
AI summaryⓘ
The authors developed a method called CaPTURe to help robots better predict where they will be in the future, especially when they touch or interact with obstacles. Their approach uses past data to adjust uncertainty estimates and create prediction regions that reliably include the robot's next position with a certain confidence. This is important because robot motion can be unpredictable, especially with contacts, which can make their possible future positions more complex. They tested their method on tasks like navigating a marble through a maze and fitting a peg into a hole, showing it worked better than other methods and improved success rates.
uncertainty representationrobot motion planningcontact-rich manipulationconformal predictionparticle-based modelscalibration datasetmultimodal distributionrobot configuration spacetrans-dimensional uncertaintypeg-in-hole insertion
Authors
Luís Marques, Kristian Popov, Dmitry Berenson
Abstract
Reliable uncertainty representation is essential for deploying autonomous systems that interact with their environment, as robots must reason about how uncertainty arising from both stochasticity and model mismatch is impacted by contacts with obstacles (e.g., when navigating through a cluttered environment or inserting a part into an assembly). We propose Calibrated Particle-sets for Trans-dimensional Uncertainty Representation (CaPTURe), a geometry-aware, conformal prediction-based algorithm that generates probabilistically valid prediction regions of the unknown future system configuration using particle-based models of arbitrary fidelity. While calibrated uncertainty predictions are essential for safe and efficient planning, analytical or learned motion models are often inaccurate - due to limited data, simplifying assumptions, unmodeled effects, etc. - which can lead to unsafe executions or task failure. Additionally, when a robot contacts an obstacle, the distribution of its future configurations can become multimodal or disjoint, or lie along manifolds of lower intrinsic dimension than the space of possible robot configurations. Our method uses a calibration dataset of system transitions to locally calibrate motion uncertainty estimates, constructing regions guaranteed to contain the future robot configuration at a user-set probability. Our calibration procedure captures how motion uncertainty varies between contact-rich and contactless motions, leading to sufficient coverage in both cases. We evaluate our method on two simulated planning tasks: controlling a marble around a labyrinth and performing tight-tolerance peg-in-hole insertion with a manipulator. Compared to relevant baselines, CaPTURe achieves the user-specified coverage requirement both in and out of contact and achieves up to a 30% absolute improvement in task success rate over the best baseline.