CoCo-InEKF: State Estimation with Learned Contact Covariances in Dynamic, Contact-Rich Scenarios

2026-05-14

AI Summary

This work addresses the limited robustness of state estimation for legged robots in highly dynamic, multi-contact scenarios, where traditional binary contact models fail to capture partial contacts or directional slipping. To overcome this, the authors propose CoCo-InEKF, a method that replaces discrete contact assumptions with continuous contact-velocity covariances predicted end-to-end by a lightweight neural network. This enables dynamic adjustment of contact confidence and is integrated into a differentiable invariant extended Kalman filter (InEKF) for joint optimization. Notably, the approach requires no ground-truth contact labels and employs an automated contact candidate selection strategy insensitive to precise foot placement. Experiments demonstrate that CoCo-InEKF significantly improves accuracy, efficiency, and filter consistency of linear velocity estimation in challenging tasks such as bipedal dancing and complex terrain interaction, exhibiting robust performance in both simulation and real-world environments.

Abstract

Robust state estimation for highly dynamic motion of legged robots remains challenging, especially in dynamic, contact-rich scenarios. Traditional approaches often rely on binary contact states that fail to capture the nuances of partial contact or directional slippage. This paper presents CoCo-InEKF, a differentiable invariant extended Kalman filter that utilizes continuous contact velocity covariances instead of binary contact states. These learned covariances allow the method to dynamically modulate contact confidence, accounting for more nuanced conditions ranging from firm contact to directional slippage or no contact. To predict these covariances for a set of predefined contact candidate points, we employ a lightweight neural network trained end-to-end using a state-error loss. This approach eliminates the need for heuristic ground-truth contact labels. In addition, we propose an automated contact candidate selection procedure and demonstrate that our method is insensitive to their exact placement. Experiments on a bipedal robot demonstrate a superior accuracy-efficiency tradeoff for linear velocity estimation, as well as improved filter consistency compared to baseline methods. This enables the robust execution of challenging motions, including dancing and complex ground interactions -- both in simulation and in the real world.

Problem

• state estimation
• legged robots
• contact-rich scenarios
• dynamic motion
• contact uncertainty

Innovation

• continuous contact covariance
• differentiable InEKF
• learned contact model
• state estimation
• legged robots