TRACE: Learned Proprioceptive Odometry for Legged Robots under Unreliable Contact Conditions

2026-08-06

AI Summary

This work addresses the challenge of drift and low accuracy in proprioceptive odometry for legged robots under unreliable foot contact conditions by proposing TRACE, an end-to-end learning-based approach. TRACE leverages a Transformer architecture to model temporal sequences of inertial and joint data and introduces a novel foot-aware cross-attention mechanism that adaptively fuses multi-source information without requiring predefined contact or slip thresholds. Robustness and sim-to-real transferability are further enhanced through physics-inspired auxiliary losses, policy randomization in simulation, and fine-tuning on real hardware. Experimental results demonstrate that TRACE significantly reduces pose drift across diverse indoor and outdoor terrains, outperforming conventional filtering, hybrid, and purely learning-based baselines. Ablation studies confirm the contribution of each proposed component to the overall performance.

Abstract

In this paper, we present TRACE (Tokenized Robust Attention for Contact-Aware Estimation), an end-to-end learned proprioceptive odometry estimator for legged robots under unreliable contact conditions. The proposed estimator directly predicts relative displacement, relative rotation, and body-frame velocity from a recent history of onboard inertial and joint measurements. To improve robustness under unreliable contact conditions, we introduce a foot-aware cross-attention module that adaptively weights IMU and leg-wise kinematic tokens without relying on manually defined contact or slip thresholds. The estimator is trained with direct supervision and two physics-inspired auxiliary losses that promote kinematic consistency and reliable use of leg information. To reduce policy-specific overfitting and consequently improve sim-to-real transfer, simulation training incorporates policy randomization, followed by partial real-world fine-tuning of the temporal encoder and prediction head. Experiments across diverse indoor and outdoor terrains demonstrate consistent reductions in position drift compared with classical filtering-based, hybrid, and purely learning-based baselines. Ablation studies further validate the contributions of the proposed training objectives, policy randomization, and real-world fine-tuning, particularly under unreliable contacts and sim-to-real mismatch.

Problem

• proprioceptive odometry
• legged robots
• unreliable contact
• sim-to-real transfer
• position drift

Innovation

• proprioceptive odometry
• cross-attention
• contact-aware estimation
• sim-to-real transfer
• legged robots