Learning Locomotion on Discrete Terrain via Minimal Proximity Sensing

2026-06-30

AI Summary

This work addresses the limitations of quadrupedal locomotion over unstructured discrete terrains—such as stairs and gaps—where conventional approaches suffer from delayed, occluded, or computationally expensive visual perception and reactive-only proprioception. To overcome these challenges, the authors propose a minimalist near-field sensing strategy that embeds low-cost, high-frequency infrared proximity sensors in the robot’s feet to provide pre-contact feedback. This sensory input is integrated into a reinforcement learning–based control framework, enabling real-time gait adaptation with minimal computational overhead. The approach requires only sparse sensor placement yet effectively mitigates occlusion and state estimation drift, achieving robust, low-latency, and energy-efficient terrain negotiation. Sim-to-real experiments demonstrate a significant improvement in traversal robustness over discrete obstacles, offering a lightweight perceptual paradigm for agile mobility in complex environments.

Abstract

Learning-based control has revolutionized dynamic locomotion, yet navigating unstructured terrain remains limited by a robot's incomplete awareness of imminent ground contact. While global perception systems such as LiDARs and depth cameras provide environmental context, they are frequently plagued by latencies, occlusions, and the high computational cost of dense geometric reconstruction. On the other hand, proprioceptive feedback is purely reactive, initiating corrections only after impact has occurred. This work explores embedding a minimal suite of low-cost, high-frequency infrared proximity sensors directly into the feet of a quadrupedal robot. These sensors provide "pre-contact" feedback that is robust to self-occlusions and significantly less computationally demanding than conventional vision-based pipelines. By integrating these localized signals into a reinforcement learning framework, we enable the robot to anticipate terrain discontinuities such as gaps and stepping stones that are problematic for traditional perception stacks due to occlusions or state estimation drift. We demonstrate that such sparse, near-field sensing can be reliably modeled in simulation and transferred to the real world with high fidelity. Experimental results show that local proximity sensing substantially improves traversal robustness over discrete terrain and offers a low-power, low-latency alternative or complement to complex global perception suites in unpredictable environments. For more information about results and methods, please see the project website: https://sites.google.com/view/foot-tof/home.

Problem

• locomotion
• discrete terrain
• proximity sensing
• perception latency
• terrain awareness

Innovation

• proximity sensing
• reinforcement learning
• quadrupedal locomotion
• pre-contact perception
• sim-to-real transfer