aiqbench analysis
This research introduces PGTT, a deep reinforcement learning framework that enhances legged robot locomotion by using terrain perception to guide gait phases without relying on restrictive kinematic priors. By encoding swing phase as cubic Hermite splines and adapting height based on local LiDAR data, the method achieves superior robustness against disturbances across different robot morphologies like Unitree Go2 and ANYmal-C. The approach demonstrates that reducing inductive bias through perception-aware reward shaping allows for more adaptable policies without extensive re-tuning for new platforms.
PGTT: Phase-Guided Terrain Traversal for Perceptive Legged Locomotion
arXiv:2510.18348v2 Announce Type: replace-cross Abstract: State-of-the-art perceptive Reinforcement Learning controllers for legged robots typically either (i) impose oscillator-or IK-based gait priors that constrain the action space, bias policy optimization, and limit adaptability across robot morphologies, or (ii) operate "blind," making them unable to anticipate hind-leg terrain and brittle to observation noise. We propose Phase-Guided Terrain Traversal (PGTT), a perception-aware deep-RL approach that enforces gait structure through reward shaping, thereby reducing inductive bias compared to oscillator- or IK-conditioned action priors. PGTT encodes per-leg phase as a cubic Hermite spline, adapts swing height to local heightmap statistics, and adds a swing-phase contact penalty, while the policy acts directly in joint space for morphology-agnostic deployment. Trained in MuJoCo (MJX) on procedurally generated stair-like terrains with curriculum learning and domain randomization, PGTT achieves the highest success rate among the evaluated baselines under push disturbances (median +7.5% over the next-best baseline) and on discrete obstacles (+9%), while maintaining comparable velocity tracking. We validate PGTT on a Unitree Go2 using a real-time LiDAR elevation-to-heightmap pipeline and report preliminary results on ANYmal-C using the same hyperparameters. These results provide early evidence that terrain-adaptive, phase-guided reward shaping can transfer across platforms without platform-specific policy priors or extensive re-tuning.
Abstract:State-of-the-art perceptive Reinforcement Learning controllers for legged robots typically either (i) impose oscillator-or IK-based gait priors that constrain the action space, bias policy optimization, and limit adaptability across robot morphologies, or (ii) operate "blind," making them unable to anticipate hind-leg terrain and brittle to observation noise. We propose Phase-Guided Terrain Traversal (PGTT), a perception-aware deep-RL approach that enforces gait structure through reward shaping, thereby reducing inductive bias compared to oscillator- or IK-conditioned action priors. PGTT encodes per-leg phase as a cubic Hermite spline, adapts swing height to local heightmap statistics, and adds a swing-phase contact penalty, while the policy acts directly in joint space for morphology-agnostic deployment. Trained in MuJoCo (MJX) on procedurally generated stair-like terrains with curriculum learning and domain randomization, PGTT achieves the highest success rate among the evaluated baselines under push disturbances (median +7.5% over the next-best baseline) and on discrete obstacles (+9%), while maintaining comparable velocity tracking. We validate PGTT on a Unitree Go2 using a real-time LiDAR elevation-to-heightmap pipeline and report preliminary results on ANYmal-C using the same hyperparameters. These results provide early evidence that terrain-adaptive, phase-guided reward shaping can transfer across platforms without platform-specific policy priors or extensive re-tuning.
Robotics (cs.RO); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
arXiv:2510.18348 [cs.RO]
(or arXiv:2510.18348v2 [cs.RO] for this version)
https://doi.org/10.48550/arXiv.2510.18348 Focus to learn more arXiv-issued DOI via DataCite
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 292 words
Related coverage
SABER: Learning Attention-based Semantic Affordance for Legged Locomotion
arXiv · Sep 21, 2026
Generate, Track, Improve: Perceptive Multi-Skill Humanoid Locomotion with RL-Fine-Tuned Motion Generators
arXiv · Sep 28, 2026
TAGA: Terrain-aware Active Gaze Learning for Generalizable Agile Humanoid Locomotion
arXiv · Sep 28, 2026
TAGA: Terrain-aware Active Gaze Learning for Generalizable Agile Humanoid Locomotion
arXiv · Jun 5, 2026