Learning Structured Locomotion References for Optimization-Based Whole-Body Control
We present a hierarchical controller in which reinforcement learning adapts these structured locomotion references while a gait scheduler, reactive footstep planner, and whole-body impulse controller retain responsibility for high-frequency execution.
ProofPaper ↗
Key points
- Quadruped locomotion requires coordinating body motion,contact forces, foothold placement, and gait timing.
- We train the complete hierarchy in a vectorized controller-in-the-loop environment using a six-direction disturbance curriculum.
- Training-time channel ablations identify adaptive foothold placement as the largest marginal contributor to grid survival and reveal direction-dependent effects of gait-rate adaptation.
- Together, these results demonstrate the viability of the proposed hierarchical controller architecture for integrating learned adaptation with optimization-based whole-body control through structured locomotion references.
Sources (1)
- [1]Learning Structured Locomotion References for Optimization-Based Whole-Body ControlarXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 17, 04:47 PM
We present a hierarchical controller in which reinforcement learning adapts these structured locomotion references while a gait scheduler, reactive footstep planner, and whole-body impulse controller retain responsibility for high-frequency execution.
Quadruped locomotion requires coordinating body motion,contact forces, foothold placement, and gait timing.
Extractive summary: sentences quoted from the sources.
Before this
- Oct 17, 2026ProbePO: Sparse Probes, Redistributed Credit for Critic-Free LLM Reinforcement Learning
- Oct 16, 2026Not All Actions Contribute Equally: Learning Precise Progress Estimation from Imprecise Demonstrations
- Oct 16, 2026Learning Robot Metareasoning: Adaptive Computation Budgets for Real-Time Control
- Oct 16, 2026RL-Guided Temporal Localization for Dual-Channel Retrieval in Long-Horizon Agent Memory
- Oct 16, 2026Reasoning Reranking with Semantic IDs as the Sole Item Representation
- Oct 15, 2026TIPS: A Tiny Inference-time Policy for Scaling in Diffusion Language Models