RACER: Residual-Adaptive Closed-Loop Estimation for Sampling-Based Planning in Wheeled-Quadruped Racing

2026-10-05

AI Summary

This study addresses the decoupling of planning and control, as well as the sim-to-real transfer challenges, in high-speed racing for wheeled quadrupedal robots by proposing a hierarchical control framework. The upper layer employs a Model Predictive Path Integral (MPPI) planner integrated with a neural residual dynamics model to achieve efficient motion planning, while the lower layer utilizes a reinforcement learning-based velocity tracker for closed-loop control. Furthermore, a Low-Rank Residual Adaptation (LoRRA) method is introduced to bridge the domain gap through large-scale simulation pre-training followed by few-shot fine-tuning on real-world data. Experimental results demonstrate that the proposed approach significantly improves the success rate of high-speed cornering under domain discrepancies and enhances overall racing performance.

Abstract

We present RACER, a hierarchical control framework for wheel-based quadruped racing that combines an MPPI planner with a learned residual dynamics model and a low-level RL velocity tracker. The planner augments a nominal unicycle kinematic model with a neural residual term to capture the closed-loop tracking behavior of the RL policy. To train this residual model under limited real-world data, we propose Low-Rank Residual Adaptation (LoRRA), a two-stage approach that pre-trains on large-scale simulation data for broad coverage and then fine-tunes on a small real-world dataset with a low-rank constraint. In simulation, we empirically validate our engineering choices by showing (A) Residual dynamics improve the overall performance of our pipeline by capturing the tracking error of RL velocity tracker at high-speed cornering. (B) Residual dynamics trained with both source-domain and target-domain data gives racing performance significantly better than the residual dynamics trained with only target-domain data. (C) Low-rank constraint at target-domain adaptation gives higher success rates and higher performance than full-tune and from-scratch when domain gap in ground coefficient or joint gain increases.

Problem

• wheeled-quadruped racing
• sampling-based planning
• residual dynamics
• sim-to-real gap
• limited real-world data

Innovation

• Residual Dynamics
• Low-Rank Residual Adaptation (LoRRA)
• MPPI Planner
• Reinforcement Learning
• Sim-to-Real Transfer