Integrating Trajectory Optimization and Reinforcement Learning for Quadrupedal Jumping with Terrain-Adaptive Landing

2025-09-16

AI Summary

To address insufficient robustness in quadrupedal jumping and landing on rough, unstructured terrain, this paper proposes a safety-oriented landing framework integrating trajectory optimization and reinforcement learning (RL). Methodologically, trajectory optimization generates dynamic takeoff and landing reference motions, while an RL policy—trained with a novel reward relaxation mechanism—enables compliant, adaptive landing control. The reward relaxation explicitly encodes terrain uncertainty, facilitating efficient exploration and recovery under environmental variability. Our key contribution is the first application of reward relaxation to quadrupedal jumping-landing control, enabling end-to-end co-optimization of motion planning and low-level control. Experiments demonstrate substantial improvements in tracking accuracy and landing safety across diverse rugged terrains: landing success rate increases by 32% over baseline methods, and the policy exhibits strong generalization to unseen terrain configurations and disturbances.

Technology Category

• Intelligent Robots: Learning & Optimization for ROB
• Search and Optimization: Learning to Search
• Planning, Routing, and Scheduling: Replanning and Plan Repair

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

Jumping constitutes an essential component of quadruped robots' locomotion capabilities, which includes dynamic take-off and adaptive landing. Existing quadrupedal jumping studies mainly focused on the stance and flight phase by assuming a flat landing ground, which is impractical in many real world cases. This work proposes a safe landing framework that achieves adaptive landing on rough terrains by combining Trajectory Optimization (TO) and Reinforcement Learning (RL) together. The RL agent learns to track the reference motion generated by TO in the environments with rough terrains. To enable the learning of compliant landing skills on challenging terrains, a reward relaxation strategy is synthesized to encourage exploration during landing recovery period. Extensive experiments validate the accurate tracking and safe landing skills benefiting from our proposed method in various scenarios.

Problem

• Achieving adaptive landing on rough terrains for quadruped robots
• Combining trajectory optimization and reinforcement learning for jumping
• Enabling compliant landing skills through reward relaxation strategy

Innovation

• Combining trajectory optimization and reinforcement learning
• Reward relaxation strategy for compliant landing
• Safe landing framework for rough terrains