arXiv subjects

Zhaoyuan Gu

Publications and source records attributed to Zhaoyuan Gu.

15 recordsLinked to original sources

arXiv2026

REFINE-DP: Diffusion Policy Fine-tuning for Humanoid Loco-manipulation via Reinforcement Learning

Humanoid loco-manipulation requires coordinated task-space motion planning with stable loco-manipulation command tracking under complex robot-environment dynamics and long-horizon tasks. While diffusion policies (DPs) show promise for learning from demonstrations, deploying them on humanoids poses critical challenges: the motion planner trained offline is decoupled from the loco-manipulation controller, leading to poor command tracking, compounding distribution shift, and task failures. The common approach of scaling demonstration data is prohibitively expensive for high-dimensional humanoid systems. To address this challenge, we present REFINE-DP (REinforcement learning FINE-tuning of Diffusion Policy), a hierarchical framework that jointly optimizes a DP motion planner and an RL-based loco-manipulation controller. The DP is fine-tuned via a PPO-based diffusion policy gradient to improve task success rate, while the controller is simultaneously updated to accurately track the planner's evolving command distribution, reducing the distributional mismatch that degrades motion quality. We validate REFINE-DP on a humanoid robot performing loco-manipulation tasks, including door traversal and long-horizon object transport. REFINE-DP achieves an over 90% success rate in simulation, even in out-of-distribution cases not seen in the pre-training data, and enables real-world execution without privileged state information. Our proposed method substantially outperforms pre-trained DP baselines and demonstrates that RL fine-tuning is key to reliable humanoid loco-manipulation. https://refine-dp.github.io/REFINE-DP/

cs.RO↗

arXiv2026

WT-UMI: Tactile-based Whole-Body Manipulation via Force-Supervised Contact-Aware Planning

Whole-body humanoid manipulation of bulky, deformable, and shared-load objects requires distributed contact sensing and explicit force regulation, yet most imitation policies treat contact force only implicitly. On the other hand, different demonstration sources provide complementary modalities with inherent trade-offs: human demonstrations capture natural contact forces but not robot-executable actions, while teleoperation directly records robot actions but with less natural force regulation. This paper presents \textbf{WT-UMI}, a wearable whole-body tactile interface worn by human operators or mounted on humanoids, providing accurate observations of tactile images, contact forces, and end-effector poses across both human demonstration and humanoid teleoperation modes. We introduce a force-conditioned target-pose correction module that converts measured human poses into contact-aware robot targets by learning corrections from teleoperation data. To leverage the natural force interaction in human data, we propose a force-supervised planner that predicts end-effector pose chunks and contact-force trajectories. The predicted contact force serves as the reference for a tactile-based admittance controller. Across five contact-rich tasks spanning deformable objects, bulky rigid objects, and human--humanoid collaboration, WT-UMI improves success rate and reduces contact-position tracking error over four policy baselines. Our project page is available at https://wt-umi.github.io/WTUMI/.

cs.RO↗

arXiv2025

Opt2Skill: Imitating Dynamically-feasible Whole-Body Trajectories for Versatile Humanoid Loco-Manipulation

Humanoid robots are designed to perform diverse loco-manipulation tasks. However, they face challenges due to their high-dimensional and unstable dynamics, as well as the complex contact-rich nature of the tasks. Model-based optimal control methods offer flexibility to define precise motion but are limited by high computational complexity and accurate contact sensing. On the other hand, reinforcement learning (RL) handles high-dimensional spaces with strong robustness but suffers from inefficient learning, unnatural motion, and sim-to-real gaps. To address these challenges, we introduce Opt2Skill, an end-to-end pipeline that combines model-based trajectory optimization with RL to achieve robust whole-body loco-manipulation. Opt2Skill generates dynamic feasible and contact-consistent reference motions for the Digit humanoid robot using differential dynamic programming (DDP) and trains RL policies to track these optimal trajectories. Our results demonstrate that Opt2Skill outperforms baselines that rely on human demonstrations and inverse kinematics-based references, both in motion tracking and task success rates. Furthermore, we show that incorporating trajectories with torque information improves contact force tracking in contact-involved tasks, such as wiping a table. We have successfully transferred our approach to real-world applications.

cs.RO↗

arXiv2025

PPF: Pre-training and Preservative Fine-tuning of Humanoid Locomotion via Model-Assumption-based Regularization

Humanoid locomotion is a challenging task due to its inherent complexity and high-dimensional dynamics, as well as the need to adapt to diverse and unpredictable environments. In this work, we introduce a novel learning framework for effectively training a humanoid locomotion policy that imitates the behavior of a model-based controller while extending its capabilities to handle more complex locomotion tasks, such as more challenging terrain and higher velocity commands. Our framework consists of three key components: pre-training through imitation of the model-based controller, fine-tuning via reinforcement learning, and model-assumption-based regularization (MAR) during fine-tuning. In particular, MAR aligns the policy with actions from the model-based controller only in states where the model assumption holds to prevent catastrophic forgetting. We evaluate the proposed framework through comprehensive simulation tests and hardware experiments on a full-size humanoid robot, Digit, demonstrating a forward speed of 1.5 m/s and robust locomotion across diverse terrains, including slippery, sloped, uneven, and sandy terrains.

cs.RO↗

arXiv2025

Learn to Teach: Sample-Efficient Privileged Learning for Humanoid Locomotion over Diverse Terrains

Humanoid robots promise transformative capabilities for industrial and service applications. While recent advances in Reinforcement Learning (RL) yield impressive results in locomotion, manipulation, and navigation, the proposed methods typically require enormous simulation samples to account for real-world variability. This work proposes a novel one-stage training framework-Learn to Teach (L2T)-which unifies teacher and student policy learning. Our approach recycles simulator samples and synchronizes the learning trajectories through shared dynamics, significantly reducing sample complexities and training time while achieving state-of-the-art performance. Furthermore, we validate the RL variant (L2T-RL) through extensive simulations and hardware tests on the Digit robot, demonstrating zero-shot sim-to-real transfer and robust performance over 12+ challenging terrains without depth estimation modules.

cs.RO↗

arXiv2025

Humanoid Locomotion and Manipulation: Current Progress and Challenges in Control, Planning, and Learning

Humanoid robots hold great potential to perform various human-level skills, involving unified locomotion and manipulation in real-world settings. Driven by advances in machine learning and the strength of existing model-based approaches, these capabilities have progressed rapidly, but often separately. This survey offers a comprehensive overview of the state-of-the-art in humanoid locomotion and manipulation (HLM), with a focus on control, planning, and learning methods. We first review the model-based methods that have been the backbone of humanoid robotics for the past three decades. We discuss contact planning, motion planning, and whole-body control, highlighting the trade-offs between model fidelity and computational efficiency. Then the focus is shifted to examine emerging learning-based methods, with an emphasis on reinforcement and imitation learning that enhance the robustness and versatility of loco-manipulation skills. Furthermore, we assess the potential of integrating foundation models with humanoid embodiments to enable the development of generalist humanoid agents. This survey also highlights the emerging role of tactile sensing, particularly whole-body tactile feedback, as a crucial modality for handling contact-rich interactions. Finally, we compare the strengths and limitations of model-based and learning-based paradigms from multiple perspectives, such as robustness, computational efficiency, versatility, and generalizability, and suggest potential solutions to existing challenges.

cs.RO↗

arXiv2024

Robust-Locomotion-by-Logic: Perturbation-Resilient Bipedal Locomotion via Signal Temporal Logic Guided Model Predictive Control

This study introduces a robust planning framework that utilizes a model predictive control (MPC) approach, enhanced by incorporating signal temporal logic (STL) specifications. This marks the first-ever study to apply STL-guided trajectory optimization for bipedal locomotion, specifically designed to handle both translational and orientational perturbations. Existing recovery strategies often struggle with reasoning complex task logic and evaluating locomotion robustness systematically, making them susceptible to failures caused by inappropriate recovery strategies or lack of robustness. To address these issues, we design an analytical stability metric for bipedal locomotion and quantify this metric using STL specifications, which guide the generation of recovery trajectories to achieve maximum robustness degree. To enable safe and computational-efficient crossed-leg maneuver, we design data-driven self-leg-collision constraints that are $1000$ times faster than the traditional inverse-kinematics-based approach. Our framework outperforms a state-of-the-art locomotion controller, a standard MPC without STL, and a linear-temporal-logic-based planner in a high-fidelity dynamic simulation, especially in scenarios involving crossed-leg maneuvers. Additionally, the Cassie bipedal robot achieves robust performance under horizontal and orientational perturbations such as those observed in ship motions. These environments are validated in simulations and deployed on hardware. Furthermore, our proposed method demonstrates versatility on stepping stones and terrain-agnostic features on inclined terrains.

cs.RO↗

arXiv2023

Signal Temporal Logic-Guided Model Predictive Control for Robust Bipedal Locomotion Resilient to Runtime External Perturbations

This study investigates formal-method-based trajectory optimization (TO) for bipedal locomotion, focusing on scenarios where the robot encounters external perturbations at unforeseen times. Our key research question centers around the assurance of task specification correctness and the maximization of specification robustness for a bipedal robot in the presence of external perturbations. Our contribution includes the design of an optimization-based task and motion planning framework that generates optimal control sequences with formal guarantees of external perturbation recovery. As a core component of the framework, a model predictive controller (MPC) encodes signal temporal logic (STL)-based task specifications as a cost function. In particular, we investigate challenging scenarios where the robot is subjected to lateral perturbations that increase the risk of failure due to leg self-collision. To address this, we synthesize agile and safe crossed-leg maneuvers to enhance locomotion stability. This work marks the first study to incorporate formal guarantees offered by STL into a TO for perturbation recovery of bipedal locomotion. We demonstrate the efficacy of the framework via perturbation experiments in simulations.

cs.RO↗

arXiv2023

Infer and Adapt: Bipedal Locomotion Reward Learning from Demonstrations via Inverse Reinforcement Learning

Enabling bipedal walking robots to learn how to maneuver over highly uneven, dynamically changing terrains is challenging due to the complexity of robot dynamics and interacted environments. Recent advancements in learning from demonstrations have shown promising results for robot learning in complex environments. While imitation learning of expert policies has been well-explored, the study of learning expert reward functions is largely under-explored in legged locomotion. This paper brings state-of-the-art Inverse Reinforcement Learning (IRL) techniques to solving bipedal locomotion problems over complex terrains. We propose algorithms for learning expert reward functions, and we subsequently analyze the learned functions. Through nonlinear function approximation, we uncover meaningful insights into the expert's locomotion strategies. Furthermore, we empirically demonstrate that training a bipedal locomotion policy with the inferred reward functions enhances its walking performance on unseen terrains, highlighting the adaptability offered by reward learning.

cs.RO↗

arXiv2023

Walking-by-Logic: Signal Temporal Logic-Guided Model Predictive Control for Bipedal Locomotion Resilient to External Perturbations

This study proposes a novel planning framework based on a model predictive control formulation that incorporates signal temporal logic (STL) specifications for task completion guarantees and robustness quantification. This marks the first-ever study to apply STL-guided trajectory optimization for bipedal locomotion push recovery, where the robot experiences unexpected disturbances. Existing recovery strategies often struggle with complex task logic reasoning and locomotion robustness evaluation, making them susceptible to failures caused by inappropriate recovery strategies or insufficient robustness. To address this issue, the STL-guided framework generates optimal and safe recovery trajectories that simultaneously satisfy the task specification and maximize the locomotion robustness. Our framework outperforms a state-of-the-art locomotion controller in a high-fidelity dynamic simulation, especially in scenarios involving crossed-leg maneuvers. Furthermore, it demonstrates versatility in tasks such as locomotion on stepping stones, where the robot must select from a set of disjointed footholds to maneuver successfully.

cs.RO↗

arXiv2023

Integrated Task and Motion Planning for Safe Legged Navigation in Partially Observable Environments

This study proposes a hierarchically integrated framework for safe task and motion planning (TAMP) of bipedal locomotion in a partially observable environment with dynamic obstacles and uneven terrain. The high-level task planner employs linear temporal logic (LTL) for a reactive game synthesis between the robot and its environment and provides a formal guarantee on navigation safety and task completion. To address environmental partial observability, a belief abstraction is employed at the high-level navigation planner to estimate the dynamic obstacles' location. Accordingly, a synthesized action planner sends a set of locomotion actions to the middle-level motion planner, while incorporating safe locomotion specifications extracted from safety theorems based on a reduced-order model (ROM) of the locomotion process. The motion planner employs the ROM to design safety criteria and a sampling algorithm to generate non-periodic motion plans that accurately track high-level actions. At the low level, a foot placement controller based on an angular-momentum linear inverted pendulum model is implemented and integrated with an ankle-actuated passivity-based controller for full-body trajectory tracking. To address external perturbations, this study also investigates safe sequential composition of the keyframe locomotion state and achieves robust transitions against external perturbations through reachability analysis. The overall TAMP framework is validated with extensive simulations and hardware experiments on bipedal walking robots Cassie and Digit designed by Agility Robotics.

eess.SY↗

arXiv2022

Reactive Locomotion Decision-Making and Robust Motion Planning for Real-Time Perturbation Recovery

In this paper, we examine the problem of push recovery for bipedal robot locomotion and present a reactive decision-making and robust planning framework for locomotion resilient to external perturbations. Rejecting perturbations is an essential capability of bipedal robots and has been widely studied in the locomotion literature. However, adversarial disturbances and aggressive turning can lead to negative lateral step width (i.e., crossed-leg scenarios) with unstable motions and self-collision risks. These motion planning problems are computationally difficult and have not been explored under a hierarchically integrated task and motion planning method. We explore a planning and decision-making framework that closely ties linear-temporal-logic-based reactive synthesis with trajectory optimization incorporating the robot's full-body dynamics, kinematics, and leg collision avoidance constraints. Between the high-level discrete symbolic decision-making and the low-level continuous motion planning, behavior trees serve as a reactive interface to handle perturbations occurring at any time of the locomotion process. Our experimental results show the efficacy of our method in generating resilient recovery behaviors in response to diverse perturbations from any direction with bounded magnitudes.

cs.RO↗

arXiv2019

Adversary A3C for Robust Reinforcement Learning

Asynchronous Advantage Actor Critic (A3C) is an effective Reinforcement Learning (RL) algorithm for a wide range of tasks, such as Atari games and robot control. The agent learns policies and value function through trial-and-error interactions with the environment until converging to an optimal policy. Robustness and stability are critical in RL; however, neural network can be vulnerable to noise from unexpected sources and is not likely to withstand very slight disturbances. We note that agents generated from mild environment using A3C are not able to handle challenging environments. Learning from adversarial examples, we proposed an algorithm called Adversary Robust A3C (AR-A3C) to improve the agent's performance under noisy environments. In this algorithm, an adversarial agent is introduced to the learning process to make it more robust against adversarial disturbances, thereby making it more adaptive to noisy environments. Both simulations and real-world experiments are carried out to illustrate the stability of the proposed algorithm. The AR-A3C algorithm outperforms A3C in both clean and noisy environments.

cs.LG↗

arXiv2019

Design and Implementation of a Three-Link Brachiation Robot with Optimal Control Based Trajectory Tracking Controller

This paper reports the design and implementation of a three-link brachiation robot. The robot is able to travel along horizontal monkey bars using continuous arm swings. We build a full order dynamics model for the robot and formulate each cycle of robot swing motion as an optimal control problem. The iterative Linear Quadratic Regulator (iLQR) algorithm is used to find the optimal control strategy during one swing. We select suitable robot design parameters by comparing the cost of robot motion generated by the iLQR algorithm for different robot designs. In particular, using this approach we show the importance of having a body link and low inertia arms for efficient brachiation. Further, we propose a trajectory tracking controller that combines a cascaded PID controller and an input-output linearization controller to enable the robot to track desired trajectory precisely and reject external disturbance during brachiation. Experiments on the simulated robot and the real robot demonstrate that the robot can robustly swing between monkey bars with same or different spacing of handholds.

eess.SY↗

arXiv2019

State Estimation for Legged Robots Using Contact-Centric Leg Odometry

Our goal is to send legged robots into challenging, unstructured terrains that wheeled systems cannot traverse. Moreover, precise estimation of the robot's position and orientation in rough terrain is especially difficult. To address this problem, we introduce a new state estimation algorithm which we term Contact-Centric Leg Odometry (COCLO). This new estimator uses a Square Root Unscented Kalman Filter (SR-UKF) to fuse multiple proprioceptive sensors available on a legged robot. In contrast to IMU-centric filtering approaches, COCLO formulates prediction and measurement models according to the contact status of legs. Additionally, COCLO has an indirect measurement model using joint velocities to estimate the robot body velocity. In rough terrain, when IMUs suffer from large amounts of noise, COCLO's contact-centric approach outperforms previous IMU-centric methods. To demonstrate improved state estimation accuracy, we compare COCLO with Visual Inertial Navigation System (VINS), a state-of-the-art visual inertial odometry, in three different environments: flat ground, ramps, and stairs. COCLO achieves better estimation precision than VINS in all three environments and is robust to unstable motion. Finally, we also show that COCLO and a modified VINS can work in tandem to improve each other's performance.

eess.SY↗