Lizhi Yang

Junheng Li, Liang Wu, Sergio A. Esteban et al.

In humanoid motion control, model predictive control (MPC) offers physically grounded prediction and constraint handling, while reinforcement learning (RL) enables robust whole-body skills through large-scale simulation. However, using MPC inside RL often requires time-consuming problem construction or excessive training overhead, making such frameworks difficult to justify in practice. This work studies efficient training-time MPC guidance for humanoid locomotion and manipulation, termed MPC-RL. We introduce a centroidal-dynamics MPC reward formulation that leverages guidance from MPC trajectories in training time. To make this practical in massively parallel RL, we develop $π^n$MPC, a parallel-in-horizon and construction-free batched GPU MPC solver that operates directly on time-varying dynamics to avoid high memory usage and pre-compilation. Through a variety of comparative studies and hardware validations, we have found that MPC-RL achieves superior performance in locomotion and manipulation skills. The code base is available at https://github.com/junhengl/mpc-rl.

Gilbert Feng, Hongbo Zhang, Zhongyu Li et al. · berkeley

Recent years have seen a surge in commercially-available and affordable quadrupedal robots, with many of these platforms being actively used in research and industry. As the availability of legged robots grows, so does the need for controllers that enable these robots to perform useful skills. However, most learning-based frameworks for controller development focus on training robot-specific controllers, a process that needs to be repeated for every new robot. In this work, we introduce a framework for training generalized locomotion (GenLoco) controllers for quadrupedal robots. Our framework synthesizes general-purpose locomotion controllers that can be deployed on a large variety of quadrupedal robots with similar morphologies. We present a simple but effective morphology randomization method that procedurally generates a diverse set of simulated robots for training. We show that by training a controller on this large set of simulated robots, our models acquire more general control strategies that can be directly transferred to novel simulated and real-world robots with diverse morphologies, which were not observed during training.

Xiaoyu Huang, Zhongyu Li, Yanzhen Xiang et al. · berkeley

We present a reinforcement learning (RL) framework that enables quadrupedal robots to perform soccer goalkeeping tasks in the real world. Soccer goalkeeping using quadrupeds is a challenging problem, that combines highly dynamic locomotion with precise and fast non-prehensile object (ball) manipulation. The robot needs to react to and intercept a potentially flying ball using dynamic locomotion maneuvers in a very short amount of time, usually less than one second. In this paper, we propose to address this problem using a hierarchical model-free RL framework. The first component of the framework contains multiple control policies for distinct locomotion skills, which can be used to cover different regions of the goal. Each control policy enables the robot to track random parametric end-effector trajectories while performing one specific locomotion skill, such as jump, dive, and sidestep. These skills are then utilized by the second part of the framework which is a high-level planner to determine a desired skill and end-effector trajectory in order to intercept a ball flying to different regions of the goal. We deploy the proposed framework on a Mini Cheetah quadrupedal robot and demonstrate the effectiveness of our framework for various agile interceptions of a fast-moving ball in the real world.

Chenyu Yang, Guo Ning Sue, Zhongyu Li et al. · berkeley

This paper tackles the problem of robots collaboratively towing a load with cables to a specified goal location while avoiding collisions in real time. The introduction of cables (as opposed to rigid links) enables the robotic team to travel through narrow spaces by changing its intrinsic dimensions through slack/taut switches of the cable. However, this is a challenging problem because of the hybrid mode switches and the dynamical coupling among multiple robots and the load. Previous attempts at addressing such a problem were performed offline and do not consider avoiding obstacles online. In this paper, we introduce a cascaded planning scheme with a parallelized centralized trajectory optimization that deals with hybrid mode switches. We additionally develop a set of decentralized planners per robot, which enables our approach to solve the problem of collaborative load manipulation online. We develop and demonstrate one of the first collaborative autonomy framework that is able to move a cable-towed load, which is too heavy to move by a single robot, through narrow spaces with real-time feedback and reactive planning in experiments.

Lizhi Yang, Zhongyu Li, Jun Zeng et al. · berkeley

In this paper, we propose a multi-domain control parameter learning framework that combines Bayesian Optimization (BO) and Hybrid Zero Dynamics (HZD) for locomotion control of bipedal robots. We leverage BO to learn the control parameters used in the HZD-based controller. The learning process is firstly deployed in simulation to optimize different control parameters for a large repertoire of gaits. Next, to tackle the discrepancy between the simulation and the real world, the learning process is applied on the physical robot to learn for corrections to the control parameters learned in simulation while also respecting a safety constraint for gait stability. This method empowers an efficient sim-to-real transition with a small number of samples in the real world, and does not require a valid controller to initialize the training in simulation. Our proposed learning framework is experimentally deployed and validated on a bipedal robot Cassie to perform versatile locomotion skills with improved performance on smoothness of walking gaits and reduction of steady-state tracking errors.

Lizhi Yang, Junheng Li, Nehar Poddar et al.

For a humanoid robot to be deployed in the real world, the choice of command space (i.e., the interface between task planning and whole-body control) is crucial. Existing whole-body controllers typically demand dense kinematic or spatial references that planners struggle to synthesize from task semantics. We instead propose a compact, explicit interface that is intuitive, general, modular, and expressive enough for diverse manipulation skills. To this end, we introduce HANDOFF, a single humanoid whole-body controller that follows this interface and is distilled via multi-teacher KL distillation under a context-conditioned gating scheme into a mixture-of-experts student from three complementary specialists: whole-body motion tracking with safety-filtered data, locomotion, and fall-recovery. On the Unitree G1, HANDOFF matches state-of-the-art velocity tracking and offers one of the largest robust manipulation workspaces. We further demonstrate hardware feasibility through multiple natural-language-driven task roll-outs, powered by a VLM-driven agentic planner with no task-specific data or controller fine-tuning.

Pol Mestres, Shima Sadat Mousavi, Pio Ong et al.

This paper studies the efficient implementation of safety filters that are designed using control barrier functions (CBFs), which minimally modify a nominal controller to render it safe with respect to a prescribed set of states. Although CBF-based safety filters are often implemented by solving a quadratic program (QP) in real time, the use of off-the-shelf solvers for such optimization problems poses a challenge in applications where control actions need to be computed efficiently at very high frequencies. In this paper, we introduce a closed-form expression for controllers obtained through CBF-based safety filters. This expression is obtained by partitioning the state-space into different regions, with a different closed-form solution in each region. We leverage this formula to introduce a resource-aware implementation of CBF-based safety filters that detects changes in the partition region and uses the closed-form expression between changes. We showcase the applicability of our approach in examples ranging from aerospace control to safe reinforcement learning.

Chong Zhang, Lizhi Yang

Terrain-aware locomotion has become an emerging topic in legged robotics. However, it is hard to generate diverse, challenging, and realistic unstructured terrains in simulation, which limits the way researchers evaluate their locomotion policies. In this paper, we prototype the generation of a terrain dataset via terrain authoring and active learning, and the learned samplers can stably generate diverse high-quality terrains. We expect the generated dataset to make a terrain-robustness benchmark for legged locomotion. The dataset, the code implementation, and some policy evaluations are released at https://bit.ly/3bn4j7f.

Junheng Li, Lizhi Yang, Aaron D. Ames

Real-time humanoid teleoperation requires inverse kinematics (IK) solvers that are both responsive and constraint-safe under kinematic redundancy and self-collision constraints. While differential IK enables efficient online retargeting, its locally linearized updates are inherently basin-dependent and often become trapped near joint limits, singularities, or active collision boundaries, leading to unsafe or stagnant behavior. We propose a GPU-parallelized, continuation-based differential IK that improves escape from such constraint-induced local minima while preserving real-time performance, promoting safety and stability. Multiple constrained IK quadratic programs are evaluated in parallel, together with a self-collision avoidance control barrier function (CBF), and a Lyapunov-based progression criterion selects updates that reduce the final global task-space error. The method is paired with a visual skeletal pose estimation pipeline that enables robust, real-time upper-body teleoperation on the THEMIS humanoid robot hardware in real-world tasks.

Lizhi Yang, Blake Werner, Massimiliano de Sa et al.

Reinforcement learning (RL), while powerful and expressive, can often prioritize performance at the expense of safety. Yet safety violations can lead to catastrophic outcomes in real-world deployments. Control Barrier Functions (CBFs) offer a principled method to enforce dynamic safety -- traditionally deployed online via safety filters. While the result is safe behavior, the fact that the RL policy does not have knowledge of the CBF can lead to conservative behaviors. This paper proposes CBF-RL, a framework for generating safe behaviors with RL by enforcing CBFs in training. CBF-RL has two key attributes: (1) minimally modifying a nominal RL policy to encode safety constraints via a CBF term, (2) and safety filtering of the policy rollouts in training. Theoretically, we prove that continuous-time safety filters can be deployed via closed-form expressions on discrete-time roll-outs. Practically, we demonstrate that CBF-RL internalizes the safety constraints in the learned policy -- both enforcing safer actions and biasing towards safer rewards -- enabling safe deployment without the need for an online safety filter. We validate our framework through ablation studies on navigation tasks and on the Unitree G1 humanoid robot, where CBF-RL enables safer exploration, faster convergence, and robust performance under uncertainty, enabling the humanoid robot to avoid obstacles and climb stairs safely in real-world settings without a runtime safety filter.

Blake Werner, Lizhi Yang, Aaron D. Ames

Robust humanoid locomotion in unstructured environments requires architectures that balance fast low-level stabilization with slower perceptual decision-making. We show that a simple layered control architecture (LCA), a proprioceptive stabilizer running at high rate, coupled with a compact low-rate perceptual policy, enables substantially more robust performance than monolithic end-to-end designs, even when using minimal perception encoders. Through a two-stage training curriculum (blind stabilizer pretraining followed by perceptual fine-tuning), we demonstrate that layered policies consistently outperform one-stage alternatives in both simulation and hardware. On a Unitree G1 humanoid, our approach succeeds across stair and ledge tasks where one-stage perceptual policies fail. These results highlight that architectural separation of timescales, rather than network scale or complexity, is the key enabler for robust perception-conditioned locomotion.

Lizhi Yang, Ruhang Ma, Avideh Zakhor

Object detection using aerial drone imagery has received a great deal of attention in recent years. While visible light images are adequate for detecting objects in most scenarios, thermal cameras can extend the capabilities of object detection to night-time or occluded objects. As such, RGB and Infrared (IR) fusion methods for object detection are useful and important. One of the biggest challenges in applying deep learning methods to RGB/IR object detection is the lack of available training data for drone IR imagery, especially at night. In this paper, we develop several strategies for creating synthetic IR images using the AIRSim simulation engine and CycleGAN. Furthermore, we utilize an illumination-aware fusion framework to fuse RGB and IR images for object detection on the ground. We characterize and test our methods for both simulated and actual data. Our solution is implemented on an NVIDIA Jetson Xavier running on an actual drone, requiring about 28 milliseconds of processing per RGB/IR image pair.

Scott Gilroy, Derek Lau, Lizhi Yang et al.

Quadrupeds are strong candidates for navigating challenging environments because of their agile and dynamic designs. This paper presents a methodology that extends the range of exploration for quadrupedal robots by creating an end-to-end navigation framework that exploits walking and jumping modes. To obtain a dynamic jumping maneuver while avoiding obstacles, dynamically-feasible trajectories are optimized offline through collocation-based optimization where safety constraints are imposed. Such optimization schematic allows the robot to jump through window-shaped obstacles by considering both obstacles in the air and on the ground. The resulted jumping mode is utilized in an autonomous navigation pipeline that leverages a search-based global planner and a local planner to enable the robot to reach the goal location by walking. A state machine together with a decision making strategy allows the system to switch behaviors between walking around obstacles or jumping through them. The proposed framework is experimentally deployed and validated on a quadrupedal robot, a Mini Cheetah, to enable the robot to autonomously navigate through an environment while avoiding obstacles and jumping over a maximum height of 13 cm to pass through a window-shaped opening in order to reach its goal.

Anxing Xiao, Wenzhe Tong, Lizhi Yang et al.

An autonomous robot that is able to physically guide humans through narrow and cluttered spaces could be a big boon to the visually-impaired. Most prior robotic guiding systems are based on wheeled platforms with large bases with actuated rigid guiding canes. The large bases and the actuated arms limit these prior approaches from operating in narrow and cluttered environments. We propose a method that introduces a quadrupedal robot with a leash to enable the robot-guiding human system to change its intrinsic dimension (by letting the leash go slack) in order to fit into narrow spaces. We propose a hybrid physical Human-Robot Interaction model that involves leash tension to describe the dynamical relationship in the robot-guiding human system. This hybrid model is utilized in a mixed-integer programming problem to develop a reactive planner that is able to utilize slack-taut switching to guide a blind-folded person to safely travel in a confined space. The proposed leash-guided robot framework is deployed on a Mini Cheetah quadrupedal robot and validated in experiments.

Huijuan Xu, Lizhi Yang, Stan Sclaroff et al.

Spatio-temporal action detection in videos requires localizing the action both spatially and temporally in the form of an "action tube". Nowadays, most spatio-temporal action detection datasets (e.g. UCF101-24, AVA, DALY) are annotated with action tubes that contain a single person performing the action, thus the predominant action detection models simply employ a person detection and tracking pipeline for localization. However, when the action is defined as an interaction between multiple objects, such methods may fail since each bounding box in the action tube contains multiple objects instead of one person. In this paper, we study the spatio-temporal action detection problem with multi-object interaction. We introduce a new dataset that is annotated with action tubes containing multi-object interactions. Moreover, we propose an end-to-end spatio-temporal action detection model that performs both spatial and temporal regression simultaneously. Our spatial regression may enclose multiple objects participating in the action. During test time, we simply connect the regressed bounding boxes within the predicted temporal duration using a simple heuristic. We report the baseline results of our proposed model on this new dataset, and also show competitive results on the standard benchmark UCF101-24 using only RGB input.