Research on omnidirectional seated locomotion for humanoids on passive chairs. Policies use proprioception only, outperforming standing in tracking. Analyzes regularization effects on convergence.
LAC is a general humanoid whole-body controller that simultaneously realizes commanded linear and angular compliance for wrenches applied to the upper body. Sampled force and couple events imposed on contact frames from human interaction data synthesize a large-scale compliant-response dataset; teacher-student RL then trains a single policy to track compliant motions under external wrenches, with sim and real experiments showing monotonic stiffness modulation and teleoperated loco-manipulation.
The first Behavior World Model (BWM) for humanoid whole-body control. A causal Transformer jointly predicts next action, state, and latent behavior command distribution, enabling the policy network to model how the environment shapes actions. Automatic terrain-annotation pipeline recovers 3D contact geometry from retargeted motion. At deployment, implausible commands are detected and retracted onto learned behaviors. Achieves highest success rate across all four regimes: 81.3% terrain interaction, 83.1% under implausible commands, 99.3% fall recovery. Unitree G1 checkpoint transfers to Maker L01 robot.
Human athletes demonstrate versatile and highly-dynamic tennis skills to successfully conduct competitive rallies with a high-speed tennis ball. However, reproducing such behaviors on humanoid robots is difficult, partially due to the lack of perfect humanoid action data or human kinematic motion data in tennis scenarios as reference. In this work, we propose LATENT, a system that Learns Athletic humanoid TEnnis skills from imperfect human motioN daTa. The imperfect human motion data consist only of motion fragments that capture the primitive skills used when playing tennis rather than precise and complete human-tennis motion sequences from real-world tennis matches, thereby significantly reducing the difficulty of data collection. Our key insight is that, despite being imperfect, such quasi-realistic data still provide priors about human primitive skills in tennis scenarios. With further correction and composition, we learn a humanoid policy that can consistently strike incoming balls under a wide range of conditions and return them to target locations, while preserving natural motion styles. We also propose a series of designs for robust sim-to-real transfer and deploy our policy on the Unitree G1 humanoid robot. Our method achieves surprising results in the real world and can stably sustain multi-shot rallies with human players. Project page: https://zzk273.github.io/LATENT/
Achieving both anthropomorphic naturalness and robust terrain traversal remains a fundamental challenge in humanoid locomotion. Existing Reinforcement Learning (RL) approaches typically rely on fixed motion priors, limiting their adaptability to varying environments. We propose Terrain-conditioned Generative Motion Priors (T-GMP), a module that captures a terrain-conditioned latent motion manifold from a few expert state-terrain demonstrations using a Conditional Variational Autoencoder (CVAE). The learned priors enable smooth style transitions, facilitating a unified policy that adapts to terrain variations. We integrate T-GMP into an adversarial learning pipeline with our proposed Foothold Penalty, where a discriminator dynamically modulates naturalness constraints conditioned on local terrain features, guiding the generation of versatile and human-like motions. Experimental results demonstrate that our method outperforms existing baselines in traversal success rate and motion smoothness, while preserving biomimetically natural and physically coordinated motions.
RPL uses a two-stage framework that distills terrain-specific height-map experts into a unified depth-based transformer policy, enabling robust bidirectional humanoid locomotion with payloads on slopes, stairs, and stepping stones.
Humanoid robots are expected to traverse complex terrains, where the plantar support may vary dramatically due to foot placement errors, ground properties, and transient dynamics. To achieve robust locomotion, the robots are required to adapt to uneven terrain and uncertain foot--ground interactions. Existing locomotion policies rely primarily on proprioception or exteroceptive terrain perception, where the former provides only indirect evidence of plantar support, while the latter predicts contact conditions before touchdown but cannot observe the actual support in real-time. Although some studies incorporate plantar contacts as an auxiliary perception, they rely mainly on summary statistics, overlooking the spatial topology of plantar pressure, which provides a more direct characterization of the realized contact state. To bridge this gap, we present Tac4Loco, a tactile-perceptive framework that incorporates multi-array plantar pressure as direct feedback for humanoid locomotion. We formulate a topology-preserving ordinal representation to map simulated and physical sensor signals into a shared observation space, with a dual-branch encoder for extracting their spatial and temporal representations. Subsequently, the learned spatiotemporal features are integrated with augmented proprioception including terrain estimation cues, and provided to an asymmetric actor-critic architecture for policy learning. Extensive simulation and real-world experiments demonstrate improved tracking performance and support adaptation on terrains with inclined, partial, asymmetric, and changing support. We further demonstrate its zero-shot deployment on unseen compliant and unstructured terrains, including a foam platform and a gravel road. All code and experimental configurations will be released as open-source to facilitate reproducibility.
Humanoid robots hold great promise as general-purpose agents in human-centered environments, yet generalist vision-language-action (VLA) foundation models are not readily applicable to humanoid whole-body loco-manipulation. The high dimensionality and interdependence of humanoid motions make it challenging for conventional single-stage VLA architectures to coordinate locomotion, waist posture, and dual-arm manipulation effectively. Moreover, policies trained through offline behavior cloning can remain suboptimal during real-world deployment. Although online reinforcement learning can refine policies through real-world interaction, directly tuning large VLA backbones demands excessive computation and may introduce safety risks during real-robot exploration. To address these bottlenecks, we introduce HAF (Humanoid Adaptation Framework), a two-part framework consisting of HAF-VLA and HAF-Steer that transfers off-the-shelf generalist VLA foundation models to humanoid whole-body loco-manipulation. HAF-VLA is a hierarchical action-flow generator built on a pretrained flow-matching VLA. It splits full-body action denoising into three sequential stages with stage embeddings and cross-stage KV caches that retain kinematic dependencies, avoiding incoherent whole-body actions from one-shot generation. On top of the frozen HAF-VLA, HAF-Steer is a latent offline-to-online RL pipeline that leverages flow-matching invertibility and DCT-based dimensionality reduction to restrict RL optimization to a compact noise subspace and train a regularized SAC policy. This avoids updating the large VLA backbone and enables efficient real-world policy refinement. Evaluated on seven real-world humanoid loco-manipulation tasks, HAF surpasses vanilla single-stage VLA baselines and improves whole-body coordination and task performance. Project website: https://grange007.github.io/HAF .
VLK synthesizes paired vision-language-kinematics supervision inside 3DGS-reconstructed real scenes: it generates navigation and object-interaction trajectories with privileged scene info, renders egocentric views after the fact, and produces 48,000 paired trajectories to train a policy predicting Unitree G1 whole-body motion, enabling sim-to-real perception-based humanoid loco-manipulation.
CMoE uses contrastive learning to improve MoE expert specialization for humanoid terrain navigation. Validated on Unitree G1, achieving robust gait over 20cm steps and 80cm gaps.
G0.5 is a pretrained autoregressive VLA in which a single transformer decoder emits reasoning and action tokens under one objective. A cross-embodiment action tokenizer, a native chain-of-thought stream, and a visual memory module make this tractable at foundation-model scale, letting the pretrained VLM capabilities carry over to physical behavior. It surpasses pi-0.5 and GR00T-N1.7 across 7 regimes including real-world R1 robot fine-tuning, BEHAVIOR Challenge, DROID zero-shot transfer, LIBERO, RoboTwin 2.0 and SimplerEnv.
Humans exhibit remarkable motor agility, which highlights the great potential of humanoid robots for athletic locomotion. Long rope skipping requires two rope turners to cooperatively swing the rope while adapting to a player under different jumping rhythms. We propose Marope, a multi-agent reinforcement learning (MARL) framework for cooperative long rope skipping with multiple humanoid robots. It adopts a hierarchical RL framework: the lower level learns decentralized rope manipulation policies through MARL, while the upper level trains a centralized scheduling policy to coordinate execution. Diverse jumping policies are incorporated to improve generalization across player behavioral styles. Experiments on Unitree G1 robots in simulation and the real world show that Marope outperforms baselines.
