Humanoid Rickshaw Pulling: Whole-Body Locomotion under Coupled Wheeled Loads

Summary

Humanoid robots could transport payloads substantially heavier than themselves by pulling passive wheeled vehicles instead of carrying the load.

Abstract Summary

Humanoid robots could transport payloads substantially heavier than themselves by pulling passive wheeled vehicles instead of carrying the load. This capability, however, creates a coupled locomotion problem: the robot must maintain persistent upper-body contact while adapting to unknown, configuration-dependent forces arising from the payload, vehicle, and terrain. We present a whole-body control framework for humanoid rickshaw pulling that tracks commanded vehicle motion while preserving balance and stable grasps under uncertain load dynamics. During training, a privileged teacher exploits vehicle states, interaction forces, and load properties. Its actions and latent are distilled into a history-conditioned student that implicitly infers coupled dynamics from proprioceptive responses, followed by reinforcement-learning fine-tuning. Comparisons with \emph{No History} and \emph{Only History} baselines show that the resulting policy achieves accurate vehicle tracking while reducing vehicle oscillation, torso tilt, and actuation cost. Behavioral analysis shows that Unitree G1 propels the rickshaw and generates gait-synchronized whole-body reactions that stabilize its lateral and roll motions. Moreover, pulling redistributes joint effort and yields a lower robot-normalized cost-of-transport proxy than unloaded walking over most tested load--speed conditions. On hardware, a single policy performs starting, sustained pulling, turning, and stopping with both rigid payloads and human passengers, handling a loaded rickshaw mass of up to 115~kg without load-specific retuning. These results demonstrate robust heavy-load transportation through coordinated and persistent humanoid--vehicle interaction.

Key Points

Humanoid robots could transport payloads substantially heavier than themselves by pulling passive wheeled vehicles instead of carrying the load.
This capability, however, creates a coupled locomotion problem: the robot must maintain persistent upper-body contact while adapting to unknown, configuration-dependent forces arising from the payload, vehicle, and terrain.
We present a whole-body control framework for humanoid rickshaw pulling that tracks commanded vehicle motion while preserving balance and stable grasps under uncertain load dynamics.
During training, a privileged teacher exploits vehicle states, interaction forces, and load properties.
Its actions and latent are distilled into a history-conditioned student that implicitly infers coupled dynamics from proprioceptive responses, followed by reinforcement-learning fine-tuning.

Problem Formulation

We formulate a humanoid robot’s pulling of a passive two-wheeled rickshaw through persistent bimanual grasping as a partially observed MDP. The task is to realize desired vehicle velocity while maintaining balance and grasp, without directly observing vehicle states or load parameters during deployment. Fig. 2 shows the hardware system and its kinematics.

Humanoid and Gripper : We use a Unitree G1 with $29$ actuated joints spanning the legs, waist, and arms, and two Dex1-1 grippers on its hands.

Problem Formulation — Humanoid Rickshaw Pulling: Whole-Body Locomotion under Coupled Wheeled Loads — Fig. 2 : Humanoid–rickshaw system. (a) The physical setup connects G1 to a passive rickshaw through custom grippers. (b) Relative geometry characterizes the coupled configuration. (c) G1 applies pulling forces through the handles, while the wheels support the load and constrain lateral motion.

Method

Proprioceptive responses during pulling provide cues about the unobservable vehicle dynamics. To learn a control-relevant representation from these responses, we train a privileged teacher using robot and vehicle histories together with static parameters ( S0 ). Its actions and latent are distilled into a history-conditioned student ( S1 ).

We then jointly fine-tune the student encoder and action head with PPO [ 23 ] ( S2 ; Fig. 3 ). This approach shares the motivation of prior history-based adaptation methods [ 17 , 19 ] .

Method — Humanoid Rickshaw Pulling: Whole-Body Locomotion under Coupled Wheeled Loads — Fig. 3 : Teacher–student training and deployment. We first train a teacher with privileged load and interaction information, then distill its actions and latent into a student using only onboard observation history. Joint encoder and policy fine-tuning completes training for deployment without privileged inputs.

Experimental Results

We evaluate vehicle tracking across loads and the contribution of each design choice. Beyond tracking, joint-effort and handle-interaction analyses help explain how pulling a wheeled load changes locomotion, transport costs, and balance behavior. Hardware trials evaluate different passenger configurations, revealing the remaining sim-to-real differences.

Sources and Demonstrations

Figures are reproduced from Yang et al. See the full paper for experimental details and the project page, when the paper links one, for demonstrations and videos.