HoRD: Robust Humanoid Control via History-Conditioned Reinforcement Learning and Online Distillation

The paper frames the work as humanoid-locomotion-control. The reported platform or hardware context is humanoid (simulation-based). The evaluationSimulation & Sim-to-RealEvaluationMeasuring how well a robot system performs. setting is simulationSimulation & Sim-to-RealSimulationA virtual environment where robots can be trained or tested.. Start here because it defines what success means and which assumptions the rest of the method inherits.

The method is organized around transformer-based policyCore ConceptsPolicyThe rule or model that maps observations or states to actions. with history-conditioned RLImitation & Reinforcement LearningReinforcement Learning (RL)Teaching a robot through trial and error using rewards. and online distillation. This paper addresses a critical problem in robotics: humanoid robots often fail when conditions change slightly from their trainingRobot LearningTrainingThe process of fitting a model using data or experience. environmentCore ConceptsEnvironmentThe external world the robot operates in, including objects, obstacles, people, and surfaces.. HoRD solves this by trainingRobot LearningTrainingThe process of fitting a model using data or experience. policies that learn to adapt online by observing recent movement patterns, then distilling that adaptive knowledge into a simpler student model. For developers, this means you can build humanoid robotCore ConceptsRobotA physical system with sensors and actuators that can observe the world and take actions. controllers that gracefully handle variations in physics parameters, environmental conditions, and perturbations without retraining for each new situation—making robots much more practical for real-world deploymentSimulation & Sim-to-RealDeploymentPutting the trained system on a real robot.. When reading the method section, identify the inputs, the learned or engineered representation, and the actionCore ConceptsActionA command the robot sends to its motors, controller, or low-level system. or prediction produced by the system.

For robotics work, the data story is part of the method: check whether the system depends on teleoperationImitation & Reinforcement LearningTeleoperation (teleop)A human remotely controlling the robot, often to collect demonstrations., simulationSimulation & Sim-to-RealSimulationA virtual environment where robots can be trained or tested., internet video, human labels, or robotCore ConceptsRobotA physical system with sensors and actuators that can observe the world and take actions. rollouts.

The key reported result is HoRD outperforms strong baselines in robustnessModern Robot LearningRobustnessHow well a robot keeps working despite noise, disturbances, or variation. and transfer under unseen domains and external perturbations with zero-shotModern Robot LearningZero-shotDoing a new task without task-specific training. adaptation. Look for the gap between the headline result and the deploymentSimulation & Sim-to-RealDeploymentPutting the trained system on a real robot. setting you would actually care about.