REVIEW 15 cited by
Hybrid Internal Model: Learning Agile Legged Locomotion with Simulated Robot Response
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
read the original abstract
Robust locomotion control depends on accurate state estimations. However, the sensors of most legged robots can only provide partial and noisy observations, making the estimation particularly challenging, especially for external states like terrain frictions and elevation maps. Inspired by the classical Internal Model Control principle, we consider these external states as disturbances and introduce Hybrid Internal Model (HIM) to estimate them according to the response of the robot. The response, which we refer to as the hybrid internal embedding, contains the robot's explicit velocity and implicit stability representation, corresponding to two primary goals for locomotion tasks: explicitly tracking velocity and implicitly maintaining stability. We use contrastive learning to optimize the embedding to be close to the robot's successor state, in which the response is naturally embedded. HIM has several appealing benefits: It only needs the robot's proprioceptions, i.e., those from joint encoders and IMU as observations. It innovatively maintains consistent observations between simulation reference and reality that avoids information loss in mimicking learning. It exploits batch-level information that is more robust to noises and keeps better sample efficiency. It only requires 1 hour of training on an RTX 4090 to enable a quadruped robot to traverse any terrain under any disturbances. A wealth of real-world experiments demonstrates its agility, even in high-difficulty tasks and cases never occurred during the training process, revealing remarkable open-world generalizability.
Forward citations
Cited by 15 Pith papers
-
Embedding Hybrid Systems into Continuous Latent Vector Fields
An n-dimensional hybrid system embeds into a continuous vector field in m > 2n dimensions, enabling latent Neural ODEs with consistency losses to recover hybrid flows from time series.
-
FADA: Few-Shot Domain Adaptation via Dynamics Alignment for Humanoid Control
FADA is a three-stage Planner-IDM method that achieves few-shot domain adaptation for humanoid control by distilling an oracle policy then finetuning only the IDM on short target-domain rollouts via supervised learning.
-
StairMaster: Learning to Conquer Risky Hollow Stairs for Agile Quadrupedal Robots
A three-stage RL framework with cross-attention, spatial memory, and depth-noise augmentation lets a Unitree Go2 climb 55° hollow stairs zero-shot from simulation.
-
StairMaster: Learning to Conquer Risky Hollow Stairs for Agile Quadrupedal Robots
StairMaster trains an RL policy that lets a Unitree Go2 quadruped climb hollow stairs up to 55 degrees via zero-shot sim-to-real transfer using cross-attention, SRU memory, and active-perception rewards.
-
Perceptive Humanoid Parkour: Chaining Dynamic Human Skills via Motion Matching
A modular system uses motion matching to compose long-horizon human skill chains, trains RL experts, and distills them into a depth-based policy that lets a Unitree G1 humanoid autonomously climb, vault, and roll over...
-
TOP: Time Optimization Policy for Stable and Accurate Standing Manipulation with Humanoid Robots
A reinforcement-learned time optimization policy that adaptively slows upper-body motion clips improves stability and precision of humanoid standing manipulation at a modest time cost.
-
Generalized Locomotion in Out-of-distribution Conditions with Robust Transformer
A transformer with body tokenization and consistent dropout generalizes to unseen leg damages and sensor noise while trained on limited dynamics and clean observations.
-
IMPACT: Learning Internal-Model Predictive Control for Forceful Robotic Manipulation
IMPACT decouples forceful manipulation into task-planning and internal-model predictive control, claiming higher success rates, better generalization to unseen weights, and improved safety and energy efficiency in sim...
-
Mind Your Steps: A General Learning Framework for Accurate Humanoid Foothold Tracking
A lightweight RL framework trains terrain-agnostic 3D foothold-tracking policies for humanoids that transfer directly to real-world use as standalone low-level controllers.
-
Learning to Balance Motor Thermal Safety and Quadrupedal Locomotion Performance with Residual Policy
A thermal residual policy on a pre-trained quadruped locomotion controller prevents motor overheating under payload while preserving performance, lasting over 13 minutes on a Unitree A1 versus ~5 minutes for the nomin...
-
MUJICA: Multi-skill Unified Joint Integration of Control Architecture for Wheeled-Legged Robots
A single reinforcement learning policy jointly trains multiple locomotion skills for wheeled-legged robots with DC-motor constraints and learns a proprioceptive skill selector for adaptive behavior.
-
KiVi: Kinesthetic-Visuospatial Integration for Dynamic and Safe Egocentric Legged Locomotion
A quadruped locomotion controller that explicitly separates proprioceptive and visual pathways stays stable under camera occlusion and visual corruption that destabilizes fused-vision policies.
-
Learning to Balance Motor Thermal Safety and Quadrupedal Locomotion Performance with Residual Policy
A two-stage RL method adds a residual policy on a thermal model to let a quadruped maintain performance while preventing motor overheating, shown in simulation and on a Unitree A1 carrying 3 kg for over 13 minutes ver...
-
Learning to Balance Motor Thermal Safety and Quadrupedal Locomotion Performance with Residual Policy
A two-stage RL framework with a thermal-aware residual policy enables a Unitree A1 quadruped to achieve over 13 minutes of stable locomotion under 3 kg payload versus 5 minutes before overheating with the nominal poli...
-
IDEA: Insensitive to Dynamics Mismatch via Effect Alignment for Sim-to-Real Transfer in Multi-Agent Control
IDEA elevates multi-agent policies to semantic actions with effect alignment and synchronization for improved sim-to-real robustness on navigation tasks.
Discussion (0). Sign in to comment.