Trajectory-based actuator identification via differentiable simulation

2026-04-11

AI Summary

This work addresses the challenge of constructing high-fidelity actuator dynamics models from real robot data in the absence of torque sensors, current measurements, or internal motor information. The authors formulate actuator identification as a trajectory-matching problem, relying solely on joint position and velocity provided by encoders. By integrating differentiable simulation with gradient-based optimization, their unified framework accommodates a spectrum of model classes—from compact analytical formulations to neural actuator mappings—without requiring specialized test benches or torque sensing. Experimental results demonstrate that the proposed approach reduces position error from 14.20 mrad to 7.54 mrad. Furthermore, when applied to downstream motion policy training, it yields a 46% increase in robot travel distance and a 75% reduction in heading deviation.

Technology Category

• Intelligent Robots: State Estimation
• Search and Optimization: Sampling/Simulation-based Search
• Planning, Routing, and Scheduling: Activity and Plan Recognition

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• User Modeling, Personalization and Recommendation: User modeling and simulation for interactive and conversational systems
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Accurate actuation models are critical for bridging the gap between simulation and real robot behavior, yet obtaining high-fidelity actuator dynamics typically requires dedicated test stands and torque sensing. We present a trajectory-based actuator identification method that uses differentiable simulation to fit system-level actuator models from encoder motion alone. Identification is posed as a trajectory-matching problem: given commanded joint positions and measured joint angles and velocities, we optimize actuator and simulator parameters by backpropagating through the simulator, without torque sensors, current/voltage measurements, or access to embedded motor-control internals. The framework supports multiple model classes, ranging from compact structured parameterizations to neural actuator mappings, within a unified optimization pipeline. On held-out real-robot trajectories under identical commands, the proposed torque-sensor-free identification achieves much tighter trajectory alignment than a supervised stand-trained baseline dominated by steady-state data, reducing mean absolute position error from 14.20 mrad to as low as 7.54 mrad (1.88 times). Finally, we demonstrate downstream impact in a real-robot locomotion study: training policies with the refined actuator model increases travel distance by 46% and reduces rotational deviation by 75% relative to the baseline.

Problem

• actuator identification
• differentiable simulation
• trajectory matching
• robot dynamics
• simulation-to-reality gap

Innovation

• differentiable simulation
• actuator identification
• trajectory matching
• sensor-free modeling
• robot dynamics learning