NeuralActuator: Neural Actuation Modeling for Robot Dynamics and External Force Perception

2026-07-13

AI Summary

This work addresses the challenges posed by friction, hysteresis, backlash, and thermal effects in low-cost robotic actuators, which violate the assumed linear current-torque relationship, leading to a significant mismatch between simulated and real-world actuator dynamics and precluding sensorless external force perception. To overcome these limitations, the authors propose NeuralActuator—a unified Transformer-based neural architecture that jointly models actuator dynamics, estimates external forces without dedicated force sensors (incorporating a contact-probability gating mechanism), and predicts motor health scores. Trained via differentiable simulation and multi-task supervised learning without requiring ground-truth torque labels, the model enables real-time inference. Evaluated across diverse platforms with 5–7 degrees of freedom and hardware costs ranging from \$500 to \$30,000, NeuralActuator substantially improves accuracy in dynamics modeling, external force estimation, health assessment, and behavior cloning pretraining.

Abstract

Differentiable simulators have advanced policy learning and model-based control, yet actuator dynamics remain an important source of sim-to-real error. This is particularly acute on low-cost platforms, where the linear current-to-torque relation $τ= K_tI$ becomes unreliable during commanded-target tracking because of friction, hysteresis, backlash, and thermal effects. We present NeuralActuator, a neural actuator model that jointly predicts (i) a simulator-equivalent generalized-effort surrogate for trajectory propagation on low-cost servo platforms, (ii) external force with a contact-probability gate for sensorless force perception, and (iii) a motor-condition score for the supervised joint. We also introduce the Neural Actuation Dataset (NAD), collected with a twin-arm teleoperation system that records robot states and actuator telemetry together with external-force labels. The torque-surrogate head is trained through differentiable simulation from pose trajectories without direct generalized-effort labels, while the force, gate, and motor-condition heads receive direct supervision. A Transformer captures temporal dependencies while supporting real-time inference. We evaluate NeuralActuator on a 5-DoF OpenManipulator-X, a 6-DoF SO-101, and a 7-DoF Franka Emika Panda, spanning three actuator families and platforms costing approximately USD 500 to over USD 30,000. The low-cost platforms support dynamics and force evaluation, while the offline Franka experiment provides an additional payload-force-estimation benchmark. Experiments further demonstrate its application for motor condition estimation on OpenManipulator-X and improved behavior-cloning performance when NeuralActuator is used as a pretrained module.

Problem

• actuator dynamics
• sim-to-real gap
• low-cost robots
• torque modeling
• external force perception

Innovation

• Neural Actuator Modeling
• Differentiable Simulation
• Sensorless Force Perception
• Motor Condition Monitoring
• Transformer-based Dynamics