SMAT: Staged Multi-Agent Training for Co-Adaptive Exoskeleton Control

• URL: http://arxiv.org/abs/2603.07618v1
• Date: Sun, 08 Mar 2026 13:06:36 GMT
• Title: SMAT: Staged Multi-Agent Training for Co-Adaptive Exoskeleton Control
• Abstract summary: We propose a four-stage curriculum designed to mirror how users naturally acclimate to a wearable device.<n>We implement SMAT in the MyoAssist simulation environment using a 26-muscle lower-limb model and an attached hip exoskeleton.<n>Our musculoskeletal simulations demonstrate that the learned exoskeleton control policy produces an average 10.1% reduction in hip muscle activation relative to the no-assist condition.
• Score: 1.7061322026474812
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Effective exoskeleton assistance requires co-adaptation: as the device alters joint dynamics, the user reorganizes neuromuscular coordination, creating a non-stationary learning problem. Most learning-based approaches do not explicitly account for the sequential nature of human motor adaptation, leading to training instability and poorly timed assistance. We propose Staged Multi-Agent Training (SMAT), a four-stage curriculum designed to mirror how users naturally acclimate to a wearable device. In SMAT, a musculoskeletal human actor and a bilateral hip exoskeleton actor are trained progressively: the human first learns unassisted gait, then adapts to the added device mass; the exoskeleton subsequently learns a positive assistance pattern against a stabilized human policy, and finally both agents co-adapt with full torque capacity and bidirectional feedback. We implement SMAT in the MyoAssist simulation environment using a 26-muscle lower-limb model and an attached hip exoskeleton. Our musculoskeletal simulations demonstrate that the learned exoskeleton control policy produces an average 10.1% reduction in hip muscle activation relative to the no-assist condition. We validated the learned controller in an offline setting using open-source gait data, then deployed it to a physical hip exoskeleton for treadmill experiments with five subjects. The resulting policy delivers consistent assistance and predominantly positive mechanical power without the need for any explicitly imposed timing shift (mean positive power: 13.6 W at 6 Nm RMS torque to 23.8 W at 9.3 Nm RMS torque, with minimal negative power) consistently across all subjects without subject-specific retraining.

Related papers

• ATP: Anatomical Torque with Passivity-based Control Framework for Safe Upper-Limb Exoskeleton Assistance [8.72367908819341]
  This paper proposes Anatomical Torque with Passivity-Based Control for safe upper-limb exoskeleton assistance.<n>An online torque-refinement scheme adapts the reference to diverse movements, suppresses tendon-induced spikes, and incorporates a learned anomaly score for safe and comfortable assistance.<n> Simulations and real-world experiments show accurate tracking of long-duration motion sequences and generalization to real-time human movements.
  arXiv  Detail & Related papers  (2026-08-06T08:06:40Z)
• Staged Multi-Agent Training (SMAT) for Hip Exoskeletons: Metabolic and Biomechanical Validation of a Simulation-Trained Co-Adaptive Controller [1.5617797712963508]
  This article provides the first physiological validation of SMAT.<n>A single simulation-trained SMAT policy delivers a significant metabolic benefit on real users while remaining robust beyond the conditions it was trained on.
  arXiv  Detail & Related papers  (2026-08-01T15:31:14Z)
• A Replay-Constrained Simulation Framework for Personalization of Powered Knee--Ankle Prosthesis Controllers [11.765906112645466]
  Personalization of impedance controllers for powered prosthetic legs is critical to accommodating individual gait but remains challenging.<n>MuJoCo-based simulator reproduces prosthetic knee-ankle dynamics while replaying recorded hip kinematics and feedback-based ground reaction forces.<n>Deep reinforcement learning policy personalizes phase-dependent stiffness, damping, and equilibrium angle at both joints simultaneously.
  arXiv  Detail & Related papers  (2026-07-24T18:59:48Z)
• Learn Weightlessness: Imitate Non-Self-Stabilizing Motions on Humanoid Robot [56.377161933570335]
  Humans naturally exploit a "weightless" state during non-self-stabilizing (NSS) motions.<n>Inspired by this biological mechanism, we propose a weightlessness-state auto-labeling strategy for dataset annotation.<n>Our work bridges the gap between precise trajectory tracking and adaptive environmental interaction, offering a biologically-inspired solution for contact-rich humanoid control.
  arXiv  Detail & Related papers  (2026-04-23T07:10:05Z)
• Exoskeleton Control through Learning to Reduce Biological Joint Moments in Simulations [0.0]
  Data-driven joint-moment predictors offer a scalable alternative to laboratory-based inverse-dynamics pipelines for estimation and exoskeleton control.<n> physics-based reinforcement learning enables simulation-trained controllers to learn dynamics-aware assistance strategies.<n>This paper presents (1) an RL framework to learn exoskeleton assistance policies that reduce biological joint moments, and (2) a validation pipeline that verifies the trained control networks.
  arXiv  Detail & Related papers  (2026-03-08T13:31:06Z)
• Learning Hip Exoskeleton Control Policy via Predictive Neuromusculoskeletal Simulation [0.0]
  We present a physics-based neuromusculoskeletal learning framework that trains a hip-exoskeleton control policy entirely in simulation.<n>A reinforcement learning teacher policy is trained using a muscle-synergy action prior over a wide range of walking speeds and slopes.<n>In simulation, exoskeleton assistance reduces mean muscle activation by up to 3.4% and mean positive joint power by up to 7.0% on level ground and ramp ascent.
  arXiv  Detail & Related papers  (2026-03-04T15:23:50Z)
• ResMimic: From General Motion Tracking to Humanoid Whole-body Loco-Manipulation via Residual Learning [59.64325421657381]
  Humanoid whole-body loco-manipulation promises transformative capabilities for daily service and warehouse tasks.<n>We introduce ResMimic, a two-stage residual learning framework for precise and expressive humanoid control from human motion data.<n>Results show substantial gains in task success, training efficiency, and robustness over strong baselines.
  arXiv  Detail & Related papers  (2025-10-06T17:47:02Z)
• HHI-Assist: A Dataset and Benchmark of Human-Human Interaction in Physical Assistance Scenario [63.77482302352545]
  HHI-Assist is a dataset comprising motion capture clips of human-human interactions in assistive tasks.<n>Our work has the potential to significantly enhance robotic assistance policies.
  arXiv  Detail & Related papers  (2025-09-12T09:38:17Z)
• Reinforcement learning-based motion imitation for physiologically plausible musculoskeletal motor control [47.423243831156285]
  We present a model-free motion imitation framework (KINESIS) to advance the understanding of muscle-based motor control.<n>We demonstrate that KINESIS achieves strong imitation performance on 1.9 hours of motion capture data.<n>KINESIS generates muscle activity patterns that correlate well with human EMG activity.
  arXiv  Detail & Related papers  (2025-03-18T18:37:49Z)
• Adaptive Torque Control of Exoskeletons under Spasticity Conditions via Reinforcement Learning [0.0]
  We describe a novel adaptive torque controller via deep reinforcement learning (RL) for a knee exoskeleton under joint spasticity conditions.<n>Results for a simulated knee extension movement showed that the agent learns to control the exoskeleton for individuals with different levels of spasticity.
  arXiv  Detail & Related papers  (2025-03-14T14:22:09Z)
• Natural and Robust Walking using Reinforcement Learning without Demonstrations in High-Dimensional Musculoskeletal Models [29.592874007260342]
  Humans excel at robust bipedal walking in complex natural environments. It is still not fully understood how the nervous system resolves the musculoskeletal redundancy to solve the multi-objective control problem.
  arXiv  Detail & Related papers  (2023-09-06T13:20:31Z)
• Continuous Decoding of Daily-Life Hand Movements from Forearm Muscle Activity for Enhanced Myoelectric Control of Hand Prostheses [78.120734120667]
  We introduce a novel method, based on a long short-term memory (LSTM) network, to continuously map forearm EMG activity onto hand kinematics. Ours is the first reported work on the prediction of hand kinematics that uses this challenging dataset. Our results suggest that the presented method is suitable for the generation of control signals for the independent and proportional actuation of the multiple DOFs of state-of-the-art hand prostheses.
  arXiv  Detail & Related papers  (2021-04-29T00:11:32Z)
• Online Body Schema Adaptation through Cost-Sensitive Active Learning [63.84207660737483]
  The work was implemented in a simulation environment, using the 7DoF arm of the iCub robot simulator. A cost-sensitive active learning approach is used to select optimal joint configurations. The results show cost-sensitive active learning has similar accuracy to the standard active learning approach, while reducing in about half the executed movement.
  arXiv  Detail & Related papers  (2021-01-26T16:01:02Z)
• Reinforcement Learning of Musculoskeletal Control from Functional Simulations [3.94716580540538]
  In this work, a deep reinforcement learning (DRL) based inverse dynamics controller is trained to control muscle activations of a biomechanical model of the human shoulder. Results are presented for a single-axis motion control of shoulder abduction for the task of following randomly generated angular trajectories.
  arXiv  Detail & Related papers  (2020-07-13T20:20:01Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.