Neuromuscular Modeling for Locomotion with Wearable Assistive Robots -- A primer

TL;DR

The paper surveys the neuromechanical foundations of human locomotion, the modeling approaches that link neural commands to muscle-driven movement, and the controllers that integrate these insights into wearable lower-limb robotics. It highlights central pattern generators, spinal reflexes, synergies, and EMG-informed muscle models as key building blocks, and reviews neural-, muscle-, human-in-the-loop-, and data-driven control strategies for wearable robots. By synthesizing foundational neurophysiology with computational models and control methods, the primer aims to foster cross-disciplinary collaboration to develop intuitive, robust, and personalized WRs for locomotion and balance. The work emphasizes the need for real-time, physiologically faithful NMSK representations and scalable frameworks to translate laboratory insights into daily-life assistive technologies, with future directions including high-density neural measurements, vision-based inputs, and cloud-supported big data.

Abstract

Wearable assistive robots (WR) for the lower extremity are extensively documented in literature. Various interfaces have been designed to control these devices during gait and balance activities. However, achieving seamless and intuitive control requires accurate modeling of the human neuromusculoskeletal (NMSK) system. Such modeling enables WR to anticipate user intentions and determine the necessary joint assistance. Despite the existence of controllers interfacing with the NMSK system, robust and generalizable techniques across different tasks remain scarce. Designing these novel controllers necessitates the combined expertise of neurophysiologists, who understand the physiology of movement initiation and generation, and biomechatronic engineers, who design and control devices that assist movement. This paper aims to bridge the gaps between these fields by presenting a primer on key concepts and the current state of the science in each area. We present three main sections: the neuromechanics of locomotion, neuromechanical models of movement, and existing neuromechanical controllers used in WR. Through these sections, we provide a comprehensive overview of seminal studies in the field, facilitating collaboration between neurophysiologists and biomechatronic engineers for future advances in wearable robotics for locomotion.

Figures (3)

• Figure 1: High-level representation of the neural control for human locomotion. A) The central control (i.e., supra-spinal circuits, inter-neurons, etc.) integrates neural information from sensory systems (e.g., vestibular, visual, propioceptive, etc.) and generates a motor command. This is transmitted to pools of $\alpha$-motoneurons, which transform the final common signal and send it to the mechanical actuator (i.e., the muscle-tendon units). B) Sensory feedback from the propioceptive system is provided by Golgi tendon organs (GTOs) and muscle spindles. This is transmitted to the central control via Ia, Ib, and II afferent fibres. Here, only fibres from flexor muscle are depicted for clarity. The sensitivity of the Ia and II fibers (K) is regulated by $\gamma$-motoneurons (i.e., fusimotor system). C) Mechanical actuation results from the contraction of flexor and extensor muscles. These contract in response to the motor code send by $\alpha$-motoneurons. Thicker lines represent cumulation of the thinner pathways.
• Figure 2: Modeling a) synergies and b) reflex Geyer2010 of select muscles of the lower leg. Synergies model contains a primitive which receives the gait phase information and a weight block which receives sensory input to adjust the weights. In the reflex model, gastrocnemius and soleus muscle activations during the stance phase result from a positive force feedback loop derived from the respective muscles. Tibialis anterior activation results from offset length feedback and suppressive force feedback from the soleus muscle. Feedback loops are saturated between 0 and 1.
• Figure 3: Possible approaches to integrate neural models and musculoskeletal models to control wearable assitive robots.