Periodic Reporting for period 1 - SMARTSENS (SMARTSENS: Smart wear for sensing the neuromusculoskeletal system during human movement in vivo)
Periodo di rendicontazione: 2023-08-01 al 2025-01-31
Millions of people across the world suffer from neuromuscular disorders due to conditions such as stroke or spinal cord injury. Recovery is often slow and incomplete, primarily because doctors and therapists lack tools to monitor how spinal motor neurons and muscles work together during movement.
Currently, clinicians cannot observe the spinal neuron activity and the resulting forces produced by muscles and tendons during everyday movements. This limits rehabilitation and long-term health monitoring.The SMARTSENS project addresses this major healthcare gap by developing a system that allows continuous and detailed tracking of neuromuscular health. SMARTSENS combines soft sensor-equipped grids, with advanced digital models of the human body, to create a personalized “digital twin” that mirrors a person’s real-time neuromuscular state during daily activities like walking. This opens new pathways for continuous neuromuscular monitor and personalized healthcare. In the future, this technology could allow clinicians to detect early signs of functional decline, monitor recovery progress, and design more effective rehabilitation strategies tailored to each individual. It also offers people living with movement impairments the opportunity to track their own progress and make informed decisions about their care.
SMARTSENS builds upon the European Research Council’s Starting Grant INTERACT. INTERACT achieved breakthroughs in decoding motor neuron activity using high-density electromyography (HD-EMG) and advanced biomechanical models during muscles isometric contractions. SMARTSENS takes this research one step further by enabling this technology to be used during muscle dynamic contractions (with direct validation), to monitor functional movement like locomotion. Moreover, SMARTSENS embedded this technology in wearable systems including soft trunk exosuits and bionic legs to achieved volitional robot control.
Currently, clinicians cannot observe the spinal neuron activity and the resulting forces produced by muscles and tendons during everyday movements. This limits rehabilitation and long-term health monitoring.The SMARTSENS project addresses this major healthcare gap by developing a system that allows continuous and detailed tracking of neuromuscular health. SMARTSENS combines soft sensor-equipped grids, with advanced digital models of the human body, to create a personalized “digital twin” that mirrors a person’s real-time neuromuscular state during daily activities like walking. This opens new pathways for continuous neuromuscular monitor and personalized healthcare. In the future, this technology could allow clinicians to detect early signs of functional decline, monitor recovery progress, and design more effective rehabilitation strategies tailored to each individual. It also offers people living with movement impairments the opportunity to track their own progress and make informed decisions about their care.
SMARTSENS builds upon the European Research Council’s Starting Grant INTERACT. INTERACT achieved breakthroughs in decoding motor neuron activity using high-density electromyography (HD-EMG) and advanced biomechanical models during muscles isometric contractions. SMARTSENS takes this research one step further by enabling this technology to be used during muscle dynamic contractions (with direct validation), to monitor functional movement like locomotion. Moreover, SMARTSENS embedded this technology in wearable systems including soft trunk exosuits and bionic legs to achieved volitional robot control.
SMARTSENS has led to 8 scientific publication outputs, grouped into 2 achievements.
Achievement 1 - Modelling and decoding motor units during locomotion with direct validation:
This involved the development of a novel methodology to decode motor unit activity adaptively during walking. This approach enables non-invasive tracking of individual motor unit discharges from surface HD-EMG signals. The method was validated against intramuscular EMGs [see publication #6]. Based on the availability of in vivo motor neuron activity, the project introduced EMG-driven models of spinal motor neurons that simulate entire motor neuron pools and their recruitment behavior as a function of external EMG inputs. This addresses the problem of incomplete motor neuron decoding in vivo and enables estimation of joint-level and muscle-level forces with higher fidelity, than using experimentally decoded (incomplete) motor neuron activity only [see publications #1-2].
Achievement 2 - Control of ankle exoskeletons, soft trunk exosuits and bionic legs:
Building upon these models, SMARTSENS contributed to the development of predictive control strategies for wearable robots. A key result was the implementation of a model predictive control (MPC) method for leg exoskeletons aimed at managing loads on the Achilles tendon. The system dynamically adjusts ankle exoskeleton assistance to modulate tendon force across time, promoting controlled tendon loading, a key factor in the treatment and prevention of tendon injuries [see publication # 3-4]. SMARTSENS methodologies were used to create adaptive control systems for occupational soft trunk exosuits. These systems reduce lower back load during lifting tasks by continuously adjusting support based on movement intent and biomechanical demand [see publication #5].
The SMARTSENS framework also enabled the design of a speed-adaptive controller for a powered bionic ankle prosthesis. This controller uses decoded neuromechanical information to continuously adjust ankle torque based on walking speed, offering more natural and efficient gait patterns for individuals with lower-limb amputation. The solution demonstrates how wearable neuromusculoskeletal models can translate into improved mobility and autonomy for prosthesis users [publication # 7]. These methods were incorporated into an open source software to maximize scientific impact [publication #8].
Achievement 1 - Modelling and decoding motor units during locomotion with direct validation:
This involved the development of a novel methodology to decode motor unit activity adaptively during walking. This approach enables non-invasive tracking of individual motor unit discharges from surface HD-EMG signals. The method was validated against intramuscular EMGs [see publication #6]. Based on the availability of in vivo motor neuron activity, the project introduced EMG-driven models of spinal motor neurons that simulate entire motor neuron pools and their recruitment behavior as a function of external EMG inputs. This addresses the problem of incomplete motor neuron decoding in vivo and enables estimation of joint-level and muscle-level forces with higher fidelity, than using experimentally decoded (incomplete) motor neuron activity only [see publications #1-2].
Achievement 2 - Control of ankle exoskeletons, soft trunk exosuits and bionic legs:
Building upon these models, SMARTSENS contributed to the development of predictive control strategies for wearable robots. A key result was the implementation of a model predictive control (MPC) method for leg exoskeletons aimed at managing loads on the Achilles tendon. The system dynamically adjusts ankle exoskeleton assistance to modulate tendon force across time, promoting controlled tendon loading, a key factor in the treatment and prevention of tendon injuries [see publication # 3-4]. SMARTSENS methodologies were used to create adaptive control systems for occupational soft trunk exosuits. These systems reduce lower back load during lifting tasks by continuously adjusting support based on movement intent and biomechanical demand [see publication #5].
The SMARTSENS framework also enabled the design of a speed-adaptive controller for a powered bionic ankle prosthesis. This controller uses decoded neuromechanical information to continuously adjust ankle torque based on walking speed, offering more natural and efficient gait patterns for individuals with lower-limb amputation. The solution demonstrates how wearable neuromusculoskeletal models can translate into improved mobility and autonomy for prosthesis users [publication # 7]. These methods were incorporated into an open source software to maximize scientific impact [publication #8].
SMARTSENS has led to breakthroughs in neuromechanical sensing, modeling, and human-machine interfacing.
1) Novel HD-EMG Decomposition for Dynamic Movements:
Prior to SMARTSENS, fully validated decomposition algorithms existed only for isometric (static) contractions, significantly limiting their applicability to real-world human movement. SMARTSENS is the first to validate that individual motor unit discharges—previously measurable only in tightly controlled conditions—can now be extracted during functional, whole-body activities like locomotion. This is a fundamental step forward in human neurophysiology, as it enables, for the first time, the study of how populations of motor neurons coordinate muscle contractions in the context of natural tasks such as walking, running, or stair climbing. This capability is essential for understanding movement disorders, optimizing rehabilitation strategies, and designing assistive technologies that respond intelligently to the user’s neural intent.
2) New Spinal Motor Neuron and Reflex Models for Torque Estimation:
SMARTSENS introduced sensor-driven models of spinal motor neuron pools, capable of simulating how thousands of spinal neurons collectively generate musculoskeletal force. These models integrate information from HD-EMG signals to estimate the activity of individual neurons, including their recruitment thresholds, firing behavior, and reflex contributions. Importantly, this improves the accuracy of reconstructing ankle torque during movement. This is crucial for understanding how the nervous system translates cellular-level activity into joint-level function.
3) Non-Invasive Neural Control of Assistive Robots
SMARTSENS validated these models within fully wearable systems, including soft exosuits, powered orthoses, and bionic leg prostheses. By embedding EMG sensors and neuromuscular models into wearable robots, the project demonstrated real-time, volitional control of assistive devices. This establishes a new class of non-invasive, neural human-machine interfaces (HMIs), capable of decoding intent from spinal neurons and converting it into movement assistance in real-world settings. This development is especially relevant for people with limb loss, neuromuscular diseases, or age-related mobility decline, offering them the possibility of enhanced autonomy and function through wearable neuroprostheses.
1) Novel HD-EMG Decomposition for Dynamic Movements:
Prior to SMARTSENS, fully validated decomposition algorithms existed only for isometric (static) contractions, significantly limiting their applicability to real-world human movement. SMARTSENS is the first to validate that individual motor unit discharges—previously measurable only in tightly controlled conditions—can now be extracted during functional, whole-body activities like locomotion. This is a fundamental step forward in human neurophysiology, as it enables, for the first time, the study of how populations of motor neurons coordinate muscle contractions in the context of natural tasks such as walking, running, or stair climbing. This capability is essential for understanding movement disorders, optimizing rehabilitation strategies, and designing assistive technologies that respond intelligently to the user’s neural intent.
2) New Spinal Motor Neuron and Reflex Models for Torque Estimation:
SMARTSENS introduced sensor-driven models of spinal motor neuron pools, capable of simulating how thousands of spinal neurons collectively generate musculoskeletal force. These models integrate information from HD-EMG signals to estimate the activity of individual neurons, including their recruitment thresholds, firing behavior, and reflex contributions. Importantly, this improves the accuracy of reconstructing ankle torque during movement. This is crucial for understanding how the nervous system translates cellular-level activity into joint-level function.
3) Non-Invasive Neural Control of Assistive Robots
SMARTSENS validated these models within fully wearable systems, including soft exosuits, powered orthoses, and bionic leg prostheses. By embedding EMG sensors and neuromuscular models into wearable robots, the project demonstrated real-time, volitional control of assistive devices. This establishes a new class of non-invasive, neural human-machine interfaces (HMIs), capable of decoding intent from spinal neurons and converting it into movement assistance in real-world settings. This development is especially relevant for people with limb loss, neuromuscular diseases, or age-related mobility decline, offering them the possibility of enhanced autonomy and function through wearable neuroprostheses.