Smart Control of a Soft Exoskeleton with a Textile-based Actuation
Motivation & Goal
Work-related musculoskeletal disorders affect 44 million Europeans, costing the EU 240B € annually and highlighting the urgent need for preventive solutions in Industry 5.0. Active exoskeletons offer a promising solution by enhancing workers’ musculoskeletal capabilities while preserving their agility and productivity.
The SmartExo project aims to develop a functional active exoskeleton prototype with a soft human-robot interface to support the lower back and hip-knee joints, promoting worker well-being and efficiency.
Approach
The project is structured around six technical foundations over 36 months: a fully textile-based garment (exosuit) for a soft human-robot interface; a bioinspired multi-joint force transmission system driven; Shape Memory Alloy (SMA) actuation informed by human musculoskeletal models; active forced-air cooling to improve SMA bandwidth; AI-driven real-time perception of worker intention and physical effort and an assist-as-needed control loop that personalizes assistance continuously based on human state.
Application Scenarios
The primary target is industrial logistics, replicating physically demanding tasks such as squatting, lifting, lowering, and carrying heavy loads. A proof-of-concept validation at laboratory scale will compare four conditions: no assistance, passive exosuit, active exosuit assistance, and rigid exoskeleton assistance. Beyond logistics, the technologies transfer readily to manufacturing, agriculture, construction, healthcare, and aeronautics — wherever physically demanding work and human-robot interaction matter.
Innovations & Contributions
- First fully textile-based active exosuit targeting multiple joints with SMA actuation, offering clothing-like usability without bulky motors or pumps
- SMA actuator design modeled on human musculature, enabling multi-joint actuation with reduced energy consumption
- Active air-cooling to overcome SMA bandwidth limitations and achieve low response times
- AI perception toolchain for real-time estimation of locomotion intention and effort using minimal sensors
- Assist-as-needed, human-in-the-loop control that dynamically balances assistive force with dexterity preservation — adapting to each worker’s age, fatigue level, and task requirements in real time
Grant
This project is funded by national funds through FCT- Fundação para a Ciência e Tecnologia, I.P., as part of project 2023.17737.ICDT. https://doi.org/10.54499/2023.17737.ICDT
