Smart knee exoskeleton delivers rehab and motion assistance in real time
The device was created to address the growing global need for accessible, scalable rehab solutions—particularly in underserved regions, overburdened healthcare systems like the NHS, and for individuals without consistent access to medical professionals, said Yassin El Salamouny, who was a final-year Robotics Engineering student at UWE Bristol at the time.
His invention stands out for its combination of automated calibration, muscle-adaptive control, natural biomechanics, and wearable usability, all achieved through a novel, lightweight design that operates safely and independently in real-world rehabilitation scenarios.
The research was one of the winning poster presentations at the Minoritised Life Scientists Future Forum at the ICC in Birmingham in April 2025.
Public health burden
“Musculoskeletal conditions—particularly knee osteoarthritis and knee inflammation—represent a major public health burden in the UK. These two conditions alone account for a significant proportion of rehabilitation needs, with a combined estimated occurrence rate of nearly 25 people per minute,” he said.
“Additionally, other conditions such as stroke recovery, ACL/MCL injuries, muscular dystrophy, and SCI can all benefit from a knee exoskeleton at some point in their treatment or recovery process.
“Despite this demand, current solutions like passive braces or physiotherapy often fail to offer real-time adaptability, are not user-intent driven, and typically lack comfort or customizability—especially for elderly users or those undergoing unsupervised recovery.”
Limitations in exoskeleton design
Key limitations in existing knee exoskeleton designs include misalignment with the natural joint axis, bulky or tethered configurations, open-loop control, and limited sensor integration. These flaws hinder patient compliance and reduce rehabilitation effectiveness.
Yassin’s work directly addresses these issues through an adaptive, lightweight, and biologically inspired system that provides:
- Real-time EMG + IMU closed-loop control
- Antagonistic dual-motor string-based actuation for natural movement
- A compact, low-power PCB (printed circuit board) with fully wireless operation and auto-calibration
- A hybrid HRI system with audio guidance for intuitive use
This combination enables safe, personalized, and accessible rehabilitation—bridging a critical gap in current assistive technology.
Detecting muscle intent in real time
“In 2025, I designed and developed a smart EMG-controlled knee exoskeleton at UWE Bristol that goes beyond traditional assistive devices. What makes this work publication-worthy is its dual capability as both a rehabilitation and assistance system, enabled through an intelligent closed-loop control architecture and fully automated calibration pipeline,” Yassin explained.
“The system detects the user’s muscle intent in real time using surface EMG signals and provides precisely the level of support needed to complete the motion.
“Crucially, the automated calibration and setup process is fully completed in under 90 seconds, guiding the user through audio prompts. Once initialized, the device monitors changes in muscle strength over time, allowing it to dynamically adjust assistance levels. As strength improves, it automatically reduces support—providing continuous, adaptive rehabilitation without needing a clinician.”
Mimicking natural muscle behaviour
Mechanically, the system uses a dual-motor string-driven configuration that mimics natural antagonistic muscle behaviour.
This avoids the common design flaw of placing bulky motors to the side of the knee, which risks joint misalignment, discomfort, and even injury during use. The chosen layout provides safer, more biomechanically accurate motion, enhancing both comfort and functional realism.
The device also includes built-in safety mechanisms, such as mechanical stops that prevent knee movement beyond the natural anatomical range (0° to 120°), ensuring joint protection even in unexpected scenarios. Performance testing shows the system achieves angular tracking accuracy within ±10° and a response delay of under 2 seconds, making it suitable for real-time motion therapy and gradual muscle reactivation.
Reducing physio needs
This approach significantly reduces the need for traditional physiotherapy sessions, offering daily, passive rehabilitation integrated into everyday activities—without overloading weakened muscles as some aggressive rehab programs do.
It is particularly suited to patients recovering from knee osteoarthritis, inflammation, stroke, or neuromuscular conditions, and has the potential to expand with stronger actuators to support standing, sitting, or even running in future iterations.
“Ultimately, this system offers a low-cost, self-sustaining alternative to conventional clinical rehabilitation, making it highly accessible in rural or remote areas, where patients often live far from hospitals, clinics, or physiotherapy centres,” Yassin said.
Scalable solution
“It holds particular value in countries with limited access to medical professionals or advanced healthcare infrastructure, providing a scalable solution that delivers continuous, personalized therapy without requiring ongoing clinical supervision.
“Within national healthcare systems like the NHS, where rehabilitation services are under immense pressure, this device could significantly reduce the workload on overburdened practitioners by enabling autonomous, at-home therapy. By transforming rehabilitation into a wearable, everyday experience, this system closes the gap between assistive technology and long-term recovery—bringing advanced care to the people who need it most, wherever they are.”
Yassin admits he was surprised by how much automated calibration improved both the ease of use and consistency of the results.
“Early manual calibration attempts were unreliable, time-consuming, and error-prone. Switching to fully automated, voice-guided calibration not only streamlined the process but significantly improved the robustness of the angle predictions and overall user experience,” he said.
Accessible experience
“This project directly addresses the global challenge of delivering consistent, effective knee rehabilitation to patients who lack access to traditional physiotherapy—whether due to financial barriers, geographic isolation, or overstretched healthcare systems. By enabling real-time, personalized support based on the user’s actual muscle condition, the exoskeleton transforms rehabilitation from a clinic-bound activity into an accessible, daily wearable experience.
“It reduces dependency on medical professionals for routine therapy, allowing patients to recover independently at home while still benefiting from responsive and adaptive treatment. This has significant implications for public health systems like the NHS, where rehabilitation demand far exceeds available resources, and in developing countries or rural areas where medical infrastructure and expertise are limited. Ultimately, the device helps close the care gap—empowering recovery, reducing costs, and improving quality of life for millions who would otherwise go underserved.”
Next steps
Yassin also outlined the next steps that are needed to build on his work.
1- Real-Time Data Analysis Interface:
While the current system logs all session data for offline review, future development should introduce a real-time graphical user interface (GUI) for analysis purposes. This tool would enhance the research & development phase—not end users—by visualising EMG signals, IMU angles, and predicted targets live, supporting model debugging, system tuning, and future clinical evaluations.
2- Mechanical Fit and Comfort Enhancement:
The current rigid design can be further refined by adopting a hybrid mechanical structure combining soft and hard materials. This would improve comfort, allow the device to better conform to the user’s body shape, and reduce localized pressure during movement—making the exoskeleton more suitable for extended wear in daily rehabilitation use.
3- Improved Wireless Connectivity for Broader Accessibility:
While the exoskeleton operates fully as a standalone device, a wireless connection to a laptop is currently used for optional data logging and real-time monitoring. To enhance portability and user accessibility, future versions should replace this laptop-based monitoring system with a dedicated mobile or desktop app. This would retain wireless functionality while simplifying the interface—especially for non-technical users—without compromising the device’s core functionality or independence.
Robotics engineering dissertation
The project is Yassin’s individual final year (dissertation) project of his bachelors of robotics engineering study and is supervised by Dr. Appolinaire Etoundi, Senior Lecturer in Mechatronics, Head of the Mechatronics Lab at UWE Bristol & the Programme Leader for Mechanical Engineering.
It was presented at the first ever MLSFF (Minoritised Life Science Future Forum) conference through a poster presentation and a presentation talk at ICC Birmingham between 31st of March & 2nd of April. Yassin also presented it through another poster presentations at the DMHLS (Decolonising Research in Medicine , Health, & Life Sciences) conference at the University of Exeter on the 7th of May, at UWE’s School of Engineering 2025 Degree Showcase on the 5th of June and at UWE’s School of Engineering Advanced Engineering & Aerospace Employer Showcase on 20th November.
Yassin is currently pursuing his masters (MSc) in Health Technology at UWE Bristol as a continuity of his academic journey.
Registrations is now LIVE for #MLSFF26 in Edinburgh (23–25 March 2026) - visit mlsfutureforum.com to find out more.
