Last Updated on July 7, 2026 by Staff
Millions of people around the world struggle with severe hand movement problems caused by conditions such as stroke, spinal cord injuries, nerve damage and neurodegenerative diseases like amyotrophic lateral sclerosis (ALS). For many patients, even simple daily activities such as eating, holding a cup, buttoning a shirt or picking up small objects become nearly impossible.
Scientists have been searching for technologies that can restore independence and improve quality of life for these individuals. Now, researchers from the Technical University of Munich and the rehabilitation center Passauer Wolf have developed a lightweight soft robotic exoskeleton that may provide a breakthrough.
The new wearable device, described in Nature Machine Intelligence, combines soft robotics, muscle-sensing technology and artificial intelligence (AI) to help people with severe hand paralysis regain the ability to grasp and manipulate objects.
Soft Robotic Design
Unlike traditional robotic exoskeletons made from rigid materials, the new system is built using lightweight fabric that resembles ordinary clothing.
The glove-like device is manufactured using a standard sewing machine and is powered by air pressure rather than heavy electric motors. This makes it flexible, comfortable and much safer for patients to wear during daily activities. Because the exoskeleton is soft, it naturally adapts to the shape of the user’s hand instead of restricting movement. This reduces discomfort while allowing more natural finger and hand motions.
The lightweight design also makes the device easier to wear for longer periods, increasing its potential for everyday rehabilitation and personal use.
AI Reads Intentions
One of the biggest challenges in assisting people with severe paralysis is understanding when they actually want to move.
The researchers solved this problem by equipping the exoskeleton with sensors capable of detecting extremely weak electrical signals produced by the user’s remaining hand muscles. Although these muscle signals may be far too weak to create movement on their own, they still contain valuable information about the person’s intended action.
An artificial intelligence system analyzes these tiny muscle signals in real time. After recognizing the user’s intention to grasp an object, the AI immediately sends commands to the soft robotic glove, which performs the required hand movement. This allows users to remain in control of their actions rather than having the exoskeleton move automatically.
According to the researchers, maintaining user intention was one of the most important aspects of the entire system.
Real-Life Success
The research team evaluated the new exoskeleton with one patient suffering from advanced ALS and six stroke survivors who had severe hand impairments. Participants completed two standard clinical assessments: the Box and Block Test and the Action Research Arm Test (ARAT), both widely used to measure hand function.
The results were especially encouraging for individuals with the most severe disabilities.
One ALS patient, who had been unable to move his right hand for approximately four years, successfully used the exoskeleton to pick up a fork and feed himself a piece of cake. This marked an important milestone because it restored a daily activity that had become impossible without assistance.
Interestingly, researchers found that patients with moderate impairments did not always experience the same level of improvement. The greatest benefits were seen in those who had almost completely lost hand function.
Future Possibilities
The researchers believe this technology represents an important step toward restoring independence for people living with severe motor impairments.
Beyond ALS and stroke, similar systems could eventually help individuals recovering from spinal cord injuries, traumatic brain injuries, nerve disorders and other neurological conditions that affect hand movement.
An important lesson from the project came directly from the ALS patient, who actively participated in improving the system. Rather than allowing the robotic glove to move automatically, he wanted to personally decide when each movement should occur. This feedback helped researchers refine the AI intention prediction system so users remain fully involved in controlling their own actions.
The research team is already working on expanding this technology. Their next goal is to develop a lower-body soft exoskeleton capable of helping people regain walking ability.
As artificial intelligence, wearable robotics and sensor technologies continue advancing, future exoskeletons may become smaller, smarter and more affordable. Such innovations could transform rehabilitation by allowing people with severe paralysis to perform everyday activities independently, improving both their physical abilities and overall quality of life.
The success of this soft AI-powered hand exoskeleton demonstrates how combining robotics with human intention can create assistive technologies that truly work alongside the user rather than simply replacing movement.