A recent study revealed that the combination of artificial intelligence technologies and advanced proximity and pressure sensors allows a commercial bionic hand to hold objects in a natural and intuitive way, which contributes to reducing the mental burden borne by amputees. By training an artificial neural network for different grip positions, each finger was able to independently recognize objects and automatically move to the correct position, which enhanced grip stability and accuracy.
The results showed that the participants were able to complete daily tasks, such as lifting cups and picking up small objects, smoothly and with much less mental effort and without the need for intensive or long-term training.
Study details
When stretching out a hand to take a cup, a pen, or even to shake someone's hand, a person usually does not need to think about the position of each finger separately to make a proper fist. But the loss of this innate ability is one of the most prominent challenges faced by users of artificial arms and hands. Even the most sophisticated robotic limbs make everyday activities an additional mental burden; the user is forced to deliberately open and close the fingers around the target body.
To address this problem, researchers at the University of Utah (University of Utah) have developed a solution based on artificial intelligence; they integrated proximity and pressure sensors into a commercial bionic hand, and then trained an artificial neural network on different holding positions, which helped simulate the natural and intuitive way a human grasps objects. When this AI-powered system was turned on, participants showed higher grip stability, better accuracy, and less mental effort.
Participants were also able to perform a wide range of everyday tasks, such as picking up small objects and lifting a cup, using multiple grip patterns, all without the need for lengthy training or practice.
The study was led by Dr. Jacob A. George, in addition to Marshall Trout, is a researcher at the laboratory (Utah neurobotics) Utah Neurobotics Lab. The results of the study were published on December 9, 2025 in the journal Nature Communications.
Although bionic arms have become closer to reality in terms of form, their control is still complicated and not intuitive,”Marshall Trout said. About half of users give up their prostheses, and the reason for this is often poor control systems,he added.
One of the main problems with most commercial bionic arms and hands is that they cannot simulate the sense of touch that normally provides intuitive gestures for grasping objects. The second problem is that the brain has unconscious models that predict and simulate natural hand interactions with objects, which makes the bionic hand unable to fully simulate natural hand movement.
To address the first problem, Utah researchers equipped an artificial hand produced by TASKA Prosthetics with custom finger tips. In addition to pressure sensors, these limbs were equipped with optical proximity sensors designed to simulate the most precise degrees of touch, so the fingers were able, for example, to sense a very light cotton ball falling on them.
As for the second problem, the researchers trained an artificial neural network model on proximity data, which enabled the fingers to automatically move to the exact distance necessary to form an ideal grip around the body. Since each finger has its own sensor and independent ability to sense the object in front of it, the fingers all work in parallel to form a firm and perfect grip on any object.
But even after solving these problems, a challenge remained related to how to control the method of gripping, so what if the user does not want to hold the body in that way that the sensors have determined Or he wanted, for example, to open his hand to drop the body
To counter this, researchers have developed an approach based on sharing control between the user and the AI agent, while finding a delicate balance between human and machine control.
The researchers also conducted experiments on four participants with an amputation of the arm between the elbow and the wrist. In addition to improving their performance on standardized tests, they tried to perform multiple daily activities that required fine motor control. Even simple tasks, such as drinking from a plastic cup, can be very difficult for an amputee; light pressure can cause the cup to fall, and excessive pressure can cause it to break.