A research team from the University of Utah in the United States has proposed a new method of bionic hand control that integrates artificial intelligence (AI), which is expected to significantly reduce the brain burden during the use of arm prosthetics. The relevant results were published in the latest issue of the journal Nature Communications. Although the current high-end replica hand is very close to a real arm in appearance and driving method, the wearer still needs to deliberately control the opening and closing of the fingers and the degree of force during use. The lack of direct control and heavy operational burden are the main reasons why nearly half of users ultimately give up prosthetic limbs. One key issue is that most commercial bionic hands are unable to replicate tactile feedback, which is the key source of human intuitive and reflective grasping of objects. To address this issue, the research team installed a customized fingertip module on top of a commercial bionic hand. In addition to sensing pressure, these fingertips are also equipped with optical proximity sensors that can simulate the most delicate tactile sensations. For example, they can even feel the sensation of almost weightless cotton balls falling onto them. In response to the issue of tactile feedback, the research team trained an artificial neural network model using sensor data to enable fingers to automatically move to the appropriate position to form a perfect grip with objects. Due to each finger having an independent sensor that can "sense" the situation ahead, multiple fingers can work in parallel to form a perfect and stable grip on objects. The bionic method created by the research team does not completely hand over control to AI, but adopts a "human-machine sharing" strategy. The user is responsible for issuing the overall intention of grasping or relaxing, while the AI system completes fine adjustments on its own to avoid mutual interference between humans and machines, making the operation process smoother. In the experiment, four participants with wrist amputations under the elbow showed improved stability and accuracy in standard grasping tasks after using the system, and a significant decrease in subjective perceived cognitive burden. More importantly, participants can complete daily actions such as picking up small objects and holding plastic cups to drink water without undergoing prolonged training. (New Society)