A revolutionary bionic hand has been developed by scientist that can see objects and instantly decides what kind of grip to adopt. This smart prosthetic hand can be taught to grab objects including a cup and a television remote.
This will be the first time a robotic hand can select from four types of grips depending on whether it is picking up an apple or holding a pen between thumb and forefinger.
In order for a person to pick up an object, he would just have to glance at the object and picture clenching their fist. This technology lets amputees grab objects without thinking.
The bionic hand is fitted with a camera that feeds information to an electronic brain. It uses a 99p ($1.30) webcam to work out the grasp needed. The computer automatically assesses the shape and size of the object within milliseconds to give correct movements whether a light pinch or firm grip needed to pick it up an object. The bionic hand is 10 times faster than a traditional artificial hand.
In order to give precise gripping action, the system must learn up to a dozen of trigger actions. The new system is based on artificial intelligence that involves teaching a computer how to recognize the grip needed for different objects. The computer is not just matching an image; it is learning to recognize objects and grouping them according to the grasp type the hand needed to successfully pick up the object.
Dr. Kianoush Nazrpour, co-author of the study on the prosthetic hand, said using prosthetic arm and leg seem slow and cumbersome to some amputees, the new technology can react without thinking.
Biomedical engineers at Newcastle University trained a computer to recognize nearly 500 objects photographed from more than 70 different angles. This will allow it to decide which grip it will use whether a pinch between the thumb and forefinger or the palm wrist to pick up a television remote.
This new technology is expected to be available for NHS patients within two years.