Robotics & AI
There are over 3 million people in the world who have lost at least one of their hands. Robotic prosthetic hands exist, but biologically equivalent robotic hands for amputees are still science fiction. My team seeks to change this and restore hand dexterity in amputees by giving robotic hands a brain of their own. Then amputees can signal for their robotic hand to grasp, and the hand figures out the best way to pick up the object in its path.
We have invented a new robotic hand that is both soft and strong. Similarly to a human hand, the robot hand has skin that enables it to be flexible and grippy and internal bone-like joints that provide structure and strength. Ten embedded smart motors drive the hand to allow realistic movements. A camera is integrated into the palm so that the hand can see and perceive objects in its path.
Borrowing from techniques used to train self-driving cars, we modeled our robot hand in video game simulation software to teach it to grasp objects. The software simulates real word physics so we can run realistic simulations millions of times in a small fraction of the time it would take in real life. We are implementing the latest in artificial intelligence methods inside the virtual environment to train the hand, thus giving it a mind of its own.
We are on the 4th generation of the robotic hand design now. Experiments are in progress and showing promising results.
2020 Poster Competition Winner in Georgia Tech’s Career, Research, and Innovation Development Conference
2019 Poster Competition Winner in Georgia Tech’s Career, Research, and Innovation Development Conference