Trategy based on human reflexes may keep legged robots and prosthetic legs from tripping

Trips and stumbles too often lead to falls for amputees using leg prosthetics, but a robot leg prosthesis being designed at Carnegie Mellon University promises to support users improve their balance by using approaches based on the way real human legs are operated.

Hartmut Geyer, asst prof. of robotics, explained a control strategy devised by learning human reflexes and some other neuromuscular control systems has displayed promises in simulation and in laboratory testing, producing stable walking gaits over uneven terrain and better recovery from trips and shoves.

Geyer has learned the mechanics of legged walking and motor handle for the past 10 years. Among his observations is the role of the leg extensor muscle tissue, which in general work to straighten joint parts. He says the force feedback from these muscles automatically responds to ground disturbances, quickly slowing leg movements or extending the leg further, as necessary.

Geyer's team has examined the neuromuscular product by using computer system simulations and a cable-driven device about half the size of a human leg, called the Robotic Neuromuscular Leg 2. The leg test bed was funded by the Eunice Kennedy Shriver National Fondation of Child Health & Human Improvement.