New robotic legs accurately mimic human walking

(CBS News) The non-stop march of scientific progress gets a fresh set of legs thanks to research from the University of Arizona. Roboticists there have constructed what they call the most biologically-accurate robotic legs ever.

In a study published in the Journal of Neural Engineering, the team describes the science behind the robot. As the journal's title suggests, the advancements come mostly in the form of sensors that more accurately mimic human neurophysiology. The robot is able to replicate the central pattern generator (CPG) - a cluster of nerve cells in the lower back that generates rhythmic muscle signals to facilitate walking.

The robot itself is a complex creation of motors and sensors. Artificial "muscles" made of Kevlar straps are controlled by the motors and regulated by sensors, mainly the CPG system.

"This robot represents a complete physical, or 'neurobotic' model of the system," researchers wrote in the study.

By focusing on the neural side of walking, the University of Arizona team approached the process of walking from a new angle. The simplest form of CPG is known as the "half-center" system - a simple rhythmic system of two neurons which alternate moving the legs.

The study suggests human babies are born with this "half-center" system, and the team set out to replicate that in robotic form.

"Interestingly, we were able to produce a walking gait, without balance, which mimicked human walking with only a simple half-center controlling the hips and a set of reflex responses controlling the lower limb," Dr. Theresa Klein, who worked on the study, told BBC News.

Advancements in robotic walkers could more accurately model how babies learn to walk. The research could also aid in spinal-injury treatment.