A research spearheaded by scientists at Tokyo Tech (Tokyo Institute of Technology) has discovered new methods of controlling multi-legged robots by using a two-level controller. The projected device employs a network of supposed non-linear oscillators that allows the creation of diverse postures and gaits, which are define by just a handful of high-level parameters. The research motivates new study into how multi-legged bots can be managed, comprising in the future employing interfaces of computer with brain.

In the normal world, most of the species can walk over irregular surfaces and slopes, reaching spots unreachable even to the most enhanced rover bots. It stays a mystery how complicated actions are managed so flawlessly by even the smallest creatures.

What we are aware of is that even the most normal brains have pattern-generator circuits, which are connected up particularly for generating patterns for walking. Efforts to artificially duplicate such circuits have so far had restricted success, owing to deprived flexibility.

Now, scientists in Italy and Japan suggest a new method to generation of walking pattern, on the basis of a hierarchical arrangement of electronic oscillators set up over 2 stages, which they have shown employing a hexapod robot similar to an ant. The accomplishment unlocks fresh avenues for the management of legged bots. Posted in IEEE Access, the study is the outcome of teamwork between researchers from Tokyo Tech, partially supported by the Polish Academy of Sciences in Krakow, Poland, the World Research Hub Initiative, and the University of Catania, Italy.

The biologically-stimulated controller has 2 levels. At the top, it contains a pattern-generator circuit, in charge for managing the general progression of leg movements, dubbed as gait. At the bottom, it has 6 local pattern generators, in charge for managing the routes of the separate legs. This together is responsible for a well programmed walk of the robot.