抄録
This paper proposes a distributed intelligent controller for multi-legged walking robots. Since living things have nerve center to control basic exercise motion, walking robots should also have such distributed low level controllers to govern the fundamental local motion. The proposed controller tries to simulate the architecture of the nerve center; its control structure consists of multiple controllers, where each controller corresponds to each leg and it determines the motion of its leg by common information and its own local information. The central controller needs not to consider the detail of each leg motion; it only does the global path planning and adjust the total synchronization. Distributed intelligent Control Structure is like a intelligence cluster and consists of Intelligent Clock Generator (ICG), Distributed Intelligent Controller (DIC) and Common Data Field (CDF) . Distributed intelligent control structure generates some gaits naturally. It is also easy to add other function to this system. This system does not depend on number or structure of robot's leg. This paper describes the details of each functions of controllers, and explains that the whole system works well only by joing these functions. Several numerical simulations verify the validty of the proposed control structure.
