A Novel Hybrid Control Method for the Underwater Manipulator

Abstract

Manipulator is the most important part for underwater flexible robot manipulation. It is difficult to design the underwater manipulator control system due to the particularity of the work environment, high nonlinearity and strong coupling of the dynamic model. In order to improve fine performance capacity, a novel hybrid control method based on the cerebellar model articulation controller is proposed. The neural network structure of the fuzzy CMAC was designed, and the input and output formula of each layer was deduced. Based on BP arithmetic, the associative strength, core and width of Gaussian membership function were obtained through training. Simulation experiments have been developed to verify the effectivity of the novel hybrid control method. The results show that the novel method has an excellent system performance including high precision trajectory tracking ability, good real-time performance and strong anti-interference. The novel hybrid control method can be used in the underwater manipulator control system.