S Plane Control of Underwater Vehicle and Its Hybrid Training Algorithm

Li Ye,Tang Xu-dong,Pang Yong-jie
DOI: https://doi.org/10.1109/ccdc.2009.5192715
2009-01-01
Abstract:For the particular control object of automatic underwater vehicles (AUV), an S Plane Control based on the analysis of fuzzy control and combined with the form of PID control is introduced which is a simple and effective method, but parameters must be manually set for its controller. In order to improve the adaptability of AUV, the adaptive S Plane Control algorithm based on single neuron cell is proposed as well, in which parameters self adjustment come true. But the neural network has many defects, such as learning slowly, easy get into local convergence and learning results depend on initial condition. This improved method is proposed on which improved PSO algorithm is used offline first for optimizing which can avoid the phenomenon of precocity and stagnation during evolution, followed by online adjustment with single neuron cell algorithm. The feasibility and advantages of this method are demonstrated by simulation test results.
What problem does this paper attempt to address?