Semi-fuzzy CMAC and PD hybrid controller with compressed memory and semi-regularisation for electric load simulator

Cerebellar model articulation controller (CMAC) and proportional derivative hybrid controller is widely used for torque control in electric load simulator. However, due to some uncertain factors and especially the strong external interference of the surplus torque, the control stability could not be guaranteed. Here, a novel semi-regularised semi-fuzzy CMAC is proposed to get an efficient control effect. This study expands the weight smoothing into a semi-regularisation algorithm, and proves the stability of the hybrid control system based on the stability of the torque motor. To save storage space and reduce computation, the memory space is compressed and a semi-fuzzy mapping rule is proposed to smooth the output of CMAC with a high calculation speed. Both of simulation and experiment results demonstrate that this novel controller could keep stable and reach a good loading accuracy.