An Efficient Procedure For Activating Bi-State Actuator Arrays Using Neuro-Fuzzy Network

A novel approximation procedure based on hybrid neuro-fuzzy (NF), called Takagi-Sugeno multi-real-input multi-binary-output (TS-MIMO) neuro-fuzzy network, is proposed to control the bi-state actuator arrays. This efficient procedure is used in static force mechanism of 2D-rigid body manipulator. The proposed NF model employed off-line neuro-mechanism using Levenberg-Marquardt algorithm (LMA) and Takagi-Sugeno model as inference system in fuzzy systems. Additional hill climbing method as optimal searched procedure improved the minimum error in the learning computation. Simulation results are provided not only to demonstrate the efficiency of the NF model in activating the bi-state in arrays, but also to explain how to find the minimum number of actuators that should be put in the rigid body system. These first results will lead the NF mechanism as a universal force control for bi-state actuator arrays.