Intelligent Adaptive Control of Machining Process Based on Hybrid Recurrent Neural Network

Aiming at the feasibility of intelligent adaptive control for a machining process, a new network architecture, called a hybrid recurrent neural network (HRNN) is first presented based on the diagonal recurrent neural network (DRNN). Considering the uncertain information in the machining process, a generalized entropy square error (GESE) criterion is then proposed. The learning algorithm of the HRNN and the mathematic model of the machine tool for experiments are also explained. Finally, the HRNN is applied to the constant force control of the machining process. Simulated results verify the effectiveness of the proposed control schemes. And the experimental results also confirm the applicability of the described controller in practice.