Permanent magnet synchronous motor (PMSM) is one of the most viable motion control products due to the inherent benefits of low rotor inertia, high efficiency and high-power density in industrial applications. The speed control for PMSM is a significant task. Many researchers carried out their research for improving performance of PMSM through speed control. But, the performance and efficiency of PMSM are reduced due to the external load disturbances and parameter deviation like nonlinear, time-varying, strong coupling of PMSM. In order to address these problems, Particle Swarm Maxpooling Fully Connective Deep Convolutional Neural Learnt Sugeno-Takagi Fuzzy Controller (PSMFCDCNLSTFC) model is introduced. The key objective of PSMFCDCNLSTFC model is to regulate the speed of PMSM for obtaining the highest current value. PSMFCDCNLSTFC model comprises two processes, namely Particle Swarm Weight and Hidden Neuron Optimization process and Maxpooling Fully Connective Deep Convolutional Recurrent Neural Network-based Takagi-Sugeno Fuzzy Controller process. In the former process, weight parameters and number of hidden neurons are optimized to design efficient deep convolutional neural network. In the latter process, four layers are used to regulate the speed of PMSM through Takagi-Sugeno Fuzzy Controller. After that, the soft sign activation function is used to find the minimum mean square error for attaining the rated current value of PMSM. Finally, the performance of PMSM gets improved. The performance of PSMFCDCNLSTFC model is performed with PMSM data and measured in terms of rise time, peak time, peak value, peak overshoot and settling time. The simulation results show that the PSMFCDCNLSTFC model increases the performance of PMSM with higher output current value when compared to state-of-the-art works.

永磁同步电机 (PMSM) 是最可行的运动控制产品之一，因为它在工业应用中具有低转子惯量、高效率和高功率密度的固有优势。PMSM 的速度控制是一项重要任务。许多研究人员开展了通过速度控制来提高 PMSM 性能的研究。但是，由于PMSM的非线性、时变、强耦合等外部负载扰动和参数偏差，PMSM的性能和效率降低。为了解决这些问题，引入了粒子群Maxpooling全连接深度卷积神经学习Sugeno-Takagi模糊控制器（PSMFCDCNLSTFC）模型。PSMFCDCNLSTFC 模型的主要目标是调节 PMSM 的速度以获得最高电流值。PSMFCDCNLSTFC 模型包括两个过程，即粒子群权重和隐藏神经元优化过程和基于 Maxpooling 全连接深度卷积循环神经网络的 Takagi-Sugeno Fuzzy Controller 过程。在前一个过程中，优化权重参数和隐藏神经元数量以设计高效的深度卷积神经网络。在后一个过程中，使用四层通过 Takagi-Sugeno Fuzzy Controller 来调节 PMSM 的速度。之后，使用软符号激活函数来寻找达到 PMSM 额定电流值的最小均方误差。最后，PMSM 的性能得到改善。PSMFCDCNLSTFC 模型的性能由 PMSM 数据执行，并根据上升时间、峰值时间、峰值、峰值过冲和稳定时间进行测量。