PID control, widely used in industrial automation, is not ideal for controlling objects with nonlinearity, large time delay, and strong coupling. In other words, PID controllers require different parameters for different objects, are inconvenient to adjust, have poor anti-interference ability, and large overshoot. Fuzzy control is limited by its ability to process control actions in gears, is a nonlinear control, has low control accuracy, exhibits static errors, and generally oscillates when the deviation of the linguistic variable approaches zero. Therefore, a composite controller, the fuzzy PID controller, is formed by combining the advantages of PID control (high control accuracy) and fuzzy control (not dependent on the mathematical model of the controlled object, simple algorithm design, easy implementation, can be directly derived from operator experience and optimized, and has good adaptability, strong anti-interference ability, and good robustness). Experiments have shown that it has achieved good results in high-speed servo systems. I. Fuzzy PID Control Principle (I) PID Parameter Fuzzy Self-Tuning Control Principle PID parameter fuzzy self-tuning control utilizes a fuzzy controller to perform online self-tuning of the parameters of the PID controller. The process is as follows: First, find the fuzzy relationship between the three PID parameters, error, and error change rate. During operation, continuously detect the error and error change, and then modify the three parameters online according to the fuzzy control principle to meet the different requirements of the controller parameters under different errors and error changes. The conventional PID control algorithm is: where and represent the input variables deviation and deviation change, respectively, and kp, ki, and kd are the parameters characterizing its proportional, integral, and derivative actions, respectively. A Fuzzy self-tuning PID parameter controller is a Fuzzy controller that, based on the conventional PID controller, applies Fuzzy ensemble theory to establish a bivariate continuous function relationship between parameters kp, ki, and kd with the absolute value of the deviation |E| and the absolute value of the deviation change |EC|: and self-tunes parameters kp, ki, and kd online according to different |E| and |EC|. (II) PID Parameter Fuzzy Tuning Strategy Generally, the self-tuning requirements of parameters kp, ki, and kd for the controlled process under different |E| and |EC| can be summarized as follows: When |E| is large, to ensure good fast tracking performance and avoid differential saturation caused by large instantaneous changes in E, kp should be large and kd should be small. To avoid large overshoot in the system response, the integral action should be limited, usually ki = 0; P control can also be directly adopted depending on the actual situation. When |E| is at a moderate level, to reduce system overshoot while ensuring system response speed, kp should be small; the value of ki should be appropriate; in this case, the value of kd has a greater impact on the system response. Generally, the empirical values ​​are: when |EC| is large, kd can be slightly smaller; when |EC| is small, kd can be slightly larger. PD control can also be directly adopted in practice. When |E| is small, to ensure good steady-state performance, improve system anti-interference ability, and avoid system oscillation, both kp and ki should be large. Meanwhile, to avoid system oscillations near the setpoint, the choice of the kd value is crucial. It can be determined based on |EC|. When |EC| is large, kd can be slightly smaller; when |EC| is small, kd can be slightly larger. II. P-FUZZY-PI Multi-Mode Multi-Segment Controller The P-FUZZY-PI multi-mode multi-segment controller uses different modes for segmented control based on different conditions and requirements. Specifically, when the error exceeds a certain threshold, proportional control is used to improve the system's response speed and accelerate the response process; when the error is less than a certain threshold, fuzzy control is switched to improve the system's damping characteristics and reduce overshoot in the response process; when the error reaches near the equilibrium point, PI control is used, utilizing its integral action to ultimately eliminate the error. III. Application of Fuzzy PID Control in Sewing Machines (I) Introduction to Industrial Sewing Machine Servo Control Systems Industrial sewing machine control systems require precise positioning. A stepless speed-regulating servo system is used. The entire system consists of an electromagnetic clutch motor (also known as an electromagnetic speed-regulating asynchronous motor), functional solenoid valves related to the mechanical parts of the sewing machine, a main control circuit board, a speed feedback board, a foot pedal control command interface board, and a power supply board. Figure 2 shows the system structure block diagram. Among them, the inner closed loop is the speed loop, and the outer closed loop is the position loop. Control algorithm 1 is used for speed regulation of the electromagnetic clutch motor, and control algorithm 2 is used for precise positioning of the electromagnetic clutch motor. The basic speed regulation method of the controlled object electromagnetic clutch motor is PWM speed regulation. The most important technical indicator is positioning accuracy. The maximum speed of a typical industrial sewing machine can reach 5000 rpm. At such a high speed, it is required to position quickly and accurately, and the maximum positioning accuracy cannot exceed -5mm - 5mm. Because the subsequent thread cutting and thread take-up operations depend on its accuracy, if the positioning is inaccurate, the phenomenon of needle punching may occur. (II) Industrial sewing machine system controller algorithm When only PID control is used, the following problems exist: 1. The dead zone is too large. When the integral constant is very small, the control accuracy is poor and the control is not stable. If the integral time constant is too large, it will cause large overshoot, oscillation, long transient response, and difficulty in guaranteeing control accuracy. 2. Poor positioning accuracy. This makes parameter determination very difficult. When using piecewise PID control, the system speed is divided into 5 segments, each with different PID parameters. The control effect can achieve satisfactory results, but it cannot fully meet the accuracy requirements for all points. Moreover, the accuracy becomes worse as the hardware temperature rises. For the above reasons, the final algorithm scheme for the controller is as follows: when the system deviation is small, parameter self-tuning PID control is used; when the system has a large deviation, fuzzy control is switched. The specific control adopts incremental digital PID control. The formula is: Where kp, ki, and kd are the proportional, integral, and derivative coefficients, , and T is the sampling period. During the dynamic response process, based on the characteristic information of the deviation E and the deviation change EC, a corresponding reasonable α change is generated by fuzzy inference and applied to Kp, Ti, and Td. Where γ and β are the adjustment coefficients of proportional gain and integral time, respectively, and their values ​​are given empirically according to different systems. kpo, kio, and kdo are the initial setpoints of the PID controller. Under these conditions, the PID parameters are dynamically adjusted according to the deviation E and the deviation change EC to meet dynamic performance requirements. The online adjustment of α(t) is generated by a fuzzy inference. Based on the current E and EC, the characteristics of the controlled object, and practical experience, a fuzzy variable H is generated, which reflects the fuzzy decision of the change trend that α should have in the dynamic process. After H is defuzzified, h(t) is obtained, and α is adjusted online by it, that is: α(t) = α(t-1) + ηh(t) α(t)∈[0,1] η is the rate of change of α. (III) Comparison of Experimental Results The positioning accuracy of the sewing machine using the three control strategies is compared in Table 1. At this time, the speed of the sewing machine is 2000 r/min. IV. Conclusion The comprehensive experimental results show that: fuzzy PID control can make the system respond quickly and without overshoot, has stronger resistance to load disturbances, and has stronger robustness. After debugging and operation, the fuzzy PID controller has been shown to have the functions of fast response and elimination of steady-state error, thus meeting the positioning accuracy requirements of high-speed servo systems.