Abstract: Common motion control chips employ traditional PID control strategies, which cannot meet high-performance requirements. This paper adopts a single-neuron control strategy, modifying the traditional PID control parameters and methods to construct a single-neuron PID control system. Practice has proven that the improved system has stronger adaptive capability to parameter changes and exhibits strong robustness. Keywords: neuron; PID; DC servo 1 Introduction Motion control chips provide a complete PID control strategy, which can be easily interfaced with a computer to form a multi-axis motion control system, thus gaining wide application in control engineering, especially in robot control. However, classic PID control cannot meet the needs of applications with varying system parameters. This paper takes the commonly used LM628 motion control chip as an example, pointing out that the host computer provides the possibility of PID parameter setting for the LM628 (which actually provides the implementation of the single-neuron control strategy). The paper uses a simple method to transform the traditional PID control strategy into single-neuron PID control and achieves good results. 2 Classic PID Position Servo System The classic PID position servo system is shown in Figure 1. Figure 1 consists of four parts: a host computer, an LM628 control card, drivers, and servo motor and position sensor components. In robot control systems, many similar position servo drive axes are often used simultaneously. The environmental parameters of different drive axes are different. Debugging multi-axis robot systems is difficult because the control parameters and their ranges for each axis require repeated adjustments to determine their final values. To improve the adaptability or robustness of the control system for each axis to environmental parameters, it is desirable to transform the control strategy of each axis into a neural network control strategy. [Click for details]