Actuator control system and method based on deep learning and deep neural network

Abstract

The invention is an actuator control system and method based on deep learning and deep neural network, and belongs to actuator control technology. The system includes a machine vision recognition module, a PID controller, a motor module, a camera and an upper computer. The machine vision recognition module uploads video information obtained by the camera to the upper computer to obtain image feature data after image processing; the upper computer inputs target angle value of motor, and the PID controller obtains error value according to current actual angle value and target angle. Its change rate and cumulative value are input into the PID neural network with the data of image features, the current actual angle value of the last motor and the output value of the last PID controller. Afteranalysis and calculation, the three coefficients of the PID algorithm are output and input into the PID controller. The output value is calculated and transmitted to the motor to control the angle change of the motor. The invention combines machine vision recognition and PID algorithm, effectively solves the problem of unstable performance of traditional steering gear in various or variable environments.

Description

technical field [0001] The invention belongs to the control technology of steering gear, and in particular relates to a steering gear control system and method based on deep learning and deep neural network. Background technique [0002] In recent years, the development of robot technology has increasingly become the mainstream of modern industry. Among many robot products, the motor, as one of the core and most critical power transmission devices, has received more and more attention. Among them, the most prominent one is digital control. The motor is the steering gear. [0003] The steering gear is mainly composed of a motor, a power output control circuit, an angle detection module, and its internal control algorithm. The working method of the steering gear is: first, the power output control circuit sends a signal to control the rotation direction and output power of the motor, and then passes through the gear The mechanical transmission mechanism of the group is transm...

AI Technical Summary

Problems solved by technology

However, there are disadvantages in such a design, because its internal control algorithm is usually only PID algorithm or fuzzy algorithm, and its algorithm parameters are unchanged, which means that in a nonlinear environment, the change of load will affect the performance of the steering gear. Because the internal control algorithm is only optimized for a specific range of load, but when the load exceeds the range, its accuracy will decrease, the steady-state error will increase, and abnormal jitter will appear

Images