Abstract : This paper proposes a model reference adaptive motor speed identification method based on fuzzy neural network, and combines it with a frequency converter system controlled by SVPWM modulation technology to form a DSP-based sensorless vector control system for asynchronous motors. The structure and hardware/software design are described in detail. Simulation results show that the system has good dynamic performance and can accurately track changes in motor speed. Keywords : Asynchronous motor, Sensorless, SVPWM, Vector control, Digital signal processor 1 Introduction The mathematical model of an asynchronous motor consists of voltage equations, flux linkage equations, torque equations, and motion equations, forming a high-order, nonlinear, strongly coupled, multivariable system. Traditional control strategies result in poor dynamic control performance. Current research on asynchronous motor control focuses on several aspects: employing novel power electronic devices and pulse width modulation control technology; applying vector control technology and modern control theory and intelligent control technology; widely applying digital control systems and computer technology; and sensorless control technology. This paper uses the MS320F240 dedicated motor control chip as the core, adopts a flux and speed closed-loop vector control strategy, and utilizes SVPWM pulse width modulation technology, sensorless operation, and intelligent control technology to design a motor control system. Simulation results show that the control system has strong anti-interference capability and good dynamic performance. 2. Speed ​​Estimation Strategy Model Reference Adaptive Method (MRAS) is a widely used speed estimation method. This paper designs a model reference adaptive speed estimation system that uses an instantaneous reactive power model to reduce the influence of stator resistance. Furthermore, to effectively address the speed instability problem in speed identification when the instantaneous reactive power model reference adaptive method has a negative step speed, its original adaptive mechanism is replaced with a fuzzy neural network with online learning capabilities. [b][align=center]For more details, please click: Design of Sensorless Vector Control System for Asynchronous Motor[/align][/b]