Surface electromyography(sEMG) as a non-intrusive technology is used to estimate muscle strength because it is easy to collect and contains relevant information of human muscles.It has broad research prospects in evaluating and treating muscle diseases.To achieve the accurate estimation of muscle strength

the current research is mainly divided into two categories:one is to improve the sEMG signal processing methods

and the other is to improve the sEMG-muscle strength model.This review summarizes the research development of sEMG muscle strength estimation models in detail.Firstly

the relationship between muscle strength and sEMG signals is summarized.Secondly

the acquisition methods of sEMG signals are summarized from the aspects of sensors and datasets

and the processing methods of preprocessing and feature extraction of sEMG signals at present are analyzed.Then

aiming at the different research methods of sEMG-muscle strength model

it is divided into deep learning

hybrid network and other muscle strength estimation algorithms

and their respective advantages

limitations and practical applications are compared and summarized.Finally

the challenges and future trends of muscle strength estimation are discussed.