Robot control method based on Bayesian classifier

Abstract

The invention provides a robot control method based on a Bayesian classifier. The robot control method comprises the steps that surface electromyography signals generated by arm movement are collected, and the collected surface electromyography signals generated by arm movement are subjected to amplification filtering preprocessing; the preprocessed surface electromyography signals generated by arm movement are divided into a training set and a testing set; the characteristics of the surface electromyography signals of the training set and the testing set are obtained; according to the characteristics of the surface electromyography signals of the training set, the prior probability of the characteristics of the surface electromyography signals of the training set is obtained; and the Bayesian classifier is utilized for conducting movement mode classification on arm movement. According to the robot control method, aiming at the characteristics of the surface electromyography signals, asurface electromyography signal collecting scheme is designed, the surface electromyography signals amplified by 240 times are obtained, and the randomness of the surface electromyography signals isreduced. On the basis of analyzing the corresponding relation between upper limb movement and related muscles, the biceps brachii and the triceps brachii of the upper limb on the right side are determined as the collecting positions of the two-channel surface electromyography signals, and the surface electromyography signals with good consistency are obtained.

Description

technical field [0001] The invention relates to a robot control method. In particular, it relates to a robot control method based on Bayesian classifiers. Background technique [0002] The advantages and disadvantages of the EMG control method mainly depend on the two evaluation indicators of recognition accuracy and calculation speed. [0003] Among related technologies, the more representative technologies include the myoelectric control method based on threshold decision-making and the myoelectric control method based on pattern recognition. [0004] The basic principle of the EMG control method based on threshold decision-making is that after the EMG signal is corrected, filtered and modulated, a corresponding signal peak is generated for a contraction of the muscle. Through the comparison between the peak value and the threshold value, the output grasp or Stretch. This method is relatively simple and runs fast. However, when the number of degrees of freedom is more t...

AI Technical Summary

Problems solved by technology

[0007] Although the myoelectric control method based on pattern recognition has been greatly improved in terms of recognition accuracy and calculation speed, there is still room for improvement.

Images