| With the development of human-computer interaction,gesture recognition technology based on data gloves has attracted more and more attention.This method has the advantages of less environmental interference,high recognition rate and strong real-time.In this paper,the key technologies of gesture recognition based on data gloves are deeply studied.According to the idea of software engineering,each module is designed to realize real-time gesture recognition and display of data gloves.The main work and achievements of this paper are as follows:(1)Wireless data glove communication based on RS-232 serial port and WiFi is realized.That is,based on the RS-232 serial port protocol and WiFi wireless protocol,the RS-232 interface of the computer is programmed to receive WiFi wireless information from the MCU on the data glove.(2)A large number of experimental data were collected to complete the calibration of data glove sensor data and the conversion of corresponding Angle values.(3)The most effective part of gesture feature extraction and the selection of threshold value were determined through the normalization and filtering pretreatment of experimental data.(4)Different algorithms for gesture recognition of data gloves are realized,and their effects are compared.Firstly,matlab calculation package of radial basis function is used to realize gesture recognition by adjusting the central value and variance of radial basis neural network.Then the gesture recognition based on BP(Back Propagation)algorithm is realized through the C++program written by myself.Finally,by comparing the gesture recognition results based on the traditional method(without using any algorithm),radial basis neural network algorithm and BP algorithm,the most effective BP algorithm is selected to realize the real-time recognition of gesture of data gloves. |
