Application Of Human-computer Interaction Technology Based On VR In Wearable Devices

In recent years,VR(virtual reality)devices are more and more popular,and the world's leading technology companies have also laid out the VR market.As one of many intelligent wearable devices,data gloves have become a hot research topic at home and abroad.However,the current data gloves have the shortcomings of low precision,low real-time performance and long training time.The focus of this paper is to design a wireless data glove recognition algorithm.Hand recognition technology is based on RBF neural network,combined with LM(Levinberg)algorithm and GA(genetic)algorithm.Firstly,the traditional neural network is improved by LM(Levinberg)algorithm,supplemented by differential derivation,so that the convergence speed is faster,the real-time performance of data gloves can be enhanced and the training time can be reduced.According to the global characteristics of GA(genetic)algorithm,the weight optimization of neural network is realized,that is,the evolutionary principle of "survival of the fittest" in the traditional sense is replaced by gambling wheel,and the screening of individual data is completed according to the corresponding biological principles.For example,heredity,crossing,variation and so on,the specific steps are to compare the new individuals with other individuals,based on the similarity comparison results,to eliminate the individuals with low fitness,to retain the individuals with good fitness,and to output the optimal solution as a new neural network weight through continuous population recombination,screening,output and new neural network weights,until the accuracy consistent with the setting requirements is obtained.A new GL-RBF neural network algorithm is designed,which has the characteristics of high precision,strong real-time performance and short training time.In this paper,Opengl is used to realize 3D modeling,and the drawing of animation is completed with the help of basic graphic elements.System According to the principle of Opengl illumination,a three-dimensional sphere is drawn.The lighting effect is very beneficial to the simulation of real scene,and the human joint can be clearly simulated.At the end of this paper,the designed data gloves are tested and tested.In order to improve the stability of the system,the sensor data drift,filtering and magnetic deflection angle correction are improved and optimized.in the process of communication between the upper computer and the lower computer,the accurate joint adjustment test and level feedback are carried out to verify the working situation of the whole system.The GL-RBF neural network algorithm in this paper is reasonable,effective and effective,and can reduce the complex coefficient of data processing in gesture recognition to a certain extent.The accuracy,real-time and training time have been greatly improved,and it has not appeared in the display of 3D picture.It's a real gesture.The research work in this paper has achieved good results,and a data glove with higher precision,strong real-time performance and easy to be trained is designed,and the data glove is greatly improved in performance compared with the previous product.