基于小波神经网络的远程教学系统设计  

Design of distance teaching system based on wavelet neural network

作　　者：乔文增[1] 罗玉强 QIAO Wenzeng;LUO Yuqiang(Shanghai Ocean University,Shanghai 201306,China;University of Shanghai for Science and Technology,Shanghai 200093,China)

机构地区：[1]上海海洋大学,上海201306 [2]上海理工大学,上海200093

出　　处：《现代电子技术》2021年第6期34-38,共5页Modern Electronics Technique

基　　金：国家自然科学基金项目(61903254)。

摘　　要：远程教学系统对通信质量要求极高,传统远程教学系统中由于信道失衡,存在通信收敛速度慢、误码数高的问题。为此,设计基于小波神经网络的远程教学系统。通过ATmega128开发板的USART接口实现数据的通信,并在以太网控制芯片上完成通信网络的接入。在硬件设计的基础上,利用DirectShow提供的API函数处理终端产生的音视频数据,从I/O通道进行数据的接收与发送,选择合适的小波神经网络输出层传递函数,经过多次迭代平衡通信信道,完成软件设计。测试结果表明,与传统的远程教学系统相比,设计的基于小波神经网络的远程教学系统收敛速度快、误码数低,适合推广使用。The distance education system has extremely high requirements for the communication quality.In the traditional distance education system,there are problems of slow communication convergence and high bit errors caused by the channel unbalance.Therefore,a distance teaching system based on wavelet neural network is designed.The data communication is realized through the USART interface of the ATmega128 development board,and the access of the communication network is completed on the Ethernet control chip.On the basis of the hardware design,the API functions provided by the DirectShow are used to process the audio and video data generated by the terminal,and the data is received and sent through IO channel;the appropriate wavelet neural network output layer transfer function is selected,and the communication channel is balanced after multiple iterations,so that the software design is completed.The testing results show that,in comparison with the traditional distance teaching system,the designed distance learning system based on wavelet neural network has faster convergence speed and lower bit error number,and is suitable for promotion and popularization.

关 键 词：远程教学系统 小波神经网络 远程通信 数据传输 系统设计 系统测试 

分 类 号：TN926-34[电子电信—通信与信息系统] TP37[电子电信—信息与通信工程]