Data acquisition system of Internet of Things based on Kohonen multi-protocol adaptation

Abstract

With the development of highway information, according to the actual demand of highway electromechanical equipment management, this paper designs a multi-protocol internet of things data acquisition system to realize instant messaging and data sharing of electromechanical equipment. The Kohonen neural network is used to train the adaptive module of the existing input protocol. According to the eigenvalues of the header, the shaping number of the end bytes and the length of a single packet of the protocol, the function of automatically selecting the appropriate protocol is realized. In the later stage, the ability to learn more protocol adaptation independently can be realized by updating the protocol knowledge base on the internet of things platform. The results show that the average processing time of Kohonen network for each protocol data is about 109ms, and the average recognition rate reaches 95.45%. Kohonen network can be applied in traffic engineering field and realize the conversion of various information data of electromechanical equipment with different protocols into unified information data through protocol conversion rules.