Research Of Car Prewarning Middleware Of Iov Based On Multi-agents

With the continuous improvement of people’s living conditions, cars gradually enter the ordinary people’s living. Internet of vehicle(IoV) technology as an important branch of the Internet of things(IoT)has been developing rapidly in recent years. On-board middleware technology is increasingly mature, and more vehicle terminal has installed on new generation automotive products. As a good solution of car traffic safety aspects of the networking, it is becoming vital to design an on-board warning middleware which can meet the demand of modern car networking.Because of the limitation of existing network middleware in perception vehicle data and deficiency of decision fusion, this paper puts forward a kind of vehicle network monitoring middleware based on multi Agent technology which can implement car state perception, judgment and risk early warning. It is a connection of car networking interface layer and application layer. First of all, according to the characteristics of the automobile complex composition and various states, this paper designs the multi-sensor mounted on the vehicle key performance parts by using multi-sensor vehicle data sensing technology. Then Based on the characterics of limited embedded system performance and the demand of real-time on car network, this paper puts forward the multi-Agent technology of distributed system to manage and control the sensors in order to improve the concurrency and system performance optimization. Moreover, this paper makes decisions based on vehicle sensor data processing.In this paper, the decision tree ID3 algorithm of decision level data is used to fuse and provide decision support for early warning in perfect condition which can realize the transformation from vehicle data to the state of vehicle. Eventually the warning function is completed through the interface to the application layer monitoring platform when the vehicle is abnormal in a timely manner.Experiments show that the proposed multi-sensor perceptual scheme can collect data comprehensively and accurately. The high-level decision algorithm can quickly identify abnormal state which greatly reduces the rate of false warning. The middleware has good practicability and meet the need of car networking warning.