Research And Implementation Of An Intelligent Vehicle Terminal For Vehicular Network

Vehicular network is one of the important research directions in intelligent transportationfield. And the intelligent vehicle terminal is the core of vehicular network. It obtains self andother information to assist drivers with safety driving and improve the active safety ofvehicles. The existing active safety equipments are developed independently by manufactures,which have disadvantages of higher price, nonuniform communication interface and lowpopularity. So, an intelligent vehicle terminal system for vehicular network was researchedand developed. This paper solved several key problems of the system, including:1. An intelligent vehicle terminal for vehicular network was developed. Camera, GPSreceiver, wireless network card were connected to an embedded chip. The terminal usedARM11as the main control unit, with Linux as the operating environment to realize dataacquisition and wireless transmission.2. A method of vehicle safety distance warning based on machine vision was designed.Firstly, the lane line extraction was realized with the five features of the field image, whichare target information distribution, gray, edge, texture and shape. Then, the vehicle targetwithin the lane line was identified using the shadow and rectangular feature under the vehicle.Finally, the distance between the current vehicle and target vehicle was calculated based onthe geometric model of monocular imaging.3. A JPEG codec algorithm of field images based on graphic processor(GPU: GraphicProcessing Unit) was designed, which makes full use of task parallelism and GPU multi-coreresources. The discrete cosine transform and its inverse transform were ported to GPU, whichenhance the real-time performance of the system.4. The problem of mutually information isolation among cars under highwayenvironment was solved. Information was uploaded to the server and could be obtained byother terminals from the server. A back-stage management system was designed, whichachieves the view of scene images, vehicle tracking and trace playback.Experiment results show that the system is stable and reliable. It can meet the needs ofengineering application in vehicle identification, image compression&transmission andvehicle driving information sharing.