The Question And Answer System Based On Vehicle Networking Telematics Research

With the popularity of the Internet, people can get a variety of information through a search engine, but many drawbacks of search engines, such as natural language questions instead of keyword query and returns a large number of relevant pages rather than accurate answers, quiz system in order to overcome these drawbacks have emerged, but because of the particularity and complexity, as well as more difficult Chinese characters Chinese information processing technology, so Chinese question answering system coincided with the development of the study period, the application of QA systems also appeared in various fields, the main content of this article is a Q telematics system in the vehicle networking.With the development of vehicle networking, car is more and more applications in the telematics system environment included growing surge, such as roadside assistance, car theft, condition grasp, receive personalized information, multimedia entertainment and information receiving emergency vehicle warning system, POI query, navigation destinations, etc., but if you want to get content, need to directly manipulate related applications, which users, very convenient, especially in the driving process, the existence of directly manipulate application security risks. There is an urgent need to convert directly manipulate, voice control, by understanding the user’s intent, automation-related applications, and to get information back to the user, the process and answering system is very similar, but different from the traditional question answering system, in the automotive environment, quiz system has two unique characteristics: strong field, the application answers and responses related to the connected vehicle; personalized strong, answers and personal habits love problems are related. So in this article the user’s problem into three categories: car machine command issues, information inquiries, vehicle navigation problem. Designed a new car answering system architecture, combined with the unique characteristics of geographic information, make geographic level, geographic level feature words, the concept of word types; unknown words and propose a web search engine based on the judgment and to expand thesaurus finishing; and by geographic level of unified word processing features, through hidden Markov model to calculate the word type, and proposed three categories of problem determination method: question classification model based question classification based on information retrieval based K- nearest neighbor classification problem. Three categories of ways to determine complement each other, a good analysis of the user’s problem by testing a large number of users, the system achieved a very good practical results.