The Research Of Abnormal Driving Behavior Recognition Technology Based On Vehicle Dynamic Monitoring Data

With the development of road transport industry, the number of operational vehicle is growing rapidly year by year. However, at the same time, serious traffic accidents usually happen in China. According to the analysis, the main reasons of traffic accidents are speeding, fatigue driving, and other illegal driving behaviors. For a long time, because of the low informationization level of transportation enterprise or industry administration, and lacking the support technology of analyzing for a large amount of data, supervisors cannot analysis drivers’driving behavior in real time and effectively. So that, the safety management will be lacking of pertinence and effectiveness.At present, finding abnormal driving behavior usually needs other type of data such as video data or driver’s physiological parameters data. The method need extra physiological parameters sensors to get the data. However, the cost of those systems are much larger, and the technology is not fully mature. According to the relevant laws and actual condition in China, the research of identification, statistic and analysis for abnormal driving behavior using the basis satellite positioning data back from dynamic monitoring platform is relatively few. The dynamic monitoring platform can give alarm based on satellite positioning data. But there are still some limitations on abnormal driving behavior. And the recognition mechanism is relatively simple, the accuracy still need to improve.According to the above problems, the abnormal driving behavior recognition algorithms have been researched in this paper. The main work are as follows:(1) The paper proposes a new fatigue driving behavior recognition algorithm based on time schedule of the industry. The algorithm can calculate the rest time and effective driving time of drivers based on the satellite positioning data. So, the algorithm can find whether fatigue driving behaviors have been happened in the law.(2) The paper proposes a new recognition algorithm respectively for snap acceleration and deceleration behavior. The algorithm can identify the snap acceleration or deceleration behaviors, and combine two behaviors that the time intervals are tiny. Then, the algorithm can calculate various kinds of parameters.(3) The paper proposes a modified algorithm based on the existing algorithm of speeding behavior and not according to stipulations route driving behavior to enhance the anti-jamming ability from noise point because of the coordinate drift.(4) After researching of the algorithm theory, this paper uses Java related programming technology to imply the algorithm. Using the typical simulation data and real satellite positioning data from the dynamic monitoring platform the paper tests algorithms, and verifies the effectiveness of algorithms. The paper also gives the cost time of the algorithm.(5) The paper integrates all algorithms together, and set up an abnormal driving behavior identification system. Then, the abnormal driving behavior identifying system has been illustrated and displayed in the paper.After completing all research, the paper has concluded the research. By finding the merit and demerit, the future work is proposed.