Study On The Key Technologies Of Driving Behavior Recognition Facing To Automotive Active Safety

The usual automotive safety measures are passive security and defense, which can only reduce the injury and loss caused by accidents. Automotive active safety technology is aimed to avoid accidents or reduce the possibility of the accidents and stifle the dangerous behavior in driving intention. In this paper, the analysis of driving behavior and intention was made by collecting a variety of drivers'operating signal and environmental information. The weight relationship for identifying the typical driving behavior and intention was built according to the different proportion of variety operating information on identifying common intention of driving behavior and intention. The typical driving behavior and intention can be judged and the dangerous behavior can be predicted or reminded by the weight relationship. Through programming emulation with Matlab, the weight relationship under the eight typical given state of driving was established.In this paper, the basic idea of active safety technology is presented, and the active safety technology involved in"human"and"vehicle","environmental"factors and the interactions between them is exposited. The driving and driver intent is considered by acquaintance of vehicles environmental information. The composition of road traffic system and the human role or status in the system is discussed of, and the importance of person in the active safety analysis is presented. From the three types of general behavior patterns, the general behavior of people is discussed. A brief overview and research of the driving behavior is given.Have done some research on sensor technology and camera calibration, including the methods of various data collection, image processing, lane detection and the detection of road occupation. Resolved the problems that how can the active safety system provide the basis of decision-making by analysis the data which were collected and extracted.Established a human-centered behavior and intentions based on the person driving– automobile– environmental active safety system model. The model emphasizes the importance of driving behavior and intentions and tries to achieve the goal of making early warning of security and safe operation. Realized the automotive active safety by doing the research of overall framework of the system. After that, the early warning information can be made by the system more close to the idea of"driver-centric".Finally, through the simulation of artificial neural network research, the article gives the result of the weight relationship about signal matrix which comes from vehicle status and environment. The result can be used to judge the driver's driving behavior and intentions successfully. A testing experimental platform based on Arm-Linux for the system model was designed. The experiment simulates the functions and the early warning capabilities of the active safety system successfully.