Application Of Weight Fuzzy Identificaiton Method In Vehicle Occupant Classification Algoirthm

The use of automobile airbag has greatly reduced the vehicle occupant casualty ratewhen the collision occurred. The main role of the airbag is, during a collision process,quickly open an air cushion. So that, the occupant can crash into the air cushion to preventthe secondary impact.However, in some special cases, the inappropriate deployment of the airbag will leadoccupant to the unnecessary harm, even to death. This is because the traditional airbag isdesigned according to the average height and weight of adult male and the detonation modeis fixed. When the airbag system is working, the airbag will explode to open the airbag cover.It is likely to impact on the occupant, and children, short stature adults or occupant out ofposition may be got serious injury.With the increasing requirements on vehicle safety, traditional airbag system can notbeen able to meet the modern traffic requirements on safety. Intelligent airbag system canuse different sensors installed in the vehicle, such as pressure sensors, ultrasonic sensors orvideo cameras, to detect the type of the occupant and send this type information to the airbagECU, so that it can dynamically determine the initiation time and inflatable strength of theairbags, providing the best protection for the occupant. Therefore, the accurate identificationof the type of occupant is an important prerequisite for intelligent airbag systems.Today, the world-renowned auto parts suppliers and automobile manufacturers havecarried out extensive research into vehicle occupant classification system，and some relatedproducts have been commercialized. Domestic research into vehicle occupant classificationsystem has been carried out only in some universities and research institutions, and there isno complete design and commercial production. Therefore, it is necessary to carry outin-depth research. Combining with the research and development requirement of Chongqing ChanganAutomobile Co., Ltd. cooperative project “Vehicle Occupant Characteristic IntelligentRecognition”, after researching the presented occupant classification algorithm, and based onthe laboratory’s early occupant classification algorithm, this paper carried out extensiveresearch into occupant pressure collection system, occupant weight estimation model, andoccupant classification system. The work of this paper can be concluded as:Firstly, this paper analyzed the pressure variation and pressure distribution, whendifferent occupants sitting in different posture, to find the pressure sensitive points; after thatthe vehicle occupant pressure collection system based on the pressure sensitive points wasestablished.Secondly, according to the presented occupant classification algorithm of ourinnovation team and the highly nonlinear relationship between the pressure data of the seatsurface and the real weight of the occupant, this paper applied fuzzy identificationtechnology to design an occupant weight estimation model, which can calculate the weightof the occupant according to the pressure data. And this weight estimation model caneffectively resist the interference of different sitting postures.Thirdly, by dividing the weight of the occupant, the vehicle occupant classificationsystem based on the occupant weight estimation model is established. This system canidentify the type of the occupant and give the weight of the occupant. At last, this paper usedthe proposed system to identify the type of the occupant sitting in different posture. Thetesting results show that this system can effectively identify the type of the occupant sittingin different posture.