Real-time Identification of the Dynamic Model of a Vehicle Driver's Steering Control

• Main Authors: Jeng, Jr-Feng, 鄭至峰
• Other Authors: Yeh,Tse-Liang Alan
• Format: Others
• Language: zh-TW
• Published: 1996
• Online Access: http://ndltd.ncl.edu.tw/handle/98170662369656778150

碩士 === 國立中央大學 === 機械工程學系 === 84 === Most of the time, an on-load vehicle is operated to stay in its current lane on the road. Therefore, the dynamics of a driver steering his vehicle to follow the lane reflects the driver's capability to control the vehicle. If we treat the driver and the vehicle as a whole dynamically controlled system, then the path of the road is the reference input of the steering control and the position of the vehicle is the control result. If the vehicle is properly controlled by the driver, the dynamic relationship between this system's input and output is stable. If we apply digital image processing techniques on the video of the road scene ahead of the vehicle to measure the input and output of this driver-vehicle system, we can apply system identification techniques to monitor the dynamic characteristics of the steering controls and the performance of the driver. If any dangerous condition occurs from the variation of the system stability and safety margin, we can either provide warning to the driver or restrain further operation of the vehicle. The technical key point to implement such a driver performance monitoring system is on the identification of the driver steering dynamics and the detection of dangerous performances. This dissertation research is to develop a test platform and the identification of proper dynamic models for driver reactions. We found that a fourth order ARMA discrete dynamic model with 240ms pure delay is appropriate. The sampling rate of the discretization was 50 Hz. We also found that the dynamic characteristics of the identified model reflect the changes in the psychological situation of the subject under test.