Prediction model time-to-line-crossing during lane change process
Author(s): WANG Chang, DU Kang, ZHENG Chu-qing, School of Automobile, Chang’an University
Pages: 737-744
Year: 2016 Issue: 3
Journal: Journal of Jilin University(Engineering and Technology Edition)
Keyword: vehicle engineering; time to line crossing; geometry analysis; yaw angle estimate; vehicle active safety;
Abstract: To solve the time-to-line-crossing(TLC)prediction problem during lane change process,the geometrical motion feature of vehicle during lane change was analyzed and a yaw angle estimation algorithm between vehicle and lane mark was proposed.Based on this,TLC prediction models for straight and curve roads were established.A test platform was developed which consists of vision sensor,vehicle gyro and onboard CAN bus data capture card.Real-road driving test of several drivers was carried out,and representative data of lane change state were collected.The TLC prediction models were tested using the real lane change data.Results show that 87.0% prediction errors on straight road were less than or equal 0.1s,and the corresponding rate on curve road was 93.7%.The prediction errors of two models meet the normal distribution.
Pages: 737-744
Year: 2016 Issue: 3
Journal: Journal of Jilin University(Engineering and Technology Edition)
Keyword: vehicle engineering; time to line crossing; geometry analysis; yaw angle estimate; vehicle active safety;
Abstract: To solve the time-to-line-crossing(TLC)prediction problem during lane change process,the geometrical motion feature of vehicle during lane change was analyzed and a yaw angle estimation algorithm between vehicle and lane mark was proposed.Based on this,TLC prediction models for straight and curve roads were established.A test platform was developed which consists of vision sensor,vehicle gyro and onboard CAN bus data capture card.Real-road driving test of several drivers was carried out,and representative data of lane change state were collected.The TLC prediction models were tested using the real lane change data.Results show that 87.0% prediction errors on straight road were less than or equal 0.1s,and the corresponding rate on curve road was 93.7%.The prediction errors of two models meet the normal distribution.
Related Articles