Enhancement of Lane Departure Warning System in Short Term Autonomous Warning using Inertial Sensors Assistance

摘要

This paper devotes to design and integrate a hybrid and innovative LDWs scheme to increase its availability. The scheme presents a solution of discontinuous lane marking for markings detection using camera and sensor fusion, including image processing, dead-reckoning and vehicular signals. To extract lane markings from image catching, DSP perform the task after image calibration and coordinate transform. The vision is applied by a camera which plays the role of displaying preceding image and warning information with buzzer. When operating at failure mode, inertial signals apply and provide available track. In dead reckoning technology, hybrid sensors address a global and all weather assistant design based on odometry and inertial sensor. From inertial data and vehicular information, this paper provides an extended solution with learning parameters to do remote warning. The proposed system is carried out with theoretical application and hardware integration, and furthermore the result shows extended LDWs approach applicability.