Font Size: a A A

Study On Algorithms Of Automotive Highway Lane Departure Warning System

Posted on:2005-02-06Degree:DoctorType:Dissertation
Country:ChinaCandidate:Y P DongFull Text:PDF
GTID:1102360152456674Subject:Vehicle Engineering
Abstract/Summary:PDF Full Text Request
Lane Departure Warning System (LDWS) is mainly focus on help driver to keep vehicle in the lane under fatiguing or monotonous driving scenarios on highways and highway like roads. Accordingly, a warning is provided to alert the driver in case of lane departure caused by inattention. Its object is to reduce traffic accident and augment driving safety and vehicle's active safety.In recent years, LDWS has gotten great development abroad, and has equipped to top class cars and freight trucks. While in China, the research in the field is at very beginning. Hence, to study LDWS has great significance and also has broad application prospect.After a comprehensive review of R&D history, present status of LDWS, following can be summarized :(1) LDWS is directly based on the viewpoints of vehicle's motion. Its core idea is first about how to preview vehicle's motion and its time to lane departure. Secondly is to select appropriate warning timing. (2) As to vehicle's motion, three typical ideas can be concluded. One considers vehicle's motion of constant velocity no matter it is longitudinal or lateral motion. The second idea regards the motion of constant acceleration, vehicle's trajectory is curved. And the other idea emphasize that the motion can follow the change rate of lane curvature. Hence the trajectory should be the same with the lane curve, that is if the lane curve is clothoid and the vehicle's trajectory will be the same. (3) Any idea mentioned above has common about hypothesis on steering angle; they all consider the steering angle is constant. From the principle of plane kinematics, vehicle's motion should include transformations and rotary around the gravity. Its acceleration is variable; the trajectory is determined by velocity and acceleration both. Change of The steering angle may influence the trajectory's curvature. Therefore, It is necessary to take the change of steering angle into account when preview the trajectory.Further more, as a driver assistant system, the LDWS must consider the driver's decision process when controlling vehicle's speed and steering. Namely the system must resemble driver's behavior in traffic. Hence, the driver's is a system to simulate safety decision process of a driver. Therefore consistence is acceptable.Finally, several developments have been got after series of research on driver operation characteristics at the automobile state key lab of Jilin University. It is possible to adopt the production to realize Lane Departure Warning.Based on above mentioned reason, the dissertation proposes a research subject: study on highway vehicle lane departure warning system and its algorithm implementation. Its aims are mainly about (1) to analysis and propose LDWS's decision making model and algorithm based on the optimal preview acceleration decision making, on the previous research on driver handling behavior characteristics and modeling in the State Key Laboratory of Automobile Dynamic Simulation, through application of the hypothesis of steady preview. (2) to establish road sensing model and implant its visual sensing algorithm. (3) to implement vehicle's future motion steady preview algorithm based on steady preview hypothesis. (4) to propose a novel LDWS uniform decision algorithm with consideration of complicated factors influencing Lane Departure through application of the multi-object fuzzy decision-making theory and method.First, the paper reviews the basic principle and architecture of Driver Optimal Preview Acceleration Model. And then establishes the steady preview theory for LDWS based on steady preview hypothesis. Later, the paper establishes visual sensing model of road drivable area and implements visual sensing algorithm and vehicle future motion steady preview algorithm. At the end, the paper utilizes vehicle dynamic model of 29 degrees of freedom and developing driving simulator in the State Key Lab of Automobile Dynamic Simulation of Jilin University respectively. From the scheme of the paper, the content can be divi...
Keywords/Search Tags:Lane Departure Warning, Lane Recognition, Optimal Acceleration, Steady Preview
PDF Full Text Request
Related items