Research Of Human-Vehicle Safety Behavior Based On Naturalistic Driving Data And Active-Passive Test

This paper mainly concentrates on the study of automotive safety and protection of the driver’s security. In order to protect the safety of the drivers and the car, the most effective way is to avoid car crash to the most extent. In all kinds of car accidents, the largest proportion are those related to drivers. Hence, based on large amount of previous studies, the paper has carried out the integrated natural active safety driving tests and the intelligent network of spontaneous combustion vehicles driving tests and has collected large amount of authentic data to research the driving behavior so as to lessen the improper driving behavior to improve driving safety(1) Researching the driving behavior of three different groups. This paper is based on the light vehicle platforms which could provide pre-warning functions like rear-end, lane departure and blind collision. In the process, 108 drivers have been chosen for the six-week period natural driving tests. An analysis has been done on the drivers’ real reaction and acceptance of the three different groups –the old people, the middle-aged and young people. In the meantime, the different groups travel frequency, travel time, travel speed and braking times, with the car and system warnings from driving behaviors have studied and analyzed(2) Study of lane departure warning system and driving behavior. From the perspective of lane departure warning system and driving behavior, the driving behavior of three different groups of the old, middle-aged and young people are researched under the circumstances of highway driving among which five variables of the steering wheel angle correction amount, vehicle speed, vehicle lateral speed, accelerator pedal position and vehicle location in the driveway are analyzed for the lane departure warning and driver behavior. As a result, the drivers’ reactive behavior are obtained when they were under the lane departure warning.(3) Analysis of Early warning systems and driving behavior of a frontal collision. In the forward collision warning systems and driving behavior, the paper focuses on the driving behavior under driving situation from different groups of old, middle-aged and young and different genders.Varibles including acceleration pedal position, brake status, vehicle longitudinal speed, steering wheel angle, following distance, approach speed, forward collision alert system status are used for frontal pre-alarming analysis and driving behavior such as the driver’s speed, steering wheel control and collision time control are obtained.(4) Research into the vehicle performance associated with intelligent network and driver safety and vehicles. In order to assess the impact of car networking technology in vehicle safety, transport and energy, a Safety Pilot project were carried out in Michigan University in which around 3500 cars were involved and these cars include light vehicles, heavy trucks, buses, motorcycles and bicycles and all of them were equipped with advanced safety equipment or specific DSRC communication capabilities. During the test, a complete database of car networking communications has been established and through the collection of these car to car communication, car to road communication, a relevant analysis has been made to the car to car communication performance, car to road communication performance, driving behavior and traffic conditions.(5)Study of the algorithm on early warning system for the intelligent network linked vehicles. In the research of early warning system in the intelligent network, a framework of the vehicle collision pre-warning has been put up for the intelligent network vehicle. When the predicted collision time(Time To Collision, TTC) reaches at a set threshold, the warning information to the driver is provided to the drivers. Firstly, the vehicle would pre-process the received basic information from the other car; then relocate the coordinate of the other car and calculate its relative position and velocity. Thirdly, an estimation was made on the trajectory of the remote vehicles and the track rails of the distant car by way of Kalman Filter(KF) and an error analysis was made to the position of for the latitude and longitude of the main car. It is found that there are positive predictions on curved corners and lanes; however, there are relative larger errors on the latitude direction. And finally, a radar map has been established for the relative positions of the distant car and the main car. Information of the relative position and the velocity and predictive collision time between the distant car and the main car can be shown on the map.Therefore, risk of vehicle collision can be detected in real time, through the use of this safety information, earlier collision warning information are provided to the driver, you can develop more advanced vehicle collision avoidance control system can provide unmanned vehicle collision auxiliary warning information to prevent the occurrence of a vehicle collision risk potential.