Adaptive cruise control system, front valid target recognition algorithm

Adaptive cruise control system is developed on the basis of traditional cruise control them, and therefore has both the ability of traditional cruise control cruise control, while other automotive radar sensors perceive through the front of the vehicle environment, identify the front of the valid target. By automatically applying throttle or brake control vehicle longitudinal speed so that the vehicle in front of the vehicle to maintain appropriate safety distance. Thereby reducing the burden of pilot operation, reduce driver fatigue driving, improving vehicle active safety and driving comfort. In a complex multi-objective traffic scene in front, the adaptive cruise control algorithm followed by a valid target in front of the target is known, specifically, the vehicle traveling in the region next nearest target. The vehicle driving area is the next expected locus of the vehicle to the center, the width of one lane wide effective target detection area. Valid target recognition algorithm is the key to determining the area, which is expected to determine the vehicle travel path. Conventional algorithm current state of motion of the vehicle such as speed and the yaw angular velocity to predict the target track. This algorithm in some typical conditions expose its flaws, for example, in the corners of the front entrance of the goal has already begun into the corner, and the car is still straight, through the current state of the vehicle can not predict the trajectory of the vehicle will react into the corner of the facts into the corner of the valid target is not correctly identified, may occur serious collisions. Therefore, for the deficiencies and defects of the traditional algorithm, this paper carried out an effective target tracking based object recognition algorithm. As the history of each object in front of the point where the road ahead reveals the shape characteristics of radar targets through continuous tracking data estimate the location and movement of the target state, and then estimate the shape of the road ahead and the expected travel path of the vehicle, Compared with traditional methods, this algorithm has better predictability. Paper were carried out straight, bend, bend entrance, corner exit, change lanes before the car under various conditions such as the multi-objective simulation to verify the effectiveness of the algorithm. This article also expected for the vehicle running track as well as stationary target precise handling done a brief discussion, the expected running track is divided into two sections, the first paragraph of the proximal end of the vehicle by the vehicle state estimation, the distal end of the second car segment consists of the target state estimation, so take advantage of the proximal state estimation accuracy of the vehicle as well as the remote target motion estimation predictability; stationary target does not exist because the speed and yaw rate, can not be estimated based on the target tracking methods expected running track. Finally, the development of an effective target selection algorithm in front of experimental data acquisition platform, through multi-objective straights and corners two kinds of multi-target curvature conditions of the preliminary field experiment to explore the adaptive cruise control front valid target recognition to the experimental research methods. The main contents and conclusions are as follows: First, the analysis of the traditional fixed front curvature effectively target recognition algorithm into the corners, out of the corners and other conditions of the limitations. Traffic scene in front of the target vehicle in front of a certain extent, reflect the road information in front of the target by using the historical trajectory of the vehicle may better predict the future running track and the road ahead trajectory, but also more effective precisely identify the front target. The second, carried out in front of target tracking based on a valid target recognition algorithm, and conducted Simulink Carsim co-simulation. The car in front of the historical trajectory of the target contains information about the road ahead, so consider ahead by continuously tracking the target to estimate its state of motion, combined with the state of motion of the vehicle to estimate the curvature of the road ahead and the path of the car horn, and to determine the The future trajectory of the vehicle and the lane is expected trajectory. Estimates of the target relative lateral distance of the lane in order to determine the target is the driveway or in the other lane, the final choice out of the car from the lane nearest the front of the vehicle target is a valid target. On the use of Kalman filter algorithm to estimate the target state, and the curvature of the road ahead of the vehicle path angle. The straights, corners, straight corners, bend straightening, before the car change lanes and other conditions to achieve a more effective target vehicle target recognition algorithm simulation results show that the algorithm can be a variety of conditions accurately identify the corresponding scene in front of valid targets, proven effectiveness of the algorithm. Third, the development of efficient target identification experimental data acquisition platform and conducted a preliminary study of the field experiment. Jetta car in a field experiment developed data acquisition platform, including integrated a variety of sensors, capture cards and IPC hardware platforms and software platform based on LabVIEW development. Millimeter-wave radar mounting bracket design, layout, wiring harness, adjust the mounting position and attitude, and conduct field experiments to analyze the data features. Installation of non-contact optical sensors and fiber optic gyro to obtain the longitudinal side of the vehicle speed and yaw rate and other status information. Through a variety of capture card hardware interface to access various sensors IPC, in the LabVIEW environment on a variety of data acquisition and storage. Ultimately through multi-objective straights and corners two kinds of multi-target conditions of the initial curvature field experiment to explore the adaptive cruise control front valid target recognition to the experimental research methods.