多传感器融合的视觉导航智能车避障仿真研究  

Simulations of Obstacle Avoidance for Intelligent Vehicles Based on Vision Navigation and Multi-sensor Fusion

作　　者：王檀彬[1] 陈无畏[1] 李进[1] 焦俊[1] 

机构地区：[1]合肥工业大学机械与汽车工程学院,合肥230009

出　　处：《系统仿真学报》2009年第4期1015-1019,共5页Journal of System Simulation

基　　金：安徽省自然科学基金(050420302);清华大学汽车安全与节能国家重点实验室开放基金(KF2007-06)

摘　　要：在视觉导航标志线跟踪的智能车辆上加超声波传感器进行避障。智能车辆通过CCD摄像头获取的信息可以跟踪路面标志线行驶。用D-S证据理论和模糊处理的方法对多个超声波传感器的信息进行融合。包括了超声系统的建模和信息的融合。仿真结果给出了在不同环境下,智能车辆的避障运行轨迹,表明该信息处理和融合方法能有效判断障碍物,获得良好的避障导航效果。To the vision guided intelligent vehicle, some ultrasonic sensors are added for obstacle avoidance. Intelligent vehicle can move along the lane marker through the information acquired by the CCD camera. The information of multiple ultrasonic sensors was researched by using the D-S evidential theory and the method of fuzzy logic. Information fusion and the model of ultrasonic sensor are included. The results of simulation give the tracks of the intelligent vehicle when moving and avoiding obstacles in different environments. It indicates that this method of information processing can identify obstacles efficiently and makes satisfying results of navigation and obstacle avoidance.

关 键 词：信息融合 D-S证据理论 模糊逻辑 智能车辆 

分 类 号：TP242.6[自动化与计算机技术—检测技术与自动化装置]