题名: |
AUTOMOTIVE COLLISION AVOIDANCE SYSTEM FIELD OPERATIONAL TEST PROGRAM : MULTI-SENSOR FUSION FOR COLLISION AVOIDANCE. |
作者: |
Khosla-Deepak; Matic-Roy; Schwartz-David |
关键词: |
Collision-avoidance-systems; Data-fusion; Kalman-filtering |
摘要: |
This paper describes a data fusion architecture and system for fusing multiple sensor information for automotive applications, specifically adaptive cruise control and forward collision warning applications. Both of these applications rely on accurate detection of vehicles that are in the forward path of the host-vehicle. This in turn requires accurate estimation of forward road geometry and host vehicle state. Information about road geometry in the forward path of an automobile can be obtained by using a variety of sensors such as gyro, camera GPS, etc. To reduce uncertainty as well as handle missing and/or incomplete data from single sensor systems, one approach is to combine information from multiple redundant and/or complementary sensors. We describe the interface and system for fusing multi-sensor information from gyro, radar, vision, and GPS to provide robust and accurate host vehicle path and state estimation. The core fusion algorithm is based on a Kalman filter and a module that determines the information value of a sensor prior to the fusion step. This paper also describes a novel road model that is superior to a conventional single-clothoid road model as it produces smaller road geometry estimation errors especially during sharp transitions in road curvature. The data fusion software operates at 10 Hz rate on a 266 MHz PC 104 platform. We will also describe results of performance evaluation of the data fusion system. |
总页数: |
ITS America. Meeting (12th : 2002 : Long Beach, Calif.). Securing our future : ITS America 12th Annual Meeting and Exposition 2002 : conference proceedings. 2002. pp14 |
报告类型: |
科技报告 |