Design Offorward Collision Warning System Based On Millimeter Waveradar

With the development of science technology and economy in recent years, thevehicle population in my country is increasing, followed by a growing number oftraffic accidents. According to the statistics of Traffic Management Bureau in China,more than60percent accidents belong to rear-end collisions which cause the greatestdamage, so, developing a collision avoidance warning system is quite meaningful toprotect drivers’ security. With the development of various sensors,information fusionand computer communication technology, collision warning system is not relying onsingle information sources to estimate risks, but according to the analysis ofintelligent algorithm and information fusion of self-vehicle and its surroundingenvironment.This thesis was supported by "National863projects", combining the targettracking technology and collision warning system. It both considers the self-vehicledriving state and pre-vehicle driving intention to estimate the collision probability,then builds an active safety collision warning system which based on the targettracking, including the functions of active collision avoidance and warning.Firstly, through the analysis of velocity and distance measurement ofmillimeter-wave radar, the thesis designs a specific data analysis procedure whichcould provide the warning system with accurate and valid information. Thenaccording to the analysis of vehicle braking process and car-following mode, thethesis divides the pre-vehicle state into three stations which are stationary、constantspeed or accelerated movement and decelerated movement or parking. And adriver-based minimal safety distance is established.On the basis of what mentioned above, the thesis introduces the target-trackingconcept into the collision warning system,through Karman Filter algorithm based onCV/CA vehicle modes to track the pre-vehicle; Then the two-level warning algorithmand assistant braking algorithm are set up.Finally, the thesis builds a set of Simulink-modes in Matlab software, includingthe vehicle dynamics mode、longitudinal distance measurement mode and warningmode, and sets up a simulation platform based on GUI(Graphical User Interface).Through the joint simulation experiment, the results show that targettracking technology can effectively reduce the probability of false-detecting andleaking-detecting which largely improve the accuracy and reliability of warningsystem.