Particle Filter-based Automobile Collision Avoidance Radar Study

The millimeter-wave band automotive collision warning radar, which has been developing all over the world, is an important part of vehicle safety system. Because the radar is applied in all kinds of non-Gaussian interference circumstances, it is difficult to detect and track the target by using traditional Kalman and Extended Kalman methods which were liner or approximated linear measures. Particle filter is very fit for non-linear and non-Gaussian interference circumstances. It has a high-veracity. Such a vehicle radar used particle filter in practical transportation application can provide great benefits in safety and driver convenience. It has academic and practical significance. Research works and the major contributions in this thesis are as follows:1. After discussing non-linear filter, particle filter is introduced. Both the theoretic analysis and simulation results show that the particle filter does well at a non-linear and non-Gaussian interference circumstances.2. After analyzing the functions and specifications of the system, discussing radar working method and antenna scanning mode, the scheme using chaotic FM signal as a substitute of LFMCW signal is proposed, the results of analysis and simulation show that chaotic FM signal is superior to the traditional LFMCW signal in electromagnetic compatibility. It can solve the false alarm problem to some extent.3. The principle block diagram of the vehicle radar system is given. The simulation results of correlation show that it has a good resolution of distance. The azimuth angle measurement problem of automotive collision warning radar is discussed. DBF is introduced. The simulation shows it has a good split-angle precision.4. The application of particle filter in the millimeter-wave band automotive collision warning radar is analyzed, the simulation show that the performance of particle filter is superior to extend Kalman, it can solve the problem of low- precision and low-reliability.