Radar Data Fusion And Engineering Implementation Under Complex Environment

This paper focuses on the issue of the radar data fusion under the complex environment, especially for the high frequency ground wave radar system. Some new significant theoretical methods and engineering methods are proposed to solve these problems. The main contents as follows:(1) A high frequency ground wave bi-station system is proposed to triangulate the target tracks using only range and Doppler measurements. Three key issues are discussed for this new system concept. Firstly, the locating model of this system is established and the nonlinear tracking problem is solved by EKF. Secondly, the data association problem about the plots and tracks of this system is solved by three steps: the first step is the matching method of plot-to-track and track-to-track by rule-based tracking methodology. The second step is the method of the eliminating ghost targets by estimating the speed of the target using only one radar's observations and calculating the change in target's angle by two adjacent observations. The third step is to turn the two steps ahead into a 0-1 programming problem to get the final result of the association problem. Lastly, the error model of this system is established and the position precision curve is supplied by the GDOP. The effectiveness of the methods above is given by the simulations.(2) An enhanced sequential nonlinear tracking filter with denoised pseudo measurement is proposed in this paper. The denoised pseudo measurement is established by range and range-rate measurements as a liner model. A debiased Kalman filter is proposed to reduce the error of the pseudo measurement which can enhanced the performance of the second-order EKF in the sequential nonlinear tracking filter. The Monte Carlo simulations show that the performance of the new tracking filter.(3) Many notable problems are discussed by designing an engineering application software about multiple radar data fusion using C language: regional searching, setting main station, time alignment, track number management by double-linked list, outlier rejection, track smoothing, human computer interaction and error correction processing. Some of them are new methods. The performance of the multiple radar data fusion software is shown by a program written by matlab language at last.