Multi-station Passive Localization Technology Based On Radiation Source

In the modern battlefield, with the continuous development of the encryption technology, it is hard to obtain the content of the communication. In that case, it becomes more and more important to get the trend information. How to localize the emitter quickly and accuratly is the key to obtain the trend information. With the progress of the electron disturber and missle, the active location system is facing more and more threats. The passive position system does not transmit electromagnetic wave which makes the prospect of application broad because of its advantages, such as excellent invisibility, large detectable radius and wide application. In order to locate the emitters quickly and accurately, the passive localization algorithms have been studied systematically in this dissertation. The main contributions can be described as follows:1. Aming at the problem of classical two-step weighted least squares(WLS) algorithm, the thresholding effect, that is caused by the high level noise, the reason which causes the thresholding effect is researched. The neglect of the high order terms of noise actually induces the thresholding effect. Then the constrained weighted least squares(CWLS) algorithm is applied to multi-station passive localization to solve the problem, and the CWLS algorithm based on TDOA(time differences of arrival) jointed AOA(angle of arrival) is proposed. The CWLS algorithm firstly transforms the nonlinear equation into the two pseudo-linear equations, and then adds the constraints to the pseudo-linear equations and finally obtains the location. The new algorithm can extend the operating range and overcome the thresholding effect effectively with little increase in the computation load. Through simulation results, it is proved that the novel method is robust to noise and can still obtain an accurate location when the noise rises beyond the thresholding point.2. In order to reduce the bias and the computation load, a bias reduction algorithm using constraint condition is proposed. The algorithm firstly substitutes the constraint into the cost function by applying the matrix partition theory to avoid solving the Lagrange multiplier. The proposed method can reduce the bias effectively using a little iteration numbers. Targeting at the significant bias using classical least square algorithm in a large noise background, the quantitative analysis of the theoretical bias is derived and the reason which causes the bias is found. Then the improved TDOA and AOA algorithm is proposed by adding the quadratic constraints to the expectation of the error. The new method can reduce the bias significantly and obtain the original Mean Square Error(MSE). It is able to lower the bias to the same level as the Maximum Likelihood Estimator. Furthermore, the new algorithm has little computation load because it does not require singular value decomposition (SVD) and can obtain the closed-form solution.3. A majority of algorithms is suitable for limited certain location scenarios, and their positioning accuracy decreases dramatically in the presence of systematic error, the theoretical framework for deriving the location performance of the Taylor-series iteration algorithm is systematically presented in this dissertation. Two cases, in the absence of calibration state error (a) and in the presence of calibration state error (b) are discussed. For the case (a), the corresponding Cramer-Rao lower bound(CRLB) is derived. For the case (b), the corresponding CRLB is derived and compared with the one in case (a). The theoretical MSE which neglects the calibration state error is derived. The comparison between the theoretical MSE and the CRLB illustrates that the source location accuracy cannot reach the optimal accuracy when neglecting calibration state error. Aiming to eliminate the shortcomings, a two-step optimal fusion location method based on the Taylor-series iterative algorithm is proposed, whose accuracy is proved to reach the CRLB. The proposed method can be applied to any location scenarios and has strong universality. Finally, passive localization scenarios were simulated to verify the performance of the location algorithm and the validity of the theoretical analysis.