Research On Array Calibration And Signal Source DOA Estimation And Location Algorithm

DOA estimation is an important research content of array signal processing,and has a wide range of applications in radar,sonar,communication and other systems.However,most high-resolution DOA estimation algorithms require precise information about the array manifold.These conditions are very difficult to meet in the actual system.Therefore,the paper studies the DOA estimation algorithms in two different situations,one is that the array has phase errors and another is the presence of multipath propagation.Traditional DOA estimation algorithms are invalid because the mismatch between ideal steering vector and actual steering vector caused by array phase errors.This paper proposes a new calibration method for uniform line array.Firstly,the proposed method reconstructs array snapshot data,and then according to subspace theory,the autocorrelation matrix of reconstructed data is characterized by feature decomposition.Thus establishing a series of equations on DOA and array phase errors.Finally,the algorithm use the least squares method to solve these equations.Compared with the previous array correction algorithms,simulation results show that the new algorithm has higher DOA estimation accuracy under certain SNR conditions.In addition,the paper also makes improvements to the above algorithm to make it suitable for any coplanar arrays.The resolution and positioning accuracy of the radar system is corrupted by multipath effect,time reversal method can reduce the correlation between the signal source by probing the channel again,thus improving the performance of traditional localization algorithms.This paper establishes a receiving signal model for uniform linear array in multipath environment,and points out the multipath effect increases the correlation between the signals,and makes the classic DOA estimation algorithms fail,then proves the proposed algorithm de-correlate the coherent sources in multipath environment by using time reversal method,thereby improving the performance of classic DOA estimation algorithms.Simulation results verify the correctness of the theoretical derivation and confirms the effectiveness of the time reversal method in terms of signal de-correlation.According to the arrangement of the research,this paper also studies the indoor positioning algorithm based on the received signal strength,a weighted least square localization algorithm based on the unbiased estimate of the distance is proposed.In the proposed algorithm,firstly the unbiased estimate of the distance between the target and the sensors were calculated according to the received signal strength observation model.Then linear equations were established based on the distance calculation formula.After that,the unbiased estimate of the distance was substituted into linear equations to establish the linear least squares model on the source position,simultaneously calculated the noise covariance matrix of the linear least squares mode.Finally the estimation of the signal source position was obtained by weighted linear least square method.In the paper,full computer simulations had been done,simulation results show that the algorithm proposed in this paper has higher positioning accuracy.