Research Of Amplitude And Phase Error Correction Algorithm In Array Antenna

DOA(Direction of Arrival)estimation is an important research area of array signal processing,with growing applications in engineering,such as radar,sonar,communication,biomedicine,etc. Array signal processing in a multi-channel system is with bad performance, the actual existenceof systematic errors often leads to a serious deterioration in algorithm estimated performance orsystem output. Therefore this dissertation presents a calibration algorithm for amplitude and phaseerrors in uniform linear array.First of all, this thesis reviews the main spatial spectrum estimation algorithm, throughsimulation, we know that the resolving ability of MUSIC algorithm can not attain to 5 degree underSNR =5 dB and snapshots = 1000, so we have to look for spectrum estimation algorithm with betterresolution.Then this thesis reviews the main amplitude and phase error correction algorithm, focusing onthe NSF algorithm, NSF algorithm result converges to a local optimal value after 2 iterationsthrough simulation In the case of high initial error, so we have to look for amplitude and phase errorcorrection algorithm with more stable.Finally, based on the advantages of array signal processing algorithms from our predecessors'research, using uniform linear array covariance matrix has a Toeplitz structure, the nature of the costfunction is constructed by a stable amplitude and phase error correction algorithm, and finallythrough the Root-MUSIC algorithm, the amplitude and phase error correction algorithm is with agood performance on DOA. Through simulation we can see that in the SNR = 5dB, snapshots =1000, angle difference between the two waves within five degrees , this amplitude and phase errorcorrection algorithm is better than MUSIC algorithm in the resolving ability; Through simulationwe can see that when SNR = 20dB , the initial amplitude and phase error is large, the root meansquare error algorithm of magnitude factor is only 0.01dB, close to 0, while the NSF algorithmlocally convergent to 0.1dB; The root mean square error algorithm of phase factor is only 0.8°,while the NSF algorithm locally convergence to 7.1°. This also proves this algorithm is more stable.Although the algorithm requires iterative solution, but its fast convergence, iterative convergence isonly 7, so the calculation is not complicated.