Fast Parallel Implementation Of Spatial Adaptive Algorithms For Large Arrays

The complex electromagnetic environment is a severe environment for modern radars to exert their power.Phased array radar has a good effort of interference suppression due to the use of anti-interference technologies such as adaptive beamforming.With the more complex interference forms,and the detection capability of weak targets also becomes more demanding,the array antenna size of phased array radar is larger and larger,which also makes the computational complexity of traditional adaptive beamforming technology large,causing the algorithm to converge slowly and fail to achieve real-time processing.Aiming at the large-scale arrays,combined with the Partitioned-Parallel method the traditional arrays adaptive beamforming technology is improved.And then,combined the improved method with dimensionality reduction processing technology,also with angle measurement of sum and difference tracking technology,the modified method can reduce the computational complexity largely.Finally the theoretical derivation and simulation experiment analysis are carried out.The main research contents are summarized as follows:Firstly,the basic theory of array fast adaptive beamforming is introduced.An array signal processing model is established,then the basic concepts of the adaptive array beamforming technology for full-dimensional arrays and the dimensionality reduction processing technology commonly used in large arrays are introduced.For the latter,the sub-array division,grating lobe problem,and sub-array beamforming technology are introduced.Finally,combined with the Linearly Constrained Minimum Variance(LCMV)algorithm,the technique of Partitioned-Parallel fast implementation is introduced.Secondly,the effect of Partitioned-Parallel Sample Matrix Inverse(PSMI)algorithm and different block sizes on the performance and computational complexity of the beamforming algorithm are studied.Combined with Partitioned-Parallel technology,the traditional Sample Matrix Inverse(SMI)algorithm is modifide to reduce the computational complexity,which can ensure the real-time implementation,and also the adaptive anti-interference weight obtained has better anti-interference performance under certain conditions,then the impact of different block numbers on the performance of the algorithm is analyzed.The genetic algorithm(GA)and particle swarm optimization(PSO)algorithm are used to optimize the block size to further improve the performance of the main-side ratio of the pattern.Combined with the gradient optimization method,a robust PSMI algorithm is obtained.To correct the error of the steering vector,combined with the robust processing method,the PSMI algorithm can obtain accurate beam pointing and good anti-interference performance.Then,combined with the dimension reduction processing technology,the application of PSMI algorithm at the sub-array level is studied,which further reduces the calculation amount and does not affect the anti-interference performance.Combined with sum and difference tracking technology,the weight obtained by the PSMI algorithm is taken as the sum beam weight,and the symmetrical reverse of the weight is taken as the difference beam weights.The performance of the proposed algorithm has been studied and analyzed,simulation experiments and performance analysis,demonstrate that the proposed method still has good angle tracking performance when there is interference inside and outside the main lobe.Finally,the PSMI algorithm is implemented with a multi-core DSP,which verified the feasibility of the algorithm in engineering.The actual run time and calculation results are analyzed and evaluated.The results show that the PSMI algorithm can greatly reduce the computational complexity and increase the signal Real-time processing,so it has high engineering application value.