Robust Optimization Design And DOA Estimation Of 2D Sparse Ultrasonic Phased Array In Air

To address the issue of poor performance, high cost or bulkiness of the optical and the electromagnetic imaging devices in the presence of thick fog, rain, snow, strong radiation or smoke, a solution for layout optimization, robust design of amplitude weig ht matrix, adaptive DOA estimation and beamforming of the 2D sparse ultrasonic phased array used for scenario perception of mobile robot is proposed. And the Matlab simulation and experimental tests are used for the feasibility verification of the proposed algorithms.According to the structural features and ultrasonic field characteristic of the 2D sparse ultrasound phased array in the air, the equations of the ultrasonic field directivity and the array structure parameters of the 2D sparse ultrasound phas ed array are established. The simulation analysis of the influence of the transducer size, spacing and the number of transducer on the acoustic field characteristics are carried out. The structure parameter of the array is obtained according to the results of the ultrasonic field simulation. The transmitter array number of the designed 2D sparse ultrasonic phased array is 8 ?8, and the array spacing is 2?. Considering the existence of the sparse grating lobes which affect s the scanning of the array, the transmit array and the receive array are interleavingly placed to eliminate the grating lobes and the element spacing and element number of the receive array are determined. After optimization, the horizontal and vertical resolution reaches about 4 degree. The horizontal and vertical range of scanning is expanded from 30 degree to 60 degree.Considering the influence of the actual installation errors, manufacturing errors and the element failures on the resolution of the imaging, the genetic algorithm, Monte Carlo simulation analysis and 2D golden section method are combined, and the robust optimization design is carried out on the array amplitude weighting matrix. The genetic algorithm is used to search the amplitude matrix, the Monte Carlo method is used to simulate the normal distribution N(1, 1) continuity element error, the 2D golden section algorithm is used to search the peak of the side lobes and main lobe. The ratio of the primary side lobe and the main lobe is used as the goal of robust optimization design, in order to search for a set of the optimal weights under the error condition and to reduce the sensitivity of the array to the error, so as to improve the efficiency of optimization calculation. The resulting ratio of the primary side lobe and the main lobe of the transmit array decreases from 0.28 to 0.07, and the resulting ratio of the primary side lobe and the main lobe of the receive arrays decreases from 0.29 to 0.142 under the error conditions.The reflected signals of the target object of th e 2D sparse ultrasonic phased array are easily influenced by the ground reflected signal and other objects ’ scattering signals and the reflected signal of the beam grating lobe, the space coherent signal source need to be separated to eliminate the influence of the interference signals and obtain the desired target signal. To address the issue of 2D DOA estimation with multiple coherent signal sources in the space for the 2D sparse ultrasonic phased array with a large number of elements, the 2D spatial smoothing is proposed based on MUSIC algorithm on the 2D DOA estimation, in order to deal with the coherence of the multiple coherent signal and to realize the target signal DOA estimation. To eliminate the angle uncertainty of the grating lobe of the receive array due to its sparsity, the scope of receive beam direction is limited. This method can effectively solve the problem of multiple coherent signal sources in the space. The angle resolution of the t wo coherent signals can reach 3°.