A Beamforming method based on subarray processing with delta-sigma oversampling

ABSTRACT A BEAMFORMING METHOD BASED ON SUBARRAY PROCESSING WITH DELTA-SIGMA OVERSAMPLING Hasan Şakir Bilge Ph.D. in ELECTRICAL AND ELECTRONINCS ENGINEERING Supervisor: Prof.Dr. Turhan Çiftçibaşı February 2003 In this Ph.D. thesis, a new beamforming method is presented for coherent imaging system, which is based on digital processing of the subarray signals sampled with delta-sigma analog to digital converter. The proposed method brings many new approaches to existing beamforming systems from different aspects. In this system, the beamformer architecture is completely digitally designed. In digitization, oversampling delta-sigma analog to digital converters are intended to use. With a high oversampling rate, it is possible to sample the echo signals with fine time intervals. When the oversampled data are summed after shifting with very fine delays calculated for focusing and steering, the coherency of the beamformer is increased. In delta-sigma oversampling, there are two important steps; firstly the generation of single-bit data at a high rate, and secondly the downsampling to multi-bit Vlllrepresentation. Low-pass filtering in the downsampling process is a linear operation according to the summation in the beamformation. Using this linearity in the proposed system, only a filter is applied to the output of the beamformer instead of applying many different filters for each receive channel. Thus, the number of filters and consequently the hardware complexity of the system are reduced significantly. Performing the downsampling at the output of the summation makes it possible to beamform with one-bit delta-sigma coded data. One-bit beamformer structure simplifies the hardware significantly. Although one-bit data have high frequency quantization noise, this noise is successfully filtered after summation. The delta-sigma beamforming method performs good results for fixed receive focusing. But when dynamic receive focusing delays are applied, the synchronization between delta-sigma modulator and demodulator is lost. The lost of the relation between the delta-sigma codes decreases the SNR value of the beamformer output. To solve this problem, we propose that the delta-sigma converter samples in time instances when data are needed for dynamic focusing, instead of sampling uniformly every time. Because there is no bits are omitted and repeated between the delta-sigma samples that are generated with this non-uniform sampling approach, the SNR of the beamformer output is kept to a desired level. In this study, a large transducer array is synthesized by subarray processing. By using subarrays, the number of active transducer elements is decreased. Thus, the front-end hardware is significantly simplified. This enables to reduce the cost and size of the system. In this study, to reduce the data acquisition time, we propose to scan with a small number of angles consistent with the Nyquist spatial sampling criteria. Thus, it is possible to design real-time imaging systems. As the Nyquist spatial sampling criteria is proportional to the length of the array, the number of angles needed for synthesizing a large array is greater than the number of angles needed for data IXacquisition by subarrays. In the proposed method, the limited number of the low- resolution beams acquired by subarrays is increased by lateral interpolation. Then, the upsampled low-resolution beams are summed to high-resolution beams. Interpolation filters used in upsampling should be band-pass and design appropriately for each subarray, because each subarray has a different phase in its point spread function. The proposed system is theoretically investigated and tested by experimental data. The test results show that the system has a good image quality enough. The subarray delta-sigma beamforming method reduces the cost, the size and the power consumption of the coherent imaging system whereby maintaining high quality images. This method for coherent imaging systems has many potential applications. 3D ultrasonic imagers, small portable scanners, and intravascular imaging devices are some examples of the important medical applications. On the other hand, underwater acoustic imaging systems (sonar's), smart antennas in communications, and robot vision are some other potential applications. Keywords: Array signal processing, coherent imaging, ultrasonic imaging, beamforming, dynamic focusing, synthetic aperture, subarray, acoustic transducers, delta-sigma modulation, non-uniform sampling.