Abstract: As an important equipment in marine fisheries, Multi-beam fishery sonar makes use of sound waves to detect underwater creatures, with the beamforming algorithms being the core technology in its development process. Based on the traditional shifted sideband beamforming (SSB) algorithm, this paper proposes an improved algorithm which converts parallel processing to serial processing, and increases the signal processing speed without causing performance loss. Firstly, this improved algorithm was compared with SSB. Then the improved algorithm was implemented on the field gate array (FPGA) platform, and performance evaluation and resource consumption were analyzed for the two algorithms. Finally a directive test was conducted over the FPGA realization of the improved algorithms in an anechoic sonar pond, and the simulated beam pattern was compared with the measured beam pattern through MATLAB. The results showed no difference of the two algorithms in their performance, while the resource consumption of the improved algorithm was nearly 70% lower than SSB. Meanwhile, the measured results of the improved algorithm were roughly consistent with the theoretical simulations, and the deviation was in a reasonable range.

Key words:  fishery sonar, beamforming, shifted sideband beamformer, time compensation, phase compensation, FPGA implementation

摘要： 多波束渔用声呐是海洋渔业的重要设备，可利用声波对水下生物进行探测，其波束成形的算法是研制过程的核心技术。本研究从圆柱阵多波束渔用声呐的扫海方案出发，在传统移边带算法（SSB）的基础上提出了一种将并行处理转为串行处理的方法，同时在不损失性能的前提下提高信号处理速度的改进算法。首先将传统算法，与本改进算法进行对比。其次在现场可编程逻辑门阵列（FPGA）平台对改进算法进行了软件的实现，完成了两种算法实现的性能评估和资源消耗分析。最后在消声水池对改进算法的FPGA实现进行了指向性的测试，并通过MATLAB将仿真的波束图与实测后的波束图进行了比对。结果显示：两种算法在性能上保持一致，改进算法消耗的资源比SSB算法消耗的少近70%，同时改进算法的实测结果与理论仿真大致一致，偏差在合理范围内。

关键词: 渔用声呐, 波束成形, 移边带波束成形, 时间补偿, 相位补偿, FPGA实现