提出改进的粒子群优化算法,获得波束方向图主瓣宽度和旁瓣级折中的优化立体阵形,避免基本实数粒子群算法仅采用旁瓣级或主瓣宽度一个性能指标优化而导致另一个性能指标恶化的问题,利用阵列视角限制进一步优化立体阵形并设计了声成像测量系统。改进的粒子群算法与基本粒子群算法仿真优化阵形比较表明改进粒子群算法设计的优化阵形在保持较窄的主瓣宽度的条件下具有较低的旁瓣级。阵列声成像测量系统的性能测量分析结果表明该系统的空间分辨率和旁瓣抑制能力与理论结果接近,验证了所提算法的有效性。
A modified PSO method was proposed to optimize the configuration of microphones for conflicting objects of the mainlobe width and sidelobe level to avoid that optimizing either of both performance would lead to the other performance degradation by basic particle swarm optimization (PSO). According to the limit of array viewing angle, a better cubic array was developed and an experimental device was designed. Comparison between the beampatterns by basic and modified PSO shows the array with the method proposed keeps the low sidelobe level on the condition of the narrow mainlobe width. Performance measurement results of spatial resolution and interference rejection capability for the device in a semi-anechoic chamber accords with the theoretical ones, which indicate the validity of the modified method.