坚实信号处理的半无穷维二次优化方法

SEMI-INFINITE QUADRATIC OPTIMIZATION METHOD FOR ROBUST SIGNAL PROCESSING

采用普通方法设计的传感器阵列处理器对有用信号参量的变化极为敏感,有用信号参量的微小变化均可能引起处理性能严重下降。本文讨论阵列处理器的一种坚实性设计方法。这种方法将阵列处理器设计问题转化为半无穷维二次优化模型。该模型使输出功率最小化,通过选择约束来确保有用信号响应,从而允许有用信号可在一定范围内变化以实现坚实性。最后,文中给出了一个用该方法设计的天线阵列处理器的例子,证明了该方法的有效性。

Sensor array processors designed by ordinary methods are very sensitive tothe variations of desired signal parameters.A minute deviation from ideal values of theseparameters may deteriorate the performance of the processor remarkably.This paper pres-ents a robust processor design method.The problem of designing sensor array processorsis formulated as semi-infinite quadratic optimization,which minimizes the output noisepower or the total output power,subject to hard constraints.The constraints allowcentain uncertainties of the desired input signal,resulting a robust processor.As anexample the method is applied to a narrowband antenna array system to showits effectiveness.