中文摘要：

声源定位在军事、工业噪声定位等领域具有广泛应用,基于麦克风阵列信号的波束形成是声源定位的主要方法。与传统波束形成算法相比,最小方差无失真响应（MVDR）算法具有较大优势,但MVDR算法运算量大,一般处理器无法满足实时性要求。针对波束形成算法的数据量大,运算复杂的特点,选取TI公司推出的TMS320DM642定点DSP芯片,采用C和汇编语言混合编程,实现了8个麦克风组成的均匀圆阵基于MVDR算法在TMS320DM642DSP的应用,其处理速度比在ARM上快,具有速度快,精度高,实用性强的特点。

英文摘要：

Sound source location has a wide application in the military,industrial noise location and other areas;the beam-forming based on microphone array is the main method for sound source localization.MVDR algorithm has a greater advantage compared to conventional beam-forming algorithms,however,MVDR algorithm has a large amount of calculation and general processor can not meet the real-time requirement.Considering the characteristics of huge data and complex operations in the procession of beam-forming,the fixed-point DSP chip TMS320DM642 produced by TI Company is selected.Mixed by C and Assembly Language,the application of uniform circular array of eight microphones on TMS320DM642 DSP based MVDR algorithm can be realized.The processing speed is much faster than on the Advanced RISC Machines（ARM）,with the features of high speed,high precision and strong practicability.