基于麦克风阵列的远场声源目标定位技术研究

发布时间：2018-09-19 16:23

【摘要】：随着信号处理技术的发展,基于麦克风阵列的声源定位越来越成为这个领域的研究热点。在军事、交通、智能机器人等方面都有着较为广泛的应用前景和重要的应用价值。本文设计了一套基于麦克风阵列的信号采集实验平台,该硬件平台由麦克风阵列、NI-9234数据采集卡、计算机三部分组成,声源目标所发出的声音信号由麦克风转化成电压信号,数据采集卡将采集到的模拟信号转化成为数字信号并通过网口传输到计算机,以便对其进行研究处理。在搭建了信号采集平台的基础上,深入研究了基于麦克风阵列的声源定位技术,首先对传统的利用麦克风阵列确定声源位置的方法进行了比较分析,确定了基于时延估计的声源定位方法。其次对采集到的声音信号进行滤波降噪、加窗、端点检测等预处理,使其能够达到使用时延估计算法的要求。然后对时延估计的方法进行了系统的研究,通过分析比较了几种时延估计算法和加权函数,发现了相位变换(PHAT)加权广义互相关时延估计算法的峰值比其他普通算法更加明显,抗噪能力更强,并对PHAT加权函数进行改进使其更加适用于本课题所研究的外部环境以及单一声源。最后研究了时差定位算法,分析研究了在二维平面以及三维空间上的时差定位情况并进行仿真分析,确定该方法的可行性。在理论研究的基础上,开发设计了一套声源目标定位的实验验证软件系统,并进行了模拟仿真以及实际测试,对比了仿真与实测结果,结果表明本文所采用的方法能够对远场声源目标进行定位,定位精度达到预期目标,且算法实现简单,运算量小。定位中发现可以通过抛物线拟合、插值法升采样率、温度补偿法和增加麦克风阵元数量来提高定位的精度,在此基础上对声源定位系统的误差进行了分析。
[Abstract]:With the development of signal processing technology, sound source location based on microphone array has become a hotspot in this field. It has wide application prospect and important application value in military, traffic, intelligent robot and so on. In this paper, a signal acquisition experiment platform based on microphone array is designed. The hardware platform is composed of a microphone array NI-9234 data acquisition card and a computer. The sound signal from the sound source target is converted from microphone to voltage signal. The data acquisition card converts the collected analog signal into digital signal and transmits it to the computer through the network port in order to study and process it. On the basis of building a signal acquisition platform, the sound source location technology based on microphone array is deeply studied. Firstly, the traditional method of using microphone array to determine sound source location is compared and analyzed. The method of acoustic source location based on time delay estimation is determined. Secondly, the acquired sound signal is preprocessed by filtering, windowing and endpoint detection, so that it can meet the requirements of the time delay estimation algorithm. Then the method of time delay estimation is studied systematically. By analyzing and comparing several time delay estimation algorithms and weighting functions, it is found that the peak value of phase transform (PHAT) weighted generalized cross-correlation time delay estimation algorithm is more obvious than other ordinary algorithms. The ability of anti-noise is stronger, and the PHAT weighting function is improved to be more suitable for the external environment studied in this paper and the single sound source. Finally, the algorithm of TDOA is studied, and the situation of TDOA in 2D plane and 3D space is analyzed and simulated to determine the feasibility of the method. On the basis of theoretical research, an experimental verification software system for target location of sound source is developed, and the simulation and actual test are carried out, and the results of simulation and measurement are compared. The results show that the method proposed in this paper can locate the far-field sound source target and achieve the desired accuracy. The algorithm is simple and the computation is small. It is found that the positioning accuracy can be improved by parabola fitting interpolation method increasing sampling rate temperature compensation method and increasing the number of microphone array elements. On this basis the error of sound source positioning system is analyzed.
【学位授予单位】：沈阳理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TN912.3

【参考文献】