基于环形阵列的近场多声源波达方向跟踪

Tracking DOAs of multiple near field acoustic sources using circular sensor array

针对基于可控功率响应的多声源跟踪算法的定位复杂度高,且弱声源易受强声源影响而造成无法跟踪的问题,提出了一种利用环形阵列的近场多声源波达方向跟踪方法。首先利用语音信号在时—频域的稀疏性和声源间的屏蔽效应定义了信号的相关性,并利用相关性检测主声源较强的帧,其次利用环形阵列结合SRP方法获取单帧主声源的方位角,然后集合多帧的数据从方位角的统计直方图中分析出当前声源的数目和所在方位。仿真实验结果表明,该方法对声源数目估计的准确率为92%,且具有较高的方位角跟踪精度,其误差小于1.5°,因此可有效应用于近场环境中声源数目和方位的跟踪中。

For complexity of multiple acoustic sources tracking algorithm based on steered response power（ SRP） is very high,and the weak sound sources are unable to track for the impact of the strong sources,this paper proposed a method for tracking direction of arrivals（ DOA） of multiple acoustic sources. First,it defined the correlation between signals by using the sparsity of speech signal and the masking effect of the sound sources. It used the correlations of the signals to detect frames with strong main source. Second,it obtained the DOAs of single source frames by using a circular array combined with SRP method.Finally,analyzed the number of the sound sources and their DOAs from the histogram of DOAs which derived from a set of consequent frames. The source number estimation accuracy rate of the proposed method is 92%,the DOAs of sound sources are also estimated with high accuracy,the errors are less than 1. 5°. Therefore,it can be effectively applied to tracking multiple sound sources and their DOAs in the near field environment.