基于稀疏过零点信息的抗幅值失真时延估计方法

Anti-Amplitude Distortion Time Delay Estimation Method Based on Sparse Cross-Zero Information

基于时延估计的TDOA方法广泛地应用于声学定位技术中,应用时引入的干扰会导致一定程度的幅值失真,从而影响时延估计的准确与稳定性.文中分析了不同干扰因素带来的影响,进而针对幅值失真问题,提出了一种新的基于信号过零点信息进行时延估计方法.该方法对需要进行时延估计的两路信号进行幅值归一化处理,得到仅含0、1两种特征值的稀疏化信号,在一定范围内将稀疏化信号错位相减得到误差系数函数,通过索引误差系数函数的最小值求得通道时延估计值.为了验证该方法的有效性,分别进行了数值模拟、耦合腔实验与室内混响定位实验,实验中采用广义互相关(GCC)时延估计方法作为对比,结果表明:对于模拟信号和耦合腔标准声源生成信号,该方法均有较好的抗干扰能力,在信噪比为-5 dB时依然可以准确进行时延估计计算;与GCC方法相比,可用于极端条件如信号削波时的时延估计;在低采样率实验中,该方法较GCC方法受采样率降低影响小,在低采样率的条件下依然保留一定时延估计能力;在室内混响条件下的近距离声源定位实验中,该方法受声源距离、混响影响小,并且较GCC方法具备更强的抗突发干扰能力.文中所提出的方法能很好地抑制信号削波、低采样率、混响干扰等因素导致的幅值失真带来的影响,提升复杂环境下时延估计的精度与稳定性,面对突发干扰也表现出了更好的抗干扰能力.同时该方法设计简单、计算量较小,可满足对数据实时处理的需求.

The time delay of arrival(TDOA)method based on time delay estimation is widely used in acoustic localization techniques.The interference introduced during its application can lead to a certain degree of amplitude distortion,which affects the accuracy and stability of time delay estimation.In this paper,the effects of different interference factors are analyzed,and a new method of time delay estimation based on signal cross zero information is proposed for the amplitude distortion problem.The method normalizes the amplitude of the two signals for delay estimation to obtain a sparse signal containing only two eigenvalues:0 and 1.The error coefficient function is obtained by subtracting the sparse signal from the misalignment within a certain range,and the channel delay estimation is obtained by indexing the minimum value of the error coefficient function.To verify the effectiveness of the method,numerical simulations,coupled-cavity experiments,and indoor reverberation localization experiments were carried out.The results were compared using the generalized cross-correlation(GCC)time delay estimation method.The results show that the proposed method has good immunity to interference for both analog and coupled-cavity standard source-generated signals.Moreover,the method can give accurate time delay estimation at a signal-to-noise ratio of-5 dB.Compared with the GCC method,the proposed method can be used for delay estimation under extreme condi-tions such as signal clipping.In low sampling rate experiments,the method is less affected by the reduced sampling rate than the GCC method,and it retains some delay estimation capability under low sampling rate conditions.In indoor reverberation experiments,the method is less affected by the source distance and reverberation and shows stronger resistance to sudden interference than the GCC method.The proposed method can suppress the effects of amplitude distortion caused by signal clipping,low sampling rate,and reverberation interference.Moreover,the proposed method results in improved accuracy and the stability of time delay estimation in complex environments and better immunity to sudden interferences.Additionally,the proposed method is simple in design,small in computa-tion,and can meet the demand for real-time data processing.