Distributed Sampling Rate Offset Estimation Over Acoustic Sensor Networks Based on Asynchronous Network Newton Optimization

Abstract

Sampling rate synchronization is an inevitable issue in distributed acoustic sensor networks. In this paper, an analytical sampling rate offset (SRO) estimation approach is first proposed, and then, it is extended to a distributed method that suitable for acoustic sensor networks with arbitrary communication graphs. Specifically, a linear-phase drift model in the short-time Fourier transform domain is used to approximate the SRO between each pair of microphone nodes. Next, after unwrapping the temporally averaged phase information, SROs are recovered analytically via a new weighted-sum criterion. Based on this, a distributed cost function is established at each node to obtain the SROs of all nodes simultaneously in a distributed manner. Finally, a state-of-the-art distributed algorithm named asynchronous network Newton optimization is adopted to carry out the distributed SRO estimation. The proposed method can effectively estimate the SROs among acoustic sensor nodes in noisy and reverberant environments. Compared with the existing approaches, it does not require an external central processor, and only local communications among nodes are needed. Experimental results confirm the validity of the proposed method.

Efficient Joint Optimization of Sampling Rate Offsets Using Entire Multichannel Signal

“…They are thus suitable for tracking time-varying SRO frame-by-frame [33], [34]. On the other hand, their estimation accuracy is limited compared with the following CM-based and ML-based methods [21], [35].…”

“…To exploit the full redundant relationships, a recent method estimates SROs for the M C 2 pairs and takes their consensus [35]. It minimizes the following total cost function to estimate SROs with respect to the reference channel 3 :…”

“…We expect this grid resolution to be fine enough to avoid local optima. In addition to these pair-wise methods, we evaluated the consensus-based method [35] in (15) where ζ n,m was estimated by the pairwise ML-based method. The optimization problem in (15) was solved by the Levenberg-Marquardt method [47].…”

“…The total cost function in(15) is modified from that in the original paper[35]. This is because a different definition of SROs and some additional assumptions are employed in the original paper.…”