Abstract:
Wireless sensor networks, including wireless acoustic sensor networks, have found applications in diverse areas including hearing aids, hands-free telephony and target tracking. The objective of this brief is to introduce a new sparsity regularization parameter in sparse distributed network estimation, to achieve a better estimation accuracy in comparison with existing sparse-aware algorithms. In order to further reduce the computational complexity, the algorithm has also been designed for heterogeneous sensor networks, where only a fraction of the sensor nodes use sparse-aware adaptive estimation schemes.