Abstract:

In view of both the pollution distribution of the sensor observation noise and the noise in the communication channels in a wireless sensor network,two novel robust distributed estimation schemes,namely,QME and TME,are proposed in this paper. Then,the performances of the two proposed schemes are verified by a simulation and are compared with those of the MLE scheme with the known noise covariance. The results show that the QME scheme is superior to the TME scheme and the MLE scheme with the known noise covariance in terms of estimation performance; and that the QME and TME schemes asymptotically achieve the CRLB with the increase in the number of sensor nodes.

Key words: wireless sensor networks, distributed estimation, robustness, quantized mean estimator, trimmed mean estimator