Digital processing of random signals: theory and methods
Digital processing of random signals: theory and methods
Applicable Algebra in Engineering, Communication and Computing
Source-channel communication in sensor networks
IPSN'03 Proceedings of the 2nd international conference on Information processing in sensor networks
Estimation over fading channels with limited feedback using distributed sensing
IEEE Transactions on Signal Processing
Estimation Diversity and Energy Efficiency in Distributed Sensing
IEEE Transactions on Signal Processing
Linear Coherent Decentralized Estimation
IEEE Transactions on Signal Processing
Type based estimation over multiaccess channels
IEEE Transactions on Signal Processing
A proof of the Fisher information inequality via a data processing argument
IEEE Transactions on Information Theory
Universal decentralized estimation in a bandwidth constrained sensor network
IEEE Transactions on Information Theory
Distributed Estimation Via Random Access
IEEE Transactions on Information Theory
Uncoded Transmission Is Exactly Optimal for a Simple Gaussian “Sensor” Network
IEEE Transactions on Information Theory
| Hi-index | 35.68 |
A distributed estimation scheme where the sensors transmit with constant modulus signals over a multiple access channel is considered. The proposed estimator is shown to be strongly consistent for any sensing noise distribution in the i.i.d. case both for a per-sensor power constraint, and a total power constraint. When the distributions of the sensing noise are not identical a bound on their variances is shown to establish strong consistency. The estimator is shown to be asymptotically normal with a variance (AsV) that depends on the characteristic function of the sensing noise. Optimization of the AsV is considered with respect to a transmission phase parameter for a variety of noise distributions exhibiting differing levels of impulsive behavior. The robustness of the estimator in the sense of degrading gracefully when the sensing noise deviates from the Gaussian distribution to impulsive sensing noise distributions such as those with positive excess kurtosis, or those that do not have finite moments is shown. The proposed estimator is favorably compared with the amplify and forward scheme under an impulsive noise scenario. The effect of fading and phase error at the sensors are shown to not affect the consistency of the estimator, but to degrade the asymptotic variance by terms that depend on the fading and phase error distributions. Simulations corroborate our analytical results.