A constrained MDP approach to dynamic quantizer design for HMM state estimation
IEEE Transactions on Signal Processing
| Hi-index | 35.69 |
This correspondence considers the state estimation of hidden Markov models (HMMs) by sensor networks where the sensor nodes communicate with a fusion center via bandlimited fading channels. The objective is to minimize the long-term average of the mean-square state estimation error for the underlying Markov chain. By employing feedback from the fusion center, a dynamic quantization scheme for the sensor nodes is proposed and analyzed by a Markov decision approach. The performance improvement by feedback and power control at the sensor nodes, as well as the effect of fading, is illustrated. This leads to a systematic optimization framework for distributed and collaborative information processing in wireless sensor networks.