Distributed State Estimation for Hidden Markov Models by Sensor Networks with Dynamic Quantization
Minyi Huang, Subhrakanti Dey · 2005
This paper considers the state estimation of hidden Markov models by sensor networks. We study a network structure with feedback from the fusion center to the sensor nodes, and a dynamic quantization scheme is proposed and analyzed by a stochastic control approach. The resulting dynamic programming equation is solved by the relative value iteration algorithm. Furthermore, a dynamic rate allocation method is also proposed.