The Hybrid HMM for RSS-Based Localization in Wireless Sensor Networks

We proposea method of RSS-base localization in WSN (Wireless Sensor Network), called Hybrid HMM, to improve the stabilityof node localization basedon RSS(Received Signal Strength).This model utilizesHMM(Hidden Markov Model) to takeinto account the time factor when receiving the RSS sequence, andconverts the action of ranging into an operationof classification.For the received RSS used for localization,our Hybrid HMMwill compare it withthe preset RSS threshold value, and put the result into one of two categories for subsequent processing: If the received value is higher than the threshold value, the distance value will be drawn from the signal propagation model. If lower, the information will be obtained from a trained HMM. Experimental results show that the Hybrid HMM method can greatly improve the localization accuracy.