WMSN Network Optimization Method Based On Miners’ Behavior Analysis

China’s coal mine production environment is relatively harsh with the result that the proportion of accidents is far higher than the world average. Application of wireless sensor media networks in the underground can raise awareness of their environment, detect danger and reduce the accident rate. In this case, this paper presents a wireless multimedia sensor network optimization method based on the miners’behaviors analysis, so as to improve the safety of miners down the mine from both the miners’behaviors and wireless sensor multimedia networks.This paper studies the behavior patterns of moving miners under the mine, point out their features and establish moving miners behaviors analytical model based Hidden Markov Model. In this model, the miners’ movement trajectory is perceived as the observation sequence, while state of the miners as ’work’,’rest’as a hidden sequence. Then, by machine learning methods, two parameters:the miners’position transition matrix and emission probability matrix are determined with their moving behaviors describing. Since hidden Markov models was unfit for prediction location and hidden state which was hard to find, this paper presents simplified model which set the time as a hidden state. In the study of the moving groups down the mine, this paper presents the concept of wireless multimedia sensor monitoring hot spots and base it, using Markov shift hotspots theory to propose an optimizing approach on monitoring and energy saving in use of wireless multimedia sensor down the mine.The experiments proved that behavior patterns of moving miners under the mine based on Hidden Markov model reveal the latent association between the miners moving behaviors and miners moving position. As a result, the miners daily moving and position can be predicted and through it to develop abilities of detecting abnormal behavior of the miners down the mine. Simplified model can describe higher degree of movement model associated with the time and place. Finally, based on the groups miners behaviors analysis, mine monitoring network lifetime will improve.