Abstract:

In emergency rescue, the situation map (plume) of gas propagation is normally gained by modeling method. With the development of WSN (wireless sensor network), it provides a new strategy to plot accuracy situation map due to its strengths in fast deployment and distributed sensing. However, in practice, the nodes may be deployed limitedly and erroneously, which could cause blind area and low spatial resolution. To solve this problem, a method that can improve the resolution of the situation map is proposed. The method use potential energy map yielded by Gaussian mask to generate the sensor subnode position. Then the concentration value of subnode is evaluated by Bayesian classifier. Also the vector field of environment is concerned to simplify the course of parameter learning. The simulation result shows that the method can improve the spatial resolution of the situation map by 1~3 times, and decrease the blind area by 10%~30%. Furthermore, the algorithm can be used for real time applications.

Key words: wireless sensor networks;situation map;sensor subnode;Bayesian classifier