Research Of Continuous Object Boundary Monitoring Based On WSNs

Continuous object is kind of object that its size and shape can not be ignored. It’s different from the particle object and it’s a kind of important monitoring target. Continuous objects such as water, air, fog, etc. are produced with human’s daily activities. In this paper, the objects which could be harmful for humans are the primary study targets, such as hazardous gas, haze, or even radiation etc. In convertional monitoring system, the general targets are particle objects. Hence, the primary mission of the monitoring system is to detect the path of the target, to predict the trend of the movement. However, for the continuous object, to detect the boundary of the continuous objects is the primary mission. At present, monitoring for continuous objects deeply rely on the density of the sensors and this has restricted the application of WSNs(Wireless Sensor Networks) for continuous object monitoring. Therefore, the purpose of this paper is to research the WSNs based continuous object monitoring, with limited sensor density.Firstly, the research background and start of art are presented. Secondly, the distribution and boundary pattern are defined based on the fuzzy logical analysis, and according to the characteristics of the continuous object the dynamic evolution model for continuous object is proposed. Based on the dynamic model, the distrubition feature of the continuous object is analyzed qualitatively. Finally, the optimal fusion set based clustering algorithm and BP neural network based global distribution status monitoring algorithm for large scale WSNs are proposed.To analyze and verify the perofomance of the proposed algorithms, numerical calculation and simulation are utilized. The optimal fusion set based clustering algorithm is compared to the traditional clustering algorithms and it shows that the proposed clustering algorithm have more advantages for continuous object monitoring. On the other hand, to offer the predefined principle of BP network design, the effect of network performance to network struct has been analyzed through testing the struct of single-hide-layer and two-hide-layer BP network. Finally, the BP network based global distribution status monitoring algorithm has been verified, and the restricted relationship between the density of the sensors and the monitoring accuracy has obtained.