Research On Data Collection And Interpolation Algorithms In Wireless Sensor Networks Based On Environmental Modeling

In wireless sensor networks, static sink is often used in current data collection methods, which leads to such problems such as the sensor nodes surrounding the sink die prematurely because of being overloaded, partitioning the network to form coverage holes; On the other hand, there are usually coverage holes caused by uneven deployment of sensor nodes, the inherent characteristic of sensor nodes also can lead to missing sensing data, and they bring many difficulties for various applications in wireless sensor networks. To solve these problems, the best way is to balance the network energy consumption and estimate the missing values as accurately as possible. This paper mainly opens out the research on the sensor data collection algorithm and interpolation algorithm based on Environmental Modeling. The main contributions of this paper are as follows:Firstly, the data collection problem with path controllable mobile sink is studied in this paper. Static sink can lead to the hotspots problem, so a path controllable traversal model for mobile sink is constructed. Based on it we propose a data collection algorithm with optimal path traversal mobile sink to achieve maiximizing the amount of data collection and minimizing the system energy consumption. Simulations under OMNET++show that the proposed algorithm can achieve maximizing the network throughput, balancing energy consumption of the sensor nodes, prolonging network lifetime and outperforming static sink methods in terms of energy utilization efficiency.Secondly, the coverage holes and missing sensing data are studied in this paper. Sensor data has highly temporal and spatial correlations, a time series model is constructed to exploit temporal correlation for missing sensing data of single sensor node and a new method named forcast processing algorithm based on sensing data time series model fitting is proposed. Moreover, we construct a spatial correlation model for multiple sensor nodes and propose an adaptive interpolation algorithm based on spatial correlation model to solve the coverage holes. Combining above two algorithms, an adaptive interpolation algorithm based on temporal-spatial correlation is proposed. The algorithm can adaptively select neighborhood sensor nodes with different correlations by presetting error threshold, then, the value of the interpolation point is able to estimate by our proposed algorithm according to the observations of neighbor set. Experimental results on a real-world dataset show that our proposed algorithms can estimate the value of coverage holes more accurately and robustly. Besides, it can achieve to retrieve the real-time value of a point in the sensor network.Thirdly, combining above proposed schemes in the dissertation, a simulator named ISSWSN is designed and developed based on open source OMNET++framework to evaluate the performance of the proposed algorithms. ISSWSN mainly focuses on estimating the coverage holes and missing sensing data, building protocol stacks from networks to application layer, universal interfaces between layers are provided to enhance the accessibility and expansibility of the emulator. Simulation results show that ISSWSN has good accuracy, is able to provide efficient evaluation for our proposed algorithms.