The Applications Of Distributed Fountain Codes In Wireless Sensor Networks

With the coming of Internet of Things era, wireless sensor networks have been an increasingly wide range of applications. Wireless sensor networks are used for data acquisition and processing within the region,which is a self-organizing network for data-centric. Compared to other networks,wireless sensor networks can be deployed in the harsh environment that humans can not or will not reach. However, in harsh environments nodes die easily, causing loss of monitoring data, making the network performance degradation. So how to ensure reliable and efficient data collection of sensor network has important research value.Now, the coding technology of the network received widespread attention, which has changed the traditional design ideas of reliable data collection. Digital fountain code is a non-rate encoding, and has the characteristics of simple encoding and decoding, which is very suitable for applications in wireless sensor networks and effectively improve the network performance of data collection. So studying fountain codes in data collection in sensor networks has important theoretical and engineering significance. In this paper, fountain codes in wireless sensor networks conducted in-depth research,and made some data distributed storage algorithms based on fountain code. The main work is as followsFirstly, the paper studies the distributed LT algorithm in wireless sensor networks. By improving the encoding mechanism, we propose an improved algorithm ELT algorithm,which can make the mix more evenly and improve the performance of distributed coding.Experimental results show that the algorithm improves the decoding efficiency and has better ability of anti-cliff effect than the original distributed LT code.Secondly, the paper studies the application of distributed Raptor code implemented in wireless sensor networks. By storing the pre-encoded data packets on the node forming the virtual node in the pre-encoding stage the algorithm achieves the pre-encoding process centre Raptor codes. In LT coding phase, the coded packets walk on the network. Packets may choose to the virtual node to complete the encoding. Simulation results show that the algorithm has better performance than the distributed LT code in the terms of anti-interference, decoding efficiency and anti-cliff effect.Finally, the paper studies the data collection problem in disaster scenario. This paper proposes a priority-based approach for data exchange, which achieves reliable data collection by sending the data to the relative safety of the area. Experimental results show that the algorithm can recover more raw data in the face of large node failure in the disaster environment.