基于TSK模糊系统的非均匀分簇算法在无线传感网中的应用

Application of heterogeneous clustering routing algorithm based on interval II TSK fuzzy system and efficient data fusion in wireless sensor networks

摘要:

无线传感网作为一个重要的数据调度工具,是人类同自然交互的有效途径。然而,无线传感网中的传感器数量有限,同时还极易被无关因素影响。因此,研究提出一种基于区间二型TSK模糊系统和高效数据融合的非均匀分簇路由算法,分簇是为了最大程度地减小网络损耗,延长其生存时间,一般的分簇方法,可能会导致负载失衡,为解决这种热区现象,研究使用区间二型TSK模糊逻辑算法,进行非均匀分级分簇。同时引入高效数据融合技术,将调度过程划分为几个周期,利用时间间隙进行数据采集,并进行降维操作,进一步提升数据传输效率。研究在MATLAB平台,对该算法以及自适应分簇层次等其余四种算法进行对照分析实验,并将其分为200m×200m和1000m×1000m的监测范围,实验结果表明,在不同大小的监测区域中,研究使用算法的分簇效果、网络吞吐量以及节点损耗率指标,都明显优于其他算法。其中,在小范围区域内,其剩余能力均值比自适应分簇层次算法提升了49.7%；在大范围区域内,该算法的HND指标比自适应分簇层次算法提升了98.7%。因此,研究采用算法具有极佳的性能。

Abstract:

As an important data scheduling tool, wireless sensor network is an effective way for human to interact with nature. However, the number of sensors in wireless sensor networks is limited and easily affected by irrelevant factors. Therefore, this paper proposes a non-uniform clustering routing algorithm based on the interval type-2 TSK fuzzy system and efficient data fusion. The clustering is to minimize network losses to the greatest extent and extend its life time. The general clustering method may lead to load imbalance. Heterogeneous classification and clustering were performed. At the same time, the efficient data fusion technology is introduced to divide the scheduling process into several cycles, use the time gap for data acquisition, and reduce the dimension operation to further improve the efficiency of data transmission. On the MATLAB platform, this algorithm and the other four algorithms, such as adaptive clustering level, are analyzed and tested, and they are divided into monitoring ranges of 200m×200m and 1000m×1000m. The experimental results show that in monitoring areas of different sizes, the clustering effect, network throughput and node loss rate indexes of the algorithm used in the study are as follows: Are obviously due to other algorithms. In a small range, the average residual capability of the algorithm is improved by 49.7% compared with the adaptive clustering hierarchical algorithm. In a large range, the HND index of the proposed algorithm is improved by 98.7% compared with the adaptive clustering hierarchical algorithm. Therefore, the algorithm adopted in this study has excellent performance.