﻿ 基于GABP-KF的WSN数据漂移盲校准算法
«上一篇
 文章快速检索 高级检索

 智能系统学报  2019, Vol. 14 Issue (2): 254-262  DOI: 10.11992/tis.201712003 0

### 引用本文

WU Jiawen, LI Guanghui. GABP-KF-based blind calibration algorithm of data drift in wireless sensor networks[J]. CAAI Transactions on Intelligent Systems, 2019, 14(2): 254-262. DOI: 10.11992/tis.201712003.

### 文章历史

1. 江南大学 物联网工程学院，江苏 无锡 214122;
2. 物联网技术应用教育部工程技术研究中心，江苏 无锡 214122;
3. 江苏省模式识别与计算智能工程实验室，江苏 无锡 214122

GABP-KF-based blind calibration algorithm of data drift in wireless sensor networks
WU Jiawen 1,2, LI Guanghui 1,2,3
1. School of Internet of Things Engineering, Jiangnan University, Wuxi 214122, China;
2. Research Center of IoT Technology Application Engineering (MOE), Wuxi 214122, China;
3. Jiangsu Provincial Engineering Laboratory of Pattern Recognition and Computational Intelligence, Wuxi 214122, China
Abstract: Data drifts easily occur in wireless sensor network. To solve this problem, we propose a novel algorithm for tracking and calibrating the drifts of sensor data stream. First, backpropagation (BP) neural network optimized by genetic algorithm is applied to model the spatio–temporal correlations between the target node and its neighbor nodes to predict the value of the node, and then, the data drift of the node is tracked and calibrated by a Kalman filter. The simulation results using different datasets demonstrate that this method has superior prediction accuracy and calibration performance, compared with other methods. The experimental results show that this method can accurately calibrate the sensor drift and improve the reliability of node data.
Key words: wireless sensor network    data drift    blind correction    BP neural network    genetic algorithm    Kalman filter    denoising    model building

1 预备知识 1.1 遗传算法

1.2 卡尔曼滤波器

2 基于GABP-KF的WSN数据漂移校准方法 2.1 定义测量环境

 $X = T + d + W$ (1)

 ${{\overset{\frown} x} _i} = f({\rm neighbours}\_{\rm data})$ (2)

2.2 算法框架

GABP-KF算法的设计目标是在随机误差(噪声)和系统误差(漂移)的干扰下精确地校准目标节点i的测量值 ${x_i}$ 。该算法有两个运行阶段：训练阶段和测试阶段。在训练阶段，首先对所有节点数据进行去噪处理。然后使用来自邻居节点的去噪后测量值 ${\left\{ X \right\}_{m - 1}}$ 作为GABP神经网络的输入值，并输出节点i的预测值 ${{\overset{\frown} x} _i}$ 。在测试阶段，将预测值 ${{\overset{\frown} x} _i}$ 和节点i的去噪后测量值 ${y_i}$ 输入到卡尔曼滤波器中以跟踪其数据漂移 ${d_i}$ 。最后通过测量值 ${y_i}$ 减去估计漂移值 ${d_i}$ 以得到最终校准后的读数 ${\bar x_i}$ 图1描述了本文算法的具体框架。

 Download: 图 1 GABP-KF算法的数据漂移校准框架 Fig. 1 Data drift calibration framework of GABP-KF algorithm
2.3 基于GABP神经网络的建模过程 2.3.1 数据预处理

2.3.2 GABP神经网络的建模

 Download: 图 2 GABP神经网络的建模算法 Fig. 2 Modeling algorithm of GABP neural network

2.4 基于卡尔曼滤波的漂移跟踪算法

 ${d_{i,t}} = {d_{i,t - 1}} + {w_{i,t}},{w_{i,t}} \sim N(0,{Q_{i,t}})$ (3)

 ${{\textit{z}}_{i,t}} = {d_{i,t}} + {v_{i,t}},{v_{i,t}} \sim N(0,{R_{i,t}})$ (4)

 ${d_{i,t}} = {r_{i,t}} - {T_{i,t}}$ (5)

 ${d_{i,t}} = {r_{i,t}} - {{\overset{\frown} x} _{i,t}}$ (6)

 Download: 图 3 针对目标节点的漂移跟踪算法 Fig. 3 Drift tracking algorithm for target node

 ${\bar x_{i,t}} = {r_{i,t}} - {d_{i,t}}$ (7)
3 实验结果及分析

3.1 数据集 3.1.1 IBRL数据集

IBRL数据集来自于部署在Intel Berkeley实验室内的无线传感器网络，包含54个节点，用于监控实验室环境(参见图4)。

 Download: 图 4 IBRL的节点分布 Fig. 4 Deployment scheme of IBRL sensor network

3.1.2 LUCE数据集

LUCE数据集(洛桑城市冠层实验)来自于2006年7月以来部署在洛桑联邦理工学院内的无线传感器网络。该网络共包含97个节点，根据节点之间的时空相关性分为10组传感器节点集。

3.2 去噪处理

3.3 模型对比实验

 ${\rm{MSE}} = \frac{1}{n}\sum\limits_{i = 1}^n {{{({x_i} - {{\hat x}_i})}^2}}$ (8)
 ${{\rm{R}}^2} = \frac{{{{\left( {n\sum\limits_{i = 1}^n {{{\mathord{{\hat x}} }_i}{x_i}} - \sum\limits_{i = 1}^n {{{\mathord{{\hat x}} }_i}} \sum\limits_{i = 1}^n {{x_i}} } \right)}^2}}}{{\left( {n\sum\limits_{i = 1}^n {\mathord{{\hat x}} _i^2} - {{\left( {\sum\limits_{i = 1}^n {{{\mathord{{\hat x}} }_i}} } \right)}^2}} \right)\left( {n\sum\limits_{i = 1}^n {x_i^2} - {{\left( {\sum\limits_{i = 1}^n {{x_i}} } \right)}^2}} \right)}}$ (9)

 Download: 图 6 4种算法的模型对比结果 Fig. 6 Comparison among the training models of four algorithms
3.4 基于GABP-KF的漂移校准对比实验

 Download: 图 7 卡尔曼滤波后的校准示例 Fig. 7 Example of data drifts calibration after Kalman filter
3.5 整体评估

 ${\rm MAE} = \frac{1}{n}\sum\limits_{i = 1}^n {\left| {{X_i} - {{{\hat X}}_i}} \right|}$ (10)

 Download: 图 8 5种情形下MAE值的变化情况 Fig. 8 Comparison of MAE among five cases

4 结束语

 [1] WU Mou, TAN Liansheng, XIONG Naixue. Data prediction, compression, and recovery in clustered wireless sensor networks for environmental monitoring applications[J]. Information sciences, 2016, 329: 800-818. DOI:10.1016/j.ins.2015.10.004 (0) [2] KUMAR D, RAJASEGARAR S, PALANISWAMI M. Geospatial estimation-based auto drift correction in wireless sensor networks[J]. ACM transactions on sensor networks, 2015, 11(3): 50. (0) [3] TAN Rui, XING Guoliang, YUAN Zhaohui, et al. System-level calibration for data fusion in wireless sensor networks[J]. ACM transactions on sensor networks, 2013, 9(3): 28. (0) [4] RAMANATNAN N, BALZANO L, BURT M, et al. Rapid deployment with confidence: Calibration and fault detection in environmental sensor networks[R]. Technical Report CENS TR 62, Center for Embedded Networked Sensing, 2006. (0) [5] MILUZZO E, LANE N D, CAMPBELL A T, et al. CaliBree: A self-calibration system for mobile sensor networks[C]//Proceedings of the 4th IEEE International Conference on Distributed Computing in Sensor Systems. Berlin, Heidelberg, Germany: Springer-Verlag, 2008: 314–331. (0) [6] TAYLOR C, RAHIMI A, BACHRACH J, et al. Simultaneous localization, calibration, and tracking in an ad hoc sensor network[C]//Proceedings of the 5th International Conference on Information Processing in Sensor Networks. Nashville, TN, USA: IEEE, 2006: 27–33. (0) [7] LEE B T, SON S C, KANG K. A blind calibration scheme exploiting mutual calibration relationships for a dense mobile sensor network[J]. IEEE sensors journal, 2014, 14(5): 1518-1526. DOI:10.1109/JSEN.2013.2297714 (0) [8] BALZANO L, NOWAK R. Blind calibration of sensor networks[C]//Proceedings of the 6th International Conference on Information Processing in Sensor Networks. Cambridge, MA, USA: IEEE, 2007: 79–88. (0) [9] LI Zhan, WANG Yuzhi, YANG Anqi, et al. Drift detection and calibration of sensor networks[C]//Proceedings of 2015 International Conference on Wireless Communications & Signal Processing. Nanjing, China: IEEE, 2015: 1–6. (0) [10] WANG Yuzhi, YANG Anqi, LI Zhan, et al. Blind drift calibration of sensor networks using signal space projection and Kalman filter[C]//Proceedings of 2015 IEEE Tenth International Conference on Intelligent Sensors, Sensor Networks and Information Processing. Singapore, Singapore: IEEE, 2015: 1–6. (0) [11] TAKRURI M, ABOURA K, CHALLA S. Distributed recursive algorithm for auto calibration in drift aware wireless sensor networks[M]//ELLEITHY K. Innovations and Advanced Techniques in Systems, Computing Sciences and Software Engineering. Netherlands: Springer, 2008: 21–25. (0) [12] TAKRURI M, RAJASEGARAR S, CHALLA S, et al. Online drift correction in wireless sensor networks using spatio-temporal modeling[C]//Proceedings of 200811th International Conference on Information Fusion. Cologne, Germany: IEEE, 2008: 1–8. (0) [13] KUMAR D, RAJASEGARAR S, PALANISWAMI M. Automatic sensor drift detection and correction using spatial Kriging and Kalman filtering[C]//Proceedings of 2013 IEEE International Conference on Distributed Computing in Sensor Systems. Cambridge, MA, USA: IEEE, 2013: 183–190. (0) [14] DING Shifei, SU Chunyang, YU Junzhao. An optimizing BP neural network algorithm based on genetic algorithm[J]. Artificial intelligence review, 2011, 36(2): 153-162. DOI:10.1007/s10462-011-9208-z (0) [15] 丁建立, 陈增强, 袁著祉. 遗传算法与蚂蚁算法的融合[J]. 计算机研究与发展, 2003, 40(9): 1351-1356. DING Jianli, CHEN Zhenqiang, YUAN Zhuozhi. Genetic algorithm and ant algorithm fusion[J]. Journal of computer research and development, 2003, 40(9): 1351-1356. (0) [16] WELCH G, BISHOP G. An introduction to the Kalman filter[R]. Chapel Hill, NC, USA: University of North Carolina at Chapel Hill, 1995. (0) [17] 钟伟才, 刘静, 刘芳, 等. 二阶卡尔曼滤波分布估计算法[J]. 计算机学报, 2004, 27(9): 1272-1277. ZHONG Weicai, LIU Jin, LIU Fang, et al. Second order estimation of distribution algorithms based on kalman filter[J]. Chinese journal of computers, 2004, 27(9): 1272-1277. DOI:10.3321/j.issn:0254-4164.2004.09.018 (0) [18] TAKRURI M, RAJASEGARAR S, CHALLA S, et al. Spatio-temporal modelling-based drift-aware wireless sensor networks[J]. IET wireless sensor systems, 2011, 1(2): 110-122. DOI:10.1049/iet-wss.2010.0091 (0) [19] CHANG S G, YU Bin, VETTERLI M. Adaptive wavelet thresholding for image denoising and compression[J]. IEEE transactions on image processing, 2002, 9(9): 1532-1546. (0)