基于区间证据理论的多传感器数据融合水质判断方法

doi:10.11959/j.issn.1000-436x.2016174

摘要/Abstract

摘要：

针对传感网测量水质数据过程中存在的必然不确定性和随机不确定性，提出了一种基于区间证据理论的多传感器数据融合水质判断方法。考虑传感器精度误差以及测量数据异常等问题，将每个传感器测量的水质数据用区间数表示，通过计算水质数据与每个水质等级特征值之间的距离，得到判断水质等级的区间证据。按照区间证据组合规则将多传感器的区间证据融合成综合区间证据，最后根据决策准则，由综合区间证据判断水质等级。实验表明，该方法能够从不确定水质数据中准确判断水质等级。

关键词: 水质判断, 区间证据理论, 传感网, 多传感器数据融合, 不确定性

Abstract:

For the inevitable uncertainty and random uncertainty in the process of measuring water quality data with the sensor network,a multi-sensor data fusion method for water quality evaluation based on interval evidence theory was proposed.Considering the precision error of sensor and the abnormalities of measured data,every water quality data measured by sensor was represented by interval number.By calculating the distance between the water quality data and the features of each water quality class,the interval evidence of water quality class was acquired.According to the interval evidence combining rule,a comprehensive interval evidence was obtained by combining the interval evidence of each sensor.Finally,the water quality class was determined based on the comprehensive interval evidence by the decision rule.Experiments show that the proposed method can evaluate water quality class more accurately from the uncertain water quality data.

Key words: water quality evaluation, interval evidence theory, sensor network, multi-sensor data fusion, uncertainty

周剑,马晨昊,刘林峰,孙力娟,肖甫. 基于区间证据理论的多传感器数据融合水质判断方法[J]. 通信学报, 2016, 37(9): 20-29.

Jian ZHOU,Chen-hao MA,Lin-feng LIU,Li-juan SUN,Fu XIAO. Multi-sensor data fusion method for water quality evaluation based on interval evidence theory[J]. Journal on Communications, 2016, 37(9): 20-29.

图/表 12

图1

图2

表1

表2

表3

表4

传感器S1、S2和S3修正后的各mass函数值"

信度	Ⅰ	Ⅱ	Ⅲ	Ⅳ	Ⅴ	Θ
${[m]}^{'}_{1}$	[0.072 3,0.104 8]	[0.112 1,0.196 8]	[0.279 4,0.406 4]	[0.080 3,0.178 0]	[0.041 3,0.085 4]	[0.095 3,0.158 2]
${[m]}^{'}_{2}$	[0.054 0,0.152 1]	[0.112 8,0.342 2]	[0.113 8,0.343 1]	[0.036 7,0.123 8]	[0.021 6,0.071 5]	[0.046 1,0.196 7]
${[m]}^{'}_{3}$	[0.054 7,0.099 9]	[0.081 5,0.176 1]	[0.200 5,0.382 1]	[0.064 8,0.196 0]	[0.033 7,0.091 2]	[0.090 7,0.236 3]

表4

表5

S1、S2和S3区间证据组合结果"

信度	Ⅰ	Ⅱ	Ⅲ	Ⅳ	Ⅴ	Θ
${[m]}^{'}_{1} \oplus {[m]}^{'}_{2}$	[0.035 3,0.135 6]	[0.098 1,0.334 8]	[0.224 4,0.544 8]	[0.029 5,0.171 7]	[0.012 9,0.076 8]	[0.011 2,0.069 8]
${[m]}^{'}_{1} \oplus {[m]}^{'}_{2} \oplus {[m]}^{'}_{3}$	[0.016 3,0.117 6]	[0.059 2,0.345 6]	[0.250 7,0.711 5]	[0.018 4,0.196 1]	[0.005 7,0.070 3]	[0.002 8,0.036 8]

表5

表6

表7

表8

传感器S4和S5修正后的各mass函数值"

信度	Ⅰ	Ⅱ	Ⅲ	Ⅳ	Ⅴ	Θ
${[m]}^{'}_{4}$	[0.214 6,0.362 9]	[0.141 1,0.289 4]	[0.047 9,0.141 9]	[0.031 1,0.073 9]	[0.020 3,0.050 5]	[0.081 4,0.229 7]
${[m]}^{'}_{5}$	[0.806,0.117 5]	[0.122 7,0.233 5]	[0.281 7,0.415 1]	[0.080 4,0.168 7]	[0.041 7,0.085 2]	[0.067 4,0.113 3]

表8

表9

S1、S2、S3、S4、S5区间信度组合结果"

信度	Ⅰ	Ⅱ	Ⅲ	Ⅳ	Ⅴ	Θ
${[m]}^{'}_{1} \oplus {[m]}^{'}_{2} \oplus {[m]}^{'}_{3} \oplus {[m]}^{'}_{4}$	[0.018 6,0.252 2]	[0.057 0,0.513 2]	[0.152 9,0.742 9]	[0.008 6,0.195 9]	[0.002 5,0.067 4]	[0.000 8,0.026 4]
${[m]}^{'}_{1} \oplus {[m]}^{'}_{2} \oplus {[m]}^{'}_{3} \oplus {[m]}^{'}_{4} \oplus {[m]}^{'}_{5}$	[0.005 5,0.180 8]	[0.029 8,0.517 7]	[0.199 1,0.883 0]	[0.003 7,0.116 2]	[0.000 6,0.043 3]	[0.000 0,0.008 3]

表9

表10

参考文献 24

[1]	ZHOU J , LIU L F , GUO J ,et al. Multisensor data fusion for water quality evaluation using Dempster-Shafer evidence theory[J]. International Journal of Distributed Sensor Networks, 2013,(2): 1-6.
[2]	HALL J , SZABO J . Water sentinel online water quality monitoring as an indicator of drinking water contamination[R]. New York:Environmental Protection Agency, 2005.
[3]	卞贺明, 边超, 佟建华 ,等. 用于氨氮检测的三维微纳氨气传感芯片及系统研究[J]. 电子与信息学报, 2012,34(5): 1258-1262. BIAN H M , BIAN C , TONG J H ,et al. Miniaturized 3D micro/nano ammonia sensor chip and system for measurement of ammonia nitrogen[J]. Journal of Electronics & Information Technology, 2012,34(5): 1258-1262.
[4]	张珏 . 基于无线传感器网络的水质在线监测系统研究[D]. 重庆:重庆大学, 2010. ZHANG J . Research on water quality online monitoring system based on wireless sensor network[D]. Chongqing:Chongqing University, 2010.
[5]	闫光 . 基于多传感器信息融合的动态目标检测与识别[D]. 北京:北京理工大学, 2015. YAN G . Moving target detection and recognition based on multi-sensor information fusion[D]. Beijing:Beijing Institute of Technology, 2015.
[6]	赵小敏, 梁学利, 蒋双双 ,等. 安全的 WSN 数据融合隐私保护方案设计[J]. 通信学报, 2014,35(11): 154-161. ZHAO X M , LIANG X L , JIANG S S ,et al. Design of secure privacy-preserving data aggregation scheme for wireless sensor network[J]. Journal on Communications, 2014,35(11): 154-161.
[7]	乐俊, 张维明, 肖卫东 ,等. 无结构动态适应无线传感器网络数据融合算法[J]. 通信学报, 2012,33(9): 53-65. YUE J , ZHANG W M , XIAO W D ,et al. Structure-free and dynamicadaptive data fusion algorithm for wireless sensor networks[J]. Journal on Communications, 2012,33(9): 53-65.
[8]	康兵义, 李娅, 邓勇 ,等. 基于区间数的基本概率指派生成方法及应用[J]. 电子学报, 2012,40(6): 1902-1906. KANG B Y , LI Y , DENG Y ,et al. Determination of basic probability assignment basedon interval numbers and its application[J]. Acta Electronica Sinica, 2012,40(6): 1902-1906.
[9]	BREHARD T , KRISHNAMURTHY V . Optimal data incest removalin Bayesian decentralized estimation over a sensor network[C]// The IEEE International Conference on Acoustics,Speech and Signal Processing. 2007: 173-176.
[10]	WILSON D H , ATKESON C . Simultaneous tracking and activity recognition (STAR) using many anonymous,binarysensors[C]// The 3rd International Conference on Pervasive Computing. 2005: 62-79.
[11]	MASKELL S . A Bayesian approach to fusing uncertain,impreciseand conflicting information[J]. Information Fusion, 2008,9(2): 259-277.
[12]	于美婷, 赵林靖, 李钊 . 基于 DS 证据理论的协作频谱感知改进方法[J]. 通信学报, 2014,35(3): 168-173. YU M T , ZHAO L J , LI Z . Improved cooperative spectrum sensing scheme based ondempster-shafer theory in cognitive radio network[J]. Journal on Communications, 2014,35(3): 168-173.
[13]	BASIR O , YUAN X . Engine fault diagnosis based on multisensory information fusion using dempster-shafer evidencetheory[J]. Information Fusion, 2007,8(4): 379-386.
[14]	ZHU D , GU W . Sensor fusion in integrated circuit faultdiagnosis using a belief function model[J]. International Journalof Distributed Sensor Networks, 2008,4(3): 247-261.
[15]	LI J M , LUO S H , JIN J S . Sensor data fusion for accuratecloud presence prediction using Dempster-Shafer evidencetheory[J]. Sensors, 2010,10(10): 9384-9396.
[16]	ALEFELDA G , MAYERB G . Interval analysis:theory and applications[J]. Journal of Computational and Applied Mathematics, 1999,121(1/2): 421-464.
[17]	曾文艺, 罗成忠, 肉孜阿吉 . 区间数的综合决策模型[J]. 系统工程理论与实践, 1997,17(11): 48-50. ZENG W Y , LUO C Z , ROZI H . Comprehensive decision model of iInterval-number[J]. System Engineering Theory and Practice, 1997,17(11): 48-50.
[18]	DEMPSTER A P . Upper and lower probabilities induced by amultiple valued mapping[J]. The Annals of Mathematical Statistics, 1967,38(2): 325-339.
[19]	SHAFER G A . Mathematical theory of evidence[M]. Princeton: Princeton University PressPress, 1976.
[20]	DENOEUX T . Modelling vague belief using fuzzy-valued beliefstructures[J]. Fuzzy Sets and Systems, 2000,116(2): 167-199.
[21]	WANGY M , YANGJ B , XU D L ,et al. The evidential reasoning approach for multiple attribute decision analysis using interval belief degrees[J]. European Journal of Operational Research, 2006,175(1): 35-66.
[22]	冯海山, 徐晓滨, 文成林 . 基于证据相似性度量的冲突性区间证据融合方法[J]. 电子与信息学报, 2012,34(4): 851-857. FENG H S , XU X B , WEN C L . A new fusion method of conflicting interval evidencebased on the similarity measure of evidence[J]. Journal of Electronics & Information Technology, 2012,34(4): 851-857.
[23]	曾文艺, 赵宜宾 . 基于区间数度量的区间值模糊集合的归一化距离、相似度、模糊度和包含度的关系研究[J]. 模糊系统与数学, 2012,26(2): 81-90. ZENG W Y , ZHAO Y B . Relationship among the normalized distance,the similarity measure,the entropy and the inclusion measure of interval-valued fuzzy setsbased on interval-number measurement[J]. Fuzzy Systems and Mathematics, 2012,26(2): 81-90.
[24]	徐祖信 . 我国河流单因子水质标识指数评价方法研究[J]. 同济大学学报(自然科学版), 2005,33(3): 321-325. XU Z X . Single factor water quality ident ification index for environmentalquality assessment of surface water[J]. Journal of Tongji University (Natural Science), 2005,33(3): 321-325.

水质参数	Ⅰ	Ⅱ	Ⅲ	Ⅳ	Ⅴ
DO	[7.5,10]	[6,7.5]	[5,6]	[3,5]	[2,3]
NH₃-N	[0,0.15]	[0.15,0.5]	[0.5,1.0]	[1.0,1.5]	[1.5,2.0]
TP	[0,0.01]	[0.01,0.025]	[0.025,0.05]	[0.05,0.1]	[0.1,0.2]
TN	[0,0.2]	[0.2,0.5]	[0.5,1.0]	[1.0,1.5]	[1.5,2.0]

传感器	DO	NH₃-N	TP	TN
S₁	[5.0,6.0]	[0.70,0.80]	[0.037,0.047]	[0.71,0.81]
S₂	[5.5,6.5]	[0.45,0.55]	[0.020,0.030]	[0.45,0.55]
S₃	[4.8,5.8]	[0.75,0.85]	[0.035,0.045]	[0.77,0.87]

信度	Ⅰ	Ⅱ	Ⅲ	Ⅳ	Ⅴ
[m]₁	[0.084 1,0.121 9]	[0.130 4,0.228 8]	[0.324 9,0.472 6]	[0.093 4,0.207 0]	[0.048 0,0.099 4]
[m]₂	[0.064 3,0.181 0]	[0.113 5,0.407 3]	[0.040 7,0.408 5]	[0.043 7,0.147 3]	[0.025 7,0.085 2]
[m]₃	[0.068 3,0.124 9]	[0.101 9,0.220 1]	[0.160 0,0.477 6]	[0.081 1,0.245 0]	[0.042 1,0.114 0]

传感器	DO	NH₃-N	TP	TN
S₄	[7.3,8.3]	[0.08,0.18]	[0.004,0.014]	[0.13,0.23]
S₅	[5.3,6.3]	[0.70,0.80]	[0.027,0.037]	[0.69,0.79]

信度	Ⅰ	Ⅱ	Ⅲ	Ⅳ	Ⅴ
[m]₄	[0.093 5,0.453 6]	[0.139 9,0.361 8]	[0.059 9,0.177 4]	[0.038 9,0.092 3]	[0.025 3,0.063 1]
[m]₅	[0.089 6,0.130 6]	[0.136 3,0.259 5]	[0.312 6,0.461 2]	[0.089 3,0.187 4]	[0.046 4,0.094 7]