智能环境下基于边缘设备规则推理的数据预部署研究

doi:10.12263/DZXB.20200891

摘要/Abstract

摘要：

在现有的规则推理机制下，大量的传感器数据导致的过大规则匹配期间的实时特征计算量降低了推理实时性，同时边缘设备受限的内存资源难以应对如此庞大的数据量.为此，本文设计了数据预部署方案（Data Pre-Deployment Scheme，DPDS）.利用规则解析与预处理模块解析规则集得到的规则网络和轻量级特征表（Light-weight Characteristic Table，LCT），该方案无需进行实时特征计算，使推理效率和实时性得到显著提高，并大大降低了规则匹配期间的内存占用量.实验表明，即使在规则、数据规模很大的情况下，DPDS仍然具有较高的时间效率和空间效率.

关键词: 智能环境, 边缘设备, 规则推理, 规则匹配, 无线传感器网络

Abstract:

Under the existing rule inference mechanism, the amount of real-time feature calculation during rule matching caused by a large amount of sensor data reduces the inference real-time performance. At the same time, the limited memory resources of edge devices cannot cope with such a huge amount of data. For this reason, this thesis designs the Data Pre-Deployment Scheme(DPDS). By utilizing the rule network and Light-weight Characteristic Table(LCT) obtained by the rule analysis and preprocessing module, this scheme enables the rule network to directly reference the characteristic values in the LCT during inference without real-time feature calculations, which significantly improves efficiency and real-time and greatly reduces the memory usage of the inference process. The experimental results show that even in the case of a large amount of data and rules, DPDS still has high time efficiency and space efficiency.

Key words: smart environment, edge device, rule inference, rule matching, wireless sensor network

中图分类号:

TP181

汪成亮, 赵凯, 刘嘉敏. 智能环境下基于边缘设备规则推理的数据预部署研究[J]. 电子学报, 2022, 50(10): 2347-2360.

Cheng-liang WANG, Kai ZHAO, Jia-min LIU . Study on Data Pre-Deployment Based on Inference and Computing of Edge Devices in Smart Environment[J]. Acta Electronica Sinica, 2022, 50(10): 2347-2360.

图/表 13

图1 智能环境系统与边缘设备

图2 数据预部署策略架构图

图3 将规则转化为规则网络

表1 典型统计特征

统计运算符	编号	可增量式计算	可引用特征
MAX	0	1	null
MIN	1	1	null
RANGE	2	0	MAX, MIN
AVG	3	1	null
VAR	4	1	AVG
STD	5	1	VAR
SKEWNESS	6	0	STD
KURTOSIS	7	0	VAR

图4 滑动窗口与基本窗口

图5 轻量级特征表

图6 轻量级特征表预部署方案

图7 不同数据流流速下数据预部署策略的性能

表2 不同最大窗口长度下的实验参数

实验编号	1	2	3	4	5	6
最大窗口长度	8	16	32	64	128	256

图8 不同窗口长度下数据预部署策略的性能

表3 不同规则数目下的代价分析

n(million)	0.01	0.1	1	2	4	8	16
T_LCT/s	0.074	0.72	7.48	16.4	39.6	100.1	246.3
T_NET/s	0.006	0.637	0.762	1.42	12.8	37.6	63.7

图9 不同规则数目下数据预部署策略的性能(规则重复率为75%)

图10 不同规则数目下数据预部署策略的性能(规则重复率为25%)

参考文献 29

1	COOK D. Smart Environments: Technology, Protocols and Applications[M]. Hoboken, NJ: Wiley, 2005.
2	汪成亮, 王小均. 基于三轴传感器的老年人日常活动识别[J]. 电子学报, 2017, 45(3): 570-576.
	WANG C L, WANG X J. Daily activity recognition based on triaxial accelerometer of elderly people[J]. Acta Electronica Sinica, 2017, 45(3): 570-576. (in Chinese)
3	LONG X, FONSECA P, HAAKMA R, et al. Actigraphy-based sleep/wake detection for insomniacs[C]//2017 IEEE 14th International Conference on Wearable and Implantable Body Sensor Networks. Eindhoven: IEEE, 2017: 1-4.
4	汪成亮, 赵凯. 一种智能环境下基于边缘设备规则推理的数据预部署方法: CN112651506A[P]. 2021-04-13.
	WANG C L, ZHAO K. Data Pre-Deployment Method Based on Edge Device Rule Reasoning in Intelligent Environment: CN112651506A[P]. 2021-04-13. (in Chinese)
5	FORGY C L. Rete: A fast algorithm for the many pattern/many object pattern match problem[J]. Artificial Intelligence, 1982, 19(1): 17-37.
6	ZHOU R L, PAN J F, TAN X B, et al. Application of CLIPS expert system to malware detection system[C]//2008 International Conference on Computational Intelligence and Security. Suzhou: IEEE, 2008: 309-314.
7	CHEN W, OUYANG D T, YE Y X. RIF2Jess: inferencing RIF rules via translation to jess rules[C]//2009 International Conference on Computational Intelligence and Software Engineering. Wuhan: IEEE, 2009: 1-6.
8	PROCTOR M. Relational declarative programming with JBoss drools[C]//Ninth International Symposium on Symbolic and Numeric Algorithms for Scientific Computing(SYNASC 2007). Timisoara: IEEE, 2007: 5.
9	HERRING C, MILOSEVIC Z. Implementing B2B contracts using Bi _z Talk[C]//Proceedings of the 34th Annual Hawaii International Conference on System Sciences. Maui: IEEE, 2001: 6859651.1-6859651.10.
10	ROTTENSTREICH O, COHEN R, RAZ D, et al. Exact worst case TCAM rule expansion[J]. IEEE Transactions on Computers, 2013, 62(6): 1127-1140.
11	WANG Y M, YANG L H, FU Y G, et al. Dynamic rule adjustment approach for optimizing belief rule-base expert system[J]. Knowledge-Based Systems, 2016, 96: 40-60.
12	BATORY D. The LEAPS algorithm[C]//Computer Science, Austin: University of Texas at Austin, 1994: 60693236.
13	CHEN J F, LI Y H. Improved rule engine based on RETE algorithm for smart-home environment[C]//2019 IEEE 19th International Conference on Communication Technology. Xi'an: IEEE, 2019: 344-349.
14	GUO C, XIONG W, HAO L Y. An improved rete algorithm with branch filtration[J]. Procedia Engineering, 2017, 174: 767-772.
15	VAN WOENSEL W, ABIDI S S R. Optimizing semantic reasoning on memory-constrained platforms using the RETE algorithm[C]//European Semantic Web Conference. Cham: Springer, 2018: 682-696.
16	YAY E, MADRID N M, RAMÍREZ J A O. Using an improved rule match algorithm in an expert system to detect broken driving rules for an energy-efficiency and safety relevant driving system[J]. Procedia Computer Science, 2014, 35: 127-136.
17	KIM M, LEE K, KIM Y, et al. RETE-ADH: An improvement to RETE for composite context-aware service[J]. International Journal of Distributed Sensor Networks, 2014, 10(4): 507160.
18	PENG X L, ZHENG L J, LIAO T. RCEDM: A rete-based RFID complex event detection method[C]//International Conference on Data Mining and Big Data. Cham: Springer, 2017: 137-147.
19	LIU D, GU T, XUE J P. Rule Engine based on improvement Rete algorithm[C]//The 2010 International Conference on Apperceiving Computing and Intelligence Analysis Proceeding. Chengdu: IEEE, 2010: 346-349.
20	YANG P L, YANG Y L, WANG N. IRETE: An improved RETE multi-entity match algorithm[C]//2011 International Conference on Electronics, Communications and Control(ICECC). Ningbo: IEEE, 2011: 4363-4366.
21	XUE L L, LIU Y, TONG X, et al. A two-level reasoning method based on SVM_RETE algorithm in industrial environments[C]//2017 IEEE 17th International Conference on Communication Technology. Chengdu: IEEE, 2017: 1920-1925.
22	HUBRECHTS M, KAMBONA K, RENAUX T, et al. Mete: A meta rete interface for distributed rule-based systems[C]//Proceedings of the 17th Belgium-Netherlands Software Evolution Workshop. Delft: CEUR, 2018: 29-32.
23	ARAUJO D A, HENTGES A R, RIGO S J, et al. Applying parallelization strategies for inference mechanisms performance improvement[J]. IEEE Latin America Transactions, 2018, 16(12): 2881-2887.
24	LIU Y, ZHANG T S, LI S C, et al. A real-time reasoning engine for injection-production optimization of water and oil wells on account of bitmap[J]. Acta Automatica Sinica, 2017, 43(6): 1007-1016.
25	KIM K, KIM B W, LIM S, et al. Two-step filtering for rules without overlapping conditions[C]//2019 34th International Technical Conference on Circuits/Systems, Computers and Communications(ITC-CSCC). JeJu: IEEE, 2019: 1-4.
26	WALZER K, SCHILL A, LÖSER A. Temporal constraints for rule-based event processing[C]//Proceedings of the ACM First Ph.D. Workshop in CIKM. New York: ACM, 2007: 93-100.
27	WALZER K, BREDDIN T, GROCH M. Relative temporal constraints in the Rete algorithm for complex event detection[C]//Proceedings of the Second International Conference on Distributed Event-Based Systems. New York: ACM, 2008: 147-155.
28	BERSTEL B. Extending the RETE algorithm for event management[C]//Proceedings of the Ninth International Symposium on Temporal Representation and Reasoning(TIME'02). New York: ACM, 2002: 49-51.
29	WU E, DIAO Y L, RIZVI S. High-performance complex event processing over streams[C]//Proceedings of the 2006 ACM SIGMOD International Conference on Management of Data. New York: ACM, 2006: 407-418.

[1]	陈嘉兴, 程杰, 董云玲, 刘志华. 基于弯曲声线和测距修正的水下节点定位算法[J]. 电子学报, 2022, 50(7): 1567-1572.
[2]	王义君, 陈忠野, 缪瑞新, 宋忠炎. 基于三角剖分的WSNs元胞感知覆盖算法[J]. 电子学报, 2022, 50(10): 2443-2451.
[3]	苏志刚, 武瑞, 郝敬堂. 面向无源定位的直达波传感器识别算法[J]. 电子学报, 2021, 49(6): 1178-1186.
[4]	汪成亮, 郑诚, 曾卓. 基于软件定义智能的睡眠动作识别[J]. 电子学报, 2021, 49(1): 85-89.
[5]	王海东, 刘云敬, 康琳, 武迎春. 射频能量捕获传感网移动能量源的布置策略研究[J]. 电子学报, 2020, 48(12): 2367-2375.
[6]	程杰, 董云玲, 陈嘉兴, 刘志华. 一种具有连续跳数值的三维DV-Hop改进算法[J]. 电子学报, 2020, 48(11): 2122-2130.
[7]	汪成亮, 黄心田. 智能环境下基于雾计算的推理节点优化分配研究[J]. 电子学报, 2020, 48(1): 35-43.
[8]	黄美根, 郁滨. 软件定义WSN规则一致更新研究[J]. 电子学报, 2019, 47(9): 1965-1971.
[9]	王頲, 白桦, 唐晓铭, 黄庆卿. 分布式全局精确时钟同步状态追踪:能观测性观测器分析[J]. 电子学报, 2019, 47(9): 1855-1862.
[10]	丁旭, 吴晓蓓, 王力立, 张新宇, 夏昕. 无线传感器网络中基于最优转发集的协作式编码传输协议[J]. 电子学报, 2019, 47(7): 1584-1590.
[11]	邱树伟, 石海燕, 池凯凯, 程珍, 朱艺华. 射频供能无线传感器网络高能效数据收集方案[J]. 电子学报, 2018, 46(9): 2238-2244.
[12]	田文飚, 芮国胜, 董道广, 康健. 基于盲自适应KLT的蒸发波导压缩感知方法[J]. 电子学报, 2018, 46(9): 2068-2074.
[13]	房卫东, 张武雄, 杨旸, 张传雷, 陈伟. 基于生物特征标识的无线传感器网络三因素用户认证协议[J]. 电子学报, 2018, 46(3): 702-713.
[14]	李国瑞, 王颖, 王聪. 一种基于矩阵补全的无线传感网数据收集方案[J]. 电子学报, 2018, 46(12): 2950-2956.
[15]	田贤忠, 祝驿楠, 何佳存, 郭敏, 刘高. 一种射频能量捕获网络移动能量源均衡化充电策略[J]. 电子学报, 2018, 46(12): 2985-2992.