面向可追溯的物联网数据采集与建模方法研究

发布时间：2018-05-05 02:18

本文选题：可追溯 + 物联网　；参考：《中国农业大学》2014年博士论文

【摘要】：可追溯系统是保证生鲜农产品质量安全的重要手段,当前技术构架下可追溯系统面临的感知数据采集能力欠缺、追溯数据粒度输出单一、追溯平台体系结构薄弱等瓶颈,阻碍了系统的规模化应用。物联网技术的发展,使生鲜农产品质量安全可追溯系统突破技术与应用的瓶颈成为可能。本研究从可追溯系统的三个技术瓶颈出发,紧紧围绕物联网“无处不在的数据采集、可靠的数据传输与信息处理、智能化的信息应用”三个核心内涵,以动、植物源性农产品可追溯供应链为研究对象,构建了物联网环境下可追溯系统数据采集与建模方法,研发了基于WSN的生鲜农产品质量安全可追溯感知数据采集硬件与嵌入式软件,设计了基于SPC的可追溯感知数据时域压缩方法,针对用户数据粒度需求的差异,进行了可追溯数据兼容建模和面向粒度分级的规约,设计和实现了基于云计算的可追溯综合服务平台。研究的主要贡献与创新之处是： (1)提出了基于WSN的可追溯感知数据采集方法和基于SPC的时域压缩方法,提高了感知数据采集效率,并延长了监测网络寿命。基于WSN所研发的可追溯感知数据采集方法软、硬件原型,测试结果表明通信链路可靠,感知节点对生鲜农产品供应链保鲜工艺环境的兼容性、传感器硬件兼容性好：基于SPC所设计的改进X-Rs感知数据时域压缩算法与阈值、K-滑动均值算法对比,能耗在同一数量级,平稳时间序列的Se为最优,2种时间序列平稳性上tc值均接近最优,算法的平衡性和适应性好。 (2)提出了面向粒度分级的可追溯系统建模方法,满足了不同用户的数据粒度需求。基于结构模式识别,构造了描述追溯单元转化的12种模式基元；基于关系代数,设计了模式基元的数据存储结构与数据采集算法；构建了基于2型文法的可追溯数据形式化描述文法和文法句子生成算法；基于改进下推自动机建立了粒度分级规约方法；以冻罗非鱼片加工、半滑舌鳎养殖、肉牛养殖与屠宰加工业务流程为实例进行了方法验证,结果表明在以上供应链,数据分级规约强度为44.8-99.4%,在供应链结构信息少的初级农产品生产流程中,规约强度最高。 (3)设计了基于云计算的可追溯综合服务平台,实现了平台级可追溯服务。识别了可追溯数据在生鲜农产品供应链上各阶段的潜在价值,包括文档标准化、危害溯源、精确召回、物流监控、关键点预警、质量预测、货架期管理和库存优化；基于Hadoop设计了平台的技术架构、服务引擎、体系结构,基于Map/Reduce实现了决策模型并行化；在Ubuntu10.0.4操作系统和Hadoop0.20.0并行计算环境上进行了平台实现；以工厂化水产养殖、水产品冷链物流为例的系统评价表明平台在数据采集、信息追溯和智能决策等方面改善了生鲜农产品供应链管理水平。
[Abstract]:Traceability system is an important means to guarantee the quality and safety of fresh agricultural products. Under the current technical framework, traceability system is faced with the bottleneck of lacking of perceptual data collection ability, single output of traceability data granularity, weak architecture of traceability platform, etc. It hinders the large-scale application of the system. With the development of Internet of things, it is possible to break through the bottleneck of technology and application in the traceability system of fresh agricultural products. This research starts from the three technical bottlenecks of traceability system, tightly revolves around "everywhere data collection, reliable data transmission and information processing, intelligent information application" three core connotations, in order to move, Based on the traceability supply chain of plant-derived agricultural products, the data acquisition and modeling method of traceability system in Internet of things environment is constructed, and the hardware and embedded software of traceability sensing data acquisition for fresh agricultural products based on WSN are developed. A time-domain compression method of traceability perceptual data based on SPC is designed. According to the difference of user's data granularity requirements, the compatible modeling of traceability data and granulity-oriented hierarchical specification are carried out. Design and implementation of a cloud-based traceability integrated service platform. The main contributions and innovations of the research are: 1) A method of traceable perceptual data acquisition based on WSN and a time-domain compression method based on SPC are proposed to improve the efficiency of perceptual data acquisition and prolong the lifetime of monitoring network. Based on the software and hardware prototype of traceable perceptual data acquisition developed by WSN, the test results show that the communication link is reliable and the perception node is compatible with the fresh agricultural product supply chain fresh-keeping technology environment. The sensor hardware compatibility is good: the improved X-Rs perceptual data time domain compression algorithm based on SPC is compared with the threshold K- sliding mean algorithm. The energy consumption is in the same order of magnitude. The se of stationary time series is optimal and the values of TC are close to optimal on the stationarity of two kinds of time series, and the algorithm has good balance and adaptability. (2) the modeling method of traceability system for granularity classification is proposed to meet the data granularity requirements of different users. Based on structural pattern recognition, 12 kinds of pattern primitives are constructed to describe the transformation of tracing units, and data storage structures and data acquisition algorithms of pattern primitives are designed based on relational algebra. A formal description grammar and sentence generation algorithm for traceable data based on 2 type grammar is constructed, a granularity classification method based on improved push-down automata is established, and frozen tilapia slices are processed, and semi-smooth tongue sole is cultured. The results showed that the strength of the data classification specification was 44.8-99.4 in the supply chain, and the intensity of the specification was the highest in the primary agricultural production process with little information on the supply chain structure. The platform of traceability integrated service based on cloud computing is designed, and the platform level traceability service is realized. The potential value of traceable data in all stages of supply chain of fresh agricultural products was identified, including document standardization, hazard traceability, accurate recall, logistics monitoring, key point early warning, quality prediction, shelf life management and inventory optimization. The technology architecture, service engine, architecture of the platform are designed based on Hadoop, and the decision model is parallelized based on Map/Reduce. The platform is implemented on the Ubuntu10.0.4 operating system and Hadoop0.20.0 parallel computing environment. The systematic evaluation of the cold chain logistics of aquatic products shows that the platform improves the supply chain management level of fresh agricultural products in the aspects of data collection, information tracing and intelligent decision making.
【学位授予单位】：中国农业大学
【学位级别】：博士
【学位授予年份】：2014
【分类号】：TS205

【参考文献】