【作者】 赵江滨；

【导师】 史铁林；

【作者基本信息】 华中科技大学 ， 机械制造及其自动化， 2007， 博士

【摘要】 随着微电子与微系统技术、计算机与网络通讯技术的发展,制造出高度集成、开机即用、具有自识别与自描述能力的自主智能传感器成为可能,这种自主智能传感器组成自组织传感器网络,能够大大简化基于传感器的测试测量系统的部署、配置和应用开发。本文研究了开发基于传感器的测试测量系统的一般过程、所需要的技术与开发工具。对比了基于传感器网络与基于普通模拟传感器的测试系统的不同特点,研究了普通模拟传感器的配置管理,说明了IEEE 1451.4智能传感器标准的自识别与自描述能力如何简化普通模拟传感器测试系统的配置管理;研究了传感器网络的节点管理,指出传感器网络需要专门设计的应用层协议,以及这种协议需要具有能够自动组网与自动解析网络数据的特点。参考IEEE 1451系列标准,提出了一套轻量级的传感器网络应用层协议Agile-SN,易于理解与实现,具有节点自动发现、支持任意的网络传输协议、支持任意的传感器类型,具有高性能与高可扩展性的特点。详细描述了AgileSN协议的模型、工作过程与数据类型,对比了与IEEE 1451的异同,介绍了一种具体实现,并给出了一个以软件模拟实现的AgileSN传感器节点实例。适应传感器网络节点的自主智能特性,开发了一套灵活高效的即时可重构虚拟仪器软件平台IsVIP,具有如下特点:1)采用开放的组件化体系结构,用户可以任意扩展其功能,能够支持任意的硬件平台,包括传感器网络、DAQ、以及示波器、万用表等各种测试仪器;或者通过控件实现任意的软件功能。2)高度灵活的即时可重构,能够在运行时修改虚拟仪器软件系统的运行逻辑与人机界面。3)采用数据流驱动的工作方式,多条数据流能够并行工作。4)支持高性能的流水线模式,能够大幅度的提高应用系统的吞吐率。详细研究了IsVIP的性能,特别针对步骤时间不等的流水线过程,深入研究了吞吐率和时延2个重要的性能指标,理论分析、模拟验证和实际测试显示,流水线模式能够大幅度乃至成倍的提高吞吐率,能够处理通常模式不能处理的高速数据处理请求;对于时延指标,只有在数据处理请求超过峰值吞吐率以及瓶颈置后的情况下才会变坏,其它情况下与标准模式相同。给出了利用IsVIP开发应用系统的2个实例,通过简单的鼠标操作就能够快速的开发测试系统。第1个实例是利用软件仿真实现的AgileSN自主智能传感器节点采集PC的性能参数;第2个实例利用IsVIP开发温度测量系统,硬件系统为支持NI-DAQmx驱动的DAQ设备和1款集成模拟温度传感器芯片。实例证明,IsVIP能够作为一套通用易用的测试测量应用系统开发平台。更多还原

【Abstract】 With technology advances in microecletronics, MEMS, Computer, and network communications, it’s possible to build highly intergrated, ready-to-run autonomous smart sensors, with features of self-identification and self-description. This type of autonomous smart sensors can make up self-oranizational sensor networks, greatly simplify sensor network system deployment, set-up and application development.Detailed Inspects the general procedure, technical infrastructure and tools needed to build sensor based test and measurement systems. Compares sensor networks with tranditional sensors, describes tranditional sensor based test system set-up, point out how IEEE 1451.4 can simplify this set-up procedure. Describes node management issues of sensor networks, point out that sensor networks need specially designed application layer protocol, and this protocol should automatically makeup a network, and automatically parse network traffic data.With the reference of IEEE 1451 series standards, this paper presents a lightweight application layer protocol–AgileSN, for autonomous smart transducers. It is simple to understand and easy to implement, with the properties of automatic node detection, high performance and high scalability. This paper further describes the AgileSN model and working procedure in detail, makes comparisons with IEEE 1451, introduces an implementation, and gives a software simulated AgileSN node example.Presents a run-time reconfigurable virtual instrumentation software platform-IsVIP, it adopts component based open architecture, provides two extension points: device console and software control, can supports any device and can realize any function; it also adopts data flow model, multiple data flow processing tasks can run simultaneously; the virtual instrument function in IsVIP is realized throw software controls and the wires between them, controls and wires can be added or removed at run time, and take effects immediately, then the instrument system visual interface and running logic changed; it also supports the high throughput pipeline execution model, achieving parallel performance through serial architecture, can process higher volume incoming data.Study the performance of IsVIP detaily, especially the 2 performance indicators: throughput and dealy. Through theoretical analysis, software simulation and real permance measurement, shows that the pipeline execution model can greatly improve peak throughput, some times can even improve several times. And the dealy indicator get worse only when the bottle neck stage placed back and data supply rate larger than the process capability.Two examples of building IsVIP applications were given. The first example is using AgileSN sensor node to acquire PC performance parameter, the second example is to apply a DAQ device supporting NI-DAQmx driver and an intergrated temperature sensor to build temperature measurement system, showing how through simple point and click operation to build test and measurement system.更多还原