Design Space Exploration of Embedded Systems for Intelligent Maintenance

Design Space Exploration of Embedded Systems for Intelligent Maintenance
复制标题

智能维护嵌入式系统的设计空间探索

DOI:
--
复制
发表时间:
2013
期刊:
影响因子:
--
通讯作者:
R. V. Henriques
R. V. Henriques
中科院分区:
--
文献类型:
--
作者:
E. Lazzaretti;Marcos Zuccolotto;C. Pereira;R. V. Henriques

文献摘要

被引文献

相似文献

随着智能维护技术的发展,嵌入式系统将越来越多地与这些算法一起使用,以呈现更大的灵活性,结合高处理速度和低功耗。在这种情况下,基于模型的编程与自动特定于平台的代码生成能力的极大兴趣。这项工作进行了设计空间探索的一些算法通常用于智能维护应用程序,在这种情况下,小波包能量和逻辑回归,通过分析性能和所需的足迹不同的实现智能维护算法时,在硬件和软件中执行。比较的起点是所谓的Watchdog Agent™ IMS系统,该系统目前在MATLAB™和LabVIEW™环境中都可用。使用可用的代码生成工具,部署不同的硬件和软件版本,并且比较生成的系统的性能以及FPGA实现中使用的资源的一些能量和存储器度量。使用从由电动机械致动器组成的测试台收集的振动数据,并且获得的结果证实了所生成的解决方案在评估的指标,清楚地表明最佳解决方案可能会因应用程序要求而异。
With the development of intelligent maintenance techniques, the embedded systems that will be used with such algorithms will need increasingly to present more flexibility, combined with high processing speed and low power consumption. Within this context, model based programming associated with automatic platform-specific code generation capabilities are of great interest. This work performs a design space exploration of some algorithms commonly used on intelligent maintenance applications, in this case wavelet package energies and logistic regression, by analyzing the performance and required footprint of different implementations for intelligent maintenance algorithms when executed in hardware and software. Starting point for the comparison was the so called Watchdog Agent™ IMS system, which is currently available both in MATLAB™ and LabVIEW™ environments. Using available code generation tools distinct hardware and software versions are deployed and both the performance of the generated systems as well as some energy and memory metrics of the resources used in FPGA implementations are compared For the validation tests, vibration data collected from a test bench composed by an electric mechanical actuator was used and obtained results confirmed a great variability of the generated solutions in terms of the assessed metrics, clearly indicating that best solution may vary depending on the application requirements.