论文标题
城市声音感测的可扩展的物联网框架
A Scalable IoT-Fog Framework for Urban Sound Sensing
论文作者
论文摘要
物联网（IoT）是一个相互关联的设备系统，可用于允许大规模收集和分析数据。但是，随着它的增长，物联网网络无法管理这些服务的数据。结果，引入了云计算以满足对物联网网络数据中心的需求。随着技术的发展，对支持和管理众包和实时数据的适当方法的需求增加了，云服务器无法再跟上大量传入数据。这种需求引起了雾计算。它成为云的扩展，并允许有效地围绕网络分配资源。它与物联网的集成减少了向云服务器的应变。但是，最终设备和数据管理约束的高功耗问题浮出水面。本文提出了两种减轻这些问题以保持雾计算的方法，作为与物联网相关的应用程序的可靠选择。我们创建了一个基于IoT的传感框架，该框架使用了城市声音分类模型。通过主动低功率和高功率状态以及资源重新分配，我们创建了一个网络配置。我们针对使用默认雾和云设置的IoT框架测试了这种配置。结果改善了框架的最终设备功耗和服务器延迟。总体而言，通过拟议的框架，雾计算被证明能够支持可扩展的IoT框架以进行城市声音感测。
Internet of Things (IoT) is a system of interrelated devices that can be used to allow large-scale collection and analysis of data. However, as it grew, IoT networks were not capable of managing the data from these services. As a result, cloud computing was introduced to address the need for datacentres for IoT networks. As the technology evolved, the demand for a proper means of supporting and managing crowdsensing and real-time data increased, and cloud servers could no longer keep up with the large volumes of incoming data. This demand brought rise to fog computing. It became an extension to the cloud and allowed resources to be allocated around the network effectively. Its integration to IoT reduced the strain towards the cloud servers. However, issues in high power consumption at the end device and data management constraints surfaced. This paper proposes two approaches to alleviate these issues to keep fog computing remain as a reliable option for IoT-related applications. We created an IoT-based sensing framework that used an urban sound classification model. Through active low and high power states and resource reallocation, we created a network configuration. We tested this configuration against IoT frameworks that use the default fog and cloud setups. The results improved the framework's end device power consumption and server latency. Overall, with the proposed framework, fog computing was proven to be capable of supporting a scalable IoT framework for urban sound sensing.
