Energy-efficient and Reliable Wearable Internet-of-Things through Fog-Assisted Dynamic Goal Management
Energy-efficient and Reliable Wearable Internet-of-Things through Fog-Assisted Dynamic Goal Management
复制标题
DOI：
10.1016/j.procs.2019.04.067
复制
发表时间：
2019
期刊：
影响因子：
--
通讯作者：
A. Anzanpour;H. Rashid;A. Rahmani;A. Jantsch;N. Dutt;P. Liljeberg
中科院分区：
文献类型：
--
作者：
A. Anzanpour;H. Rashid;A. Rahmani;A. Jantsch;N. Dutt;P. Liljeberg
能量耗散和电池寿命的管理是健康监测可穿戴设备的一个挑战。低质量的数据收集、不可靠的监测过程以及错过重要的健康事件是单一目标固定策略解决方案的后果。在这项研究中，基于物联网的可穿戴系统的能量耗散是通过动态多目标方法进行管理的。可穿戴设备用户的健康状况、监控的连续性以及收集数据的准确性是我们在目标层次结构中考虑的参数，以便在运行时选择适当的系统管理策略，以在给定时间实现最重要的目标。在我们的方法中，动态观察过程评估用户和系统数据，雾辅助控制引擎检测状态，执行适当的策略，并重新配置可穿戴传感器。为了演示我们的解决方案，我们开发了一个真正的可重新配置的无线传感器节点，该节点能够遵循一组参数化定义的策略，并进行了一组实验来找到最有效的设置。我们的评估表明，与没有目标管理器的基准系统相比，所提出的系统能够在总数据收集时间的 0.78% 内降低 44% 的功耗并防止由于电池短缺而导致的数据丢失。
Management of energy dissipation and battery life is a challenge in health monitoring wearables. Low-quality data collection, non-reliable monitoring process, and missing important health events are consequences of single-goal fixed-policy solutions. In this research, energy dissipation of IoT-based wearable systems is managed through a dynamic multi-goal approach. Health status of the user of a wearable device, the continuity of monitoring, and the accuracy of collected data are parameters we consider in our goal hierarchy to select a proper system management policy at run-time to achieve the most significant goal at a given time. In our approach, a dynamic observation process assesses the user and system data and a fog-assisted control engine detects the states, enforces the proper policy, and reconfigures the wearable sensor. To demonstrate our solution, we develop a real reconfigurable wireless sensor node with an ability to follow a set of parametrically defined policies and performed a set of experiments to find the most efficient setting. Our evaluation shows that the proposed system is able to reduce the power consumption by 44% and prevent the data loss due to battery shortage in 0.78% of total data collection time compared to a baseline system without a goal manager.