-
人体传感网络(body sensor network,BSN)集成了传感器、电子学、医学、数据分析与融合、人工智能、无线通信和其他创新应用等多学科知识[1]. BSN有着无线化、网络化、信息化的综合优势,将其与物联网(internet of things,IoT)相结合,能够实现人体健康的全程跟踪与服务,可以说由BSN构成的智慧医疗是未来医疗保健低成本化的发展方向之一[2].
每个人都是一个独立的个体样本,其身体状况属性在符合一定规律的同时彼此之间也有着很大的差异性,这些属性在生活习惯、性别和遗传等方面随年龄增长一直在动态变化[3].借助BSN,IoT和云计算开发一种针对人体数据传感的网络架构[4],对个人医疗保健有着非常重要的意义.
现在已经有人提出了基于云系统的各种医疗保健架构方法和服务技术.文献[5]提出一种基于移动的Android应用程序,用于监测心电图信号.文献[6]提出一种多层应用程序架构,集成了云计算平台和人体传感器网络,并允许云端收集、处理、存储和分析采集到的人体数据,但该解决方案不支持边缘计算,也无法满足实时性要求.换言之,在将监测数据从传感器传输到云之后,需要等待云系统计算出结果并做出决定.文献[7]提出一种移动设备的任务卸载方法,任务卸载是基于物联网和雾计算的一种有吸引力的技术,它可以用在传感器设备、边缘节点和云节点之间,但移动设备物理尺寸小,且计算能力非常有限.虽然架构可以将任务卸载到云,但是当网络延迟不可容忍时,这种思路不可行.文献[8]提出一种基于多代理的云监控模型,通过结合人工智能算法来检测错误任务.但是多代理的云监控模型使用代理的主从架构,因此可能会遇到大规模基础架构中的可伸缩性问题[9-11].
在研究了基于中央云的人体感知架构基础上,针对人体健康状态这个领域对于实时性和可移动性特定的要求,本文提出了一种新的人体数据传感架构,在无线传感器网络和云平台之间加入了边缘云服务器,并重新构建了框架和传输协议,设计了能够支持多种类型应用程序的存储架构,优化了传感架构.本文提出的传感架构既可以通过应用程序接口(API)安全地将人体传感数据聚合到中央云,并通过任务调度和资源整合充分利用云计算来处理数据,还能将中央云计算反馈的信息实时发送给处于移动状态的用户.该架构降低了由虚拟机容量限制而导致的故障率,同时降低了整体服务器的利用率,减少了服务时间和处理时间.
Human Body Data Sensing Architecture Based on Edge Cloud Computing
- Received Date: 18/07/2019
- Available Online: 20/07/2020
-
Key words:
- cloud platform /
- edge cloud computing /
- human body data sensing architecture /
- virtual machine scheduling
Abstract: In order to solve the problem that the cloud platform-based architecture method can not support real-time and mobility when dealing with human body monitoring sensor data, a human body data sensing architecture based on edge cloud computing has been proposed in this paper and been used for big data analysis. It constructs a three-layer cloud computing human body data stream architecture, which is composed of wireless body area network, edge cloud system and central cloud system respectively. Then, a mobile service architecture supporting real-time and mobility is proposed for the human body data stream architecture. The architecture uses the edge server between the central cloud data center and the user, and the edge cloud data sensing architecture is designed. The software architecture consists of three layers of application programming interface for sensor devices, edge clouds, and central clouds. In addition, a storage architecture has been designed in this paper that supports multiple types of applications, optimizes the sensing architecture and improves data storage reliability. The experimental results show that the edge cloud-based architecture reduces the failure rate and overall server utilization caused by virtual machine capacity, and is superior to the existing architecture in terms of service time and data processing time.