Real-time scheduling for energy harvesting sensor nodes

Energy harvesting has recently emerged as a feasible option to increase the operating time of sensor networks. If each node of the network, however, is powered by a fluctuating energy source, common power management solutions have to be reconceived. This holds in particular if real-time responsiveness of a given application has to be guaranteed. Task scheduling at the single nodes should account for the properties of the energy source, capacity of the energy storage as well as deadlines of the single tasks. We show that conventional scheduling algorithms (like e.g. EDF) are not suitable for this scenario. Based on this motivation, we have constructed optimal scheduling algorithms that jointly handle constraints from both energy and time domain. Further we present an admittance test that decides for arbitrary task sets, whether they can be scheduled without deadline violations. To this end, we introduce the concept of energy variability characterization curves (EVCC) which nicely captures the dynamics of various energy sources. Simulation results show that our algorithms allow significant reductions of the battery size compared to Earliest Deadline First scheduling.

无线传感器网络能量管理研究

传感器节点能量受限的特点使能量管理成为无线传感器网络(WSNs)设计的一个重要内容。对近年来无线传感器能量管理的研究进展进行了讨论和综述,指出了WSNs能量管理面临的问题与挑战。在分类阐述和比较当前具有代表性的各种休眠机制的同时,介绍了动态电压调节、动态功率管理、数据融合等能量管理策略。最后,指出了WSNs能量管理的未来发展方向。     

应用RTOS的TD-LTE无线综测仪的实现

针对通信软件设计中实时性和可靠性两大难题,提出了一种第三代合作伙伴(3GPP)长期演进(LTE)网络端协议栈通信软件的实现方案,即基于Nucleus PLUS实时操作系统的新一代宽带无线通信软件的设计。重点介绍了Nucleus PLUS实时多任务操作系统的主要组件及其相关功能,并论述了在Nucleus PLUS操作系统下对通信软件进行设计和实现的方法。实际测试表明,采用NucleusPLUS操作系统的终端测试仪确保了通信的可靠性和实时性,可应用于其他通信软件的开发过程中。     

Non-preemptive scheduling to maximize the minimum global inter-completion time

Temporal load-balancing—“spreading out” the executions of tasks over time—is desirable in many applications. A form of temporal load-balancing is discussed, scheduling to maximize minimum minimum global inter-completion time (MGICT-scheduling). It is shown that MGICT-scheduling is, in general, NP-hard. A number of restricted classes of task systems are identified, which can be efficiently MGICT-scheduled. The technique is applied to a Defense Network System. Simulation results indicate that the proposed strategy achieves higher communication performance in multiprocessor systems. Specifically, our strategy significantly reduces average message delay and percentage of delayed messages.     

基于流量的自组网休眠管理算法

休眠管理作为自组网中节省能量的一种有效方法,正在被广泛的关注.休眠管理通过关闭节点射频,减少空闲状态能量消耗,有效延长了网络生存时间,提高了网络吞吐量;但是常常带来端到端延时增加和递交成功率降低等问题.针对这些问题,提出一种休眠管理算法TSM,各节点根据历史数据流量独立判断是否休眠,利用RTS丢失发现无法通信的邻居节点.NS-2仿真实验表明,相比基于MAC 802-11的AODV,在网络低负载情况下,TSM大幅延长网络生存时间,提高网络吞吐量;无论网络负载高低,TSM对端到端延时和递交成功率的影响很小.     

无线传感器网络中基于网络覆盖的节点睡眠调度机制

在无线传感器网络中,节点睡眠-唤醒调度机制对于延长其生命周期至关重要。在现有的节点睡眠调度机制中,节点能耗是调度机制的唯一考虑目标。然而这种以节省节点能耗为目的调度机制很容易导致传感网络无法完全覆盖监测区域,导致严重后果。本文基于现有调度机制的不足,提出一种考虑到网络覆盖面的节能调度机制。这种机制既能降低网络能耗也能够兼顾到网络的监测覆盖情况。仿真结果表明该机制能在保证较高的网络覆盖面的前提下有效提高网络生命周期。     

基于Android系统的多设备协同交互实验系统设计

随着计算机及Internet技术的飞速发展,便携式智能设备的应用变得普遍,而Android系统由于其开源性,更是得到了广大开发者的青睐。针对多设备的协同交互问题,设计了无线局域网内多Android设备协同交互系统。首先分析了交互系统功能及其涉及的Android开发技术,接着进行了人机交互设计,最终实现了Android设备间的图片、视频及控制命令的无线传输,从而完成不同Android设备之间的协同控制与交互功能。实际应用表明,该实验系统能够完成多设备的协同交互与控制,增强教学的趣味性和吸引力。     

面向边缘设备的高能效深度学习任务调度策略

近年来,深度学习在图像和自然语言处理等诸多领域表现出色,与深度学习相关的各类移动应用发展迅速,但由于移动网络状态的不稳定性及网络带宽的限制,基于云计算的深度模型任务可能出现较大响应延迟,严重影响用户体验.与此同时,深度模型对设备的计算及存储能力有较高的要求,无法直接在资源受限的移动设备中进行部署.因此,亟须设计一种新的计算模式,使得基于深度模型的移动应用能够满足用户对快速响应、低能耗及高准确率的期望.本文提出一种面向边缘设备的深度模型分类任务调度策略,该策略通过协同移动设备与边缘服务器,充分利用智能移动终端的便捷性和边缘服务器强大的计算能力,综合考虑分类任务的复杂度和用户期望,完成深度模型在移动设备和边缘服务器中的动态部署,并对推理任务进行动态调度,从而提升任务执行效率,降低深度学习模型推理开销.本文以基于卷积神经网络的图像识别应用为例,实验结果表明,在移动环境中,相比于准确率最高的深度模型,本文提出的高能效调度策略的推理能耗可降低93.2%、推理时间降低91.6%,同时准确率提升3.88%.     

一种太阳能传感器网络的自适应休眠调度算法

通过从外界获取太阳能,传感器网络节点的能量限制得到缓解。提出一种太阳能传感器网络的自适应休眠调度算法。当观测场景从区域观测到目标跟踪转变时,节点自适应地转换活跃和休眠状态,同时剩余能量低于能量阈值的休眠节点从外界获取太阳能。与不考虑太阳能获取的休眠调度算法相比,提出的算法延长了网络生命周期。     

浅谈分布式光伏发电

分布式光伏发电作为新型能源,在未来发展具有很广阔的前景。文章结合目前分布式光伏发电实际发展情况,介绍了光伏发电原理、光伏电源对配电网的影响及系统保护配置等,并具体介绍了分布式发电并网调度管理的运行步骤。     

Fault-Tolerant Real-Time Scheduling

We use competitive analysis to study how best to use redundancy to achieve fault-tolerance in online real-time scheduling. We show that the optimal way to use spatial redundancy depends on a complex interaction of the benefits, execution times, release times, and latest start times of the jobs. We give a randomized online algorithm whose competitive ratio is O( logΦ log Delta ( log ² n log m/ log log m)) for transient faults. Here n is the number of jobs, m is the number of processors, Φ is the ratio of the maximum value density of a job to the minimum value density of a job, and Δ is the ratio of the longest possible execution time to the shortest possible execution time. We show that this bound is close to optimal by giving an Ω(( log ΔΦ/ log log m) ( log m log log m) ² ) lower bound on the competitive ratio of any randomized algorithm. In the case of permanent faults, there is a randomized online algorithm that has a competitive ratio of O( log Phi log Δ (log m/log log m)). We also show a lower bound of Ω((log ΔΦ/ log log m) ( log m/log log m)) on the competitive ratio for interval scheduling with permanent faults. Received October 1997; revised January 1999.     

基于Linux的便携嵌入式设备电源管理解决方案

本文分析了嵌入式设备现有的节电技术和Linux系统的电源管理机制,并以iPAQ为例提出了一套使用以上机制来实现便携嵌入式设备的电源管理解决方案.     

多灾害点应急资源调度研究与实现

针对多灾害点、多点出救、多目标应急调度问题,建立一种以时间最短、成本最低为目标的数学模型。考虑到调度过程中各灾害点对资源的竞争,提出利用表上作业算法对该模型进行优化求解,实现了在整体的优化方案中时间和成本总体消耗最少。提出的方法简单、实用、易用,并成功应用在省级应急管理信息示范平台上。最后,通过一个实例验证该方法的有效性。     

无线传感器网络的拓扑控制

拓扑控制是无线传感器网络研究中的核心问题之一.拓扑控制对于延长网络的生存时间、减小通信干扰、提高MAC(media access control)协议和路由协议的效率等具有重要意义.全面阐述了拓扑控制技术的研究进展,首先明确了拓扑控制研究的问题和设计目标,然后分别从功率控制和睡眠调度两个方面介绍代表性的研究工作,并加以分析和比较,同时指出了这些工作存在的不足.最后分析和总结了研究现状中存在的问题、需要进一步研究的内容以及拓扑控制研究的发展趋势.     

Ubiquitous service access through adapted user interfaces on multiple devices

The Ubiquitous Interactor (UBI) addresses the problems of design and development that arise from services that need to be accessed from many different devices. In the UBI, a service can present itself with different user interfaces on different devices. This is done by a separation of the user–service interaction and presentation. The interaction is kept the same for all devices, and different presentation information is provided for different devices. This way, tailored user interfaces for many different devices can be created without multiplying the development and maintenance work. In this paper, we describe the design of the UBI, the system implementation, and two services implemented for the system: a calendar service and a stockbroker service.