Reference Materials-Aoke (Beijing) Biotechnology Co., Ltd.

A fast and accurate energy source emulator for wireless sensor networks

Time of publication:2020-05-09 14:28 author:Aoke
1

Wireless sensor networks (WSNs) consist of a set of communicating nodes (motes) capable of sensing, processing, and storing physical quantities. One of the most critical issue in designing and developing WSNs is the limited amount of energy usually available in the nodes.

The typical way of addressing this issue is by means of dynamic power management (DPM) strategies exploiting multiple low-power states of the motes. In such states, consumption drops from tens of milliwatts to a fraction of a microwatt increasing the system lifetime [1–3]. Recent advances in energy harvesting (EH) have introduced the opportunity of granting unbounded lifetime to sensor nodes, thus overcoming the limitations of battery-operated WSNs and introducing a new class of WSNs, called EH wireless sensor network (EH-WSNs), made of motes powered by renewable energy sources (RESs). This new opportunity has begun to

shift the research paradigm from energy-constrained lifetime maximization (typical of battery-operated systems) to power-constrained workload maximization. EH motes can be deployed in remote environments and they can run potentially for an unlimited amount of time by gathering solar, kinetic, radio frequencies, or thermal energy [4, 5]. However, designing hardware and software solutions for EH-WSNs is a difficult task due to the main characteristics of RESs. In particular, RESs show a high degree of unpredictability while EH devices provide an extremely variable efficiency. In fact, the amount of energy harvested depends not only on environmental conditions, which can vary in a rapid and uncontrolled way, but also on the energy drawn at runtime by the powered device. The combination of these two main factors makes practically impossible to produce repeatable experiments while testing a EH-WSN hardware-software platform using traditional strategies. For instance, the same node with a particular hardware-software configuration can generate extremely different results while EH conditions change. Furthermore, even when it is possible to accurately reproduce in lab the behavior of an energy harvester, replicating it over a testing network of tens or hundreds of nodes can be unfeasible for complexity and economic reasons. A viable alternative can be represented by simulating the network and the energy harvester to predict the behavior of motes possibly using real-world traces. In WSNs, the use of simulation has been deeply explored by several authors in the last ten years [6–12]. However, realistic simulation requires an accurate model not only of the EH device but also of every hardware component involved in the system plus a fast and accurate model of the communication channel. In fact, the actual energy efficiency of communication protocols in WSNs depends on many factors which are very difficult to model and simulate, like interferences, collisions, re-transmissions, and over-hearing. Handling these issues requires to trade off realism (and accuracy) for simulation time and complexity, so that accurate energy simulations of large WSNs is still to be regarded as an open research problem. Emulation provides a viable alternative. In this paper, we present an embedded hardware-software platform implementing an energy source emulator capable to run different energy source models for WSNs. The general architecture of the emulator allows the designer to model any energy source and to connect it to any physical mote. The emulator adjusts the voltage levels of the power supply according to the model of choice, while also measuring the actual current drawn by the mote in order both to collect a trace, and to provide a real-time feedback to the energy source model. Models can be changed in the field either by loading them from the on-board non volatile memory or by uploading them through the network. Finally, the small size of the platform (about 73×57 millimeters), together with its low cost (less than $50) and power consumption (about 100 mA at 12 V) make it ideal to be used in large testbeds. In summary, the proposed system (i) represents a flexible solution for emulating a set of heterogeneous energy sources, targeting WSN motes’ power supply, (ii) provides a tunable level of realism through a model-agnostic architecture, and (iii) features low-power consumption levels, a small factor form and reduced production costs, which make it particularly suitable for emulation of many real-world application scenarios even in large prototyping testbeds. Extensive characterization experiments confirm the validity of the proposed emulator both in terms of accuracy and in terms of computational performance. A representative case study regarding the emulation of a lithium battery shows that the hardware-software platform does not introduce any measurable reduction of the accuracy of the model. The rest of the paper is organized as follows: Section 2 provides a survey of existing approaches to the energy source modeling and emulation; Section 3 provides a description of the proposed modeling framework; Section 4 presents the hardware-software emulation platform; Section 5reports and discusses experimental results; Section 6 concludes the work.

Thumb up-pictures

Sign in to comment

Log In
Recommend
分享到WeChat 关闭-Reference Materials
请告知您的Phone number号码,我们将立即回电

通话对您免费,请放心接听

温馨提示:

1.手机直接输入,座机前请加区号 如13803766220,010-58103678

2.我们将根据您提供的Phone number号码,立即回电,请注意接听

3.因为您是被叫方,通话对您免费,请放心接听

关闭
大抽奖
Online Service系统
live chat