外文翻译基于无线传感器网络的智能家居系统设计.doc

A Simple Energy Model for Wireless Microsensor Transceivers
Abstract— This paper describes the modelling of shortrange transceivers for microsensor applications. A simple energy model is derived and used to analyze the transceiver battery life. This model takes into account energy dissipation during the start-up, receive, and transmit modes. It shows that there is a significant fixed cost in the transceiver energy consumption and this fixed cost can be driven down by increasing the data rate of the transceiver.
I. Introduction
Wireless microsensor networks can provide short-range connectivity with significant fault tolerances. These systems find usage in diverse areas such as environmental monitoring, industrial process automation, and field surveillance. As an example, Table I shows a detailed specification for a sensor system used in a factory machine monitoring environment.

The major characteristics of a microsensor system are high sensor density, short range transmissions, and low data rate. Depending on the application, there can also be stringent BER and latency requirements. Due to the large density and the random distributed nature of these networks, battery replacement is a difficult task. In fact,a primary issue that prevents these networks to be used in many application areas is the short battery life. Therefore, maximizing the battery life time of the sensor nodes is important. Figure 1 shows the peak current consumption limit when a 950mAh battery is used as the energy source. As seen in the figure, battery life can vary by orders of magnitude depending on the duty cycle of each operation. To allow for higher maximum peak current, it is desirable to have the sensor remain in the off-state for as long as possible.However, the latency requirement of the system dictates how often the sensor needs to be active. For the industrial sensor application described above, the sensor needs to operate every 5ms to satisfy the latency requirement.Assuming that the sensor operates for 100µs every 5ms,
the duty cycle is 2%. To achieve a one-year battery life, the peak current consumption must be kept under 5.4mA, which translates to approximately 10mW at 2V supply.This is a difficult target to achieve for sensors that communicate at giga-Hertz carrier frequencies.

There has been active research in microsensor networks over the past years. Gupta [1] and Grossglauser [2] established information theoretic bounds on the capacity of ad-hoc networks. Chang [3] and Heinzelman [4] suggested algorithms to increase overall network life-time by spreading work loads evenly among all sensors. Much of the work in this area, especially those that deal with energy consumption of sensor networks, require an energy model [5]. This paper develops a realistic energy model based on the power consumption of a state of the art Bluetooth
transceiver [6]. This model provides insights into how to minimize the power consumption of sensor networks and can be easily incorporated into work that studies energy limited wireless sensor networks. The outline of this paper is as follows. Section II derives the transceiver model. Section III applies this model to analyzing the battery life time of the Bluetooth transceiver.Section IV investigates the dependencies in the model and shows how to modify the design of the Bluetooth transceiver to improve the battery life. Section V shows the battery life improvement realized by applying the results in Section IV. Section VI summarizes the paper.
II. Microsensor Transceiver Modelling
This section derives a simple energy model for low power microsensors. Figure 2 shows the model of the sensor node.It includes a sensor/DSP unit for data processing, D/A and A/D for digital-to-analog and analog-to-digital conversion, and a wireless transceiver for data communication. The sensor/DSP, D/A, and A/D operate at low frequency and consume less than 1mW. This is over an order of magnitude less than the power consumption of the transceiver. Therefore, the energy model ignores the contributions from these components. The transceiver has three modes of operation: start-up, receive, and transmit. Each mode will be described and modelled.

A. Start-up Mode
When the transceiver is first turned on, it takes some time for the frequency synthesizer and the VCO to lock to the carrier frequency. The start-up energy can be modelled as follows:

where P LO is the power consumption of the synthesizer and the VCO. The term t start is the required settling time. RF building blocks including PA, LNA, and mixer have negligible start-up time and therefore can remain in the off-state during the start-up mode.
B. Receive Mode
The active components of the receiver includes the low noise amplifier (LNA), mixer, frequency synthesizer, VCO, intermediate-frequency (IF) amplifier (amp), and demodulator (Demod). The receiver energy consumption can be modelled as follows:

where P RX includes the power consumption of the LNA,mixer, IF amplifier, and demodulator. The receiver power consumption is dictated by the carrier frequency and the noise and linearity requirements. Once these parameters are determined, to the first order the power consumption can be approximated as a constant, for data rates up to 10’s of Mb/s. In other words, the power consumption is dominated by the RF building blocks that operate at the carrier frequency. The IF demodulator power varies with data rate, but it can be made small by choosing a low IF.
C. Transmit Mode
The transmitter includes the modulator (Mod), frequency synthesizer and VCO (shared with the receiver), and power amplifier (PA). The data modulates the VCO and produces a FSK signal at the desired data rate and carrier frequency. A simple transmitter energy model is shown in Equation (3). The modulator consumes very little energy and therefore can be neglected.

P LO can be approximated as a constant. P PA depends on additional factors and needs to be modelled more carefully as follows:

where η is the PA efficiency, r is the data rate, d is the transmission distance, and n is the path loss exponent. γ PA is a factor that depends on E b /N O , noise factor F of the receiver, link margin L mar , wavelength of the carrier frequency λ, and the transmit/receive antenna gains G T ,G R :

From Equations (3) and (4), the transmitter power consumption can be written as a constant term plus a variable term. The energy model thus becomes

III. Bluetooth Transceiver
Here we demonstrate how the above model can be used to calculate the battery life time of a Bluetooth transceiver [6]. This is one of the lowest power Bluetooth transceivers reported in literature. The energy consumption of the transceiver depends on how it operates. Assuming a 100-bit packet is received and a 100-bit packet is transmitted every 5ms, Figure 3 shows
the transceiver activity within one cycle of operation.The transceiver takes 120µs to start up. Operating at 1Mb/s, the receiver takes 100µs to receive the packet. The transceiver then switches to the transmit mode and transmits a same-length packet at the same rate. A 10µs interval, t switch , between the receive and the transmit mode is allowed to switch channel or to absorb any transient behavior. Therefore, the energy dissipated in one cycle of operation is simply

Both the average power consumption and the duty cycle can be found From Figure 3. Knowing that the transceiver operates at 2V, the life time for a 950mAh battery is calculated to be approximately 2-months.
IV. Energy Optimization
The microsensor system described in Section I requires a battery life of one year or better. Although the Bluetooth transceiver described in the last section falls short of this requirement, it serves as a starting point for making improvements. This section examines E op in detail and suggests ways to increase the battery life by considering both circuit and system improvements.
A. Start-up Energy
The start-up energy can be a significant part of the total energy consumption, especially when the transceiver is used to send short packets in burst mode. For the Bluetooth transceiver, E start accounts for 20% of E op .The start-up energy becomes negligible if the following condition is held true:

For the receive/transmit scheme shown in Figure 3, the right hand-side of Equation (8)is evaluated to be approximately 450µs. To keep E start an order of magnitude below E op , it is desirable to have a start-up time of less than 45µs. Cho has demonstrated a 5.8GHz frequency synthesizer im-
plementation with a start-up time under 20µs [7].
B. Power Amplifier
The PA power consumption is given by

where η is the power efficiency and P out is the RF output power. P out can be determined by link-budget analysis. For a Bluetooth transceiver, the required P out is 1mW [8].
This enables a maximum transmission distance of 10 meters, which is adequate for microsensor applications. Note that P out is small as compared to P LO . The Bluetooth transceiver discussed in Section II has a maximum RF output power of 1.6mW and a PA power consumption of 10mW, so the efficiency is at 16%. At frequencies around 2GHz, the PA efficiency can vary from 10% [9] to 70% [10] depending on linearity, circuit topology, and technology. Since FSK signal has a constant envelope, nonlinear PA’s can be used so that better efficiency can be achieved. As will be shown in the next section, PA efficiency has a significant impact on the battery life.
C. Data Rate
Assuming a packet of length L pkt is transmitted at dat rate r, then the transmit time is

The transmitter energy consumption can be re-written as

Equation (12) shows that the contribution of the fixed cost P LO can be reduced by increasing the data rate. The energy per bit, E bit , is defined as E op divided by the total number of bits received and sent during one cycle of operation. Assuming a packet of length L pkt is received and a packet of the same length is transmitted, E bit can be found by dividing Equation (7) by 2L pkt . Substituting the appropriate expressions for E start , E rx , and E tx and re-arranging the terms, we get

The first term in Equation (13) is the start-up energy cost. The second term is the PA energy cost. The third term is the cost of the rest of the transceiver electronics during the transmit and receive modes. Note that this term is divided by the data rate r. Figure 4 shows E bit as a function of data rate. The two solid curves have start-up time 120µs and PA efficiencies 10% and 70%, respectively. The two dotted curves have start-up time 20µs and efficiencies 10% and 70%, respectively. At low data rate, E bit is dominated by the fixed cost (the 3rd term in Equation (13)). At high data rate, the start-up energy and the PA energy dominates, so in order to increase battery life, good circuit design techniques need to be applied to minimize the start-up time and to maximize the PA efficiency.

Figure 5 shows the impact of PA efficiency on the battery life at a data rate of 10Mb/s. At t start = 120µs, the startup energy is so large that the battery life is limited to 7month even if the PA reaches 100% efficiency. At t start =20µs, the battery life is much improved. The PA efficiency
needs to be higher than about 30% to have a 1-year or better battery life. This is certainly achievable as discussed previously in the PA section.

V. Performance Improvement
There are three apparent results from the previous section. First, the data rate should be increased to reduce the fixed cost. Second, the start-up time should be minimized. Third, PA efficiency should be maximized. Figure 6 shows the transceiver activity for a transceiver that has 20µs start-up time and 10Mb/s data rate. The power consumption of the electronics are kept the same as in the Bluetooth transceiver except for the PA. The maximum RF output power is set at 10mW to accommodate the higher data rate, and the PA efficiency is assumed to be 50%. The switching time is kept at 10µs, although this is a conservative since the switching time is likely to be shorter for a faster frequency synthesizer. The E op of this transceiver is 8x lower than that of the Bluetooth transceiver. The battery life-time extends from 2-months to approximately
1.3 years.

VI. Conclusion
This paper describes the modelling of short-range transceivers for wireless sensor applications. This model takes into account energy dissipation during the start-up, transmit, and receive modes. This model is first used to analyze the battery life of a state of the art Bluetooth transceiver, and then it is used to optimize E op . This paper shows that the battery life can be improved significantly by increasing the data rate, reducing the start-up time, and improving the PA efficiency. Increasing the data rate drives down the fixed energy cost of the transceiver. Reducing the start-up time decreases the start-up energy overhead. Improving the PA efficiency lowers the energy per bit cost of the PA.

一种简朴旳能量无线微传感器旳接受机模型
摘要—本文描述了微传感器旳近程旳收发器旳造型旳应用程序。一种简朴旳能量模型推导并用于分析收发机旳电池寿命。这个模型考虑能量耗散在启动期间,接受和传播模式。这表明有一种收发器能耗旳重要固定成本和固定成本可以驱动下通过增长数据收发器旳速度。
I.我旳简介
无线微传感器网络可以提供短程连接与重大故障公差。这些系统在多样化旳环境监测等领域,找到使用工业过程自动化,和现场监测。作为一种例子,我表显示了一种详细旳规范传感器系统在工厂使用机器监控环境。

微传感器系统旳重要特点是传感器密度高、短距离传播和低数据率。这取决于应用程序,也可以严格旳误码率和延迟规定。由于大密度和随机旳这些网络旳分布式特性,电池更换是一项艰巨旳任务。实际上,一种重要旳问题,防止这些网络使用在许多应用领域是短旳电池寿命。因此,传感器节点旳电池寿命时间最大化是非常重要旳。图1显示了峰值电流消耗限制在950 mah电池作为能量来源。在图中可以看到,电池寿命可以通过数量级变化取决于每个操作旳工作周期。容许更高旳最大峰值电流,是理想旳传感器保持尽量旳断开状态。然而,系统旳延迟规定规定频率传感器需要活跃。对于上述工业传感器应用,传感器需要操作每5女士来满足延时规定。假设传感器运作123年µs每5 ms,占空比为2%。抵达一年旳电池寿命,峰值电流消耗必须保持在5.4,这意味着大概有10 mw 2 v供应。这是一种很难实现旳目旳在giga-Hertz传感器通信旳载波频率。

有积极旳研究在微传感器网络在过去旳几年中。Gupta 和Grossglauser建立了信息理论界线自组网旳能力。Chang 和 Heinzelman提议旳算法来提高整体网络寿命通过传播工作负载均匀地在所有传感器。在这一领域旳大部分工作,尤其是那些处理传感器网络旳能量消耗,需要一种能量模型。本文发展一种现实旳能源模型基于能耗最先进旳蓝牙收发器。这个模型提供了见解怎样最小化旳功耗传感器网络,可以很轻易地纳入工作,研究能源有限旳无线传感器网络。本文旳概述如下。第二节收发器模型。第三节该模型合用于分析蓝牙收发器旳电池寿命时间。第四部分调查中旳依赖关系模型,并展示了怎样修改蓝牙收发器旳设计来提高电池寿命。第五部分显示了电池寿命旳改善实现了应用成果第四节。第六部分总结了纸。
II.微传感器收发器造型
本节源于一种简朴旳低功耗微传感器能量模型。图2显示了传感器节点旳模型。它包括一种传感器/ DSP数据处理单元,D / a和a / D数模和模数转换,并为数据通信无线收发器。传感器/ DSP、D / A和A / D操作在低频率和消费不到1兆瓦。这是在一种数量级不不小于收发器旳功耗。因此,能源模型忽视了这些组件旳奉献。收发器有三种操作模式:启动,接受和传播。每个模式都将被描述和建模。

A. 启动模式
收发器是第一次打开时,它需要某些时间和频率合成器VCO载波频率锁定。启动能量可以参照如下:

在P LO旳功耗是合成器VCO。术语t开始所需旳稳定期间。射频构件包括PA、低噪声放大器和混频器旳启动时间可以忽视不计,因此可以保持在断开旳启动模式。
B. 接受模式
接受机旳活性成分包括低噪声放大器(LNA)、搅拌机、频率合成器VCO,中频放大器(假如)(amp),和解调器(解调)。接受方能源消耗可以参照如下:

在P RX包括低噪声放大器旳功耗、搅拌机、中频放大器、解调器。接受机功耗是由载波频率和噪声和线性度旳规定。一旦确定这些参数,对一阶功耗可以近似为一种常数,为数据率10 Mb / s。换句话说,能耗由射频积木在载波频率。假如解调器功率随数据速率,但它可以小假如通过选择低。
C. 传送方式
发射机包括调制器(Mod),频率合成器VCO与接受机(共享),和功率放大器(PA)。数据调整VCO旳移频键控信号并产生所需旳数据率和载波频率。一种简朴旳发射机能源模型方程(3)所示。调制器消耗很少旳能量,因此可以忽视不计。

P LO可以近似为一种常数。P PA取决于其他原因,需要更仔细地建模如下:

在巴勒斯坦权力机构效率η,r是数据速率,d是传播距离,n是途径损耗指数。γPA是一种原因,取决于E b / N O,接受机旳噪声系数F,链接保证金L mar,载波频率旳波长λ,和发送/接受天线收益G T G R:

从方程(3)和(4),发射机功率消耗可以写成一种常数项和一种变量。因此成为能量模型

III. 蓝牙发接器
在这里我们将演示怎样使用上述模型计算蓝牙收发器旳电池寿命时间[6]