无线传感器网络节点的低功耗研究
摘要
本文介绍了无线传感器网络产生的背景、特点,分析其节点的能量损耗环节,提出可以采取的各种低功耗措施,主要分为硬件和软件两个方面。
硬件方面结合无线传感器网络节点的功能要求,选择了MSP430F149超低功耗单片机和CC2420射频芯片作为首选;在存储器方面使用新型超低功耗存储器FM24CL16;具体电路上给出了可以节能的MOS管开关电路,采用了零功耗WG系列磁敏韦根传感器以及改进的阀门状态检测电路和超低功耗嵌入式实时时钟电路等一系列设计;并对各种电池性能进行了研究,选择了超低功耗超低压差的TPS76927作为LDO,保证了电池电压降至接近2.7V时系统仍能获得稳定的系统工作电压。
对节点系统可以采取的低功耗软件措施主要包括单片机低占空比工作模式的选择、动态电源管理、通信协议和操作系统的使用等。
同时按照上述设计制作了节点并进行了一系列的实验,可以得到结论:依据本文所给出的低功耗措施可以使无线传感器网络节点的使用寿命提高至8年左右,极大程度的满足了各领域的应用需要。
In recent years, the appearance of wireless sensor network which is a hot subject makes sensor itself a richer, more powerful processing, more micro volume and lower device costs and power consumption device. So it has broad application prospects in military, civilian, industrial production and other fields. It is known as one of the 21st century's most influential technology and the 10 technology which will change the world.
Wireless sensor network node is a typical example of embedded systems which is composed of sensor, processor, wireless communication and energy supply modules. It follows the general design principles of embedded system. The principles are that hardware and software can be trimmed and meet the stringent requirements in functionality, reliability, cost, size, power consumption and other aspects. In practical application, the sensor nodes need to complete the collection and conversion of the monitoring data, data transmission and control, or even handle compression and decompression, encryption and decryption and other tasks. But the capacity of batteries-powered are generally not large enough and most of the nodes work in harsh environment or the area in which there is no human. It isn’t able to recharge the batteries or change them so once the battery power is used up that means the nodes will lose its role. Therefore, low-power has become the core of the research.
When the wireless sensor network nodes are running in the network, the energy comes mainly from three operations that are processing, sensor and wireless transmission. The energy consumption of processing is mainly due to the implementation of microprocessor instructions. Processor has large difference in energy consumption when it is activated or in a state of sleep. The energy consumption of sensing including converters, front-end processing, A / D conversion operation. It is different with the different sensor types. And wireless module may be at the four states: sending, receiving, leisure and sleep. The power instructions of the four states are not the same. Wireless module is the main energy source, the power instructions at leisure state almost equal to the receive state. In order to save energy, we should make wireless module sleep as far as possible.
Considering the energy loss links of wireless sensor network node, this paper gives low-power design and research mainly from two aspects hardware and software. For hardware we should try to choose low-power and high integration chips which have rich on-chip resources, choose large capacity batteries as power supply, use a lower voltage and frequency as far as possible and low-power components and structure of the circuits. This paper compares various modules and chips in the market and selects MSP430F149 ultra-low power MCU and CC2420 RF chips as the preferred according to the functional requirements of wireless sensor network nodes. MSP430 MCU uses 1.8 V to 3.6V voltage as the power supply voltage. The power consumption is only 0.1 uA in RAM data retention mode and 250 uA / MIPS in activities mode, but the traditional MCS-51 MCU is 10 uA ~ 20uA / MIPS. The maximum input leakage current of MSP430 MCU is only 50 nA, but MCS-51 is generally 1 to 10 uA. The work frequency range of CC2420 is 2.4835 ~ 2.400 GHz; It has strong ability to dispute the interference from near channel (39 db), ultra-low current (RX: 18.8Am, TX: 17.4mA) high receiver sensitivity (-94 dbm), integrated VCO, LAN, PA and power rectifier and a low power supply voltage (2.1 to 3.6 V). We choose new ultra-low power memory FM24CL16. It can be data retention when there is no power, has high read and write speed, parallel port access speed of 70 ns, 20M serial clock speed, no 10 ms’waiting time and a standby current of less than 1 uA, write current is less than 150 uA and can be erased almost unlimited number of at least 100 million times. We use energy-saving MOS switching circuit in the sensor module and a zero-power WG Series magnetic WeiGen sensor in the instrumentation field in which wireless sensor network is applied widely. For portable devices, the power supply system is essential and is composed of LDO and batteries. The paper studies on the performance of various batteries and selects the economic and applicable No.5 high-octane alkaline batteries. And chooses ultra-low-power and ultra-low dropout TPS76927 as LDO, ultra-low dropout ensures that when the batteries’voltage drop to close to 2.7 V the system can still obtain stability system work voltage, so the battery can be fully used. At the end part of the hardware design also present some specific application examples in different nodes with a number of low-power technologies. Such as application of zero-power electromagnetic valve, improvement of valve states detection circuit and ultra-low-power design of embedded real-time clock.
The software measures include the choice of low duty cycle mode and dynamic power management: when the system doesn’t work, let system enters sleep mode in order to reduce energy consumption; short-range wireless communications technology ZigBee protocol based on the IEEE 802.15.4 standard and TinyOS embed operating system which is suitable for wireless sensor networks: ZigBee protocol has some advantages such as low power consumption, low transmission rate and low cost, until now it is the one of the preferred technology using in wireless communications applications in wireless sensor networks; and reasonable design of dormancy scheduling mechanism.
The paper also do a series of experiments about nodes that made based on the design above, including MCU module, RF communications module and the measure of the battery capacity. And obtain effective and convincing experimental data: current consumption about MSP430 MCU module of the node system is 295 uA when the MCU is working and is 30 uA under sleep mode 3; the working current of CC2420 module is 17.4 mA when receiving data, is 18.8 mA when sending data and is 1 uA at sleep mode; the working current of the entire system is about 18.4 mA when collecting ,receiving and sending data, and current consumption of dormancy is 31 uA; total capacity of three batteries in tandem can be used is 2779.5 mAh. We can get the conclusions by compute these experimental data: the service life of wireless sensor network nodes can be increased to about eight years based on the low-power hardware design and reasonable software mechanisms that given in this paper, and can meet the application requirements of various filed in great extent.
硬件方面结合无线传感器网络节点的功能要求,选择了MSP430F149超低功耗单片机和CC2420射频芯片作为首选;在存储器方面使用新型超低功耗存储器FM24CL16;具体电路上给出了可以节能的MOS管开关电路,采用了零功耗WG系列磁敏韦根传感器以及改进的阀门状态检测电路和超低功耗嵌入式实时时钟电路等一系列设计;并对各种电池性能进行了研究,选择了超低功耗超低压差的TPS76927作为LDO,保证了电池电压降至接近2.7V时系统仍能获得稳定的系统工作电压。
对节点系统可以采取的低功耗软件措施主要包括单片机低占空比工作模式的选择、动态电源管理、通信协议和操作系统的使用等。
同时按照上述设计制作了节点并进行了一系列的实验,可以得到结论:依据本文所给出的低功耗措施可以使无线传感器网络节点的使用寿命提高至8年左右,极大程度的满足了各领域的应用需要。
In recent years, the appearance of wireless sensor network which is a hot subject makes sensor itself a richer, more powerful processing, more micro volume and lower device costs and power consumption device. So it has broad application prospects in military, civilian, industrial production and other fields. It is known as one of the 21st century's most influential technology and the 10 technology which will change the world.
Wireless sensor network node is a typical example of embedded systems which is composed of sensor, processor, wireless communication and energy supply modules. It follows the general design principles of embedded system. The principles are that hardware and software can be trimmed and meet the stringent requirements in functionality, reliability, cost, size, power consumption and other aspects. In practical application, the sensor nodes need to complete the collection and conversion of the monitoring data, data transmission and control, or even handle compression and decompression, encryption and decryption and other tasks. But the capacity of batteries-powered are generally not large enough and most of the nodes work in harsh environment or the area in which there is no human. It isn’t able to recharge the batteries or change them so once the battery power is used up that means the nodes will lose its role. Therefore, low-power has become the core of the research.
When the wireless sensor network nodes are running in the network, the energy comes mainly from three operations that are processing, sensor and wireless transmission. The energy consumption of processing is mainly due to the implementation of microprocessor instructions. Processor has large difference in energy consumption when it is activated or in a state of sleep. The energy consumption of sensing including converters, front-end processing, A / D conversion operation. It is different with the different sensor types. And wireless module may be at the four states: sending, receiving, leisure and sleep. The power instructions of the four states are not the same. Wireless module is the main energy source, the power instructions at leisure state almost equal to the receive state. In order to save energy, we should make wireless module sleep as far as possible.
Considering the energy loss links of wireless sensor network node, this paper gives low-power design and research mainly from two aspects hardware and software. For hardware we should try to choose low-power and high integration chips which have rich on-chip resources, choose large capacity batteries as power supply, use a lower voltage and frequency as far as possible and low-power components and structure of the circuits. This paper compares various modules and chips in the market and selects MSP430F149 ultra-low power MCU and CC2420 RF chips as the preferred according to the functional requirements of wireless sensor network nodes. MSP430 MCU uses 1.8 V to 3.6V voltage as the power supply voltage. The power consumption is only 0.1 uA in RAM data retention mode and 250 uA / MIPS in activities mode, but the traditional MCS-51 MCU is 10 uA ~ 20uA / MIPS. The maximum input leakage current of MSP430 MCU is only 50 nA, but MCS-51 is generally 1 to 10 uA. The work frequency range of CC2420 is 2.4835 ~ 2.400 GHz; It has strong ability to dispute the interference from near channel (39 db), ultra-low current (RX: 18.8Am, TX: 17.4mA) high receiver sensitivity (-94 dbm), integrated VCO, LAN, PA and power rectifier and a low power supply voltage (2.1 to 3.6 V). We choose new ultra-low power memory FM24CL16. It can be data retention when there is no power, has high read and write speed, parallel port access speed of 70 ns, 20M serial clock speed, no 10 ms’waiting time and a standby current of less than 1 uA, write current is less than 150 uA and can be erased almost unlimited number of at least 100 million times. We use energy-saving MOS switching circuit in the sensor module and a zero-power WG Series magnetic WeiGen sensor in the instrumentation field in which wireless sensor network is applied widely. For portable devices, the power supply system is essential and is composed of LDO and batteries. The paper studies on the performance of various batteries and selects the economic and applicable No.5 high-octane alkaline batteries. And chooses ultra-low-power and ultra-low dropout TPS76927 as LDO, ultra-low dropout ensures that when the batteries’voltage drop to close to 2.7 V the system can still obtain stability system work voltage, so the battery can be fully used. At the end part of the hardware design also present some specific application examples in different nodes with a number of low-power technologies. Such as application of zero-power electromagnetic valve, improvement of valve states detection circuit and ultra-low-power design of embedded real-time clock.
The software measures include the choice of low duty cycle mode and dynamic power management: when the system doesn’t work, let system enters sleep mode in order to reduce energy consumption; short-range wireless communications technology ZigBee protocol based on the IEEE 802.15.4 standard and TinyOS embed operating system which is suitable for wireless sensor networks: ZigBee protocol has some advantages such as low power consumption, low transmission rate and low cost, until now it is the one of the preferred technology using in wireless communications applications in wireless sensor networks; and reasonable design of dormancy scheduling mechanism.
The paper also do a series of experiments about nodes that made based on the design above, including MCU module, RF communications module and the measure of the battery capacity. And obtain effective and convincing experimental data: current consumption about MSP430 MCU module of the node system is 295 uA when the MCU is working and is 30 uA under sleep mode 3; the working current of CC2420 module is 17.4 mA when receiving data, is 18.8 mA when sending data and is 1 uA at sleep mode; the working current of the entire system is about 18.4 mA when collecting ,receiving and sending data, and current consumption of dormancy is 31 uA; total capacity of three batteries in tandem can be used is 2779.5 mAh. We can get the conclusions by compute these experimental data: the service life of wireless sensor network nodes can be increased to about eight years based on the low-power hardware design and reasonable software mechanisms that given in this paper, and can meet the application requirements of various filed in great extent.
引文
[1] 孙雨耕,张静,孙永进.无线自组传感器网络.传感技术学报.2004.6:53～56
[2] Akyildiz IF,Su W,Cayirei E. Wireless Sensor Networks. A Survey. Computer Networks. 2002.38(3):393～422
[3] 叶驰,孙利民,廖勇.传感器网络的能量管理.计算机工程与应用.2004.8:196～198
[4] Jerry Zhao,Ramesh Govindan,Deborah Estrin. Residual Energy Scans for Monitoring Wireless Sensor Networks. In:IEEE Wireless Communications and Networking Conference WCNC02. Orange County Convention Center,Orlando,FL,USA. 2003-03:17～21
[5] 吴光斌.无线传感器网络能量有效性的研究.传感器技术.2004.23(7):74～76
[6] 童诗白.模拟电子技术基础.高等教育出版社.2000,第四版:168～176
[7] 阎石.数字电子技术基础.高等教育出版社.1998,第四版:93～116
[8] 沈建华,杨艳琴.MSP430 系列 16 位超低功耗单片机原理与应用.清华大学出版社.2004.11,第一版:131～138
[9] 吴少军,刘光斌.实用低功耗设计——原理、器件与应用.人民邮电出版社.2003.3,第一版:168～172
[10] 孙利民,李建中.无线传感器网络.清华大学出版社.2005.5,第一版:21～23
[11] 甘学温.低功耗 CMOS 逻辑电路设计综述.微电子学.2000. 30(4):263～267
[12] 李佳,徐勇军.体系结构级功耗分析方法.系统仿真学报.2004.12:2821～2824
[13] 李曦,王志刚,周学海.面向低功耗优化设计的系统级功耗模型研究.电子学报.2004.32(2):205～208
[14] Sinhua A,Chandrakasan A. Dynamic power management in Wireless Sensor Network. IEEE Design and Test of Computer.2001.18(2):62-74
[15] Lm C,Kim H,Ha S. Dynamic voltage scheduling technique for low-powermultimedia application using buffers. In : Proceedings of the International Symposium on Low Power Electronics and Design. California. ACM Portal Press. 2001:34～39
[16] Luca Benini, Giovanni De Micheli. System-Level Power Optimization.Techniques and Tools. ACM Transactions on Design Automation of Electronic Systems. 2000. 5(2):115～192
[17] 梁宇,韩奇,魏同立.低功耗数字系统设计方法.东南大学学报(自然科学版).2000.30(5):136～142
[18] 罗旻,杨波,高德远.寄存器传输级低功耗设计方法.小型微型计算机系统.2004.25(7):1207～1211
[19] 胡俭波.单片机系统的超低功耗设计.工业控制计算机. 2001.14(11):66～68
[20] 胡大可.MSP430 系列超低功耗 16 位单片机原理与应用.北京航空航天大学出版社.2001:44～54
[21] Shih E,Cho S,Ickes N. Physical layer driven protocol and algorithm design for energy-efficient wireless sensor networks. Italy. Porc ACM MobiCom’01. Romc.2001:272～286
[22] 张大踪,杨涛,魏东梅.无线传感器网络低功耗设计综述.传感器与微系统.2006.25(5):10～14
[23] 桂淑红.无线传感网络路由协议的节能研究.计算机与信息技术.2006.11:33～35
[24] 徐太忠,邵高平.便携式电子系统的低功耗设计.单片机与嵌入式系统应用.2001.8:11～13
[25] 何毅,王勇,蒋鸿宇.基于 ZigBee 的远程无线抄表系统设计.电子设计应用.2005.6:102～105
[26] 杜娟.基于 MSP430F149 低功耗模式的设计与应用.低压电器.2006.10:21～24
[27] 陈俊峰,杨向萍.基于 MSP430F149 的便携式仪器的低功耗设计.现代医学仪器与应用.2004.16(4):6～8
[28] 周胜海,黄文霞.智能仪表的低功耗设计.信阳师范学院学报(自然科学版).2003.4:199～202
[29] 王伟,冯良.低功耗 IC 卡燃气表的研究与开发.煤气与热力.2006.4:31～34
[30] 徐建平.超低功耗电子电路系统设计原则.集成电路应用. 2003.2:78～80
[31] 周华,徐华.低功耗仪表设计初探.仪器仪表学报 2005.4:374～377.
[32] Agustin Barberis, Leonardo Barboni, Maurizio.Evaluating Energy Consumption in Wireless Sensor Networks Application.10th Euromicro Conference on Digital System Design.2001.6:307～321
[33] Sinem Coleri, Anuj Puri, Pravin Varaiya. Power Efficient System for Sensor Networks.Eighth IEEE International Symposium on Computer.
[2] Akyildiz IF,Su W,Cayirei E. Wireless Sensor Networks. A Survey. Computer Networks. 2002.38(3):393～422
[3] 叶驰,孙利民,廖勇.传感器网络的能量管理.计算机工程与应用.2004.8:196～198
[4] Jerry Zhao,Ramesh Govindan,Deborah Estrin. Residual Energy Scans for Monitoring Wireless Sensor Networks. In:IEEE Wireless Communications and Networking Conference WCNC02. Orange County Convention Center,Orlando,FL,USA. 2003-03:17～21
[5] 吴光斌.无线传感器网络能量有效性的研究.传感器技术.2004.23(7):74～76
[6] 童诗白.模拟电子技术基础.高等教育出版社.2000,第四版:168～176
[7] 阎石.数字电子技术基础.高等教育出版社.1998,第四版:93～116
[8] 沈建华,杨艳琴.MSP430 系列 16 位超低功耗单片机原理与应用.清华大学出版社.2004.11,第一版:131～138
[9] 吴少军,刘光斌.实用低功耗设计——原理、器件与应用.人民邮电出版社.2003.3,第一版:168～172
[10] 孙利民,李建中.无线传感器网络.清华大学出版社.2005.5,第一版:21～23
[11] 甘学温.低功耗 CMOS 逻辑电路设计综述.微电子学.2000. 30(4):263～267
[12] 李佳,徐勇军.体系结构级功耗分析方法.系统仿真学报.2004.12:2821～2824
[13] 李曦,王志刚,周学海.面向低功耗优化设计的系统级功耗模型研究.电子学报.2004.32(2):205～208
[14] Sinhua A,Chandrakasan A. Dynamic power management in Wireless Sensor Network. IEEE Design and Test of Computer.2001.18(2):62-74
[15] Lm C,Kim H,Ha S. Dynamic voltage scheduling technique for low-powermultimedia application using buffers. In : Proceedings of the International Symposium on Low Power Electronics and Design. California. ACM Portal Press. 2001:34～39
[16] Luca Benini, Giovanni De Micheli. System-Level Power Optimization.Techniques and Tools. ACM Transactions on Design Automation of Electronic Systems. 2000. 5(2):115～192
[17] 梁宇,韩奇,魏同立.低功耗数字系统设计方法.东南大学学报(自然科学版).2000.30(5):136～142
[18] 罗旻,杨波,高德远.寄存器传输级低功耗设计方法.小型微型计算机系统.2004.25(7):1207～1211
[19] 胡俭波.单片机系统的超低功耗设计.工业控制计算机. 2001.14(11):66～68
[20] 胡大可.MSP430 系列超低功耗 16 位单片机原理与应用.北京航空航天大学出版社.2001:44～54
[21] Shih E,Cho S,Ickes N. Physical layer driven protocol and algorithm design for energy-efficient wireless sensor networks. Italy. Porc ACM MobiCom’01. Romc.2001:272～286
[22] 张大踪,杨涛,魏东梅.无线传感器网络低功耗设计综述.传感器与微系统.2006.25(5):10～14
[23] 桂淑红.无线传感网络路由协议的节能研究.计算机与信息技术.2006.11:33～35
[24] 徐太忠,邵高平.便携式电子系统的低功耗设计.单片机与嵌入式系统应用.2001.8:11～13
[25] 何毅,王勇,蒋鸿宇.基于 ZigBee 的远程无线抄表系统设计.电子设计应用.2005.6:102～105
[26] 杜娟.基于 MSP430F149 低功耗模式的设计与应用.低压电器.2006.10:21～24
[27] 陈俊峰,杨向萍.基于 MSP430F149 的便携式仪器的低功耗设计.现代医学仪器与应用.2004.16(4):6～8
[28] 周胜海,黄文霞.智能仪表的低功耗设计.信阳师范学院学报(自然科学版).2003.4:199～202
[29] 王伟,冯良.低功耗 IC 卡燃气表的研究与开发.煤气与热力.2006.4:31～34
[30] 徐建平.超低功耗电子电路系统设计原则.集成电路应用. 2003.2:78～80
[31] 周华,徐华.低功耗仪表设计初探.仪器仪表学报 2005.4:374～377.
[32] Agustin Barberis, Leonardo Barboni, Maurizio.Evaluating Energy Consumption in Wireless Sensor Networks Application.10th Euromicro Conference on Digital System Design.2001.6:307～321
[33] Sinem Coleri, Anuj Puri, Pravin Varaiya. Power Efficient System for Sensor Networks.Eighth IEEE International Symposium on Computer.