Towards Predictable and Reliable Wireless Communication in Harsh Environments

Full text

(1)

(2)

(3)

(4) .

(5) . !"#.

(6)

(7)

(8) !

(9)

(10) "

(11) !

(12)

(13) !.

(14) #$ !%

(15) "& '( )* +,-.,-./*,.0/).

(16) 10lODUGDOHQ8QLYHUVLW\ 3 4 5

(17).

(18)

(19)

(20)

(21) . !"#$%!!%"$%!"!%! # "#""& &!'"!.

(22) !( )*. ( *(+

(23)

(24) , *-) ,,

(25) (

(26) , ( .*

(27) (

(28) ( ( *

(29) -)

(30)

(31)

(32) / ,

(33) 0+ (

(34) ((** --

(35) ,

(36) -) - ,

(37).

(38) 123

(39) 4 15/52%*6 /

(40) +),(/' 7 8(4 99

(41)

(42) :9 - ,

(43)

(44) 6/; 6

(45) . ( *

(46) -)

(47)

(48)

(49) / ,

(50) 0+ (

(51) (.

(52) 6 0 0**4

(53) 0

(54)

(55)

(56) 4 / 0

(57) -0

(58) * 0 990

(59) + 6

(60) - + *

(61) 6

(62) -<+

(63) 4

(64) ,< 0+

(65) ,

(66) 04 - 9*

(67) /+ *90

(68) 40

(69) <4

(70)

(71) +

(72) < (

(73) *6&< 94+,+ *

(74)

(75) + 0**4

(76) 0

(77) /

(78) 04

(79) , 6

(80) 9 06 0*9 < 0**4

(81) 0

(82) + 6 4 0 + 0**4

(83) 0

(84)

(85) + 96

(86) 4 + + 4 0 /

(87) <+

(88) + * - * - + * =4 *

(89) + 0 + 9

(90)

(91) *9 0 * 4 *

(92) 99 0+

(93) ,

(94) * + 6 +4

(95) ( 6

(96) 9 06 + -04 - + < ( 9

(97)

(98) , 0

(99) 4*9

(100) / 90(

(101)

(102) 0 4 + 9 - * - 4 0+

(103) ,

(104)

(105)

(106) > ?@+ +

(107)

(108) *

(109) + *

(110) 0

(111) 64

(112) -++ + * 4 *

(113) 9- *

(114) 9 0 4 ++ 6

(115) 9 * + =4 *

(116)

(117) , *

(118) 0+

(119) ,

(120) *-9 06

(121) 6 + + +< + 96 9 0 < 0**4

(122) 0

(123)

(124) + +

(125)

(126) *

(127) &<

(128) 0 *

(129) + 0+ 0 0 - +

(130)

(131) *

(132) 6 6 0+ 46 0+

(133) ,+ * 4 *

(134) 9 0 4 9

(135)

(136) ++

(137) , + 9- *. 9

(138) 6 + 9 -

(139) ,

(140) "

(141) ++ * -+

(142) ,0

(143) 4*9

(144) - $4 +

(145) < 400 990

(146) <+9

(147) *49

(148) 0**4

(149) 0

(150) 9

(151) + * 4 *

(152) +<+

(153) 0

(154)

(155) *

(156) 9< <

(157) -- 0+

(158) ,0

(159) 4*9

(160) - $4 +

(161) A,$ * 4 &< + 90(

(162)

(163) 4*6 -

(164) *

(165) < -- 0 +

(166) , 0

(167) 4*9

(168) /

(169) + 9 * 0

(170) 6 0

(171) 0

(172). - *+

(173)

(174) *

(175) 0+ 0 0++ 9

(176) <990

(177) 6 4

(178) / *

(179) 0+

(180) B !#C

(181) < (<+ 6 0**4

(182) 0

(183) < <

(184) 0< "D - + 9

(185) . " $EFGEFG25EGF " H21G.

(186) To my family and friends.

(187) Abbrevations AFH. Adaptive frequency-hopping spread spectrum. BACnet. Building Automation and Networking. DSSS. Direct Sequence Spread Spectrum. EDR. Enhanced Data Rate. FSP. Frequency Spectrum Analyser. FHSS. Frequency-Hopping Spread Spectrum. FTS4BT. Frontline’s Bluetooth Sniffer. GFSK. Gaussian Frequency Shift Keying. HAN. home area networks. ISM. Industrial, Scientific and Medical. LEACH. Low Energy Adaptive Clustering protocol. OQPSK. Offset quadrature phase shift keying. PER. packet-error-rate. PEWIT. Power Efficient WIreless Technology. PHY. Physical layer. POMPOM Programmable Micro Power Meter Testbed for Radio Modules RQ. research question. RSSI. Received Signal Strength Indicator. SIG. Bluetooth Special Interest Group. SmartRF. Texas Instrument SmartRF packet sniffer. UCLA. University of California, Los Angeles. WSN. Wireless Sensor Network.

(188) Sammandrag Tr˚ adl¨ os kommunikation för industriella, forsknings och medicinska applikationer har flera fördelar. N˚ agra av fördelarna att använda tr˚ adl¨ os teknik innefattar enkelheten att placera ut och l¨ agga till enheter till det tr˚ adl¨ osa nätverket samt mobiliteten. Att ersätta kablarna i dessa miljöer st¨ aller h¨ ogre krav p˚ a kommunikationen, s˚ asom tillförlitlighet och f¨ orutsägbarhet. Tillf¨ orlitligheten korrelerar till radiokommunikationen och m¨ ojligheten att f¨ orutsäga att all data kommer fram inom den tidsram som systemet kr¨ aver. Denna avhandling presenterar empiriska m¨ atmetoder f¨ or att undersöka och modulera beteendet hos en radio kopplat till tillförlitlighet och f¨ oruts¨ agbarhet. Avhandlingen fokuserar p˚ a energiförbrukning, paketfel och f¨ ordr¨ ojning. Dessa mätmetoder a¨r tillämpade f¨ or olika radioteknologier och miljöer. Det huvudsakliga vetenskapliga bidraget är de m¨ atplattformar och m¨ atmetoder som har utvecklats för att undersöka moderna radioteknologier g¨ allande tillförlitlighet och f¨ orutsägbarhet. Avhandlingen visar att det är m¨ ojligt att f¨ orutsäga tr˚ adl¨ os kommunikation i tuffa miljöer. En modell f¨ or att förutsäga energif¨ orbrukningen f¨ or en Bl˚ atandsradio i tr˚ adl¨ osa nätverk. Mätresultaten visar att varken avst˚ and eller sändningseffekt p˚ averkar energiförbrukningen, varken för en Bl˚ atand- eller ZigBeeenhet. Däremot har paketfel och omskickningar en signifikant p˚ averkan p˚ aenergiförbrukningen och dessa parametrar kan andet mellan s¨ andare och mottagare men kanske främst korreleras till avst˚ miljön. Denna avhandling presenterar a¨ven tv˚ a applikationsbaserade lösningar, ett tidsynkroniserat EKG n¨ atverk och en tillf¨ orlitlig tr˚ adl¨ ost I/O f¨ or ett vattenkraftverk med kort tidsf¨ ordröjning..

(189)

(190) Abstract Wireless communication in industrial, scientific and medical applications have several benefits. The main benefits when using wireless technologies include ease-of-deployment, the simplicity to introduce new units into the network and mobility. However it also put higher demands on the communication, including reliability and predictability compared to wired communication. The reliability issues correlate to the radio communication and the possibility to ensure that the user data is received, and within the time frame of the system requirements. This doctoral thesis presents an empirical measurement approach to investigate and model the behaviour linked to reliability and predictability. The focus of the work presented is energy consumption, packet-errorrate and latency studies. This is performed for various radio technologies and standards in harsh environments. The main contributions of this thesis are the measurements platforms and procedures that have been developed to meet the requirements to investigate modern radio technologies in terms of predictability and reliability. This thesis show that it is possible to predict wireless communication in radio harsh environments. However it is necessary to determine the characteristics of the environment to be able to choose a suitable radio technology. The measurement procedures presented in this thesis alongside the platform developed enable these types of investigations. In this thesis a model of the energy consumption for a Bluetooth radio in low-duty-cycle applications with point-to-multipoint communication is presented. The measurements show that distance and transmission power will not effect the energy consumption for a Bluetooth nor ZigBee module. However the packet-error-rate and number of retransmissions will affect the overall energy consumption, and these parameters can be correlated to distance and foremost the environmental characteristics. This thesis also presents two application-based solutions, a time synchronized ECG network with reliable data communication as well as a low-latency wireless I/O for a hydro plant..

(191)

(192) Preface I have during my time as a doctoral student had the pleasure to get to know a lot of people that I today can call my colleagues as well as friends. I would like to thank all of you for making all of this possible. First I will thank my supervisors, Maria Lindén, Mats Björkman, Mikael Ekstr¨ om and Javier Garcia Casta˜ no. I also would like to thank the colleagues whom I have had the pleasure to share an office with; Marcus Bergblomma, Nikola Petrovic, Jimmie Hagblad, Gregory Koshmak and Christer Gerdtman. Furthermore I would like to thank all colleges for the support and interesting discussions during these years. I especially would have to thank Lars Asplund, Fredrik Ekstrand, J¨ orgen Lidholm, Carl Ahlberg, Magnus Otterskog, H¨ useyin Ayhan Aysan, Caroline Upps¨ all, Joakim Wangborn and Conny Collander. This would not have been possible without the support from my family. I thank you all for the support and encouragement, my parents, Britt and Bengt, and my brother and sister Mikael and Malin with their families and of course my mother in-law Eva Lind. I dedicate my doctoral thesis to my wife Ulla, my son Samuel and his future baby brother..

(193)

(194) Thesis.

(195)

(196) List of Publications The following is a list of publications that form the basis of the thesis (in reverse chronological order): Paper A Comparison study between ZigBee and Bluetooth with regard to power consumption, packet-error-rate and distance Martin Ekstr¨ om, Marcus Bergblomma, Maria Lindén, Mats Björkman, Mikael Ekström Submitted to The Fourteenth International Symposium on a World of Wireless, Mobile and Multimedia Networks, IEEE WoWMoM 2013 Paper B Development of Programmable Micro Power Meter Testbed for Radio Modules Martin Ekstr¨ om, Marcus Bergblomma, Maria Lindén, Mats Björkman, Mikael Ekström Submitted to IEEE Transactions on Instrumentation and Measurement Paper C A Bluetooth Radio Energy Consumption Model for Low-Duty-Cycle Applications Martin Ekström, Marcus Bergblomma, Maria Lindén, Mats Björkman, Mikael Ekström IEEE Transactions on Instrumentation and Measurement Volume: 61 , Issue: 3 Page(s):609-617, Date of publication: March 2012 Paper D A Wireless Low Latency Control System for Harsh Environments Marcus Bergblomma, Martin Ekstr¨ om, Mats Bj¨ orkman, Mikael vii.

(197) viii. Ekstr¨ om, Christer Gerdtman 11th IFAC/IEEE International Conference on Programmable Devices and Embedded Systems PDES 2012 Paper E Wireless ECG Network Marcus Bergblomma, Martin Ekstr¨ om, Mikael Ekstr¨ om, Javier Garcia Casta˜ no, Mats Bj¨ orkman, Maria Lindén, World Congress on Medical Physics and Biomedical Engineering Information and Communication in Medicine, Telemedicine and E-health, p 244-247, Munich 2009 Paper F Bluetooth energy characteristics in wireless sensor networks Marcus Blom, Martin Ekström, Javier Garcia Casta˜ no, Maria Lindén, 3rd International Symposium on Wireless Pervasive Computing, 2008. ISWPC 2008. Date of Conference: 7-9 May 2008 Page(s): 198 - 202 Product type Conference Publications.

(198) Contents 1 Hypotheses. 15. 2 Problem formulation. 17. 3 Introduction 3.1 Wireless sensor networks . . . . . . . . . . . . . 3.2 Wireless communication in harsh environments 3.3 Wireless communication standards . . . . . . . 3.3.1 Bluetooth . . . . . . . . . . . . . . . . . 3.3.2 Zigbee . . . . . . . . . . . . . . . . . . . 3.4 Testbed development . . . . . . . . . . . . . . .. . . . . . .. . . . . . .. . . . . . .. . . . . . .. . . . . . .. . . . . . .. 19 19 20 20 21 23 26. 4 Related work 4.1 Harsh environments . . . . . . . . . . . . 4.2 Testbed development . . . . . . . . . . . . 4.2.1 Radio Platforms . . . . . . . . . . 4.2.2 Wireless Sensor Network Testbeds 4.3 Radio energy consumption models . . . . 4.3.1 LEACH . . . . . . . . . . . . . . . 4.3.2 Bluetooth . . . . . . . . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. 27 27 28 28 29 30 30 31. . . . . . . .. . . . . . . .. . . . . . . .. 5 Research approach. 33. 6 Contribution 6.1 Research questions . . . . . . . . . . . . . . . 6.1.1 What behaviour need to be predicted? 6.1.2 How to develop accurate models? . . . 6.1.3 What is a harsh environment? . . . . ix. . . . .. . . . .. . . . .. . . . .. . . . .. . . . .. . . . .. 35 35 35 36 38.

(199) x. Contents. 6.2. Contribution of included 6.2.1 Paper A . . . . . 6.2.2 Paper B . . . . . 6.2.3 Paper C . . . . . 6.2.4 Paper D . . . . . 6.2.5 Paper E . . . . . 6.2.6 Paper F . . . . .. papers . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. 39 39 40 41 42 43 44. 7 Discussion 47 7.1 Research contribution . . . . . . . . . . . . . . . . . . . . 47 7.2 Future work . . . . . . . . . . . . . . . . . . . . . . . . . . 49 Bibliography. 51.

(200) Chapter 1. Hypotheses In this section the hypothesis of my thesis work is presented. The problem formulation, stated in section 2, and research approach, as described in section 5, is directly derived from the hypothesis. As it is stated by Eric Rogers: ”Hypotheses are single tentative guesses–good hunches–assumed for use in devising theory or planning experiment, intended to be given a direct experimental test when possible.” [1]. The hypothesis that form the base for the research presented in this doctoral thesis is as follows: To create accurate models that predict the behaviour of a radio these ought to be based upon empirical measurements. A good knowledge of the surroundings is equally important. The fewer assumptions needed the better. The focus of this thesis is empirical measurements, measurements testbeds and observing the results and thereby creating models used to predict the behaviour of the radio in wireless communication applications. Patterson et. al stated; ”Good data outlast bad theory” [2]. 15.

(201)

(202) Chapter 2. Problem formulation When replacing wired communication with wireless solutions some obvious problems emerge when trying to maintain a reliable and predictable communication. If the wireless application is deployed in a harsh environment, as described in the introduction 3, it is even harder to make an accurate prediction of the communication. The main research challenges is as follows; Can the behaviour of wireless communication be predicted when utilizing radio standards such as IEEE 802.15.1 and 802.15.4 in harsh environments? This is a complex question that raises several other questions that needs to be investigated. In the following part of this section the main questions are described as research question (RQ) 1-3. RQ1 What behaviours needs to be predicted? When looking upon wireless communication some main constrains needs to be investigated to make accurate prediction, these include energy consumption, latency, retransmission, packet loss and distance. RQ2 How accurate models can be developed? An assumption that a model is accurate implies that it is based upon empirical measurements with as few assumptions as possible. Which itself ironically is an assumption. My belief is that accurate 17.

(203) 18. Chapter 2. Problem formulation. models are based upon empirical measurements with good knowledge of the surrounding and how these effect the measurements. RQ3 What is a radio harsh environment? How do we define an environment as harsh, which aspects make the environment harsh for a specific radio communication? What measurements or communication tests can be made to define an environment as harsh, are there certain mechanisms included in different radio standards e.g. frequency hopping, forward error correction or modulation techniques that are more suitable in these types of environments? These questions leave out the important areas safety and security in this thesis as they are more of related to the application or radio standard chosen..

(204) Chapter 3. Introduction When replacing wired communication with wireless solutions the reliability and predictability is a big obstacle to overcome. Even more so when the communication is deployed in an what could be described as a radio harsh environment. This thesis focuses on two short distance and low power radio standards that can be bought off-the-self are investigated. These standards are Bluetooth and ZigBee, and will be introduced in this chapter. This chapter is divided into the following parts: Wireless Sensor Network (WSN) 3.1, Wireless communication in harsh environments 3.2, Wireless communications standards 3.3 and the final part of the introduction is, Testbed development 3.4.. 3.1. Wireless sensor networks. Wireless sensor network consists in most cases of many small, low power, and battery equipped sensor nodes. These nodes will have limited processing and communication capabilities that vary depending on the application. The purpose of WSNs is foremost to monitor one or more physical properties e.g. medical, environmental or industrial environments. The limited energy available due to the battery makes the research for energy preservation vital. These studies usually rely on simulations when developing e.g. more accurate routing, topology or data transmission algorithms. These studies are therefore dependent on accurate energy models to ensure correct results. 19.

(205) 20. 3.2. Chapter 3. Introduction. Wireless communication in harsh environments. Wireless communication in harsh environments is an interesting topic both as a research area and from an industrial point of view. Physical parameters such as background noise, induced from machines or other wireless standard may interfere. Fading of the communication channel due to multipath propagation or shadowing due to obstacles affecting the radio wave. J. Beutel et.al. shows in [3] that wireless sensor network may work in extreme harsh environments as the European Alps with X sense platform. J. F. Coll defines in his licentiate thesis a characterization of both industrial and hospital environments for a wireless communication point of view. Typically industrial building structures consists of a high amount of metal that both can act as an reflector meaning it can lead to delayed multipath components in the channel. However some materials in the building absorb the radio wave leading to a coverage problem for the communication. Other interfering sources that is present both in the industry as well as in medical environment are; radiated electromagnetic energy from wireless communication systems and electromagnetic noise from electronic equipment, e.g. instruments or machines [4]. Different radio standards utilize techniques to compensate for these problems e.g. different modulation techniques. The standards investigated in this thesis and what techniques used for reliable and predicable communication are presented in Section 3.3.1 and Section 3.3.2.. 3.3. Wireless communication standards. In this thesis work the ZigBee and Bluetooth technology have been investigated. Both technologies are of great interest for the industry and research because of the advantages such as low-cost implementation and low energy consumption. However both have different strengths and weaknesses since they are design for different applications. In the the ZigBee and Bluetooth comparison study for industrial applications it is suggested that real wireless networks for the industry will be hybrids of both technologies [5]. This is illustrated by the development of BlueGee, a hardware platform design to act as a gateway for Bluetooth and ZigBee protocols [6]..

(206) 3.3 Wireless communication standards. 21. This section presents both technologies with a brief revision of versions and standards included in the standards and specifications.. 3.3.1. Bluetooth. Bluetooth was developed in 1994 by telecom vendor Ericsson as wireless cable replacement. Today the development and qualifications of the IEEE 802.15.1 standard is managed by Bluetooth Special Interest Group (SIG) [7]. Bluetooth utilizes Gaussian Frequency Shift Keying (GFSK) modulation as well as Frequency-Hopping Spread Spectrum (FHSS) scheme over 79 channels in the 2.4 GHz Industrial, Scientific and Medical (ISM) band. Each Bluetooth channel is divided into time slots with a duration of 625 $mu§. Bluetooth offers three different power classes, 1,2 and 3 and are defined by the maximum output power, 100 mW, 2.5 mW and 1 mW. These correlate to maximum transmission distances of 100 m, 10 m and 1 m respectably. The basic network topology for Bluetooth is called a Piconet. A Picnet offers point-to-point or point-to-multipoint communication where A Master device can connect upto 7 Slave devices in each Piconet. Devices in a Piconet can interoperate in more than one picnet, thereby enabling Scatternet formation. Version 1.0 Early version Bluetooth v1.0 and v1.0B struggled with several problems when it was ratified as an standard in July [8] and in December 1999 [9]. These included anonymity problems in the connection process due to the mandatory hardware device address transmission during this process. A major setback for the standard was that the vendors had issues concerning the interoperability for the products. Version 1.1 Bluetooth v1.1 focused on fixing the known problem in earlier version and adding the possibility of non-encrypted channels and the Received Signal Strength Indicator (RSSI) as specified in [10]. Version 1.2 The specification for Bluetooth v1.2 [11] was introduced in 2003 and is fully backward compatible with v1.1. The major improvements in include: Adaptive frequency-hopping spread spectrum (AFH) a technique that aims to improve resistance to radio frequency.

(207) 22. Chapter 3. Introduction. interference by avoiding the crowded frequencies in an hopping sequence. Extended Synchronous Connections (eSCO Allowing retransmissions of corrupted packets to improve latency, mainly used to improve audio links. Host Control Interface (HCI) A standardized communication between the host stack, e.g. micro controller or PC, and the Bluetooth circuit. In this version the three-wired UART was implemented. Flow control and Retransmission modes was implemented into the Logical link control and adaptation protocol (L2CAP) layer. Higher transmissions speed in practice upto 721 kbit/s. Version 2.0+EDR When the specification for Bluetooth v2.0 was released in 2004 the Enhanced Data Rate (EDR) was introduced in [12]. EDR utilizes Gaussian frequency-shift keying (GFSK) combined with Phase Shift Keying (PSK) modulation with two variants, either π/4-DQPSK or 8DPSK. This addition to the standard gives a nominal transmission rate of 3.0 Mbtis/s, however a practical transfer rate at 2.1 Mbit/s. The EDR offers a reduced duty cycle leading to a lower power consumption. However the EDR addendum is not mandatory for the Bluetooth standard. Version 2.1+EDR In July 2007 the Bluetooth core specification version 2.1 was adopted by the Bluetooth special interest group (SIG) [13]. The main features that differs from previous version were: Secure Simple Pairing (SSP) that increases the strength of security Extended Inquiry Response (EIR) that provides more information before the connection phase. Version 3.0+HS Bluetooth version 3.0 + High Speed (HS) offers a theoretical data transfer rate upto 24M bit/s over a IEEE 802.11 link. The Bluetooth link handles the negotiation and establishment of the link while data transfer is performed over the IEEE 802.11 link as described in [14]..

(208) 3.3 Wireless communication standards. 23. The amendment + HS is not a mandatory part of Bluetooth version 3.0 but the core need to support the following features as specified in [15] L2CAp Enhanced Modes that includes Enhanced Retransmission Mode (ERTM) for relieable L2CAP channel, and Streaming mode (SM) without flow control nor retransmission for a unreliable channel. Alternate MAC/PHY While the Bluetooth radio perform device discovery, connection and configuration of profiles the high speed alternate MAC PHY 802.11 will be used for transmitting large quantities of data. Unicast Connectionless Data Allows sending service data with out an L2CAP channel. Intended for small amount of data in low latency applications. Enahanced Power Control is an update of the power control introduced in EDR addendum. The ambiguities introduces by the added modulation techniques in power control in EDR are solved by specifying the behaviour. Version 4.0+LE The Bluetooth Core Specification was adopted 30 June 2010 [16]. This specification included Classic Bluetooth, Bluetooth high speed and Bluetooth low energy (BLE) protocols. The BLE a new stack protocol for rapid build-ups and simple links, intended as an alternative to the Bluetooth standards version 1 to 3. The main aim is very low power consumption applications with two different chips modes; Single mode this chip version allows only the low energy protocol stack on the chip. Duel mode The low energy features are integrated into a Classic Bluetooth controller.. 3.3.2. Zigbee. ZigBee is a standard based technology, not a standard itself, aimed towards low-cost, low-power sensor- and control networks. The specification aim for easy implementation and self-organizing networks. The.

(209) 24. Chapter 3. Introduction. ZigBee Alliance has developed the ZigBee specification since 2002 is aimed to enhance the IEEE 802.15.4 standard by adding network, security layers and a application framework. ZigBee is build upon the PHY and MAC as it defined in the IEEE Standard 802.15.4 in 2003 [17]. The main characteristics for the ZigBee specification is: ZigBee uses Direct Sequence Spread Spectrum (DSSS) and Offset quadrature phase shift keying (OQPSK) with carrier sense multiple access with collision avoidance in 1 channel in 868 MHz (Europe), 10 channels in the 915MHz band (USA and Australia) and 16 channels in the 2.4GHz ISM band according to IEEE 802.15.4 [18]. Power saving mechanisms for all device classes Security key generation mechanism and utilizes industry standard AES-128 security scheme Zigbee network layer supports multiple Star topology, Tree and generic Mesh networks. The ZigBee alliance offers two specification that serve as a base for networking systems, Zigbee and ZigBee RF4CE specifcations [19]: ZigBee specification is officially named ZigBee 2007 and is available in two versions: ZigBee is aim towards smaller networks with hundreds of devices in network. ZigBee Pro supports more than 64,000 devices in a single network ZigBee RF4CE specification is designed for point-to-point communication, meaning that the mesh networking capabilities are not required. RF4CE specification is targeted towards devices-to-devices applications. ZigBee Alliance offers industry specific standards to simplify the implementation for the specific application [20]. Each standard are presented with short description of their features below. Building automation developed for large-scale deployments, supports upto thousands of devices in the same network. Simple to connect.

(210) 3.3 Wireless communication standards. 25. to internet or other networks via gateways. Combines the ZigBee Pro and Building Automation and Networking (BACnet) [21] protocol. [22] Health care enables remote patient monitoring with maintained freedom of mobility. Time stamped data for synchronisation. Full support for IEEE 11073 devices including; glucometer, pulse oximeter, electrocardiograph, weight scale. [23] Home automation developed for simple installation with self-organizing networks and embedded maintenance. This standard supports a wide range of devices such as intruder alarms, lightning control and heating/colling units. [24] Input device design for the RF4CE specification. Input Device standard is developed primarily for consumer electronics, e.g. keyboards, mice and touchpads. [25] Light link LED lighting control standard. timers, dimmers and remotes. [26]. Including light-switches,. Network devices standard for ZigBee Gateways to expand ZigBee Pro based networks. Zigbee Gateways enables Zigbee networks to connet to the Internet or other services providers. Network Devices standard complement the ZigBee standards; Building Automation, Health Care, Home Automation, Retail Services, Smart Energy and Telecom Services. [27] Remote control aim to replace the infrared (IR) technology commonly used for remote controls in consumer electronics. Claimed benifits are lower energy consumption than IR and non-line-sight communication. [28] Retail services standard is developed to enhance and automate the retail consumer market. It allows the user to control, monitor and automte the purchase and deliverance of products. [29] Smart energy standard is designed monitor, control, inform and automate the water and energy usage and deliverance. Version 2.0 is under development and will support control of plug-in vehicle charging, installation and configuration of home area networks (HAN). [30].

(211) 26. Chapter 3. Introduction. Telecom services standard is developed for Telecom consumer products and services. These include mobile payment, location based services and mobile gaming with the use of ZigBee SIM card. [31]. 3.4. Testbed development. The importance of testbeds for network research are highlighted in the National Science Foundation workshop report from 2002 [32]. Here it is stated: ”A the lack of good measurements and models of traffic, protocols, and applications hinders the network research community. Appropriately instrumented testbeds can help to solve this problem”. The need of programmable hardware for is also highlighted. The specific challenges for wireless sensor networks, such as resource constrains as mention in 3.1 is also emphasised. The proposed solution for the sensor network testbeds are that these should be stronger application focused than conventional network testbeds due to its interdisciplinary nature..

(212) Chapter 4. Related work This chapter will present the related work: within wireless communication in harsh environments in section 4.1, the platform and testbed development in section 4.2 and finally the work related to this thesis within energy consumption for wireless communication in section 4.3.. 4.1. Harsh environments. C. A. Boano et al. presented in [33] how temperature and weather conditions can affect the link quality and data delivery in low power wireless communication when deployed in an oil refinery. The radio platform used is the Tmote Sky [34] that utilizes the Texas Instrument CC2420 radio chip [35]. Abdullah Kadri presented in [36] a performance study of IEEE 802.15.4 based communication in harsh environments. The same radio chip have been used as in [33]the CC2420 from Chipcon. The sensor nodes used in this study were deployed in an industrial environment with a electromagnetic interference signal; a frequency-swept sinusoidal EMI with a peak power ranging from -45 to -61 dBm. Kadri draws the conclusion that the performance of the sensor network depend strongly on the pulse-modulated sinusoidal interference, both the pulse width and period between pulses. A wireless temperature sensor application deployed in harsh environment is presented by E. Sisinni et. al in [37]. The main scientific contributions of the paper is design , implementation and experimental 27.

(213) 28. Chapter 4. Related work. validation of a IEEE 802.15.4 based wireless solution for temperature sensing in harsh environments. The main obstacle for the solution is described as; ”...overcoming limitations imposed by the wireless medium” in order to achieve high availability and reliability for the controlled process. Delin and Jackson presented in [38] a Sensor Web based platform that aim to be used for in situ measurements gaseous biosignatures that can be deployed in space, atmosphere, aqueous, land, and artificial (i.e. buildings and spacecraft) environments.. 4.2. Testbed development. This chapter will present the most commonly used testbeds and platforms that have been developed for wireless sensor network measurements and simulations. This section is divided into two parts; radio testbeds and Wireless Sensor Network Platforms.. 4.2.1. Radio Platforms. This section will present the radio modules and embedded solutions that enable wireless sensor network applications and research. These are commonly known as motes or sensor nodes. The main components on a mote are; Controller, Transceiver, Sensor(s), Power source and in many cases an external Memory. The Smart Dust [39] and NASA Sensor Webs [40] are the recognized as the projects that started the development of modern motes. Berkeley University The Smart Dust project [39] was the first of a family of mote platforms that have been developed at Berkeley University’s lab. The initial Smart Dust aimed to develop mote small enough so that they would be able to be suspended in the air when distributed. The physical layer for the communication were based on optical transmission. The SunSpot [41], Intel iMote and iMote2 [42] are good examples of platform that have derived from the Smart Dust project that utilizes the more conventional radio platforms. Mica mote The MICA motes are developed at Berkley. The first generations of motes, MICA and MICA2, [43] utilizes the TR1000.

(214) 4.2 Testbed development. 29. radio chip from RF Monolithic that operates at 916 MHz band [44]. The MICAZ mote operates at 2.4 GHz band and with a ZigBee compliant radio [45]. EYES The Energy Efficient Sensor Networks (EYES) nodes are developed by Infineon and utilizes the MSP430 microcontroller from Texas Instrumment and Infineon TDA5250 radio. Tmote A texas Instrument MPS430 controlled mote with IEEE 802.15.4 wireless transceiver. The Tmote is equipped with integrated humidity, temperature and light sensors. The Tmote have implemented a link-layer encryption and authentication, AES-128. BTnode Jan Beutel presented in [46] the third version of the BTnode. A mote equipped with a Bluetooth version 1.2 Zeevo ZV4002 module and the CC1000 low power radio from Chipcon that operates in the range of 433 to 915 MHz [35]. Egs Ko et. al presented in [47] a ARM Cortex M3 based mote platform that utilizes both Bluetooth Version 2.0 and a Zigbee radio. This platform is primarily developed for medical sensing application. PicoRadio Jan M. Rabaey et. al present in [48] the architecture and hardware for the PicoRadio. The project that started at Berkley University in 1999 aim to develop communication design techniques and tools. iBadge as a part of the Smart Kindergarten the iBadge platform was developed to monitor the interaction of small children and their environment. The iBadge was design to communicate with the Sylph, a middleware infrastructure that capture the data and translate it to information about the child’s learning process. The iBagde was developed at University of California, Los Angeles (UCLA) and utilizes a Bluetooth chipset from Ericsson [?] for the wireless communication between person-to-person and person-tocomputer [49].. 4.2.2. Wireless Sensor Network Testbeds. This section describes the full scale WSN testbeds..

(215) 30. Chapter 4. Related work. MoteLab is a web based network testbed developed at Harvard University. The mote platforms used is the MICA2 and MicaZ motes [50], however the MoteLab is a device independent framework. It also includes a digital multimeter measurement setup for power measurements for MicaZ application. The WINTeR testbed is developed to support industrial wireless sensor network applications in radio-harsh environments. The WINTeR testbed is designed to support the development of WSN technology including Physical layer (PHY) development, propagation models and the validation of wireless solution for radio-harsh environments [51]. Sensor Web is WSN developed at NASA in 1997. The project aim towards a network where each node could act as a coordinator. The nodes would be spatial distributed without a specific router or coordinator meaning that all nodes could talk to each other [38]. μAMPS The μAMPS program focuses on low power wireless sensor networks. Wendi Heinzelman at the Massachusetts Institute of Technology presented the Low Energy Adaptive Clustering protocol (LEACH). This protocol aim to prolong the wireless sensor network lifetime by randomly changing roles in the network and thereby distributing the energy consumption throughout the network [52]. 4.3. Radio energy consumption models. This section will present the related work on energy and power consumption models and the methods used to extract them.. 4.3.1. LEACH. The first order radio model presented in the LEACH project refers to the Bluetooth initiative in 1999. This model characterises the power needed to transmit and receive data. The power effecting parameters are: number of bits (k) and distance (d) [52]. The transmit model is presented in Equation 4.1 and receive model in Equation 4.2, where Eelec is the energy dissipated to run the transmitter or receive circuit and amp is energy dissipated for the transmitter to achieve energy per bit.

(216) 4.3 Radio energy consumption models. 31. to noise power spectral density ratio Eb /N0 . In this model the Eelec = 50nJ/bit and amp = 100pJ/bit/m2 ET x (k, d) = Eelec ∗ (k) + amp ∗ k ∗ d2. (4.1). ERx (k) = amp ∗ k. (4.2). The second order radio model is an extension of the first with addition of how power control will effect the power consumption. the power control is implemented to over come the power loss due to both fading according to free space and multipath fading. A threshold distance, d0 , will determine when the transmission model use the free space or the multipath fading model [53]. The second order radio model for a transmitter is shown in Equation 4.3 and for a receiver in Equation 4.4 with l = numberof bits Eelec = 50nJ/bit, f s = 10pJ/bit/m2 and mp = 0.0013pJ/bit/m4 lEelec + f s ∗ l ∗ d2 d < d0 ET x (l, d) = (4.3) lEelec + mp ∗ l ∗ d4 d ≥ d0 ERx (l) = lEElec. 4.3.2. (4.4). Bluetooth. A finite state machine (FSM) based power consumption model for Bluetooth technology is presented by Negri et. al in [54]. The measurement set-up to obtain the characteristics for the two Bluetooth modules utilizes a digital multimeter controlled by PC with a LabView application to monitor the power consumption. This FSM based model is used for a point-to-point connection for two various implementations i presented by Negri et. el in [55]. The enhanced measurement procedure for acquiring a high-level power consumption model for point-to-point communication for Bluetooth devices is presented by Macii et. al in [56]. The same FSM model is used in the simulator for Bluetooth power consumtpion presented by Macii [57]. The power consumption for Bluetooth Scatternet formations are investigated and presented for the BTnode by Negri et. al in [58]..

(217)

(218) Chapter 5. Research approach To be able to answer the questions in section 2 empirical measurements are needed. As it is stated in [32];”...given a well instrumented testbed, the measurement data collected during its operation may be first-class research output in its own right” However there is no available platform that allows the user to control the communication as well as measure the energy consumption at the same time. It is also critical to evaluate the background noises and to monitor the radio links to be able to understand how different setting and environments influences the radio communication. The main focus of this research has been to develop hardware that suites the requirements for various application. To enable accurate energy consumption for a Bluetooth radio module used in wireless sensor networks a hardware platform as presented in Paper D has been developed. This platform enables monitoring of the power consumption in realtime as well as control the radio of communication. The second platform presented in Paper B(POMPOM) is an evolution of the first platform with enhanced resolution of the energy consumption measurements to meet the new radio standard such as Bluetooth low energy. POMPOM also offers interchangeable radio modules, enabling different radio standard and modules to be compared with the same platform. To complement the embedded platform a spectrum analyzer, Rohde & Schwartz FPS, has been used to record and monitor the background noise. To monitor packet error rate and retransmission different radio sniffer have been used for Bluetooth version 2.0 + EDR the FS4BT from 33.

(219) 34. Chapter 5. Research approach. Frontline Inc and for ZigBee and Bluetooth Low energy Texas Instrument SmartRF packet sniffer (SmartRF) have been used..

(220) Chapter 6. Contribution This chapter is divided into two parts. The first part the research questions formulated in Chapter 2 will be answered with the help of the contributions in the included papers . The second part of the chapter Paper A-F is presented in short with the specific scientific contribution and the author’s contribution for each paper.. 6.1. Research questions. The contributions of the collection of included papers are presented in the following section.. 6.1.1. What behaviour need to be predicted?. Predictability within wireless communication applications there are several parameters that need to be considered. These included energy consumption, latency and reliability of the link. Energy consumption The energy consumption is the main focus in Paper F and in Paper C. It is also a main objective in the development of the testbed in Paper B and the measurements performed in Paper A. The main contribution of Paper C is the energy consumption model for a Bluetooth version 2 + Enhanced Data Rate (EDR) radio module in Low-Duty-cycle application. The main contributions 35.

(221) 36. Chapter 6. Contribution. in this paper is mapping of the behaviour of the Bluetooth radio module in low energy sniff mode. Notable results presented is that the extra energy consumption due to additional transmission power is detectable but not significant for the overall energy consumption for the radio module. In Paper B the energy consumption for a Zigbee radio module as a function of packet-error-rate (PER) and distance is presented. Link reliability is investigate in Paper A and a PER comparison is presented for Bluetooth and Zigbee for different environments and distances. A time synchronised network for continues ECG monitoring is presented in paper Paper E. LatencyThe latency and packet error rate for an industrial application in hydro-power plant is presented in Paper D. The comparison study in Paper A focuses on the PER for Bluetooth and ZigBee in various environments.. 6.1.2. How to develop accurate models?. As stated in the Hypotheses the main work in this thesis is based upon empirical measurements. To enable the measurements needed to observe the behaviour described in previous Section 6.1.1 specialized hardware platforms have been developed. This part of the chapter includes three parts that correlates to the Research question 2 6.1.2; Testbed development, Empirical Measurements and finally Surrounding environment. Testbed development The hardware platform, called Power Efficient WIreless Technology (PEWIT), is used as a testbed in Paper C and Paper E. It is developed to investigate the Bluetooth version 2 + EDR radio modules in WSN applications. The architecture of the embedded measurement platform PEWIT is described in Paper C. Paper B presents the Programmable Micro Power Meter Testbed for Radio Modules (POMPOM). A testbed with higher resolution and measuremtn range than PEWIT adapted for the modern radio standards such as [16]. The POMPOM offers a measurement range from 0.4 μ to 155 mA and sample rate of 50kHz. The architecture of POMPOM testbed is described in Paper B. POMPOM offers the.

(222) 6.1 Research questions. 37. possibility to change radio module without any hardware modification to the testbed. POMPOM is used in Paper A to study and compare Zigbee and Bluetooth in harsch environments. Empirical measurements The measurement testbeds and instruments used for the measurement setup are presented in this section. In this thesis empirical measurements is the main focus. The three parameters presented in section 6.1.1 concerning the performance off the wireless communication are: Energy consumption The testbeds presented in the previous section have been used to collect data about the energy consumption depending on radio standard, radio configuration and environmental surroundings. The measurement setup presented in Paper F shows teh general characteristics for Bluetooth version 2 in WSN. The setup is similar to set-up presented in Section 4.3 by [54] with the exception of that oscilloscope was used instead of an digital multimeter. This set-up is sufficient to obtain general energy consumption characteristics for the Bluetooth radio module. However the embedded measurement platform presented in Paper C and Paper B enable distributed in situ measurement with full control of the communication. This simplify the synchronization of the measurements and parameter extraction for the energy consumption models. Link reliability The testbeds have been complemented with sniffers to analyse the packet error rate and retransmission. The sniffers used are Frontline’s Bluetooth Sniffer (FTS4BT) [59] and SmartRF [60] Latency To determine the latency for the radio link in Paper D the FTS4BT is to determine the PER. The cumulative probability for Bluetooth communication is determined with a Tektronix TDS3012 oscilloscope [61] was used. Surrounding environment To acquire a good knowledge of the surrounding environment when setting up the measurements, a Frequency Spectrum Analyser (FSP) from Rohde & Schwartz [62], have been used to gather data about the background noise. The FSP.

(223) 38. Chapter 6. Contribution. was used in Paper D and Paper A to investigate the background noise.. 6.1.3. What is a harsh environment?. To obtain a reliable and predictable wireless communication in radioharsh environments, it is vital to understand what defines a radio-harsh environment. Empirical measurements with the measurement setup described in the previous section makes it possible to correlate the behaviour for a radio to the environment i operates in. In Paper D the communication for a Bluetooth radio in a hydro-plant is investigated. The study shows that a wireless low latency control system with possible to deploy with the reliability needed. In Paper A a comparison study for Bluetooth and ZigBee radio is present the link reliability, energy consumption, distance between transmitter and receiver correlated to the environment. This study shows the possibility to investigate how different environments will effect the reliability of the link. It also present a method for creating models of the predictability for wireless communication in radio-harsh environments..

(224) 6.2 Contribution of included papers. 6.2. 39. Contribution of included papers. This section will present the authors and scientific contribution of teh included papers.. 6.2.1. Paper A. Title Comparison study of ZigBee and Bluetooth with regards to power consumption, packet-error-rate and distance Authors Ekstr¨ om, Martin. Bergblomma, Marcus. Linden, Maria. Björkman, Mats. Ekstr¨ om, Mikael. Source Submitted to The Fourteenth International Symposium on a World of Wireless, Mobile and Multimedia Networks, IEEE WoWMoM 2013 Publication type Conference Publications Summary This paper present a empirical measurement comparison study of ZigBee and Bluetooth. The parameters investigated are power consumption, packet-error-rate or retransmissions and distance in different environments. This study shows the differences and similarities for the two different short range radio technologies. A measurement set-up and procedure that makes it possible to investigate power consumption of the radio module, retransmissions and packet-error-rate as well as ambient noise is presented. For both the Bluetooth and the ZigBee modules used in this study the distance itself have no influence of the power consumption. However the retransmission rate and packet-error-rate have a large influence on the power consumption. This study have show that the environment has a great impact on the range of the radio modules and the behaviour concerning the retransmission rate and packet-error-rate. In this study we will try to answer a big part of the main research question; Can we predict the behaviour of wireless communication when utilizing radio standards such as IEEE 802.15.1 and 802.15.4 in harsh environments?.

(225) 40. Chapter 6. Contribution. Contribution A measurement set-up procedure for empirical measurements to determine the reliability and predictability for wireless communication in various environments. Comparison study of Bluetooth and Zigbee in various environments. Author’s contribution Measurement set-up Original idea Parameter extraction in comparison study. 6.2.2. Paper B. Title Development of Programmable Micro-Power-Meter Testbed for Radio Modules Authors Ekstr¨ om, Martin. Bergblomma, Marcus. Linden, Maria. Björkman, Mats. Ekstr¨ om, Mikael. Source Submitted to IEEE Transactions on Instrumentation and Measurement Publication type Journals and Magazines Short summary This paper presents the POMPOM testbed for high precision power consumption in situ measurements for interchangeable radio modules. The main requirements for the development have been; Interchangeable radio modules to enable the same hardware testbed to be used independently of the radio standard used make comparison studies possible. The testbed should be programmable so that the need for hardware development should be minimized. The testbed must be able to act as a controller for the communication and simultaneously make accurate in situ measurements of the energy consumption of the radio. The required sample rate must be at least 50 kSamples per second. The range of the current measurement should cover at least 0.2 μAmpere to 60 mAmpere with at least 14-bit resolution..

(226) 6.2 Contribution of included papers. 41. Mobility, low cost and small size are vital for the testbed. It must be possible to deploy several measurement testbeds to act as sensor nodes in a wireless sensor network to capture the behavior of the entire network. The results for test measurement set-up for POMPOM is presented to illustrate a typical usage of the testbed. The results presented show how the testbed can be used to investigate the correlation between distance, packet-error-rate and current consumption for a Zigbee radio. Scientific contribution A platform that can be used to evaluate commercial off-the-shelf wireless data communication technology in aspect of power consumption, packet error rate and system latency for in situ measurements. Author’s contribution Original Idea Hardware development including; design and testing Power sensor development including testing and calibration. 6.2.3. Paper C. Title A Bluetooth Radio Energy Consumption Model for LowDuty-Cycle Applications Authors Ekstr¨ om, Martin. Bergblomma, Marcus. Linden, Maria. Björkman, Mats. Ekstr¨ om, Mikael. Source IEEE Transactions on Instrumentation and Measurement Volume: 61 , Issue: 3 Page(s): 609 - 617 , Date of publication: March 2012 Publication type Journals and Magazines Short summary This paper presents a realistic model of the radio energy consumption for Bluetooth-equipped sensor nodes used in a low-duty-cycle network. The model is based on empirical energy consumption measurements of Bluetooth modules. This model will give users the possibility to optimize their radio communication with respect to energy consumption while sustaining the data rate. This paper shows that transmission power cannot always be directly related to energy consumption. Measurements.

(227) 42. Chapter 6. Contribution. indicate that, when the transmission power ranges from -5 to +10 dBm, the difference in consumed energy can be detected for each transmission peak in the sniff peak. However, the change is negligible for the overall energy consumption. The non-linear behaviour of the idle state for both master and slave when increasing the interval and number of attempts is presented. The energy consumption for a master node is in direct relation to the number of slaves and will increase by approximately 50% of the consumption of one slave per additional slave, regardless of the radio setting. Scientific contribution An energy consumption model for Bluetooth version 2 with enhanced data rate have been presented for low-duty-cycle networks. A wireless sensor network that can monitor its own energy consumption in real time as well as be deployed in situ. We have shown that transmission power do not have a significant effect on overall the energy consumption in Bluetooth radio Author’s contribution Measurement calibration and linearity testing Mathematical model extraction Hardware development of measurement platform including design and testing Measurement setup. 6.2.4. Paper D. Title A Wireless Low Latency Control System for Harsh Environments Authors Bergblomma, Marcus. Ekstr¨ om, Martin. Linden, Maria. Bj¨ orkman, Mats. Ekstr¨ om, Mikael. Source 11th IFAC/IEEE International Conference on Programmable Devices and Embedded Systems PDES2012 Publication type Conference Publications.

(228) 6.2 Contribution of included papers. 43. Short summary The use of wireless communication technologies in the industry offer several advantages. One advantage is the ability to deploy sensors where they previously could not easily be deployed, for instance on parts that rotate. To use wireless communication in industrial control loops, demands on reliability and latency requirements has to be met. This in an environment that may be harsh for radio communication. This work presents a reliable, low latency wireless communication system. The system is used in a wireless thyristor control loop in a hydro power plant generator. The wireless communication is based on Bluetooth radio modules. The work shows a latency analysis together with empirical hardware based latency and packet error rate measurements. The background noise of a hydro power plant station is also investigated. The average latency between the Bluetooth modules for the proposed system is 5.09 ms. The packet error rate is 0.00288 for the wireless low latency control system deployed in a hydro power plant. Scientific contribution Presents and evaluates a system where a commercial off the shelf wireless data communication technology is used in a harsh environment Author’s contribution Contributed to hardware design Contributed to mathematical statistics and validation of system latency. 6.2.5. Paper E. Title Wireless ECG Network Authors Bergblomma, Marcus. Ekstr¨ om, Martin. Linden, Maria. Bj¨ orkman, Mats. Ekstr¨ om, Mikael. Source World Congress on Medical Physics and Biomedical Engineering - Information and Communication in Medicine, Telemedicine and e-Health, p 244-247, Munich Publication type Conference Publications.

(229) 44. Chapter 6. Contribution. Short summary This paper presents a time synchronized wireless ECG sensor network with reliable data communication. Wireless ECG systems are a popular research area where several research groups have presented point-topoint solutions. Alongside the wireless ECG research, the wireless sensor network research has created an increasing interest for secure, low power and predictable network applications. Combining these research areas is a natural step for the evolution of secure wireless monitoring of physiological parameters. In this study the Bluetooth radio standard has been chosen for its versatility. This paper focuses on both the hardware and the software development for a functional multi-hop ECG network using Bluetooth. The presented wireless ECG network is reliable up to link loss and is easily configured to send more or different types of signals. The system has been tested and verified for secure multi-hop communication Scientific contribution Application oriented paper showing a time synchronized reliable wireless multi-hop network with Bluetooth equipped ECG sensor nodes. Author’s contribution Hardware development including design, sensor development (ECG amplifier) and testing. 6.2.6. Paper F. Title Bluetooth energy characteristics in wireless sensor networks Authors Marcus Blom, Martin Ekström, Javier Garcia Casta˜ no, Maria Lindén, Source 3rd International Symposium on Wireless Pervasive Computing, 2008. ISWPC 2008. Date of Conference: 7-9 May 2008 Page(s): 198 - 202 Publication type Conference Publications.

(230) 6.2 Contribution of included papers. 45. Short summary In this paper a measurement system to create an experimental model and a tool box for simulations concerning both the energy consumption and the time aspect when creating wireless sensor networks using Bluetooth 2.0 + enhanced data rate has been developed. Further energy and time characteristics for critical events when using Bluetooth 2.0 in wireless sensor networks are investigated experimentally, with the main events; create connection, send data, receive data, and idle state. Results show that when allowing higher latencies for the connection in the Wireless Sensor Networks the power consumption drops drastically when using low power mode as sniff. Scientific contribution General characteristics the energy consumption of a Bluetooth version 2.0 +EDR radio depending on radio configuration, payload and network set-up. Author’s contribution Hardware development including design, measurement set-up and testing..

(231)

(232) Chapter 7. Discussion This chapter will summarise and discuss the contribution of the work presented in this doctoral thesis. The continuation and future work within this topic will conclude the discussion and thesis.. 7.1. Research contribution. This thesis deals with the issues related to wireless communication in radio-harsh environments. The problem related to predictability and reliability of wireless communication is addressed in Chapter 2, whereas the scientific contribution of the included papers are presented in Chapter 6. The problem formulation is divided into three research questions and these are derived from the hypotheses that lay the ground for the work presented. The aim is of course to give the complete answer for all of the question presented in chapter 2. It is however a naive view to believe that the answers given in this thesis will be the complete truth. The first and second research questions are closely related to each other and all the included papers contribute to answer these two questions. When comparing our platforms to the platforms and testbeds presented in the related work our platforms focuses on in situ measurements with the possibility to monitor the behaviour of the radio module and its parameters. Both the PEWIT and POMPOM platforms enable us to control the communication as well as monitor the energy consumption of the radio module in real-time for all sensor nodes or motes used 47.

(233) 48. Chapter 7. Discussion. in the WSN. By extending the measurement set-up with packet sniffers and a spectrum analyser we are able to monitor the radio communication over the air as well as the background noise. These platforms and measurements procedures have been used for the studies presented in papers Paper A, Paper C, Paper D and Paper E. The application based study presented in Paper D show that it is possible to deploy and utilize wireless communication in what could be considered radio-harsh environments. This paper shows the possibility to monitor physical parameters where wired communication isn’t possible since the sensor node is placed on the rotor in a hydro-plant. The energy characteristics and models presented show that the energy consumption of modern radios may differ from the classic assumption about how a radio should behave. A good example of this is how the distance effect energy consumption, the main contribution in Paper A and Paper C is that the transmission power isn’t significant for the energy consumption. However as we have shown in Paper A and Paper B the PER have a large effect of the energy consumption, and the probability for increased PER is strongly correlated to the environment and distance between transmitter and receiver. Out of the three main research question the last:”What is a radio harsh environment?” is without a doubt the hardest to give a direct and complete answer. The main contribution of this work is the platform development and set-up procedures to enable research for empirical measurements for creation models for wireless communication in radio-harsh environments. The results and conclusions based upon the measurements presented in Paper A give a good understanding of how Bluetooth and ZigBee will behave when deployed in a radio-harsh environment. The results for how much energy the radio module will consume depending on PER or retransmission gives a good indication of the difference for the radio technologies. The correlation between distance and PER is only valid in the specific environments presented. However the measurement set-up and procedures can be used for all environments to determine whether it is radio-harsh environment..

(234) 7.2 Future work. 7.2. 49. Future work. The natural continuation of the work presented in this thesis is a continuation of empirical measurements with the aim to further explore the complex research question: What is an radio-harsh environment? This includes investigations of the properties of new radio standards and technologies. Future work could include an extension of the scope of the research towards interesting topics concerning safety and security..

(235)

(236) Bibliography [1] Eric. M Rogers. Physics for the inquiring mind. American Journal of Physics, 1962. [2] D. Karlin A. Kurose J. Lazowska E. Liddle D. McAuly D. Paxson V. Savage S. Patterson, D. Clark and E. Zegura. In Looking Over the Fence at Networks A Neighbour´s View of Network Research Committee on Research Horizons in Networking., number ISBN: 0-3090613-7. Computer Science and Telecommunication Board, National Research Counsil, National Academy Press, 2001. [3] J. Beutel, B. Buchli, F. Ferrari, M. Keller, M. Zimmerling, and L. Thiele. X-sense: Sensing in extreme environments. In Design, Automation Test in Europe Conference Exhibition (DATE), 2011, pages 1 –6, march 2011. [4] Javier Ferrer Coll. Rf channel characterization in industrial, hospital and home environments, 2012. QC 20120119. [5] N. Baker. Zigbee and bluetooth strengths and weaknesses for industrial applications. Computing Control Engineering Journal, 16(2):20 –25, april-may 2005. [6] E. Cano and I. Garcia. Design and development of a bluebee gateway for bluetooth and zigbee wireless protocols. In Electronics, Robotics and Automotive Mechanics Conference (CERMA), 2011 IEEE, pages 366 –370, nov. 2011. [7] Bluetooth Special Interest Group. www.bluetooth.org, 1998. [8] Bluetooth SIG. Specification of the bluetooth system volume 1, July 1999. 51.

(237) 52. Bibliography. [9] Bluetooth SIG. Specification of the bluetooth system volume 2, December 1999. [10] Bluetooth SIG. Ieee bluetooth standard 802.15.1-2002, 2002. [11] Bluetooth SIG. Ieee bluetooth standard 802.15.1-2005, 2005. [12] Bluetooth SIG. Bluetooth core version 2.0 + enhanced data rate, 2004. [13] Bluetooth SIG. Bluetooth core version 2.1 + edr, 2007. [14] Bluetooth SIG. Bluetooth core version 3.0 + hs, 2009. [15] Bluetooth SIG. Bluetooth core specification addendum (csa) 1, 2008. [16] Bluetooth SIG. Bluetooth Core Version 4.0, year = 2010. [17] ZigBee Alliance.. Zigbee technology. http://www.zigbee.org/ About/AboutTechnology/ZigBeeTechnology.aspx. [Online; accessed 6-October-2012].. [18] Ieee standard for information technology - telecommunications and information exchange between systems - local and metropolitan area networks specific requirements part 15.4: wireless medium access control (mac) and physical layer (phy) specifications for low-rate wireless personal area networks (lr-wpans), 2003. [19] ZigBee Alliance.. Zigbee specification. http://www.zigbee.org/ Specifications.aspx. [Online; accessed 6-October-2012].. [20] ZigBee Alliance.. Zigbee standards. http://www.zigbee.org/ Standards/Overview.aspx. [Online; accessed 6-October-2012].. [21] BACnet. Iso 16484-5:2012. http://http://www.bacnet.org/. [Online; accessed 6-October-2012]. [22] ZigBee Alliance. Zigbee building automation standard. http://www. zigbee.org/Standards/ZigBeeBuildingAutomation/Features.aspx. [Online; accessed 6-October-2012]. [23] ZigBee Alliance. Zigbee health care standard. http://www.zigbee. org/Standards/ZigBeeHealthCare/Features.aspx.aspx. [Online; accessed 6-October-2012]..

No results found