On Reliable Real Time Communication in Industrial Wireless Sensor Networks
56
0
0
Full text
(2)
(3)
(4) .
(5) . . .
(6)
(7)
(8) !" #
(9)
(10) $
(11) #
(12)
(13) #.
(14) %& #'(
(15) )*!+,+ -./ ./0 , +1 1 .+ 1
(16) 2
(17)
(18) !3 4! 5
(19).
(20) Abstract In the industrial automation, Industrial Wireless Sensor Networks (IWSNs) have been increasingly applied due to a great number of benefits such as convenient installation, flexible deployment and cost efficiency. Compared with conventional wireless systems, IWSNs have more stringent requirements on communication reliability and real time performance. However, IWSNs are frequently deployed in harsh industrial environments with electromagnetic disturbances, moving objects and non-line-of-sight (NLOS) communication. Because of the vulnerability of the wireless signal, IWSNs are under high risk of transmission failures, which may result in missing or delaying of process or control data. For industrial automation, missing the process or control deadline is intolerable, which may terminate industrial application and finally result in economic loss and safety problems. From hierarchy point of view, the high reliability and low latency can be achieved from different network layers. On MAC layer, existing protocols in IWSNs provide automatic repeat request (ARQ) to improve reliability at the cost of real time performance. An alternative method is to use Forward Error Correction (FEC) mechanism to provide more reliable transmissions and reduce the number of acknowledgement messages by recovering erroneous data. On network layer, routing protocol plays an important role in both communication reliability and latency. Traditional routing protocols in IWSNs are either hardly able to fulfill both of these requirements or overcomplicated. In this thesis, I initially explore the possibilities of introducing FEC into IWSN under the requirements of the existing standard on MAC layer. Then I propose compatible and flexible FEC schemes on MAC layer for IWSNs without violating the standard format. Routing protocols based on flooding are proved to increase the Packet Delivery Ratio (PDR) by transmission diversity. I propose reliable and robust routing protocols with respect to high reliability and real time performance for IWSNs. i.
(21)
(22) Sammanfattning I automationsindustrin har det blivit allt vanligare med tr˚adl¨osa sensorn¨at (Industrial Wireless Sensor Networks, IWSNs), p˚a grund av enkel installation, flexibel utplacering och kostnadseffekivitet. S˚adana industriella sensorn¨at har h˚ardare krav p˚a tillf¨orlitlighet och realtidsegenskaper a¨ n vad som a¨ r vanligt i konventionella tr˚adl¨osa n¨at. IWSNs anv¨ands ofta i milj¨oer med elektromangetiska st¨orningar, r¨orliga objekt och d¨ar kommunikationen sker utan fri sikt mellan enheterna. Eftersom den tr˚adl¨osa signalen a¨ r k¨anslig f¨or st¨orningar, a¨ r risken h¨og i IWSNs f¨or o¨ verf¨oringsfel, vilket kan resultera i att data blir f¨ordr¨ojt eller borttappat. Inom industriell automation a¨ r det inte tolererbart att missa deadlines, eftersom det kan inneb¨ara att processen stoppas vilket kan inneb¨ara ekonomiska f¨orluster och/eller s¨akerhetsproblem. H¨og tillf¨orlitlighet i kommunikationen kan a˚ stadkommas i olika skikt i protokollstacken. I MAC-skiktet tillhandah˚aller existerande protokoll f¨or IWSNs endast oms¨andningar (ARQ), vilket f¨orb¨attrar tillf¨orlitligheten p˚a bekostnad av s¨amre realtisegenskaper. En alternativ metod a¨ r att anv¨anda felr¨attande koder (FEC) i MAC-skiktet f¨or att f¨orb¨attra tillf¨orlitligheten och f¨or att minska behovet av oms¨andningar. I n¨atskiktet spelar routingprotokoll en viktig roll i s˚av¨al tillf¨orlitlighet som f¨ordr¨ojning. Traditionella routingprotokoll a¨ r antingen f¨or komplexa, eller uppfyller inte kraven p˚a tillf¨orlitlighet och l˚ag f¨ordr¨ojning. Denna avhandling fokuserar inledningsvis p˚a hur FEC kan anv¨andas i industriella sensorn¨at med standardiserat MAC-skikt som bivillkor. En anv¨andning av FEC som kan g¨oras utan a¨ ndring av MAC-skiktets standard f¨oresl˚as. D¨arefter visas hur routingprotokoll baserade p˚a flooding kan anv¨andas f¨or att o¨ ka andelen paket som kommer fram utan fel. Metoderna som f¨oresl˚as har h¨og tillf¨orlitlighet och goda realtidsegenskaper och l¨ampar sig d¨arf¨or f¨or IWSNs.. iii.
(23)
(24) To my family.
(25)
(26) Acknowledgement First I owe my deepest gratitude to my supervisors Mats Bj¨orkman(MDH), Maria Lind´en(MDH), Mikael Gidlund(ABB Corporate Research) and Johan ˚ Akerberg(ABB Corporate Research). It is my great fortune to have these encouraging researchers who made available their supports in a number of ways and guide me how to become a qualified and independent researcher during the past two years. I would also like to express my thanks to Elisabeth Uhlemann, Mikael Ekstr¨om, Martin Ekstr¨om and Marcus Bergblomma for all the helps they gave me from the first day of my PhD study. All the discussions with them also gave me a lot of valuable inspirations for my research. I would like to thank Thomas Nolte, Emma Nehrenheim, Paul Pettersson, Dag Nystr¨om, and Mikael Sj¨odin for helping and guiding me through my studies. I also have to thank Susanne Fronn˚a, and Carola Ryttersson to help me with all administrative stuffs. This work would not have been possible without the supports of staff from ABB Corporate Research. Especially I have to thank Peter L¨ofgren, Tomas Lennvall, Jonas Neander, Ewa Hansen, Krister Landern¨as, Zhibo Pang, Jimmy Kjellson, and Gargi Bag for supporting and encouraging me all the way from the first day. Many thanks go to all my friends both in the university and ABB Corporate Research for all the laughs and funs. Personally, I would like to thank my family for all the supports, loves and happiness that they have been giving me all the time. Kan Yu V¨aster˚as, May 18, 2012. vii.
(27)
(28) List of Publications Papers Included in the Licentiate Thesis1 ˚ Paper A K. Yu, M. Gidlund, J. Akerberg, M. Bj¨orkman, Reliable and Low Latency Transmission in Industrial Wireless Sensor Networks, the First International Workshop on Wireless Networked Control Systems (WNCS), Canada, September, 2011 ˚ Paper B K. Yu, F. Bara´c, M. Gidlund, J. Akerberg, M. Bj¨orkman, A Flexible Error Correction Scheme for IEEE 802.15.4-based Industrial Wireless Sensor Networks, the 21st IEEE International Symposium on Industrial Electronics (ISIE), Hangzhou, China, May, 2012 ˚ Paper C F. Bara´c, K. Yu, M. Gidlund, J. Akerberg, M. Bj¨orkman, Towards Reliable and Lightweight Communication in Industrial Wireless Sensor Networks, IEEE 10th International Conference on Industrial Informatics (INDIN), Beijing, China, July, 2012, Best Paper Finalist. ˚ Paper D K. Yu, M. Gidlund, J. Akerberg, M. Bj¨orkman, Reliable RSS-based Routing Protocol for Industrial Wireless Sensor Networks, the 38th Annual Conference of the IEEE Industrial Electronics Society(IECON), Canada, October , 2012. 1 The. included articles have been reformatted to comply with the licentiate layout. ix.
(29) x. Additional Papers, not Included in the Licentiate Thesis Conferences and Workshops ˚ • K. Yu, F. Bara´c, Mikael Gidlund, J. Akerberg, M. Bj¨orkman, Adaptive Forward Error Correction for Best Effort Wireless Sensor Networks, International Workshop on Wireless Sensor Actor and Actuator Networks 2012(WiSAAN), Canada, June, 2012.
(30) Contents I. Thesis. 1. 1. Introduction 1.1 Research Problem . . . . . . . . . . . . . . . . . . . . . . . . 1.2 Thesis Contributions . . . . . . . . . . . . . . . . . . . . . . 1.3 Thesis Outline . . . . . . . . . . . . . . . . . . . . . . . . . .. 2. Background 2.1 Industrial Automation . . . . . . . . . . . . . . . . . . . . 2.2 Industrial Wireless Sensor Networks . . . . . . . . . . . . 2.2.1 IWSN Standards . . . . . . . . . . . . . . . . . . 2.2.2 Industrial Wireless Channel Conditions . . . . . . 2.3 Reliable and Real Time Communication in Industrial WSN 2.3.1 Forward Error Correction . . . . . . . . . . . . . 2.3.2 Routing Protocols in WSNs . . . . . . . . . . . .. 3 5 6 6. . . . . . . .. 9 9 11 12 17 19 20 21. 3. Related Work 3.1 FEC in WSNs . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2 Routing protocols in WSNs . . . . . . . . . . . . . . . . . . .. 23 23 24. 4. Included Papers and Their Contribution 4.1 Paper A . . . . . . . . . . . . . . . . 4.2 Paper B . . . . . . . . . . . . . . . . 4.3 Paper C . . . . . . . . . . . . . . . . 4.4 Paper D . . . . . . . . . . . . . . . .. . . . .. 27 28 29 30 31. Conclusions and Future work 5.1 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.2 Future Work . . . . . . . . . . . . . . . . . . . . . . . . . . .. 33 33 35. 5. xi. . . . .. . . . .. . . . .. . . . .. . . . .. . . . .. . . . .. . . . .. . . . .. . . . .. . . . .. . . . . . . .. . . . ..
(31) xii. Contents. Bibliography. 36. II. 43. 6. 7. Included Papers Paper A: Reliable and Low Latency Transmission in Industrial Wireless Sensor Networks 6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.2 IEEE 802.15.4 and Industrial Wireless Channel Conditions . . 6.3 Applying Forward Error Correction Technology in IWSN . . . 6.3.1 Applying FEC Codes at the MAC Layer . . . . . . . . 6.3.2 Timing Requirement . . . . . . . . . . . . . . . . . . 6.4 Performance Evaluation . . . . . . . . . . . . . . . . . . . . . 6.4.1 Selection of FEC codes for evaluation . . . . . . . . . 6.4.2 Experimental Setup . . . . . . . . . . . . . . . . . . . 6.4.3 Evaluation Results . . . . . . . . . . . . . . . . . . . 6.5 Analysis and Discussion . . . . . . . . . . . . . . . . . . . . 6.6 Conclusions and Future Work . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 45 47 48 49 50 53 54 55 55 56 57 59 59. Paper B: A Flexible Error Correction Scheme for IEEE 802.15.4-based Industrial Wireless Sensor Networks 7.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.2 Preliminaries . . . . . . . . . . . . . . . . . . . . . . . . . . 7.2.1 IEEE 802.15.4 standard . . . . . . . . . . . . . . . . 7.2.2 Forward Error Correction . . . . . . . . . . . . . . . 7.3 Proposed Enhanced FEC Scheme for IEEE 802.15.4 . . . . . 7.3.1 Applying FEC Codes at the MAC layer . . . . . . . . 7.3.2 Error Performance Analysis . . . . . . . . . . . . . . 7.4 Experimental Setup . . . . . . . . . . . . . . . . . . . . . . . 7.4.1 Simulation Environment and Measurables of Interest . 7.4.2 Packet Reception Model . . . . . . . . . . . . . . . . 7.4.3 Simulation Scenarios . . . . . . . . . . . . . . . . . . 7.5 Evaluation Results and Discussion . . . . . . . . . . . . . . . 7.5.1 Scenario I . . . . . . . . . . . . . . . . . . . . . . . . 7.5.2 Scenario II . . . . . . . . . . . . . . . . . . . . . . . 7.6 Conclusions and Future Work . . . . . . . . . . . . . . . . . .. 61 63 64 64 65 65 65 67 71 71 72 73 73 73 74 76.
(32) Contents. 8. 9. xiii. References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 76. Paper C: Towards Reliable and Lightweight Communication in Industrial Wireless Sensor Networks 8.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.2 Proposed FEC-coded Lightweight Routing Scheme . . . . . . 8.2.1 The lightweight and reliable transmission scheme . . . 8.2.2 Theoretical Performance Analysis . . . . . . . . . . . 8.3 Simulation Setup and Scenarios of interest . . . . . . . . . . . 8.3.1 Channel model and the choice of FEC coding . . . . . 8.3.2 Simulation Scenarios . . . . . . . . . . . . . . . . . . 8.4 Evaluation Results . . . . . . . . . . . . . . . . . . . . . . . 8.4.1 Scenario I . . . . . . . . . . . . . . . . . . . . . . . . 8.4.2 Scenario II . . . . . . . . . . . . . . . . . . . . . . . 8.5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 79 81 82 83 87 90 90 91 91 92 94 97 97. Paper D: Reliable RSS-based Routing Protocol for Industrial Wireless Sensor Networks 101 9.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . 103 9.2 Previous Work . . . . . . . . . . . . . . . . . . . . . . . . . . 104 9.3 Proposed Reliable RSS-based Flooding Scheme . . . . . . . . 105 9.3.1 RSS-based Principle . . . . . . . . . . . . . . . . . . 106 9.3.2 RSS-based Weighting Method . . . . . . . . . . . . . 106 9.3.3 Forwarding Criteria . . . . . . . . . . . . . . . . . . . 108 9.3.4 Example: RSS-based flooding scheme . . . . . . . . . 110 9.4 Experimental Setup . . . . . . . . . . . . . . . . . . . . . . . 113 9.4.1 Simulation environment and measurables of interest . 113 9.4.2 Simulation Scenarios . . . . . . . . . . . . . . . . . . 114 9.5 Evaluation Results and Analysis . . . . . . . . . . . . . . . . 115 9.5.1 Scenario I . . . . . . . . . . . . . . . . . . . . . . . . 115 9.5.2 Scenario II . . . . . . . . . . . . . . . . . . . . . . . 117 9.5.3 Scenario III . . . . . . . . . . . . . . . . . . . . . . . 117 9.6 Conclusions and Future work . . . . . . . . . . . . . . . . . . 119 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119.
(33)
(34) I Thesis. 1.
(35)
(36) Chapter 1. Introduction Traditional wired field devices have dominated industrial automation for decades. Nowadays, as a consequence of the evolution of wireless communication along with microelectronics technology, wireless sensor networks (WSNs) have been exhibiting their attractive advantages over traditional wired counterpart for industrial automation. Due to the avoidance of cabling, installation cost is significantly reduced. An estimation is made by Emerson Process Management that up to 90% of cost saving can be achieved by applying Industrial Wireless Sensor Networks (IWSNs) in industrial automation [1]. Moreover, without constrains of wired connection, the deployment of field devices become much more flexible. Sensor nodes can be easily installed where cables hardly reach. Mobility of field devices also becomes available, such as rotating machines. Therefore, both the efficiency and flexibility of field device installation are greatly improved. Now IWSN technologies can serve a number of purposes, such as monitoring and actuation. Although the importance of WSNs for industrial automation is increasing because of benefits mentioned above, there exist a great number of challenges of fulfilling strict requirements from industrial applications. Different from conventional wireless community networks, IWSNs have much more stringent requirements on communication reliability and real time performance [2] [3]. However, IWSNs are foreseen to be deployed in harsh environments where the radio signal will be affected by random noise and channel fading which causes packet errors. Therefore, the vulnerability of wireless signal leads to high risk of transmission failure and then results in missing or delaying of process or control data. However, for industrial automation, missing the process or control 3.
(37) 4. Chapter 1. Introduction. deadline is intolerable, which may terminate industrial application and finally result in economic loss, or even more serious safety problems, such as loss of human life and damage to environment. On MAC layer, typically, a common method to improve the link reliability is to employ automatic repeat request (ARQ) procedure (a retransmission procedure) and another approach is to apply forward error correction (FEC) strategies [4], reducing the bit error rate and consequently the number of retransmissions. An inherent drawback of ARQ is the increased latency of packet delivery, due to a number of packet retransmissions. This is in direct conflict with time-critical requirements of industrial applications. The empirical results from [5] revealed that excessive retransmissions may lead to more serious consequences, such as network congestion and complete halt of the observed industrial process. An alternative approach is the concept of FEC, where by applying channel coding on transmitted data, a number of corrupted bits can be recovered. Consequently, data retransmissions are avoided and latency and Bit Error Rate (BER) are reduced. The drawback of FEC is the reduction of transmission efficiency due to introduction of redundancy. However, FEC is still a more preferable approach because it may improve both reliability and latency. On network layer, routing protocols play a vital role in reliable and real time communications. For a multihop network, appropriate paths should be chosen by routing protocols for sensor nodes which are several hops away from the sink. In IWSNs, although sensor nodes may not move frequently, the noisy industrial wireless channel makes the link quality unstable. Thus parts of the established routes may get disconnected due to the interferences. Typically, stringent deadlines are required by industrial applications. Therefore, a reliable and robust routing protocol in IWSNs should not only be able to maintain and reconstruct available paths in a timely manner, but also be responsible for guaranteeing process data arriving at the destinations before their deadlines. To maintain the connectivity of all routes, for conventional routing protocols, control message overhead is inevitable. A robust and lightweight routing protocol shall not generate excessive control message overhead, otherwise it makes the usage of bandwidth and energy consumption much less efficient. Moreover, a path failure or joining and leaving of a sensor node usually triggers a path recalculation process, which may prolong the transmission latency. A failure of delivering path maintenance message will even lead to incorrect route recalculations, which may cause industrial application halt, resulting in serious economic and safety problems. Currently, there exist a great number of routing protocols in WSNs. However, very few of them fulfill the industrial re-.
(38) 1.1 Research Problem. 5. quirements of both high reliability and low latency. Therefore, there is a great demand for a reliable and robust routing protocol in IWSNs.. 1.1. Research Problem. As we emphasized above, compared with conventional WSN, IWSNs for process and control demand more attention on communication reliability and real time performance. Several research works [5] [6] [7] have shown that industrial channel conditions for IWSNs are harsh. It is a great challenge to guarantee high reliable and low latency communications for IWSNs in such environments. Currently, the problems are: 1. In current IEEE 802.15.4-based IWSN standards, ARQ mechanism is applied on MAC layer. However, the reliability is improved at the cost of real time performance by using ARQ mechanism. Transmission failure will trigger retransmissions due to the harsh wireless channel. Excessive retransmissions prolong communication latency to the industrial application, but also may easily exhaust limited bandwidth resources, which even can result in network congestion. A serious congestion of the network may stop industrial applications, finally lead to economic loss or serious safety problems. Therefore, a more advanced solution on MAC layer being able to improve both reliability and real time performance is urgently required. 2. On network layer, existing routing protocols in WSNs are either designed for energy constrained system, or involve excessive control message overhead for maintaining routing information, or are overcomplicated for sensor nodes. Although energy consumption of routing protocols is also an important issue for IWSNs, reliability and real time performance are usually more important than energy consumption for a great number of industrial applications. Overwhelming control message overhead not only makes transmissions inefficient, but also may make network resources more constrained. Overcomplicated routing protocols will lead to a high workload for embedded field devices. Currently, communication reliability and real time performance are not fully considered among previous routing protocols in WSNs. Therefore, a lightweight, reliable and robust routing protocol in IWSNs is extremely important..
(39) 6. Chapter 1. Introduction. 1.2. Thesis Contributions. The contributions presented in this thesis are: 1. We implement and benchmark several widely used FEC codes in a typical embedded device with a core ARM Cortex-M3 which can be often found in industrial devices within the automation domain. We show that it is possible to introduce FEC into IWSN protocols both under the timing requirement of the existing standard on MAC layer and the resource constraints of embedded devices. 2. A IEEE 802.15.4-compliant FEC-based scheme on MAC layer is proposed. It is shown that our proposed solution can be easily retrofitted into the standard without violating the standard format or the need for any kind of interaction with chip manufacturers. The simulation results show a significant improvement in terms of reliability of data delivery, compared to the existing uncoded approach. When compared with the traditional FEC method, a significantly lower computational load of decoding is achieved. 3. It is shown that the combination of FEC and location-based flooding scheme can achieve a substantial gain in reliability and reduced latency, compared to the uncoded transmissions, as well as common Wireless Sensor Network routing protocols. 4. A reliable and flexible Received Signal Strength-based routing scheme is proposed. The simulation results verify that the proposed routing protocol outperforms conventional routing protocols in both reliability and real time performance. A seamless transition in the event of topology change can be achieved by our solution, whereas conventional routing protocols fail to recover the network in a short time.. 1.3. Thesis Outline. The thesis is organized in two parts. In Part I the research problems are described as well as an overview of the research areas, related work. In the end of Part I conclusions and future work are presented. Part II includes four peerreviewed scientific papers that are published and presented in international conferences. The rest of Part I is organized as follows..
(40) 1.3 Thesis Outline. 7. Chapter 2: This chapter is an introduction in backgound knowledge on industrial automation, IWSNs and reliabe and real time communication in IWSNs. Chapter 3: This chapter presents the related work. Chapter 4: This chapter presents the included papers and the contribution of the thesis. Chapter 5: This chapter presents the conclusions and future work..
(41)
(42) Chapter 2. Background In this chapter, an introduction of the technologies and information used in this thesis is given. Firstly, we describe industrial automation in brief. We also give an overview of IWSNs, existing standards and current status in a more detailed manner. Finally we introduce two technologies used for improving reliability and real time performance in IWSNs. Keeping all this background knowledge in mind is necessary to understand the importance of our contributions for applying IWSNs in industrial automation.. 2.1. Industrial Automation. There are many types of automation, broadly defined. Typically, automation consists of process automation, substation automation, factory automation, building automation and home automation [8]. Since requirements from these categories differ hugely with respect to error tolerance, allowed jitter, availability, safety and security, this thesis mainly focus on automation in the industrial domain, such as process automation. Industrial automation is also a vast discipline that generally encompasses hardware, such as microcontroller, field bus, sensors and actuators, and software, such as control software and communication software. It is a technology that combines control systems and information systems to build up automated manufacturing applications without direct human interference toward a series of goals, such as low costs, increased production, improved quality and maximum flexibility. 9.
(43) 10. Chapter 2. Background. Figure 2.1: The elements of a process control loop It is important to recognize that communication technologies in industrial automation have different requirements from those in other domains. A reliable, robust and low latency communication technology is critical for industrial applications. Taking an example from process control, a block diagram of a process control loop is shown in Figure 2.1. Communication systems in process control take the responsibility of transmitting two important signals, which are from measurement output to the controller and from the controller to the actuator. Failures of signal transmission will directly affect the stability of the system. For instance, if measurement data fails to be delivered to the controller, there may lead to an unexpected behavior at the valve, which may result in economic losses or safety problems. Therefore, to guarantee reliable and robust signal delivery is vital for industrial automation. In the industrial automation domain, mediums for industrial communication are traditional wired, such as twisted pair cables, coaxial cables, fibers and even power lines. The most satisfying advantage of wired communication is that high reliability and real time performance can be effectively guaranteed. Since the size of automation networks is typically not huge, once devices are correctly connected, signal transmissions via wires can hardly be interfered, nor will network congestion happen. Therefore, wired industrial communication systems in industrial automation have been applied successfully for decades. As wireless technology develops, radio became another important option for industrial communications. Flexibility and low cost are attractive benefits for wireless communications in industrial automation. More details on wireless automation networks are discussed in next section..
(44) 2.2 Industrial Wireless Sensor Networks. 11. Figure 2.2: Example of a IWSN structure. 2.2. Industrial Wireless Sensor Networks. As we described previously, given the great benefits offered by wireless technologies compared to wired methods, such as convenient deployment and cost efficiency, more and more wireless systems in industrial automation are taking the place of traditional wired ones. Although wireless technologies have been previously used in industrial automation, such as industrial wireless LANs, most of them are on the higher levels in the communication architecture. Nowadays, Wireless Sensor Networks (WSNs), as a rapid developed technology, have been applied on the fieldbus level. A typical WSN is built of spatially distributed sensor nodes and gateways. Traditional WSNs are characterized as low power consumption, low data rate and self-organizing capability and typically used for monitoring physical or environmental conditions. Instead of reliability and communication latency, power consumption is more critical for traditional WSNs, since battery powered sensor nodes often are deployed in places off the beaten track, such as volcanoes, forests, deserts, etc. Frequently changing batteries for sensor nodes is a challenge. IWSNs share the same basic concept of traditional WSNs. A general structure of IWSNs is shown in Figure 2.2. However, the requirement for IWSNs are totally different from traditional WSNs. Firstly, centralized management is necessary, instead of self-organizing. Operators in the central control room must have the knowledge of the status of all sensors nodes and also have the full.
(45) 12. Chapter 2. Background. control of the whole network. Low power consumption is less important for IWSNs, compared with traditional WSNs, in many cases, since sensor nodes in IWSNs can easily find a power supply instead of using battery. Besides these, the major differences of IWSNs are stringent requirements to communication reliability and real time performance. As we emphasized above, transmission failure and missing the process or control deadline may lead to serious economic loss and safety problems. Thus unreliable communication is absolutely intolerable for industrial automation. Protocols designed for IWSNs shall be able to provide reliable and real time data transmission in IWSNs. Currently, there exist several global standards for IWSNs. To know the advantages and disadvantages of these standards is helpful to come up with the final appropriate solution. Since wireless environment directly affect the link quality of IWSNs, it is also necessary to have the knowledge of the practical industrial wireless channel conditions. Furthermore, applying FEC on MAC layer and reliable routing protocols on network layer are effective approaches to fulfill the requirements. All these subjects are described in the following subsections.. 2.2.1. IWSN Standards. A variety of standards as wireless solutions are going to be or have been applied according to the application scenarios. For long distance communication, longhaul wireless link has been already applied in industrial automation. Wireless LANs and optical wave systems are typically used for middle range industrial wireless communication. On fieldbus level, wireless technologies for short range communications are more concerned. Therefore, this thesis only focus on the standards approved for fieldbus level. The standards for short range communications in industrial automation can be divided into IEEE 802.15.4-based and not IEEE 802.15.4-based standards. Bluetooth [9] is a typical already applied standard in industrial automation not based on IEEE 802.15.4. Bluetooth has the advantages of high throughput and high security. However, Bluetooth is often used for peer to peer communication, having no success in a large scale of network with many sensor nodes, as well ass the WLAN IEEE 802.11 standards. Moreover, it is more complicated than other IEEE 802.15.4-based standards and consumes more energy. Therefore, most of standards applied in industrial automation are IEEE 802.15.4-based standards. There are four main IEEE 802.15.4-based standards used or going to be used in industrial automation, namely Zigbee [10], WirelessHart [11], ISA100.11a [12] and WIA-PA [13]. These four standards, as well as IEEE 802.15.4 are introduced as follow:.
(46) 2.2 Industrial Wireless Sensor Networks. 13. Figure 2.3: Successful data transmission with an acknowledgment. IEEE 802.15.4 The IEEE 802.15.4 standard specifies [14] both physical(PHY) layer and MAC layer. The physical layer is based on direct sequence spread spectrum (DSSS) to comply with the sharing rules of each band. It can operate in two different bands: in 868/915 MHz bands with the data rate 20/40 kbps and in 2.4 GHz with the data rate 250 kbps. The MAC layer defines two different modes of operations:1) the unbeaconed mode that uses an unslotted CSMA/CA; 2) the beaconed mode that used a slotted CSMA/CA with a superframe structure imposed by network coordinators. In order to improve communication reliability, the 802.15.4 standard provides acknowledgment and retransmission mechanism. The use of acknowledgments and retransmissions is described in Figure 2.3. The message sequence chart in Figure 2.3 shows the scenario for transmitting a single data frame from a sender to a receiver with an acknowledgment. A device that sends a data frame with its Acknowledgment Request subfield set to one shall wait for at most macAckWaitDuration symbols for the corresponding acknowledgment frame to be received. If a corresponding acknowledgment frame is received within macAckWaitDuration symbols, the transmission is considered to be successful, and no further action regarding retransmission will be taken by the sender. Otherwise, the sender shall conclude that the single transmission attempt has failed. If a single transmission attempt has failed, the sender shall repeat the process of transmitting the data and wait for the acknowledgment, up to a maximum of macMaxFrameRetries times. If an acknowledgment is still not received after macMaxFrameRetries retransmissions, the MAC sublayer shall assume the transmission has failed and notify the next higher layer of the failure. According to the 802.15.4 standard, the macAck-.
(47) 14. Chapter 2. Background. Table 2.1: Definitions [14] Parameter. Definition. aUnitBackoffPeriod. The number of symbols forming the basic time period used by the CSMA-CA. RX-to-TX or TX-to-RX maximum turnaround time The duration of the synchronization header for the current PHY. The number of symbols per octet for the current PHY. The maximum number of symbols to wait for an acknowledgment. The maximum number of allowed retransmissions after a transmission failure.. aTurnaroundTime phySHRDuration phySymbolsPerOctet macAckWaitDuration macMaxFrameRetries. Value (symbol) 20. 12 10 2 Equation (2.1) 0-7 (3 as default). WaitDuration can be calculated as:. macAckWaitDuration = +. aUnitBackoffPeriod + aTurnaroundTime phySHRDuration + 6 × phySymbolsPerOctet (2.1). where each parameter is defined in Table 2.1. Hence totally, macAckWaitDuration = 20 + 12 + 10 + 6 × 2 = 54 symbols time. The data rate in 2.4 MHz is 250 kbps, which equals 62500 symbols per second. Then, macAckWaitDuration = 54/62.5 = 0.864 ms. Thus, if a sender fails to get the corresponding ACK within 0.864 ms, a retransmission is triggered. In the worst case, if the transmission still fails after 7 retries, a final error notification should be announced to the upper layer. This ARQ mechanism is also used in all four IEEE 802.15.4-based IWSN standards. The timing requirements from four standards are almost the same as IEEE 802.15.4..
(48) 2.2 Industrial Wireless Sensor Networks. 15. IEEE 802.15.4-based Standards As we mentioned, Zigbee, WirelessHart, ISA100.11a and WIA-PA are four main IEEE 802.15.4-based standards have been or going to be used in industrial automation. • Zigbee is a specification that serves for short range communication and targeted at radio-frequency applications requiring low data rate, low power consumption, low costs and secure transmissions. The ZigBee specification was first ratified in 2004 and typically used for home automation, monitoring and smart metering. • WirelessHART, as the first approved open standard for IWSNs, is based on Highway Addressable Remote Transducer Protocol (HART) and formulated by HART Communication Foundation (HCF). Thus, it is backward compatible with core HART technology. It is primarily designed as a reliable, secure, self-organizing, self-healing and time synchronized mesh networking technology for industrial process automation and control systems. • ISA100.11a is proposed as a standard by the International Society of Automation (ISA) to enable a single, integrated wireless infrastructure platform for plants and delivers a family of standards defining wireless systems for industrial automation and control applications. • WIA-PA (Wireless Networks for Industrial Automation/Process Automation) is the Chinese standard of industrial wireless communication architecture and specification for industrial automation. It was developed by Chinese Industrial Wireless Alliance (CIWA) under the urgent requirements of process automation. In 2008, WIA-PA also became a Public Available Specification (PAS) of International Electrotechnical Commission (IEC) via voting. Authors in [15] have revealed that Zigbee is not appropriate for industrial automation, due to no frequency diversity, no path diversity and the lack of robustness. Therefore, Zigbee had very limited success in the industrial automation marketplace. A comparison of main features from other three standards is illustrated in Table 2.2. WirelessHART is mainly designed for supporting existing HART technology (e.g., HART commands, configuration tools, etc). Although it has fewer functions, the implementation of wirelessHART is much simpler than other.
(49) 16. Chapter 2. Background. Table 2.2: IWSN Standard Comparison Feature Set. WirelessHART PHY layer. ISA100.11a. WIA-PA. PHY layer. TDMA, CSMA. TDMA, CSMA. Frequency hopping Latency determinism Timeslot duration Superframe structure Error Correction Management. channel ping Yes. slow,fast and hybrid hopping Yes. PHY and MAC layer CSMA, TDMA, FDMA AFH, AFS, and TH Yes. 10ms. flexible. 10ms. No. No. No centralized. No centralized, distributed. Manager assignment Routing. fixed. dynamic. Based on IEEE802.15.4 No hybrid centralized and distributed fixed. Source, graph, hybrid, superframe medium symmetric. Source, graph, hybrid. static redundant. high symmetric, asymmetric object-oriented (UFO) HART, Profibus, Modbus, FF, ect complicated. high symmetric, asymmetric object-oriented (UAO) HART, Profibus, Modbus, FF, ect medium. IEEE802.15.4 compatibility Radio Channel. Security Key Application definition Support for other protocols. commands. Implementation. simple. HART. hop-.
(50) 2.2 Industrial Wireless Sensor Networks. 17. two IWSN standards and it has already been deployed in a great number of industrial applications to prove its success. ISA100.11a aims for high security and supporting more legacy protocols than wirelessHART at the cost of high complexity. WIA-PA is a newly merged standard with the similar purposes of ISA100.11a, so it is currently less mature than the other two standards. Since all three standards are based on IEEE 802.15.4, only ARQ mechanism is applied on MAC layer to improve reliability. As we emphasized previously, high communication latency may be caused by retransmissions. It is notable that from the table, no error correction mechanism is used in any of three standards. Moreover, although three standards mention routing technologies those should be used, no specific routing protocol is determined by those standards. To our knowledge, existing routing protocols used for WSNs are not reliable, robust and lightweight enough, or overcomplicated.. 2.2.2. Industrial Wireless Channel Conditions. It is common knowledge that the reliability of wireless communication is much lower than wired counterpart because of many factors, such as attenuation, path loss, multi-path fading, shadowing, etc. Due to metallic surfaces, extreme temperature and high vibrations in industrial environments, industrial wireless channel condition becomes harsher and more dynamic [16]. Communication between two nodes is typically non-line-of-sight (NLOS). Moreover, most wireless sensor networks for industrial purpose run at the frequency 2.4 GHz. Thus, wireless signals from IEEE 802.15.4-based IWSNs may also suffer from interferences from other industrial wireless systems, such as wireless LANs and Bluetooth working in the same Industrial, Scientific and Medical (ISM) band. In [17], Sikora and Groza pointed out the impacts from other wireless systems using ISM band on IEEE 802.15.4 may result in a timeout in the PHY layer or an enlarged packet error rate. Authors from [16] also pointed out that co-existing communication systems are one of the major sources of disturbances in wireless industrial applications. Authors from [5] characterized the industrial wireless channel condition according to the measurements in a power plant. The photography of the power plant where the measurement took place is shown in Figure 2.4. In one of their measurement scenarios, one sensor node continuously transmitted data to another node with the distance of approximately 10 meters NLOS. Then the value of received signal strength indicator (RSSI) was measured on the receiver side. The result is shown in Figure 2.5. The measurement result shows that around 90% of the RSSI concentrates between -65 dBm and -55 dBm, but still more than 10% of the signal.
(51) 18. Chapter 2. Background. strength is less than -68 dBm. This may be caused by deep fading and shadowing from the harsh wireless channel. In another measurement scenario, when two nodes were apart from 30 meters away NLOS, the RSSI values drop to -71± 3.2 dBm, and the minimum RSSI reaches -79dBm. The measurement result from the second scenario is shown in Figure 2.6. Another measurement from an industrial factory [6] also shows that the fluctuations of received signal strength are about 25 dBm.. Figure 2.4: Photography from the power plant. Figure 2.5: The measured RSSI of messages transmitted in scenario 1 All these measurement results reveal two facts. Firstly, in IWSNs, received.
(52) 2.3 Reliable and Real Time Communication in Industrial WSN. 19. Figure 2.6: The measured RSSI of messages transmitted in scenario 2 signal strength may become extremely weak due to deep fading or shadowing from the harsh industrial wireless channel. If the receiver is not sensitive enough to pick up the weak signal, errors may be introduced into the output data. Secondly, it is also notable that, in industrial environments, although the RSSI varies for different packets, most of the values are still scattered in a limited range. Therefore, RSSI values still objectively indicates the link quality between two sensor nodes.. 2.3. Reliable and Real Time Communication in Industrial WSN. In IWSNs, reliability and real time performance are bounded with each other in a sense. For peer to peer communication, high reliability indicates the low probability of transmission failure. Since ARQ is used in IWSNs, the avoidance of retransmissions leads to low communication latency. To increase reliability, there exist a number of approaches from physical layer to application layer. The most straightforward approach is to increase transmission power, but the effectiveness is below expectations and it will also result in a very short battery life. For instance, another effective way is to use Multiple-Input MultipleOut-put (MIMO) technology. However, this method will hugely increase the complexity of the radio design, which turns out to much higher cost. Therefore,.
(53) 20. Chapter 2. Background. it is necessary to apply an appropriate and convenient method to achieve the goal without increasing a great deal of complexity or violating existing standards. Introducing FEC on MAC layer and applying reliable and robust routing protocols on network layer are considered as reasonable choices.. 2.3.1. Forward Error Correction. The basic idea of FEC coding is to introduce redundancy into the transmitted data, in order to solve the problem of potential packet corruption. According to different operation modes, FEC are divided into two basic types: block codes and convolutional codes. For a block code (n, k), it means that the original message length is k symbols. A FEC encoder adds n − k redundant symbols to form an n-symbol codeword i.e. block. Each block of information is treated independently from others. The n − k redundant symbols provide the codeword with the capability to overcome a certain length of errors caused by noisy channels. For instance, a Reed-Solomon (n,k) code is able to correct n−k symbol errors. 2 A convolutional code (n, k, m) behaves differently from a block code. The FEC encoder for a convolutional code also encodes k-bit original data to n-bit codeword with n − k-bit redundant. The difference is that to convolutionally encode original data, m memory registers are required, where m is the constraint length of the convolutional code. Each k-bit symbol is not treated independently and encoded with other m − 1 symbols together to a codeword. For both block and convolutional code, the code rate, R = k/n(R < 1), indicates the transmission efficiency of the FEC. Higher code rate means lighter transmission overhead and less extra transmission power. At the receiver side, the decoder can utilize the redundant symbols to detect and correct error bits. Hard-decision and soft-decision are two different methods to decode FEC codes. Hard-decision decoding methods can be used only when the signal strength, after demodulation, is quantized to two levels, 0 and 1. Otherwise only the soft-decision decoding method is available. Compared to the hard-decision method, the soft-decision decoding scheme obtains a 3 dB extra coding gain, but with an additional computational complexity. In our work, we focus on block codes with hard-decision decoding scheme due to their reasonable complexity and high flexibility. Since FEC coding allows the receiver to detect and correct a limited number of errors caused by the noisy channels, it helps to reduce the number of retransmission, but at the cost of a higher forward channel bandwidth. FEC coding is also useful where retransmissions are impossible, such as broadcast-.
(54) 2.3 Reliable and Real Time Communication in Industrial WSN. 21. ing. If we integrate FEC into existing IWSN standards without turning off ARQ mechanism, the timing requirement described above should be fulfilled. Taking IEEE 802.15.4 as an example, the decoding time shall be less than macAckWaitDuration = 0.864 ms; otherwise, an unexpected retransmission is triggered.. 2.3.2. Routing Protocols in WSNs. As we mentioned above, routing protocols play an important role in communication reliability and latency for IWSNs. Based on different classification methods, routing protocols in WSNs can be sorted into categories. Generally, routing protocols for WSNs fall into four primary categories. Those are flooding-based, dynamic routing table-based, cluster-based, geographical routing protocols. Flooding-based routing protocols are the most straightforward routing schemes of all. Each source node broadcasts the packets to all of or a part of its neighbors. All intermediate nodes also forward the packets by broadcasting according to some certain forwarding criteria until the packets arrive at the destination. Major advantages of flooding-based routing protocols are fast network recovery mechanism from topology changes and forwarding packets without requiring any control overhead or forwarding table, at the cost of additional bandwidth resources. Moreover, traditional flooding approaches suffer from broadcast storm and energy inefficiency problems, which were thoroughly studied in [18] and [19]. Many routing protocols also use flooding mechanism, but not for data transmission, only for route maintenance and discovery, such as Open shortest path first (OSPF) [20] and optimized link state routing (OLSR) [21]. These routing protocols are not sorted into this category. Dynamic routing table-based protocols are evolved from wired network routing protocols. It is a huge category including a great number of routing protocols. They can be further divided into many subclasses, such as linkstate-based, distance-vector-based and so on. Packet transmissions in this category are often based on unicast. To forward data packets, each sensor node should maintain a forwarding table containing the information of the next hop. Routing protocols in this category can make the usage of network bandwidth more efficient, but additional control overhead to maintain forwarding tables are inevitable. A major drawback of traditional dynamic routing table-based protocols is that, if the network topology changes due to unexpected disconnected nodes or broken links, routing tables in all nodes should be updated timely and new routes should be discovered as soon as possible; otherwise,.
(55) 22. Chapter 2. Background. parts of the communications in the network may halt. To discover the changes of the network topology in time indicates more frequent control overheads, but excessive control overhead may lead to exhaustion of bandwidth resources. Therefore, network recovery time and efficiency of network resource usage can hardly be solved in the same time by traditional dynamic routing table-based protocols. Cluster-based routing protocols are members from a family of routing protocols that use network clustering. Some research works [22] sort in as hierarchical routing protocols. All sensor nodes in the network are partitioned in clusters according to a certain negotiation method. Each cluster has a cluster head that have some responsibilities of collecting and aggregating packets from respective clusters and transmitting the aggregated data to the base station. Typically, the nodes with the higher residual energy are usually chosen as the cluster head. Since everything is processed in a distributed manner, clusterbased routing protocols are often used in ad hoc networks and can achieve better scalability. Cluster-based routing protocols often serve for a network with huge size, but due to the refresh rate of process control applications the network size of IWSNs can hardly be huge [23]. Furthermore, cluster-based routing protocols are hard to support time-critical applications due to the continuously cluster head evaluation procedure. Therefore, cluster-based routing protocols are considered not suitable for IWSNs. The most famous cluster-based routing protocols are low-energy adaptive clustering hierarchy (LEACH) [24] and Threshold sensitive energy e?cient sensor network protocol (TEEN) [25]. Different from previous routing protocols, for geographical routing protocols, each sensor node takes advantage of geographical location to transmit and forward data packets. Geographical information in each node is initially set manually or by special positioning hardware, such as a global positioning system(GPS). Although a seamless transition in the event of topology change and avoidance of excessive control overhead can be achieved, the geographical distance between two nodes does not reflect the practical link quality. The information of geographic position is far from sufficiency to define the communication cost [26]. Therefore, geographical routing protocols are more than useless for indoor applications or underground mining applications..
(56) Chapter 3. Related Work This chapter presents the previous research works related to the thesis. Since we described above that introducing FEC on MAC layer and applying reliable and robust routing protocols on network layer are two appropriate approaches to improve reliability and real time performance in IWSNs, this chapter is organized as follows. First, we discuss the related research efforts in terms of using FEC for WSNs. Second, we present the existing routing protocols for WSNs to highlight current problems for industrial automation.. 3.1. FEC in WSNs. Significant research efforts were undertaken on investigating the performance of FEC coding in WSNs. A vast majority of FEC-related work for WSNs focuses on energy efficiency of the FEC coding methods or FEC-related schemes. The authors in [27] prove that using FEC in WSNs can effectively reduce the required transmission power to achieve similar communication performance. The authors in [28] compare their optimal FEC scheme and infinite ARQ scheme in WSNs and show that optimal FEC scheme outperforms infinite ARQ scheme in energy consumption. Low-density parity-check (LDPC) codes are shown to achieve better energy efficiency than BCH codes and convolutional codes in [29]. Authors in [30] identify that binary-BCH codes with applicationspecific integrated circuit (ASIC) implementation are most suitable method for WSNs in energy consumption. In [31], relationships are set between the energy efficiency of FEC scheme, communication distance and packet size. Cross23.
(57) 24. Chapter 3. Related Work. layer approaches yielding energy-efficient designs of FEC schemes are proposed in [32] and [33]. Although energy consumption is a major concern for WSNs, the additional energy cost by FEC scheme is next to insignificant, compared with other requirements such as reliability and real time performance. Adaptive FEC scheme is another frequently encountered research topic in the literature. The authors of [34] adaptively switch applied FEC code within several predefined BCH codes based on the number of received packets. However, dedicated feedback messages are required, which may limit its usage in WSNs. A Hybrid-ARQ-Adaptive-FEC scheme in [35] is based on collecting Signal-to-Noise Ratio (SNR) information. When WSNs are deployed in environments where RSSI varies frequently and intensively, the observation of short-term variations of SNR may fail in providing the accurate information about the overall trend of channel condition changes. Adaptive FEC methods aiming for low power consumption are proposed in both [36] and [37], but these schemes are based on multipath transmissions and highly depend on the network topology. Thus it may cause significant delays in case of transmission errors. A combination of FEC and routing schemes is another trend in FEC-related works. A lightweight XOR-based FEC algorithm combined with a fault tolerant routing scheme in [38] fragmentizes and sends the original data over multiple paths. XOR-based checksums for fragments are sent in separate packets. Such an approach would be highly inefficient for small IWSN packets due to header and traffic overhead. Furthermore, the checksum packet loss would stultify the entire transmission, since the integrity of data fragments could not be checked in the sink. Authors in [39] show the efficient combinations of retransmission, FEC and alternative routes in routing scheme can effectively improve the reliability. However, they put multiple independent code words into a packet without analyzing transmission latency. Therefore, the real time performance from their scheme is quite doubtable. To the best of our knowledge, very few works can be found to provide a packet-level insight of coding in WSN. The authors of [40] introduce a MAC layer FEC scheme with Reed-Solomon (15,7) code, but this implementation violates the IEEE 802.15.4 standard.. 3.2. Routing protocols in WSNs. Reliable and robust routing scheme in WSNs, as one of the most challenging research topics, has attracted a great deal of research interests. As we men-.
(58) 3.2 Routing protocols in WSNs. 25. tioned above, existing routing protocols in WSNs can be classified as four primary categories, which include flooding-based, dynamic routing table-based, cluster-based and geographical routing schemes. By flooding-based routing protocols, as we introduced above, although traditional flooding approaches may suffer from broadcast storm and energy inefficiency problems, there are still several advantages such as not requiring path discovery nor maintenance of routing information and achieving high reliability from multipath diversity. However previous flooding schemes fail to satisfy industrial requirements. The approach by Akkaya and Younis [41] aims at avoiding the drawbacks of flooding by randomizing the selection of retransmitters, but it is unacceptable in IWSN settings. A random routing strategy for WSNs based on flooding in [42] aims for low energy consumption. However, this results in even higher latency, which makes it less useful for industrial purposes. Dynamic routing table-based routing schemes are more often applied in current WSNs. Since local routing tables are required to forward packets in each node, in order to maintain updated routing tables, there exist two basic approaches. One approach is to discover all available paths by broadcasting route discovery messages. Many conventional routing protocols, such as Ad hoc On Demand Distance Vector (AODV) routing algorithm, Dynamic Source Routing (DSR), are sorted as this type. The second method is that all nodes periodically exchange messages for being aware of existing neighbors, such as Fisheye State Routing (FSR). However, both approaches require excessive control messages to maintain the connectivity of all routes. Reliability and real time performance can be hardly achieved by conventional routing scheme in this category as we described above. Although a number of extensions of conventional routing protocols were proposed, such as AOMDV [43], an extension of AODV, and SMR [44] as a multipath version of DSR, primary drawbacks of routing table-based routing schemes still exist. To avoid the drawbacks of dynamic routing table-based routing schemes and decrease energy consumption by not using control messages, geographical routing schemes have been studied. Examples of protocols in this category include greedy approach and the combination of geographical and floodingbased scheme. However, there exist two fatal drawbacks of this approach. First drawback is that this solution is based on the assumption that better link quality can be obtained by shorter geographical distance, which is quite questionable in IWSNs. Second disadvantage is that obtaining accurate position information for each node makes network installation very inconvenient and inefficient. Cluster-based [45] routing schemes [46] are mostly used for adhoc network, so.
(59) 26. Chapter 3. Related Work. they are not appropriate solutions for centralized IWSNs. Therefore, to find a reliable, robust and efficient routing scheme for IWSNs is extremely important..
(60) Chapter 4. Included Papers and Their Contribution There are four papers included in the thesis. The first paper addresses the feasibility of introducing FEC into IWSNs since there exists both timing requirement from existing IWSN standards and resource constrains from embedded end nodes. We also discuss the use of FEC codes in IWSN in order not only to improve the link reliability but also to reduce the number of retransmissions in harsh industrial environments. We propose a compatible FEC scheme for the IEEE 802.15.4 standard and implemented different FEC codes in a typical IWSN chip to evaluate memory consumption and to ensure that we are not violating the strict timing rules for acknowledgment according to the standard. However, the performance improvement of introducing FEC in IWSNs is not evaluated yet. Therefore, the second paper extends the concept of applying FEC in IWSNs for more flexible and multihop purposes. Based on the results from the first paper, we propose and evaluate an enhanced IEEE 802.15.4compliant FEC based approach on a link- and network-level basis to show that a significant improvement in terms of reliability can be achieved by our solution. The goal of improving reliability and reducing communication latency is partially achieved on MAC layer by previous two papers, so the next step is to attempt to solve the problem on network layer. In the third paper, by combining the concepts of FEC and flooding schemes, we propose a reliable and lightweight location-based routing scheme as a cross layer solution for IWSNs. The simulation results show that our solution outperforms the uncoded trans27.
(61) 28. Chapter 4. Included Papers and Their Contribution. missions, as well as common WSN routing protocols. Due to the inherent drawbacks of location-based routing schemes, in the last paper, we propose a new reliable received signal strength based routing protocol for IWSNs for more flexible sensor node installations and suitability for different application scenarios.. 4.1. Paper A. Reliable and Low Latency Transmission in Industrial Wireless Sensor Net˚ works, K. Yu, M. Gidlund, J. Akerberg, M. Bj¨orkman, Accepted for the First International Workshop on Wireless Networked Control Systems (WNCS), Sept. 2011. Short Summary. The major advantages with Industrial Wireless Sensor Networks (IWSNs) in process automation are cable cost reduction, enhanced flexibility and enabling new emerging applications such as wireless control. However, trans-mission over the wireless channel is prone to noise and interference which causes packets to be erroneous received at the receiver node. To improve the link reliability in lossy channels, error correcting codes are commonly used. In this paper we first theoretically discuss the feasibility of introducing FEC into IWSNs. We also discuss the use of forward error correction (FEC) codes in IWSN in order not only to improve the link reliability but also to reduce the number of retransmissions in harsh industrial environments. We propose a FEC scheme suitable for MAC level protection where the packet is divided into groups and encoded using systematic FEC codes. We have implemented different FEC codes in a typical IWSN chip to evaluate memory consumption and to ensure that we are not violating the strict timing rules for acknowledgment. Our results show that some FEC codes are suitable to be implemented in a typical IWSN node while several fails due to large memory footprint or to long encoding and decoding time. Contribution. This paper aims for solving research problem 1. The main contributions in this paper are summarized as follows: • Several basic requirements of applying FEC in IWSNs are discussed and analyzed. • It is shown that the compatible FEC scheme for the IEEE 802.15.4 standard can be implemented in existing IWSN devices without violating the IWSN standard..
(62) 4.2 Paper B. 29. • According to the evaluation, RS(15,7) is proved to be suitable for IWSNs due to the memory footprint and processing timing consumption. My Contribution. I proposed the FEC scheme for IWSNs and implemented the scheme based on several FEC code in the evaluation board. I evaluated the memory footprint and FEC processing time according to the requirements from the IEEE 802.15.4 standard.. 4.2. Paper B. A Flexible Error Correction Scheme for IEEE 802.15.4-based Industrial Wire˚ less Sensor Networks, K. Yu, F. Bara´c, M. Gidlund, J. Akerberg, M. Bj¨orkman, Accepted for the 21st IEEE International Symposium on Industrial Electronics (ISIE2012), 2012. Short Summary. Noise and interference make a substantial impact on wireless transmissions in industrial environments, resulting in frequent erroneous packet deliveries. Existing industrial communication standards adopt the IEEE 802.15.4 specification, which provides no means to correct the detected errors. In this paper, we propose an IEEE 802.15.4-compliant Forward Error Correction-based approach that can be easily retrofitted into the standard without the need for any kind of interaction with chip manufacturers or standardization bodies. We evaluate the approach on link- and network-level scenarios. Improvement of reliability by using FEC can yield multiple benefits: a reduced number of retransmissions, and lower average latency, to name a few. With respect to the uncoded system, the proposed solution provides identical coding gain as the traditional FEC method, at a significantly lower computational load of decoding. Contribution. This paper aims for solving research problem 2. The main contributions in this paper are described as follows • An enhanced flexible IEEE 802.15.4-compliant FEC-based approach is proposed. • It is shown that our proposed scheme significantly improves the communication reliability in both link- and network-level scenarios, as well as a huge reduction of decoding function calls. My Contribution. I proposed the enhanced IEEE 802.15.4-compliant FECbased scheme and analyzed the theoretical performance. I partly implemented.
(63) 30. Chapter 4. Included Papers and Their Contribution. the scheme in the simulation and evaluated the reliability performance. In order to exhibit the advantages of our proposed solution, different simulation scenarios were created based on the cooperation with Mikael Gidlund and J. ˚ Akerberg.. 4.3. Paper C. Towards Reliable and Lightweight Communication in Industrial Wireless Sen˚ sor Networks, F. Bara´c, K. Yu, M. Gidlund, J. Akerberg, M. Bj¨orkman, Accepted for IEEE 10th International Conference on Industrial Informatics (INDIN), 2012. Short Summary. In this paper we endeavor to improve the performance of Industrial Wireless Sensor Networks (IWSN) in terms of timeliness and reliability of data transmission. We introduce improvements on Medium Access Control (MAC) and Network layer in order to address certain shortcomings common for the existing industrial communication standards. We combine FEC coding on the MAC layer with a lightweight routing protocol, targeting better real-time performance in IWSN. Furthermore, we analyze and provide the bounds on network size, under relevant assumptions and performance constraints. The simulation results indicate that the proposed solution gives way to significant real-time performance improvement, compared to the uncoded transmissions, as well as common WSN routing protocols. Contribution. This paper aims for solving both research problem 2 and 3. The main contributions in this paper are summarized as follows: • A reliable and lightweight routing protocol based on the combination of FEC and flooding schemes is proposed. • It is shown that a substantial improvement in both communication reliability and real-time performance can be achieved by the proposed protocol. My Contribution. The initial idea of combining FEC with flooding was ˚ formed in cooperation with Mikael Gidlund and J. Akerberg. Based on the flooding part previously from Filip Bara´c, I proposed the reliable and lightweight routing protocol and analyzed the theoretical performance in both reliability and transmission latency. The evaluations to reveal the significant improvement in both reliability and real time performance was partly done by me..
(64) 4.4 Paper D. 4.4. 31. Paper D. Reliable RSS-based Routing Protocol for Industrial Wireless Sensor Networks, ˚ K. Yu, M. Gidlund, J. Akerberg, M. Bj¨orkman, Submitted to the 38th Annual Conference of the IEEE Industrial Electronics Society(IECON), 2012. Short Summary. High reliability and real-time performance are main research challenges in Industrial Wireless Sensor Networks (IWSNs). Existing routing protocols applied in IWSNs are either overcomplicated or fail to fulfill the stringent requirements. In this paper, we propose a reliable and flexible Received Signal Strengthbased routing scheme. Our proposed solution can achieve a seamless transition in the event of topology change and can be applied in different industrial environments. The proposed solution is verified by simulations. The simulation results show that our solution outperforms conventional routing protocols in both reliability and latency. Furthermore, the result also proves that the changes of network topology have no impact on data transmissions of other nodes by our scheme, whereas conventional routing protocols are shown to fail to recover the network in a short time. Finally, due to dynamic weighting mechanism, the proposed scheme is verified to achieve significantly higher reliability in scenarios with obstacles and avoid installation troubles, compared to location-based flooding scheme. Thus, our proposed scheme is considered to be more suitable for IWSNs than other routing protocols. Contribution. This paper aims for solving research problem 3. The main contributions in this paper are summarized as follows: • A reliable and robust routing protocol based on RSS and flooding is proposed. • It is shown that the proposed solution outperforms conventional routing protocols in reliability and real time performance, and provides a seamless transition in network topology changes. • Compared with location-based flooding scheme, the proposed solution greatly simplifies the installation operations and achieves better performance in the industrial scenarios with obstacles. My Contribution. I proposed the reliable and robust RSS-based routing protocol due to the inherent drawbacks of previous location-based routing protocol. I implemented the proposed solution in the simulation and evaluated the.
(65) 32. Chapter 4. Included Papers and Their Contribution. improvement of reliability, real time performance and network recovery time, compared our solution with several conventional routing scheme and locationbased flooding scheme..
(66) Chapter 5. Conclusions and Future work 5.1. Summary. In this thesis, we described the increasing interest of deploying IWSNs in industrial automation due to significant cost reduction and flexible installation. We state the current research problems due to the stringent requirements of industrial automation. Thus, to highlight these problems, we give an overview of industrial automation to show the necessity of reliable and low latency communications. After giving a brief introduction and comparison of three existing IWSN standards, namely WirelessHART, ISA100.11a and WIA-PA, we reveal that no error correction mechanism is used on MAC layer in any IEEE 802.15.4-based standard, and only very generic routing categories are specified in all these standards, instead of specific reliable and robust routing protocols. We also study the state-of-the-art measurements in real industrial environment to show that it is a challenge to fulfill the requirement due to the dynamic nature of the wireless link and harsh industrial wireless channel conditions. According to our description, FEC coding is able to recover corrupted error bits caused by the noisy channel in a packet by adding redundancy and a reliable routing protocol can improve the reliability and real time performance by guaranteeing multi-path packet delivery and a seamless transition in the event of topology change. Therefore, introducing FEC into IWSNs on MAC layer and applying reliable and robust routing protocols on network layer are considered as 33.
(67) 34. Chapter 5. Conclusions and Future work. two appropriate and reasonable approaches to effectively achieve the goal. The study of previous research works on these two topic shows that few works can be found to provide a packet-level insight of FEC coding in IWSN without violating existing standards and most of routing protocols for WSNs fail to fulfill the requirements due to the overcomplexity or excessive control overhead. In our work, in order to show the feasibility of introducing FEC into IWSNs, we theoretically analyze the timing requirement from existing IWSN standards. Then we benchmarked different FEC codes in terms of memory consumption and processing time in a typical IWSN chip. Our results show that ReedSolomon (15,11), as one example, can fulfill the requirements given in process automation. In the next step, we propose a MAC layer FEC scheme for IEEE 802.15.4-based WSN. The most notable advantage of this solution over the traditional FEC scheme is the full compatibility with the IEEE 802.15.4 specification and its implementation by no means violates the standard. Since this solution is applied on the MAC layer with software implementation, it does not require any hardware support for executing FEC nor interaction with the radio chip manufacturers. We evaluate this solution by simulation to show a significant packet delivery rate improvement compared with uncoded system and greatly reduction of decoding function calls compared with traditional FEC schemes. After successfully applying FEC on MAC layer in IWSNs, we continue our work on network layer. We combine a FEC scheme on MAC layer with a lightweight location-based routing protocol to form an IEEE 802.15.4-conformable solution, with the aim to address some deficiencies of the existing industrial communication standards, and improve the latency and reliability of data transmission. This proposed solution is fully IEEE 802.15.4-compliant. Results obtained in simulation a substantial increase in Packet Delivery Ratio and reduced latency, compared to both uncoded and coded conventional WSN routing protocols. However, in many scenarios, the geographical distance does not reflect the practical link quality between two devices and the installation of sensor nodes is quite inconvenient by using a location-based routing protocol. Therefore, we proposed a reliable and flexible RSS-based routing protocol without requiring any positioning information. Although this approach is also based on flooding, we defined a number of forwarding criteria to avoid the drawbacks of flooding. The simulation results show a significant improvement of both transmission reliability and latency, compared with conventional routing protocols and achieve similar performance as the previous solution. Typically, when obstacles usually exist in the scenario, this solution significantly outperforms the location-based flooding scheme..
(68) 5.2 Future Work. 5.2. 35. Future Work. It is notable that all three main standards for IWSNs use TDMA on MAC layer to provide deterministic transmission and avoid communication conflict. For mesh network, although how to assign specific time slots to each node and gateway in the network is tightly related to the routing protocol, it directly affect the real time performance and the maximum allowed network size. In this thesis, we only apply a simple scheduling scheme without fully considering the need of each node. To find an appropriate scheduling scheme to lower the latency and maximum the network size is one of the subjects for the future work towards a PhD thesis. In addition, the two routing protocols we proposed focus on uplink in IWSNs, which means the transmission from sensor nodes to a gateway. The routing of downlink is not fully addressed in this thesis. For the integrity of our final solution, it is also important to provide reliable and low latency downlink. Finally, most of our previous performance evaluations are based on simulation. It is interesting to implement our solutions in real industrial devices, so that we can evaluate the performance of our solutions in real industrial environments for more convincing results..
(69) Bibliography. Bibliography [1] Srinivas Medida. Pocket Guide on Industrial Automation For Engineers and Technicians 1st Edition. IDC Technologies, 2012. [2] J. Akerberg, M. Gidlund, and M. Bjorkman. Future research challenges in wireless sensor and actuator networks targeting industrial automation. In Industrial Informatics (INDIN), 2011 9th IEEE International Conference on, pages 410 –415, july 2011. ¨ [3] Javier Ferrer-Coll, Per Angskog, Jos´e Chilo, and Peter Stenumgaard. Requirements of wireless systems in industrial areas. In RFMTC Radio Frequency Measurement Technology Conference, Oct. 2011. [4] Robert H. Morelos-Zaragoza. The Art of Error Correcting Coding, Second Edition. JOHN WILEY and SONS, LTD, OCT 2006. ˚ [5] J. Akerberg, M. Gidlund, F. Reichenbach, and M. Bj¨orkman. Measurements on an industrial wirelesshart network supporting profisafe: A case study. In IEEE Conference on Emerging Technologies and Factory Automation (ETFA’11), Sept. 2011. [6] D. Sexton, M. Mahony, M. Lapinski, and J. Werb. Radio channel quality in industrial wireless sensor networks. In Sensors for Industry Conference, 2005, pages 88 –94, feb. 2005. [7] Javier Ferrer Coll. Rf channel characterization in industrial, hospital and home environments. Licentiate thesis, February 2012. ˚ [8] Johan Akerberg. On security in safety-critical process control. Licentiate thesis, November 2009. 36.
Related documents
The thesis concludes that fountain coding in combination with braided multi- path routing, and proportionally fair packet scheduling is an e cient solution for a wireless sensor
The current implementation allows the user to trigger sending of a synchronization message, which we used to send these messages based on HMAC configuration and which phase the
överensstämmer med att 36,4 % sjuksköterskor ansåg att ett hinder för kommunikation vid palliativ vård uppstod när patienter inte hade fått tillräckligt med information samt
Även om körfältet inte hela körfältet var fritt från hinder byttes fält efter hindret och det genomfördes ett stort antal fältbyten uppströms upphörsignalen och
Man menar att detta också leder till att de kommunala skolorna tvingas till förändring, både genom att ändra sitt sätt att driva skola och sitt sätt att marknadsföra skolan på
1994) berättar Dan Quayle sin historia, från ungdomen i Indiana till åren som George Bushs vicepresident.. Den ger en annan bild av Quayle än den gängse och
lika stor utsträckning av staten blir resultatet att en enhetlig stats befolkning de facto lever under ojämlikheter. Hänsyn bör därför tas till att särskilda rättigheter för en
Hur uppfattade olika ansvariga chefer företagets krisberedskap i detta sammanhang, hur väl gick den att genomföra för att hjälpa de drabbade kunderna och har MyTravel
