Wireless sensor networks

Introduction to wireless sensor networks (WSNs). Cover topics like WSN platforms, transducers/sensors, standards, protocols, powering nodes, and other issues like privacy concerns.

1. Wireless Sensor Networks CS4492 Wireless and Broadband Networking Dilum Bandara [email protected] Slides extracted from Prof. Anura Jayasumana , Colorado State University

2. Outline Platforms Transducers/Sensors & Standards Protocols MAC , routing, & transport Powering nodes Other issues 2

3. Emerging World of Sensor Networks Emerging information infrastructure that extends reach of networks to physical world Dense collections of smart sensors, actuators, & processors that self-configure to network & process Now comes under the umbrella of IoT May be viewed as the next step in evolution of networking, following WWW & Internet “sensor-network”  de scribe “wireless-sensor networks” Consist tiny computing & sensing devices equipped with wireless communication capability It has a much bigger scope & reach Sensor-actuator networks , wireless sensor networks, distributed sensing networks, high & low speed collaborative adaptive systems, … 3

4. Sensor-Actuator Networks – Bridging Physical & Digital Worlds RFIDs ……….Motes ……Cellphones ….. Cameras …..…. Radars, Observatories Small Low data rates Moderate data rates High data rates Power limited Not power limited Not power limited Processing limited Not processing limited Storage limited Not storage limited Very small Extremely low data rates No Power Little or no processing Little or no storage 4

5. Wireless Sensor Networks (WSNs) Processor Sensor Wireless Tx /Rx 5 Source: www.cs.iit.edu /~ winet/testbeds.html

6. Wireless Sensor Networks (Cont.) Wireless Small, cheap Can deploy in Tens, hundreds, & thousands Fine grained picture of phenomena (e.g., climate) Unattended operation Ad-hoc deployment Self powered Self organizing Spontaneously create impromptu network Dynamically adapt to device failure, degradation, movement Act cooperatively Organize themselves & sharing computations Provide usable chunks of predigested information than a bunch of numbers 6

7. Mote/ iMote / iBean 7 Low data rates Power limited Processing limited Storage limited Source: Technology Review (April 2004) EISTEC Mulle iMote2

8. 8

9. Platforms – Mote Evolution 9 The Mote Revolution: Low Power Wireless Sensor Network Devices, J. Polastre , R. Szewczyk , C. Sharp, D. Culler Hot Chips 16: A Symposium on High Performance Chips. August 22-24, 2004.

10. Platforms – Motes / Smart Dust Wireless Communication Power Limitations Limited CPU Power Limited Memory 10 CPU 8-bit, 4MHZ Storage 8kB Flash (Instructions) 512B RAM, 512B EEPROM Communication 10kBps over 916 MHz radio Operating System Tiny OS (3500B) Available code space 4500 B Source: Wireless Networks 8, 521-534, 2002

11. Platforms – Telos (Contd.) 11

12. Platforms – IBM & Libelium Microcontroller: ATmega1281 Frequency: 14MHz SRAM: 8KB EEPROM: 4KB FLASH: 128KB Weight: 20gr Dimensions: 73.5 x 51 x 13 mm Clock: RTC (32KHz) Power On - 15 mA, Sleep – 55 uA 12 Chipset: AT86RF231 Frequency : 2.4GHz Link Protocol: IEEE 802.15.4 Sensitivity : - 101dBm Output Power: 3dBm Encryption: AES 128b Send IPv6 packs over 802.15.4

13. Radios Telos : Enabling Ultra-Low Power Wireless Research, J. Polastre , R. Szewczyk , D. Culler Proc. 4 th Int. Conf. Information Processing in Sensor Networks: Track on Platform Tools and Design Methods for Network Embedded Sensors (IPSN/SPOTS), April 25-27, 2005 13

14. Microcontrollers Telos : Enabling Ultra-Low Power Wireless Research, J. Polastre , R. Szewczyk , D. Culler Proc. 4 th Int. Conf. Information Processing in Sensor Networks: Track on Platform Tools and Design Methods for Network Embedded Sensors (IPSN/SPOTS), April 25-27, 2005 14

15. Telos Telos : Enabling Ultra-Low Power Wireless Research, J. Polastre , R. Szewczyk , D. Culler Proc. 4 th Int. Conf. Information Processing in Sensor Networks: Track on Platform Tools and Design Methods for Network Embedded Sensors (IPSN/SPOTS), April 25-27, 2005 15

16. Telos Telos : Enabling Ultra-Low Power Wireless Research, J. Polastre , R. Szewczyk , D. Culler Proc. 4 th Int. Conf. Information Processing in Sensor Networks: Track on Platform Tools and Design Methods for Network Embedded Sensors (IPSN/SPOTS), April 25-27, 2005 Packet yield ( left ), link quality indicator ( LQI, center ), & received signal strength ( RSSI, right ) outdoors with the Telos mote and internal antenna 16

17. Application – Microweather Station 17 Source: Overview of Sensor Networks, D. Culler, D. Estrin , M. Srivastava, Computer, Aug 2004, pp . 41-9

18. Application – Microweather Station 18 Source: Overview of Sensor Networks, D. Culler, D. Estrin , M. Srivastava, Computer, Aug 2004, pp . 41-9

19. Air Quality 25-mm Bluetooth air-quality monitor Onboard sensors monitor critical carbon-monoxide & volatile-organic-compound levels, and ambient parameters such as humidity, temperature ,& vibration Source: IEEE Pervasive Computing, Oct.-Dec. 2010. [email protected] 19

20. ZebraNet Source: Hardware Design Experiences in Zebranet , P. Zhang, C.M.Sadler , S . Lyon, M. Martonosi , Sensor Systems 2004. 20

21. Cattle Herding Source: Z . Butler et al. , “Networked Cows: Virtual Fences for Controlling Cows ,” WAMES 2004 , Boston, MA, June 2004. 21

22. OS Options TinyOS Event, component oriented OS Designed for severe memory, processing limitations Hardware, software components represented as interfaces & implementations, glued together using NesC constructs, allows for some abstraction SOS Dynamic reconfiguration, modify software on nodes after deployment NutOS Simple RTOS, originally for ATmega128, expanded to other AVR & more Includes full TCP/UDP stack & device drivers for several NICs Needs a good bit of memory Contiki Pre-emptive multitasking, dynamic memory management, TCP/IP stack, windowing system / GUI, VNC, web server, web browser, telnet Under 50K memory 22

23. WSN vs. Ad-Hoc Wireless Networks Much higher no of sensor nodes Densely deployed sensor nodes Sensor nodes are more prone to failures Topology may change frequently Sensor nodes limited in power, computation capability, & memory Sensor nodes may not have global IDs Sensor nodes deployed for a specific application 23

24. Physical Layer Minimum output power required to transmit over distance d n is closer to 4 for low-lying antennae & near ground channels (as is typical for WSNs) For small devices to cover large distances, need hop-by-hop Each bit of transmission ~1,000 instructions Process within node whenever possible 24

25. IEEE 802.15.4 Wireless Medium Access Control (MAC) & Physical Layer (PHY) Specifications for Low Rate Wireless Personal Area Networks (LR-WPANs) Extremely low-cost operation Limited power & relaxed throughput requirements Short-range operation Simplicity, flexibility Over-the-air data rates 250 kbps, 40 kbps, 20 kbps Star or peer-to-peer operation 16-bit short or 64-bit extended addressing Fully acknowledged protocol for transfer reliability 25

26. IEEE 802.15.4 / ZigBee – Topologies 26 Source: http://wireless.arcada.fi/MOBWI/material/PAN_5_2.html

27. Clusters 27

28. IEEE 802.15.4 – Cluster Tree 28 IEEE 802.15.4 Wireless Medium Access Control (MAC) & Physical Layer (PHY) Specifications for Low Rate Wireless Personal Area Networks (LR-WPANs)

29. Cluster Tree 29

30. Power Consumption Microcontroller Examples StrongArm ~400mW/50mW/0.16mW ATmega103L 16.5mW Modes – Active, Idle, & Sleep Radio Modulation scheme Data rate Transmit power (distance) Duty cycle Modes – Transmit, Receive, Idle, & Sleep 30

31. Conserving Power Process locally, communicate only when necessary Compress data Avoid explicit protocol messages Piggy back on data Overhear packets destined to others Schedule transmissions to avoid contentions & time radio has to remain alive Assign specific responsibilities to certain nodes E.g., retransmission, aggregation Reject uninteresting packets by turning off radio after receiving only a fraction 31

32. Conserving Power (Cont.) Minimize waste due to Idle listening Over transmissions Overhearing Collisions Parameters Power consumption in Rx, Tx , & sleep modes Wake-up time Bit & frame synchronization time Receive strength indicator (RSSI) Message filtering Switching time between Tx & Rx Receive sensitivity & maximum Tx power 32

33. Power Harnessing Source: IEEE Pervasive Computing, Jan-March 2005, Jan-March 2007 33

34. Typical Sensor Network Architecture 34 Source: http:// www.syssoft.uni-trier.de/systemsoftware/Download/Fruehere_Veranstaltungen/Ubiquitous_Computing/2004/

35. Network Protocol Considerations Routing is likely to be data centric Power efficiency plays a key role Maximum available power route Minimum energy route Minimum hop route Maximum minimum available power route 35

36. Data Centric Routing Conventional schemes Address individual or set of nodes Unicast, broadcast, multicast Data-centric communication paradigm Anycast , geocast , marketplace communication Approaches Disseminate interest about data Advertise availability of information, wait for request 36

37. Data Centric Routing (Cont.) How to address nodes with no IDs? Attribute-based naming Query an attribute than an individual node “Rooms where temperature is over 60° ?” Data aggregation/fusion often needed to merge data from many nodes Some specifics may not be left out (e.g ., location of the data) 37

38. Flooding & Gossiping Flooding When receiving packet for 1 st time , repeat forwarding, until maximum hop count or destination not reached Reactive technique Doesn’t require costly topology maintenance Deficiencies – Implosion, Overlap, Resource Blindness Gossiping Forward to 1 random selected neighbor only Avoids implosion problem Message propagation takes a long time 38 Source: http :// www.syssoft.uni-trier.de/systemsoftware/Download/Fruehere_Veranstaltungen/Ubiquitous_Computing/2004/

39. Flooding & Gossiping (Cont.) Place offer in a geographic area with high node density Send request towards same area Code execution at marketplace to reduce message complexity Background dissemination of marketplace locations Moving towards marketplace Geographic routing Communication on the marketplace Topology-based routing Efficient flooding 39 Source: http ://www.syssoft.uni-trier.de/systemsoftware/Download / Fruehere_Veranstaltungen / Ubiquitous_Computing /2004 /

40. Directed Diffusion 40 Source: C . Intanagonwiwat et al., “Directed Diffusion for Wireless Sensor Networking,” IEEE/ACM Trans. Net., vol. 11 , no . 1, Feb. 2002, pp. 2–16.

41. Topology Control Topology – Which node is able to/allowed to communicate with which node Networks can be too dense for efficient operation Too many collisions, high complexity (MAC) Too many paths to choose from (Routing) Topology Control Make topology less complex subject to restrictions (e.g., maintain connectivity) Control node activity (turn on/off nodes) Control link activity (turn on/off links) Flat networks (all nodes have same role) Hierarchical networks – Backbones, clustering Control power 41

42. 3-D Radio Connectivity Source: A . Savvides , Yale (ISPN/SPOTS 2005) 42

43. Challenges Limited computation & data storage Low-power consumption Wireless communication Medium, ad hoc vs. infrastructure, topology & routing Data/sensor-related issues, e.g., calibration Continuous operation Inaccessibility – network adjustment & retasking Robustness & fault tolerance Wireless Sensor/Actuator networks (WSANs) 43

44. WSN – Options Deployment – Random vs. manual, One-time vs. iterative Mobility – Immobile, some mobile, all mobile Size – brick, matchbox, grain, dust Heterogeneity – Homogeneous vs. Heterogeneous Communication – Radio vs. light vs. inductive vs. capacitive vs. sound Infrastructure vs. ad hoc Topology – Flat vs. hierarchical, single hop, multi-hop, clusters, trees, graphs, etc. Coverage – Sparse vs. dense vs. redundant Connectivity – connected vs. intermittent vs. sporadic 44

45. Design Space 45 Design Space of WSNs, K. Romer , F. Mattern , E. Zurich, IEEE Wireless Communications, Dec. 2004

46. Security of Sensor Networks Application Requirements Data Confidentiality Data Authenticity Data Integrity Data Freshness . . . Vulnerabilities Physical attacks Eavesdropping Inject bits into the channel Replay previous packets Introduction of malicious hardware Outsider attacks Insider attacks Hello floods Sink holes & wormholes . . . 46

47. Privacy Concerns Sensor networks may (will) be used for surveillance Track people, vehicles, etc ., over long periods of time Abuse of existing networks for illegal purposes Low cost  easy to deploy Data collection & coordinated analysis 47 Source: IStockPhoto.com