Principles of Wireless Sensor Networks
advertisement
Principles of Wireless Sensor Networks http://www.ee.kth.se/~carlofi/teaching/wsn_course.shtml Lecture 1 Stockholm, September 9 2009 Carlo Fischione Royal Institute of Technology - KTH Stockholm, Sweden e-mail: carlofi@ee.kth.se Fall 2009 Principles of Wireless Sensor Networks Carlo Fischione Today’s lecture Course Overview Introduction to WSNs Node Architecture Network Architecture Fall 2009 Principles of Wireless Sensor Networks Carlo Fischione Practical Information Office: Osquldas väg 10, floor 6. Meeting times: Wednesdays, 13.15-15.00. Exception: Tuesday September 22, 13.15-15:00, and Tuesday October 21, 10:00-11:00. Office Times: By appointment. Class Room: Himmelriket, floor 8, Osquldas väg 10, KTH, Stockholm. Work load: 2h per lecture + research work. Prerequisites: Familiarity with linear algebra and analysis. ¾ Knowledge of optimization theory may be useful. Fall 2009 Principles of Wireless Sensor Networks Carlo Fischione Course Content Lec 1: Introduction (course overview, applications, WSNs node architecture, protocols, operating systems). Lec 2: WSN Programming: from abstractions to running code. Invited lecture, Luca Mottola, SICS Lec 3: Iterative Methods for Parallel Computation. Lec 4: Decentralized Radio Power Control. Lec 5: Medium Access Control, Routing, Duty Cycle. Lec 6: IEEE802.15.4, ZigBee, WirelessHART, ISA SP-100, ROLL. Lec 7: Invited Lecture, TBD. Lec 8: Consensus Algorithms. Lect 9: Distributed Estimation. Lect 10: Localization and Positioning. Fall 2009 Principles of Wireless Sensor Networks Carlo Fischione Course Goal After finishing the course, the attendant will ¾ Know the essential theoretical tools to cope with WSNs. ¾ Know the basics of WSNs. ¾ Know how to design WSNs. ¾ Develop a research project. ¾ Develop presentation skills. Fall 2009 Principles of Wireless Sensor Networks Carlo Fischione Evaluation Ph.d. version: ¾ The course is worth 5 credits plus 2.5 optional credits based on an original research project. ¾ Grades (pass/fail) will be based on attendance (50%) and exercises plus presentations (50%). Master students version: ¾ The course is worth 5 credits ¾ Grades (pass/fail) will be based on attendance, exercises, presentation (3 credits) and a research project (2 credit). ¾ Less exercises than the Ph.D. version. Exercises ¾ Contain theoretical as well as practical parts (mainly through simulations and possible implementation on test-bed). ¾ Have to be delivered before the next lecture following the assignment. Fall 2009 Principles of Wireless Sensor Networks Carlo Fischione Student Presentations Each students has to present a paper/book chapter during the course. ¾ See the list on the course website http://www.ee.kth.se/~carlofi/teaching/wsn_course.shtml Choose papers/book chapter from the list. ¾ ¾ ¾ Send me an e-mail with the preferred papers by tomorrow Sept. 10 You can propose to present a paper that is not in the list, but I reserve to check if it is relevant to the course flow before accepting. I will then assign the papers to the presenters by Sept 11. Fall 2009 Principles of Wireless Sensor Networks Carlo Fischione Research projects The project title must be communicated to the teacher by September 30 2009. You can propose your own project, or ask the teacher to give to you a project. A project can be developed in collaboration with max 2 students. The project has to be presented on November 18 The project has to be summarized in a paper max 5 pages long, IEEE conference format. The paper has to be delivered by November 25. Fall 2009 Principles of Wireless Sensor Networks Carlo Fischione Text books D. P. Bertsekas, J. N. Tsitsiklis, Parallel and Distributed Computation: Numerical methods, Athena Scientific, 1997. http://dspace.mit.edu/handle/1721.1/3719 H. Karl and A. Willig, Protocols and Architectures for Wireless Sensor Networks, Wiley, 2005. http://www.cs.uni-aderborn.de/index.php?id=1119&L=1 Fall 2009 Principles of Wireless Sensor Networks Carlo Fischione Useful Links http://www.hartcomm.org/ http://www.ieee802.org/15/pub/TG4.html http://www.ietf.org/dyn/wg/charter/roll-charter.html http://www.ipso-alliance.org/Pages/Front.php http://www.isa.org/ http://www.tinyos.net/ http://www.sics.se/contiki/ http://www.zigbee.org/ WSNs Standard http://www.wsnblog.com/ http://www.wsn-security.info/index.htm Blogs http://www.dustnetworks.com/ Industries http://www.sensinode.com/ http://www.sentilla.com/ http://www.xbow.com/Home/wHomePage.aspx http://www.ee.kth.se/~mikaelj/wsn_course.shtml http://www.cs.berkeley.edu/~culler/eecs194/ http://bwrc.eecs.berkeley.edu/Research/energy_efficient_systems.htm http://wsnl.stanford.edu/ http://courses.csail.mit.edu/6.885/spring06/readings.html http://www.eecs.harvard.edu/~mdw/course/cs263/fa04/ http://www.cs.sunysb.edu/~jgao/CSE595-spring09/ Fall 2009 University courses Principles of Wireless Sensor Networks Carlo Fischione Definition of WSNs Wireless Sensor Networks are networks for communication, control, sensing, and actuation. WSNs are composed by nodes equipped by reduced processing and power resources and communicating through wireless. UCB,Fall October 26, 2007 2009 Sensor Networks DesignPrinciples Challengesof in Wireless Wireless Sensor Networks Carlo Fischione History of WSNs DARPA DSN node, 1960 Mica2 mote, 2002 Tmote-sky mote, 2003 Fall 2009 Smart Dust, 200? Principles of Wireless Sensor Networks Carlo Fischione Application of Wireless Sensor Networks Industrial control Environmental monitoring Transportation Marine monitoring Health care Fall 2009 Principles of Wireless Sensor Networks Carlo Fischione Economist Applications: Building Monitoring Sensors used to measure response to traffic, tidal and seismic activity Deployed on Golden Gate Bridge ¾ http://www.cs.berkeley.edu/~binetude/ggb/ Fall 2009 Principles of Wireless Sensor Networks Carlo Fischione Habitat Monitoring Non-intrusive monitoring of animal habitat. Example: Nesting of Leach’s Storm Petrels ¾ A. Mainwaring, J. Polastre, R. Szewczyk, and D. Culler, "Wireless Sensor Networks for Habitat Monitoring," Intel Research, IRB-TR02-006, Jun. 19, 2002. Storm petrel chick Sensor node Fall 2009 Principles of Wireless Sensor Networks Carlo Fischione Smart Buildings WSNs to Control of temperature, light intensity, air and humidity. www.instablogsimages.com Fall 2009 Principles of Wireless Sensor Networks Source: Ed Arens Carlo Fischione Smart grids source: http://deviceace.com/ Smart grids: “Smart Grids: It's All About Wireless Sensor Networks” (http://stanford.wellsphere.com) Fall 2009 Principles of Wireless Sensor Networks Carlo Fischione FeedNetBack European Project The project aims to close the control loop over wireless networks by deriving and applying a co-design framework that allows the integration of communication, control, computation and energy management aspects in a holistic way. Closing the loop over wireless networks KEYWORDS: co-design, communication, control, computation, energy Fall 2009 Principles of Wireless Sensor Networks Carlo Fischione WSNs in Industrial Automation Several advantages with wireless networking control systems: Added flexibility ¾ Sensor and actuator nodes can be placed more appropriately ¾ Less restrictive maneuvers and control actions ¾ More powerful control through distributed computations Reduced installation and maintenance costs ¾ Less cabling ¾ More efficient monitoring and diagnosis UCB,Fall October 26, 2007 2009 Principles ofinWireless SensorNetworks Networks Design Challenges Wireless Sensor Carlo Fischione Distributed Camera Calibration •WSN allows one to perform distributed camera calibration •Application: massive graphic effects in film production Fall 2009 Principles of Wireless Sensor Networks Carlo Fischione Old Market Forecasts for WSNs WSNs population forecast 2010 RFID/Sensors: 1+ Trillion Location Humidity Temperature Vibration Liquid Weight etc. Microprocessors: 500 Billion 4-64+ bits CPU’s etc. Smart Devices: Appliances Machinery Vehicles 1 Billion Info Devices: Bldg. Eqpt. etc. Mobile Phones PDA’s Web Tablets etc. 2 Billion 300 Million PC’s: Forecast of installed base, 2010 The FocalPoint Group, LLC 2003 http://www.thefpgroup.com Fall 2009 Principles of Wireless Sensor Networks Carlo Fischione New Market Forecasts for WSNs On World Expert survey “WSN Market size in 2007 and Active RFID and Sensor Networks 2007-2017”: ¾ industrial applications of WSN: 4.6 Billion $ by 2011 ¾ smart building applications of WSN: 2.5 Billion $ by 2011. IDTechEx: ¾ WSNs and active RFID of about 4B$ by 2012. WSNs R&D investments: from 522Million $ in 2007 to 1.3Billion $ in 2012. 30 millions WirelessHART wireless sensors installed so far worldwide. Fall 2009 Principles of Wireless Sensor Networks Carlo Fischione Outline Course Overview Introduction to WSNs for wireless? ¾ Requirements & mechanisms ¾ Infrastructure Node Architecture Network Architecture The following slides are adapted from H. Karl’s companion book webpage http://www.cs.uni-paderborn.de/index.php?id=1119&L=1 Fall 2009 Principles of Wireless Sensor Networks Carlo Fischione Possible applications for infrastructure-free networks Factory floor automation Disaster recovery Car-to-car communication ad c ho Military networking: Tanks, soldiers, … Finding out empty parking lots in a city, without asking a server Search-and-rescue in an avalanche Personal area networking (watch, glasses, PDA, medical appliance, …) … Fall 2009 Principles of Wireless Sensor Networks Carlo Fischione Solution: (Wireless) ad hoc networks Try to construct a network without infrastructure, using networking abilities of the participants ¾ This is an ad hoc network – a network constructed “for a special purpose” Simplest example: Laptops in a conference room – a single-hop ad hoc network Fall 2009 Principles of Wireless Sensor Networks Carlo Fischione Problems/challenges for ad hoc networks Without a central infrastructure, things become much more difficult Problems are due to ¾ Lack of central entity for organization available ¾ Limited range of wireless communication ¾ Mobility of participants ¾ Battery-operated entities ? Fall 2009 Principles of Wireless Sensor Networks Carlo Fischione Wireless sensor networks Participants in the previous examples were devices close to a human user, interacting with humans Alternative concept: Instead of focusing interaction on humans, focus on interacting with environment ¾ ¾ ¾ Network is embedded in environment Nodes in the network are equipped with sensing and actuation to measure/influence environment Nodes process information and communicate it wirelessly ! Wireless sensor networks (WSN) ¾ Or: Wireless sensor & actuator networks (WSAN) Fall 2009 Principles of Wireless Sensor Networks Carlo Fischione Roles of participants in WSN Sources of data: Measure data, report them “somewhere” ¾ Sinks of data: Interested in receiving data from WSN ¾ Typically equip with different kinds of actual sensors May be part of the WSN or external entity, PDA, gateway, … Actuators: Control some device based on data, usually also a sink Fall 2009 Principles of Wireless Sensor Networks Carlo Fischione Structuring WSN application types Interaction patterns between sources and sinks classify application types ¾ ¾ ¾ ¾ ¾ Event detection: Nodes locally detect events (maybe jointly with nearby neighbors), report these events to interested sinks • Event classification additional option Periodic measurement Function approximation: Use sensor network to approximate a function of space and/or time (e.g., temperature map) Edge detection: Find edges (or other structures) in such a function (e.g., where is the zero degree border line?) Tracking: Report (or at least, know) position of an observed intruder (“pink elephant”) Fall 2009 Principles of Wireless Sensor Networks Carlo Fischione Deployment options for WSN How are sensor nodes deployed in their environment? ¾ Dropped from aircraft → Random deployment • Usually uniform random distribution for nodes over finite area is ¾ assumed • Is that a likely proposition? Well planned, fixed → Regular deployment • E.g., in preventive maintenance or similar • Not necessarily geometric structure, but that is often a convenient assumption ¾ Mobile sensor nodes • Can move to compensate for deployment shortcomings • Can be passively moved around by some external force (wind, water) • Can actively seek out “interesting” areas Fall 2009 Principles of Wireless Sensor Networks Carlo Fischione Maintenance options Feasible and/or practical to maintain sensor nodes? ¾ E.g., to replace batteries? ¾ Or: unattended operation? ¾ Impossible but not relevant? Mission lifetime might be very small Energy supply? ¾ ¾ Limited from point of deployment? Some form of recharging, energy scavenging from environment? • E.g., solar cells Fall 2009 Principles of Wireless Sensor Networks Carlo Fischione Characteristic requirements for WSNs Type of service of WSN ¾ Not simply moving bits like another network ¾ Rather: provide answers (not just numbers) ¾ Issues like geographic scoping are natural requirements, absent from other networks Quality of service ¾ Traditional QoS metrics do not apply ¾ Still, service of WSN must be “good”: Right answers at the right time Fault tolerance ¾ Be robust against node failure (running out of energy, physical destruction, …) Lifetime ¾ The network should fulfill its task as long as possible – definition depends on application ¾ Lifetime of individual nodes relatively unimportant ¾ But often treated equivalently Fall 2009 Principles of Wireless Sensor Networks Carlo Fischione Characteristic requirements for WSNs Scalability ¾ Support large number of nodes Wide range of densities ¾ Vast or small number of nodes per unit area, very applicationdependent Programmability ¾ Re-programming of nodes in the field might be necessary, improve flexibility Maintainability ¾ ¾ WSN has to adapt to changes, self-monitoring, adapt operation Incorporate possible additional resources, e.g., newly deployed nodes Fall 2009 Principles of Wireless Sensor Networks Carlo Fischione Mechanisms to meet requirements Multi-hop wireless communication Energy-efficient operation ¾ Both for communication and computation, sensing, actuating Auto-configuration ¾ Manual configuration just not an option Collaboration & in-network processing ¾ Nodes in the network collaborate towards a joint goal ¾ Pre-processing data in network (as opposed to at the edge) can greatly improve efficiency Fall 2009 Principles of Wireless Sensor Networks Carlo Fischione Outline Course Overview Introduction to WSNs Node Architecture Network Architecture The following slides are adapted from H. Karl’s companion book webpage http://www.cs.uni-paderborn.de/index.php?id=1119&L=1 Fall 2009 Principles of Wireless Sensor Networks Carlo Fischione Network Architecture Survey the main components of the composition of a node for a wireless sensor network ¾ Controller, radio modem, sensors, batteries Understand energy consumption aspects for these components ¾ Putting into perspective different operational modes and what different energy/power consumption means for protocol design Fall 2009 Principles of Wireless Sensor Networks Carlo Fischione Enabling technologies for WSN Cost reduction ¾ For wireless communication, simple microcontroller, sensing, batteries Miniaturization ¾ Some applications demand small size ¾ “Smart dust” as the most extreme vision Energy scavenging ¾ Recharge batteries from ambient energy (light, vibration, …) Fall 2009 Principles of Wireless Sensor Networks Carlo Fischione Sensor node architecture Main components of a WSN node ¾ Controller ¾ Communication device(s) ¾ Sensors/actuators ¾ Memory ¾ Power supply Communication device Memory Controller Sensor(s)/ actuator(s) Power supply Fall 2009 Principles of Wireless Sensor Networks Carlo Fischione Memory power consumption Crucial part: FLASH memory ¾ Power for RAM almost negligible FLASH writing/erasing is expensive ¾ Example: FLASH on Mica motes ¾ Reading: ≈ 1.1 nAh per byte ¾ Writing: ≈ 83.3 nAh per byte Fall 2009 Principles of Wireless Sensor Networks Carlo Fischione Controller Main options: ¾ Microcontroller – general purpose processor, optimized for embedded applications, low power consumption ¾ DSPs – optimized for signal processing tasks, not suitable here ¾ FPGAs – may be good for testing ¾ ASICs – only when peak performance is needed, no flexibility Example microcontrollers ¾ Texas Instruments MSP430 • 16-bit RISC core, up to 4 MHz, versions with 2-10 kbytes RAM, several DACs, RT clock, prices start at 0.49 US$ • Fully operational 1.2 mW • Deepest sleep mode 0.3 μW – only woken up by external interrupts (not even timer is running any more) Atmel ATMega • 8-bit controller, larger memory than MSP430, slower • Operational mode: 15 mW active, 6 mW idle • Sleep mode: 75 μW Fall 2009 Principles of Wireless Sensor Networks Carlo Fischione Transceiver states Transceivers can be put into different operational states, typically: ¾ ¾ ¾ Transmit Receive Idle – ready to receive, but not doing so • Some functions in hardware can be switched off, reducing energy ¾ consumption a little Sleep – significant parts of the transceiver are switched off • Not able to immediately receive something • Recovery time and startup energy to leave sleep state can be significant Research issue: Wakeup receivers – can be woken via radio when in sleep state (seeming contradiction!) Fall 2009 Principles of Wireless Sensor Networks Carlo Fischione Example radio transceivers Almost boundless variety available ¾ Some examples ¾ ¾ RFM TR1000 family • 916 or 868 MHz • 400 kHz bandwidth • Up to 115,2 kbps • On/off keying or ASK • Dynamically tuneable output power • Maximum power about 1.4 mW • Low power consumption Chipcon CC1000 • Range 300 to 1000 MHz, programmable in 250 Hz steps • FSK modulation • Provides RSSI Fall 2009 ¾ Chipcon CC 2400 • Implements 802.15.4 • 2.4 GHz, DSSS modem • 250 kbps • Higher power consumption than above transceivers Infineon TDA 525x family • E.g., 5250: 868 MHz • ASK or FSK modulation • RSSI, highly efficient power amplifier • Intelligent power down, “self-polling” mechanism • Excellent blocking performance Principles of Wireless Sensor Networks Carlo Fischione Energy supply of mobile/sensor nodes Goal: provide as much energy as possible at smallest cost/volume/weight/recharge time/longevity ¾ In WSN, recharging may or may not be an option Options ¾ Primary batteries – not rechargeable ¾ Secondary batteries – rechargeable, only makes sense in combination with some form of energy harvesting Requirements include ¾ Low self-discharge ¾ Long shelf live ¾ Capacity under load ¾ Efficient recharging at low current ¾ Good relaxation properties (seeming self-recharging) ¾ Voltage stability (to avoid DC-DC conversion) Fall 2009 Principles of Wireless Sensor Networks Carlo Fischione Energy scavenging How to recharge a battery? ¾ A laptop: easy, plug into wall socket in the evening ¾ A sensor node? – Try to scavenge energy from environment Ambient energy sources ¾ Light → solar cells – between 10 μW/cm2 and 15 mW/cm2 ¾ ¾ ¾ ¾ Temperature gradients – 80 μ W/cm2 @ 1 V from 5K difference Vibrations – between 0.1 and 10000 μ W/cm3 Pressure variation (piezo-electric) – 330 μ W/cm2 from the heel of a shoe Air/liquid flow (MEMS gas turbines) Fall 2009 Principles of Wireless Sensor Networks Carlo Fischione Energy scavenging – overview Fall 2009 Principles of Wireless Sensor Networks Carlo Fischione Energy consumption A “back of the envelope” estimation Number of instructions ¾ Energy per instruction: 1 nJ ¾ Small battery (“smart dust”): 1 J = 1 Ws ¾ Corresponds: 109 instructions! Lifetime ¾ Or: Require a single day operational lifetime = 24·60·60 =86400 s ¾ 1 Ws / 86400s ≈ 11.5 μW as max. sustained power consumption! Not feasible! Fall 2009 Principles of Wireless Sensor Networks Carlo Fischione Multiple power consumption modes Way out: Do not run sensor node at full operation all the time ¾ If nothing to do, switch to power safe mode ¾ Question: When to throttle down? How to wake up again? Typical modes ¾ Controller: Active, idle, sleep ¾ Radio mode: Turn on/off transmitter/receiver, both Multiple modes possible, “deeper” sleep modes ¾ Strongly depends on hardware ¾ TI MSP 430, e.g.: four different sleep modes ¾ Atmel ATMega: six different modes Fall 2009 Principles of Wireless Sensor Networks Carlo Fischione Switching between modes Simplest idea: Greedily switch to lower mode whenever possible Problem: Time and power consumption required to reach higher modes not negligible ¾ Introduces overhead ¾ Switching only pays off if Esaved > Eoverhead Esaved Example: Event-triggered wake up from sleep mode Scheduling problem with uncertainty (exercise) Eoverhead Pactive Psleep t1 Fall 2009 τdown Principles of Wireless Sensor Networks tevent τup time Start processing Carlo Fischione Transmitter power/energy consumption for n bits Amplifier power: Pamp = αamp + βamp Ptx ¾ Ptx radiated power ¾ αamp, βamp constants depending on model ¾ Highest efficiency (η = Ptx / Pamp ) at maximum output power In addition: transmitter electronics needs power PtxElec Time to transmit n bits: n / (R · R ) code R nomial data rate, Rcode coding rate To leave sleep mode ¾ Time Tstart, average power P start ¾ → Etx = Tstart Pstart + n / (R · Rcode) (PtxElec + αamp + βamp Ptx) Simplification: Modulation not considered Fall 2009 Principles of Wireless Sensor Networks Carlo Fischione Receiver power/energy consumption for n bits Receiver also has startup costs ¾ Time Tstart, average power Pstart Time for n bits is the same n / (R · R Receiver electronics needs PrxElec ) code Plus: energy to decode n bits EdecBits → Erx = Tstart Pstart + n / (R · Rcode) PrxElec + EdecBits ( R ) Fall 2009 Principles of Wireless Sensor Networks Carlo Fischione Some transceiver numbers Fall 2009 Principles of Wireless Sensor Networks Carlo Fischione Controlling transceivers Similar to controller, low duty cycle is necessary ¾ Easy to do for transmitter – similar problem to controller: when is it worthwhile to switch off ¾ Difficult for receiver: Not only time when to wake up not known, it also depends on remote partners → Dependence between MAC protocols and power consumption is strong! Only limited applicability of techniques analogue to DVS ¾ Dynamic Modulation Scaling (DSM): Switch to modulation best suited to communication – depends on channel gain ¾ Dynamic Coding Scaling – vary coding rate according to channel gain ¾ Combinations Fall 2009 Principles of Wireless Sensor Networks Carlo Fischione Computation vs. communication energy cost Tradeoff? ¾ ¾ ¾ ¾ Directly comparing computation/communication energy cost not possible But: put them into perspective! Energy ratio of “sending one bit” vs. “computing one instruction”: Anything between 220 and 2900 in the literature To communicate (send & receive) one kilobyte = computing three million instructions! Hence: try to compute instead of communicate whenever possible Key technique in WSN – ¾ in-network processing! Exploit compression schemes, intelligent coding schemes, … Fall 2009 Principles of Wireless Sensor Networks Carlo Fischione Summary For WSN, the need to build cheap, low-energy, (small) devices has various consequences for system design ¾ Radio frontends and controllers are much simpler than in conventional mobile networks ¾ Energy supply and scavenging are still (and for the foreseeable future) a premium resource ¾ Power management (switching off or throttling down devices) crucial Fall 2009 Principles of Wireless Sensor Networks Carlo Fischione Outline Course Overview Introduction to WSNs Node Architecture Network Architecture ¾ Optimization goals ¾ Design principles ¾ Gateway concepts The following slides are adapted from H. Karl’s companion book webpage http://www.cs.uni-paderborn.de/index.php?id=1119&L=1 Fall 2009 Principles of Wireless Sensor Networks Carlo Fischione Network Architecture Having looked at the individual nodes in the, we now look at general principles and architectures how to put these nodes together to form a meaningful network We will look at design approaches to both the more conventional ad hoc networks and the non-standard WSNs Fall 2009 Principles of Wireless Sensor Networks Carlo Fischione Single-hop vs. multi-hop networks One common problem: limited range of wireless communication ¾ Essentially due to limited transmission power, path loss, obstacles Option: multi-hop networks ¾ Send packets to an intermediate node ¾ Intermediate node forwards packet to its destination ¾ Store-and-forward multi-hop network Basic technique applies to both WSN and MANET Note: Store&forward multihopping NOT the only possible solution ¾ E.g., collaborative networking, network Source coding Obstacle ¾ Do not operate on a perpacket basis Fall 2009 Principles of Wireless Sensor Networks Carlo Fischione Sink Energy efficiency of multihopping? Obvious idea: Multi-hopping is more energy-efficient than direct communication ¾ Because of path loss α > 2, energy for distance d is reduced from cdα to 2c(d/2)α • c some constant However: This is usually wrong, or at least very oversimplified ¾ Need to take constant offsets for powering transmitter, receiver into account → Multi-hopping for energy savings needs careful choice Fall 2009 Principles of Wireless Sensor Networks Carlo Fischione WSN: Multiple sinks, multiple sources Fall 2009 Principles of Wireless Sensor Networks Carlo Fischione Different sources of mobility Node mobility ¾ ¾ ¾ A node participating as source/sink (or destination) or a relay node might move around Deliberately, self-propelled or by external force; targeted or at random Happens in both WSN and MANET Sink mobility ¾ In WSN, a sink that is not part of the WSN might move ¾ Mobile requester Event mobility ¾ ¾ In WSN, event that is to be observed moves around (or extends, shrinks) Different WSN nodes become “responsible” for surveillance of such an event Fall 2009 Principles of Wireless Sensor Networks Carlo Fischione Optimization goal: Quality of Service In ad-hoc networks: Usual QoS interpretation ¾ Throughput/delay/jitter ¾ High perceived QoS for multimedia applications In WSN, more complicated ¾ Event detection/reporting probability ¾ Event classification error, detection delay ¾ Probability of missing a periodic report ¾ Approximation accuracy (e.g, when WSN constructs a temperature map) ¾ Tracking accuracy (e.g., difference between true and conjectured position of the pink elephant) Related goal: robustness ¾ Network should withstand failure of some nodes Fall 2009 Principles of Wireless Sensor Networks Carlo Fischione Optimization goal: Energy efficiency Umbrella term! Energy per correctly received bit ¾ Counting all the overheads, in intermediate nodes, etc. Energy per reported (unique) event ¾ After all, information is important, not payload bits! ¾ Typical for WSN Delay/Reliability/energy tradeoffs Network lifetime ¾ Time to first node failure ¾ Network half-life (how long until 50% of the nodes died?) ¾ Time to partition ¾ Time to loss of coverage ¾ Time to failure of first event notification Fall 2009 Principles of Wireless Sensor Networks Carlo Fischione Optimization goal: Scalability Network should be operational regardless of number of nodes ¾ At high efficiency Typical node numbers difficult to guess ¾ ¾ Ad-hoc networks: 10s to 100s WSNs: 10s to 1000s, maybe more (although few people have seen such a network already …) Requiring to scale to large node numbers has for network architecture ¾ ¾ serious consequences Might not result in the most efficient solutions for small networks! Carefully consider actual application needs before looking for n → ∞ solutions! Fall 2009 Principles of Wireless Sensor Networks Carlo Fischione Distributed organization Participants in a WSN should cooperate in organizing the network ¾ E.g., with respect to medium access, routing, … ¾ Centralistic approach as alternative usually not feasible – hinders scalability, robustness Potential shortcomings ¾ Not clear whether distributed or centralistic organization achieves better energy efficiency (when taking all overheads into account) Option: “limited centralized” solution ¾ Elect nodes for local coordination/control ¾ Perhaps rotate this function over time Fall 2009 Principles of Wireless Sensor Networks Carlo Fischione In-network processing WSNs are supposed to deliver bits from one end to the other WSNs, on the other end, are expected to provide information, not necessarily original bits ¾ Gives additional options ¾ E.g., manipulate or process the data in the network Main example: aggregation ¾ Apply composable aggregation functions to a convergecast tree in a network ¾ Typical functions: minimum, maximum, average, sum, … ¾ Not amenable functions: median Fall 2009 Principles of Wireless Sensor Networks Carlo Fischione In-network processing: Aggregation example Reduce number of transmitted bits/packets by applying an aggregation function in the network 1 1 1 1 3 1 1 6 1 1 1 Fall 2009 1 Principles of Wireless Sensor Networks Carlo Fischione In-network processing: signal processing Depending on application, more sophisticated processing of data can take place within the network ¾ Example edge detection: locally exchange raw data with neighboring nodes, compute edges, only communicate edge description to far away data sinks ¾ Example tracking/angle detection of signal source: Conceive of sensor nodes as a distributed microphone array, use it to compute the angle of a single source, only communicate this angle, not all the raw data Exploit temporal and spatial correlation ¾ Observed signals might vary only slowly in time → no need to transmit all data at full rate all the time ¾ Signals of neighboring nodes are often quite similar → only try to transmit differences (details a bit complicated, see later) Fall 2009 Principles of Wireless Sensor Networks Carlo Fischione Data centric networking? In typical networks (including ad hoc networks), network transactions are addressed to the identities of specific nodes ¾ A “node-centric” or “address-centric” networking paradigm In a redundantly deployed sensor networks, specific source of an event, alarm, etc. might not be important ¾ Redundancy: e.g., several nodes can observe the same area Thus: focus networking transactions on the data directly instead of their senders and transmitters → data-centric networking Fall 2009 Principles of Wireless Sensor Networks Carlo Fischione Gateway concepts for WSN? Gateways are necessary to the Internet for remote access to/from the WSN ¾ Same is true for ad hoc networks; additional complications due to mobility (change route to the gateway; use different gateways) ¾ WSN: Additionally bridge the gap between different interaction semantics (data vs. address-centric networking) in the gateway Gateway needs support for different radios/protocols, … Internet Gateway node Remote users Wireless sensor network Fall 2009 Principles of Wireless Sensor Networks Carlo Fischione WSN to Internet communication Example: Deliver an alarm message to an Internet host Issues ¾ Need to find a gateway (integrates routing & service discovery) ¾ Choose “best” gateway if several are available ¾ How to find Alice or Alice’s IP? Alert Alice Alice‘s desktop Gateway nodes Internet Alice‘s PDA Fall 2009 Principles of Wireless Sensor Networks Carlo Fischione WSN tunneling Use the Internet to “tunnel” WSN packets between two remote WSNs Internet Gateway Gateway nodes Fall 2009 Principles of Wireless Sensor Networks Carlo Fischione Summary We have seen the most important aspects of wireless sensor networks ¾ Applications ¾ Difference from other networks ¾ Node architecture: controller, transceiver, memory, battery. Power consumption models. ¾ Design principles: distributed optimization, innetwork processing Next Lecture: Luca Mottola, SICS, “WSN Programming: from Abstractions to Running Code”. Fall 2009 Principles of Wireless Sensor Networks Carlo Fischione Homework Mandatory reading: ¾ Chapter 1 and 2, H. Karl and A. Willig, Protocols and Architectures for Wireless Sensor Networks, Wiley. ¾ Appendix A, D. P. Bertsekas, J. N. Tsitsiklis, Parallel and Distributed Computation: Numerical methods, Athena Scientific, 1997. Paper for presentation for next lecture: • A. Dunkels, B. Grönvall, T. Voigt, "Contiki - a Lightweight and Flexible Operating System for Tiny Networked Sensors", IEEE Emnets 2004. • P. Levis, S. Madden, D. Gay, J. Polastre, R. Szewczyk, A. Woo, E. Brewer, D. Culler, "The Emerging of Networking Abstractions and Techniques in TinyOS", ACM NSDI 2004. • V. Kawadia, P. R. Kumar, "A Cautionary Perspective on Cross Layer Design", IEEE Wireless Communications Magazine, Feb. 2005. ¾ ¾ Send me an e-mail with the preferred papers from the complete list by tomorrow Sept. 10 I will then assign the papers to the presenters by Sept 11. Fall 2009 Principles of Wireless Sensor Networks Carlo Fischione
