Course “Algorithmic Foundations of SensorNetworks”

Lecture 1: Introduction / Data propagationalgorithms I

Sotiris NikoletseasProfessor

Department of Computer Engineering and InformaticsUniversity of Patras, Greece

Spring Semester 2017-2018

1 / 51

Summary of this Lecture

• Introduction to Wireless Sensor Networks (characteristics,applications, critical challenges, algorithmic properties)

• Thematic structure of the 2017-2018 Course• The problem of data propagation• Two protocols for data propagation:

• Directed Diffusion (data-centric)• LEACH (cluster-based)

2 / 51

What is a Wireless Sensor Network?

• very large number of tiny “smart” sensors• severe limitations• wireless communication• densely / randomly deployed in an area• self-organization• co-operation• locality

⇒ An “ad-hoc” wireless network for:• sensing crucial events and ambient conditions• data propagation

3 / 51

A Sensor Network - Graphical Representation

Sensor nodes Sensor field

Control Center

4 / 51

“Smart” Sensor Nodes

• very small size• operate on a small battery• multifunctional sensing• wireless communication• limited computing power / limited memory• low cost

5 / 51

Technical Specifications of Current Sensors

TPR2420 Specifications (TelosB motes)

Dimensions 6.5cm × 3.1cm × 2cmCPU 8 MHz Texas Instruments MSP430Memory 10KB RAM, 48KB Program flash, 16KB EEPROM, 1MB flashPower Supply 2X AA batteriesProcessor Current Draw 1.8 mA (active current)

< 5.1 µA (sleep mode)Interface USBNetwork Wireless 250 Kbps at 2.4GHz (802.15.4)

Radio range depends on configuration and locationIntegrated Sensors Visible Light and IR, Humidity, Temperature

6 / 51

Technical Specifications of Current Sensors

TinyOS Key Facts

Current Version TinyOS 2.0.2Users Over 500 research groups and companiesSoftware Footprint 3.4 KbArchitecture Component-basedHardware Supported 8 different mote typesPeak load CPU Usage < 50 %Prog. Language Used NesCComponents provided Sensor Interface

Basic Multihop RoutingRadio Communication

7 / 51

Examples of Sensor Motes

Mica2TelosB / Tmote Sky

8 / 51

Smart vs Traditional Sensors

In traditional sensors:

• sensors perform only sensing• sensor positions are carefully engineered• deployed far from the area of interest• data processing only at some central nodes• larger in size• constant power supply

9 / 51

The First Projects (I)

• “Smart Dust” (Berkeley) - 1998• primary goal: a few mm3 size• laser communication• motes development• TinyOS

(http://robotics.eecs.berkeley.edu/∼pister/SmartDust/)

10 / 51

The First Projects (II)

• WINS (“Wireless Intergrated Network Sensors”) (UCLA) - 1993• focus : RF communication• development of low power MEMS

(http://www.seas.ucla.edu/∼pottie/papers/smallWINS_ACM.pdf )

• “Ultralow Power Wireless Sensor” (MIT)• primary goal : extremely low power consumption• power levels : 10 µW to 10 mW• RF communication

11 / 51

Current Major Projects (I) - Partial List

• WEBS (“Wireless Embedded Systems”, David Culler)(Berkeley)

• vision: “. . . highly embedded and networked devices, ofteninteract with people and their physical environment, and arehighly constrained in resources such as computation andenergy . . .”.

• several subprojects (SMAP, blip, ACme, MotescopeTestbed, tinyOS, LoCal)

(http://smote.cs.berkeley.edu:8000/tracenv/wiki)

• NEST (“Network Embedded Systems Technology”)• originally a DARPA funded project on middleware services

and novel applications of WSNs. Today, an umbrella projectencompassing diverse ongoing research activities in theWSN domain.

• vision: “Ambient sensing and computing in the real world”.• includes several major groups (Berkeley, Virginia,

Vanderbilt etc.)

12 / 51

Current Major Projects (II) - Partial List

• CENS(“Center for Embedded Networked Sensing”, DeborahEstrin) (UCLA)

• an NSF Science and Technology Center (2001-2011)• research topics: multiscaled automated sensing, terrestrial

ecology observing systems (TEOS), aquatic observingsystems, embeddable sensors

(http://www.cens.ucla.edu)

• NESL (“Networked & Embedded Systems Laboratory”, ManiSrivastava)(UCLA)

• wireless sensor and actuator networks, low powernetworking, mobile and wireless computing, pervasivecomputing

(http://nesl.ee.ucla.edu)

• NMS (“Networks and Mobile Systems”, Hari Balakrishnan)(MIT)• topics: wireless networks, P2P networks, sensor networks,

networked systems

(http://nms.csail.mit.edu)13 / 51

The Low Energy Challenge

• In a cubic millimeter volume⇒ current best battery : 1 Joule

• For continuous operation over 1 day :

⇓• power consumption ' 10 µW

⇒{

efficient power management strategies neededefficient networking algorithms required

14 / 51

RF vs Laser Communications

• Radio Frequency :• small mote size⇒ limited space for antennas⇒ extremely

short-wavelength⇒ high-frequency transmissions• relatively complex circuits

• Laser• significantly lower power consumption (cheap, narrow

beams emitted)• free line-of-sight paths required• passive (no power) optical transmission techniques

(mirror-based retroreflection)

15 / 51

Applications (I)

Sensors for a wide variety of conditions:• temperature• motion• noise levels• light conditions• humidity• object presence• mechanical stress levels on attached objects

16 / 51

Applications (II)

• continuous sensing• detection of a crucial event

• location sensing (including motion tracking)• local control of actuators (ambient intelligence)• micro-sensing

17 / 51

Applications (III)

• Environmental applications• forest fire detection• flood detection• precision agriculture• weather forecast• planetary exploration

• Health applications• telemonitoring of human physiological data• tracking and monitoring doctors/patients

18 / 51

Applications (IV)

• Home applications• home automation (local/remote management of home

devices)• smart environments (adapt to the people’s needs)

• More applications (partial list)• inventory control• security / military• vehicle tracking and detection• interactive museum• monitoring material fatigue

19 / 51

Emerging Trends (I)

The Internet of Things (IoT)• A vision: the Digital and Physical Worlds converge

• Pervasive Technologies: “The most profound technologies are those thatdisappear. They weave themselves into the fabric of everyday life until theyare indistinguishable from it” (Mark Weiser)

• Embedded Sensor Networks: “Google for the Physical World” (DeborahEstrin)

• IoT: massive, seamless, wireless inclusion of everyday objects (clothes,vehicles, glasses etc.) and devices (lights, HVAC, refrigerator etc.) intothe WEB, so that they become “smart” and can participate into smartscenaria and services.

20 / 51

Emerging Trends (II)

Mobile Sensing/Crowdsensing• Current smartphones are truly mobile

and include several sensors (light,accelerometer, gyroscope, proximityetc.).

• Sensory readings can be obtained fromeverywhere and all the time, from the“crowd”.

• Also, these sensor hints can be used toextract human features (e.g. whethersomebody is walking or lying etc.).

• The above sensor readings and hintscan be exploited to optimize smartscenaria and services, in view of thisfeedback from the crowd(crowdsourcing).

21 / 51

Sensors vs Ad-hoc Wireless Networks

• much larger number of devices• more dense deployment / interactions / complexity• frequent failures• more limited capabilities• more frequently/more dynamically changing topology• less global knowledge

22 / 51

New/Critical Challenges (I)

• Scalability• huge number of sensor nodes• high densities of nodes• many complex interactions

• Success rate: percentage of delivered dataHow does protocol performance scale with size?Even correctness may be affected by size.

23 / 51

New/Critical Challenges (II)

• Fault-tolerance

Sensors may• fail (temporarily or permanently)• be blocked / removed• cease communication

due to various reasons• physical damage• power exhaustion• interference• power saving mechanisms

Can the network tolerate failures well?

24 / 51

New/Critical Challenges (III)

• Efficiency• energy spent• time (for data propagation)

25 / 51

New/Critical Challenges (IV)

• Inherent trade-offs (e.g. energy vs time)• Competing goals / diverse aspects:

- minimizing total energy spent in the network- maximizing the number of “alive” sensors over time- combining energy efficiency and fault-tolerance- balancing the energy dissipation

• Application dependence• Dynamic changes / heterogeneity

⇓• variety of protocols needed / hybrid combinations• adaptive protocols, locality• simplicity, randomization, distributedness

26 / 51

Thematic structure of the course - I

• Introduction (technology, applications, challenges,algorithmic performance properties)

• Network deployment, connectivity, coverage• Data propagation algorithms I (data-centric, cluster-based)• Data propagation algorithms II (greedy probabilistic

routing)• Energy balance and optimization• Geographic routing and obstacle avoidance algorithms• Localization and tracking

27 / 51

Thematic structure of the course - II

• Wireless characteristics (interference, energy cost models)• Sleep-awake power saving and topology control schemes• Medium access control (MAC) aspects and protocols• Time synchronization• Mobile sensor networks• Wirelessly Rechargeable sensor networks• Implementation and engineering of algorithms for sensor

networks• Application development environments

28 / 51

A “canonical” problem: Local Event Detection andData Propagation

A single sensor, p, senses a local event E . The generalpropagation problem is the following:

“How can sensor p, via cooperation with the rest of thesensors in the network, propagate information aboutevent E to the control center?”

29 / 51

Representative data propagation protocols

• Directed Diffusion (DD): a tree-structure protocol (suitablefor low dynamics)

• LEACH: clustering (suitable for small area networks)• Local Target Protocol (LTP): local optimization (best for

dense networks)• Probabilistic Forwarding Protocol (PFR): redundant

optimized transmissions (good efficiency / fault-tolerancetrade-offs, best for sparse networks)

• Energy Balanced Protocol (EBP): guaranteeing same persensor energy (prolong network life-time)

30 / 51

Directed Diffusion

• requires some coordination between sensors• creates / maintains some global structure (set of paths)• a paradigm / suite of several protocols• here: a tree-based version

31 / 51

Directed Diffusion elements

• Interest messages (issued by the control center)• Gradients (towards the control center)• Data messages (by the relevant sensors)• Reinforcements of gradients (to select “best” paths)

32 / 51

Interests

• An interest contains the description of a sensing task.• Task descriptions are named e.g. by a list of attribute-value

pairs.• The description specifies an interest for data matching the

attributes.

type = wheeled vehicle // detect vehicle locationinterval = 10 ms // send events every 10 msduration = 10 minutes // for the next 10 minutesrect = [-100, 100, 200, 400] // from sensors within rectangle

Example of an interest

• Interests are injected into the network at the control center

33 / 51

Gradients

For each interest message received, a gradient is created

Gradients are formed by local interaction of neighboring nodes,establishing a gradient towards each other.

Gradients store a value (data rate) and a direction (towards thesink) for “pulling down” data.

34 / 51

Data propagation

• A sensor that receives an interest it can serve, beginssensing.

• As soon as a matching event is detected, data messagesare sent to the relevant neighbors using the gradientsestablished.

• A data cache is maintained at each sensor recordingrecent history.

35 / 51

Reinforcement for Path Establishment

The sink initialy repeatedly diffuses an interest for a low-rateevent notification, through exploratory messages.

The gradients created by exploratory messages are calledexploratory and have low data rate.

As soon as a matching event is detected, exploratory eventsare generated and routed back to the sink.

After the sink receives those exploratory events, it reinforcesone (or more) particular neighbor in order to “draw down” realdata.

The gradients that are set up for receiving high data rateinformation are called data gradients.

36 / 51

Path Establishment Using Positive Reinforcement

To reinforce a neighbor, the sink re-sends the original interestmessage with a higher rate.

Upon reception of this message a node updates thecorresponding gradient to match the requested data rate.

The selection of a neighbor for reinforcement is based on localcriteria, i.e.:• the neighbor that reported first a new event is reinforced.• the higher data rate neighbor is reinforced.• more than one neighbors are reinforced.

The data cache is used to determine which criteria are fulfilled.

37 / 51

Negative Reinforcement

Negative reinforcement is applied when certain criteria are meti.e. a gradient does not deliver any new messages for anamount of time, or a gradient has a very low data rate, etc.

38 / 51

Summary Evaluation of Directed Diffusion

• improves over flooding a lot• significant energy savings• performance drops when network dynamics are high

39 / 51

Low Energy Adaptive Clustering Hierarchy (LEACH)

• The network is partitioned in clusters.• Each cluster has one cluster-head.• Each non cluster-head sends data to the head of the

cluster it belongs to.• Cluster-heads gather the sent data, compress it and send

it to the sink directly.

40 / 51

Dynamic Clusters

41 / 51

Randomized rotation of cluster-heads

• Node n decides with probability T(n) to elect itselfcluster-headP: the desired percentage of cluster heads.r : the current round.G: the set of nodes that have not been cluster-heads in thelast 1

P rounds.• T(n) is chosen so as to get on the average the same

number of cluster-heads in each round.

T (n) =

{P

1−P∗(rmod 1P )

if n ∈ G

0 otherwise

42 / 51

Cluster-head advertisement

• Each cluster-head broadcasts an advertisement messageto the rest nodes using a CSMA-MAC protocol.

• Non cluster-head nodes hear the advertisements of allcluster-head nodes.

• Each non cluster-head node decides its cluster-head bychoosing the one that has the stronger signal.

43 / 51

Cluster set-up phase

• Each cluster-head is informed for the members of itscluster.

• The cluster-head creates a TDMA schedule• The cluster-head broadcasts the schedule back to the

cluster members.

44 / 51

Steady Phase

• Non cluster-heads• Sense the environment.• Send their data to the cluster head during their

transmission time.• Cluster-heads

• Receive data from non cluster-head nodes.• Compress the data received.• Send their data directly to the base station.

45 / 51

Summary Evaluation of LEACH

• reduces energy dissipation through compression of data atcluster-heads.

• in large networks, direct transmissions are very expensive.• performance drops when the network traffic is high (e.g.

many events generated).

46 / 51

Papers related to this Lecture

• W.R. Heinzelman, A. Chandrakasan & H. Balakrishnan,“Energy-Efficient Communication Protocol for WirelessMicrosensor Networks”, 33rd Hawaii InternationalConference on System Sciences, HICCS-2000.

• C. Intanagonwiwat, R. Govindan, & D. Estrin, “DirectedDiffusion: A Scalable and Robust CommunicationParadigm for Sensor Networks”, 6th Annual InternationalConference on Mobile Computing and Networking(MobiCOM ’00), August 2000.

47 / 51

State of the art Surveys

- Akyildiz, I.F., Su, W., Sankarasubramaniam, Y., andCayirci, E.: Wireless Sensor Networks: a Survey. In theJournal of Computer Networks 38 (2002) 393-422.

- Estrin, D., Govindan, R., Heidemann, J., and Kumar, S.:Next Century Challenges: Scalable Coordination in SensorNetworks. In Proc. 5th ACM/IEEE InternationalConference on Mobile Computing (MOBICOM), 1999.

- Kahn, J.M., Katz, R.H., and Pister, K.S.J.: Next CenturyChallenges: Mobile Networking for Smart Dust. In Proc.5th ACM/IEEE International Conference on MobileComputing (September 1999) 271-278.

48 / 51

Books (I)

• F. Zhao and L. Guibas, Wireless Sensor Networks: AnInformation Processing Approach, Morgan Kaufmann,2004.

• C. Raghavendra and K.Sivalingam (Editors), WirelessSensor Networks, Springer Verlag, 2006.

• Bhaskar Krishnamachari, Networking Wireless Sensors,Cambridge University Press, December 2005.

49 / 51

Books (II)

• S. Nikoletseas and J. Rolim, Theoretical Aspects ofDistributed Computing in Sensor Networks, SpringerVerlag, 2011.

50 / 51

Book Themes

T1. Challenges for Wireless Sensor NetworksT2. Models, Topology, ConnectivityT3. Localization, Time Synchronization, CoordinationT4. Data Propagation and CollectionT5. Energy OptimizationT6. Mobility ManagementT7. Security AspectsT8. Tools, Applications and Use Cases

51 / 51