Wireless Sensor Network (WSN)

1. Wireless Sensor Network (WSN) Computer Networking

2. Definition • Wireless network consists of spatially distributed autonomous devices • Using sensors to monitor physical or environment conditions • Originally motivated by military applications • Used in many civilian application areas • Each node equipped with radio transceiver • Small microcontroller • Sensor network constitutes a wireless ad-hoc network • Support multi-hop routing algorithm

3. Constraints and Challenges • On-board battery power • Limited network communication bandwidth • Microprocessor and small amount of memory • Fault tolerance • Transmission media • Hostile environment • Changing of topology • Costs • Scalability • Challenges faced: • Limited hardware • Limited support for networking • Limited support for software development

4. Basic view of WSN Sensor devices Multi-hop communication Link between Host Computer and Base Station Communicate with WSN External Power Supply Use base node to communicate with WSN Connected to base station Processing capability Receive/Send radio signal

5. Advantages of Sensor Networks • Improve signal-to-noise-ratio (SNR) • Reducing average distances from sensor to source • Increase energy efficiency due to multi-hop technology • Additional relevant information • Aggregated through multi-hopping • Improved robustness against individual sensor node or link failure • Improved scalability • Decentralized algorithms • Energy advantage • By multi-hop RF technology • Balancing between 2 factors • Overall cost • Energy efficiency • Detection advantage • Limited sensing range • Increase sensor density Psend(Nr) Psend(Nr) Preceive Nr r

6. Research Challenges • Sensor NetworkInformation Processing • Low power • Sampling frequency • Data aggregation and routing • Distributive, collaborative processing • Decentralized, decision, estimation • Correlated source communication, coding • Distributive data representation • Overlay data movement with underlying network routing • Robust and secure operation • Handle missing, faulty, delayed data • Privacy protection, intrusion resistant • Sensor Node System • High performance • Ability to sense • Ability to communicate • Ability to process data • Low manufacture cost • Millions can be used • Low deployment cost • Tether-free • self- configuration • Low maintenance cost • Long battery life • Fault tolerant, self diagnosis • Low disposition cost • Environment friendly

7. The Nature of WSN Architecture • Highest level • Decomposes a problem domain into set of services • Set of interfaces to its service • Lowest level • Specify its protocol • Packet formats • Communication exchange • State machine • Collaborative Processing • Process data cooperatively • Combine information from multiple sources • Data transferred from every sensor node to some other nodes • Number of sensor node increase, throughput goes to zero • Data can be processed locally to remove redundancy before shipping to remote node • Node can be more selective • Collaborative signal and information processing (CSIP) Sensor Middleware

8. Internet-based sensor network

9. WSN Abstract Tuple G • Defined as a complete set of information • G = <V, E, PV, PE> • Network graph with nodes V and links E • PV – characterizes the properties of each node in V • Location • Computational capability • Sensing modality • Acoustic • Seismic • Magnetic • IR • Temperature • Light • Sensor output type • Signal amplitude • Source direction-of-arrival (DOA) • Target range • Energy reserve • PE – properties for each link • Link capability • Link quality

10. Theory of tracking • Constrained optimization problem <G, T, W, Q, J, C> • G is the sensor network • T is a set of targets • Locations • Shape • Signal source type • W is signal model • Signal propagate • Attenuate in physical medium • Q is a set of user queries • Instances and entry points • J specifies an objective function • Task requirements • C specifies set of constraints • Limitation of time, energy

11. Tracking Scenario Reporting: Summarize tracking data and send back to query node Discovery: Detect a user, initialize tracking Communication: Hand off information to next node Collaborative processing: Estimates target location Query processing: Query enter the network

12. Collaborative localization • Amplitude measurement • Minimum of 3 measurements • TDOA (time difference of arrival) • Performance issues • Detectability • How reliably and timely • Detect object stimulus • Accuracy • System detection, localization • Scalability • Specific property of system • Survivability • Node or link failure • Resource Usage • Amount of resources consumes

13. Routing in WSN • Utilizes the radio links • Spared and used only when needed • Node talks directly to its immediate neighbors • Assume radio range a disk with radius r • Unit distance graph (UDG) • Network deployment is ad-hoc • Unstable links • Node failures • Network disconnections • Operate wirelessly with limited power • Multi-path provides energy efficiency • Limited or no mobility • Dynamic topology (sleeping, falling) • Nodes know their geographical position • Sensor network generate different MAC – specialized • Sensor network are collaborative systems • Most sensor nodes are idle much of the time • Improve bandwidth utilization • Lack of mobility taken into account • Issues of energy efficiency, scalability, robustness

14. MAC protocol • Wireless cellular networks • TDMA (time division multiple access) • FDMA (frequency division multiple access) • CDMA (code division multiple access) • Ethernet and WLAN • CSMA (carrier sense multiple access) • S-MAC (for WSN) • Reduce energy waste • Idle listening, collisions, overhearing • Major components • Periodic listen and sleep • Collision avoidance • Overhearing avoidance • Message passing

15. Strategies for Routing • Reactive protocols, constructing path on demand only • Dynamic source routing (DSR) • Ad hoc on demand distance vector routing (AODV) • Local stateless algorithms • Greedy distance routing • Pick the closet to destination among the neighbors • Compass routing • Pick the one that minimizes the angle to destination

16. Strategies for Routing • Geographical Routing • All nodes know their geographical location • Knows its immediate one-hop neighbors • Routing destination is specified • Each packet hold bounded amount of additional routing information • Attribute-based Routing • Node’s location • Type of sensors • Certain range of values in a certain type of sensed data • Used attribute value pairs to describe the data

17. Time Synchronization in WSN • Latency time • Send time • Access time • Propagation time • Receive time • Three message exchange • Delay D • Phase difference d • Using spanning tree favoring direct connections with reliable delays P1 = t5 – t4 I1 = t2 – t1 Communication delay d

18. Different Types of Queries • Queries for relations among set of events • “sound an alarm whenever two sensors within 10 meters of each other simultaneously detect an abnormal temperature” • Time-dependent query • Long-running, continuous queries • Snapshot queries • Historical queries • In summary • Conditions restricting set of sensors from contributing data • Correlate data from different sensors • Trigger data collection or signal processing • Generate sub-queries • Able to use outputs from past queries as inputs for further queries • Accurate time synchronization • Node geographical location information

19. Database Organization in WSN • Centralized warehousing approach • Sensor sends data to central server • Central server connected to network via access point • Nodes near access point become traffic hot spots • Sampling rates set to highest • In-network storage approach • Rendezvous (meeting) points between data and queries selected • Overhead to store and access data is minimized • Overall load is balance across network • Reduce communication overhead • Data can be aggregated • TinyDB query processing • SQL-style declarative query interface • Sensitive to resource constraints and lossy communication • Uses an epoch-based mechanism (divided into time intervals) • Data Centric Storage (DCS) • Names and stores data by its (physical) attributes • Translate attributes into a node location or ID • Distribute storage load across the entire network • Geographical Hash Table (GHT) utilized hash functions

20. Sensor Network Programming • Operating System : TinyOS • Support sensor network application on resource-constrained hardware platforms (Berkeley motes) • No file system • Support only static memory allocation • Simple task model • Minimal devise and networking abstractions • OS is compiled with applications • Lower layer closer to the hardware • Higher layer closer to application • Unique component architecture • Provides as a library set of system software components • Components include software functionalities • Thin wrappers around hardware • Wires those components together with other components • Task • Scheduler, created by components • Maintains a task queue • Run to completion, non pre-emptive • Invokes new task from queue or put CPU to sleep mode • Events • Ultimate sources of triggered executions • Execution of interrupt handler • Run to completion but can pre-empt tasks and can be pre-empted by other events

21. Samples of wireless sensor hardware

22. Sensor Network Programming • Imperative language: nesC • Extension of C to support and reflect design of TinyOS • Interfaces • Provides interface • Set of method calls exposed to the upper layers • Uses interface • Set of method calls hiding the lower layer components • Implementation • Modules • Implemented by application code (C-like syntax) • Configurations • Connecting interfaces of existing components • Asynchronous code (AC) • Code reachable from at least one interrupt handler • Synchronous code (SC) • Code reachable only from tasks

23. Sensor Network Programming • Simulators • Quickly study the performance of algorithms • Timing • Power • Bandwidth • Scalability • Sensor node model • Software execution platform • Sensor host • Communication terminal • Communication model • Model the communication at physical layer • Stimulate RF propagation delay • Collision of simultaneous transmission • Physical environment model • Key element is physical phenomenon of interest • Statistics and visualization • Results collected for analysis • Cycle-driven (CD) • Discretizes the continuous notion of real time into ticks • Discrete-event (DE) • Assume time is continuous and an event may occur at any time

24. Applications using WSN • Emerging applications • Build on WLAN or Bluetooth technology • ZigBee and Ultra-Wide Band (UWB) • Asset and warehouse management • Monitor and track assets • Collect real-time inventory and retail information by using RFID technology • Automotive applications • Dedicated short-range communication (DSRC) • Telematics and entertainment, link to databases • Building monitoring and control • Cut down energy costs • Detect biological agents or chemical pollutants • Security system • Environmental monitoring • Monitor conditions and movements of wild animals or plants • Monitor air quality and track environmental pollutants, biological or chemical hazards • Healthcare applications • Monitor vital signs of patients, remotely connected to other places • In-home elderly healthcare system • Industrial process control • Monitor manufacturing process, condition of industrial equipment • Military battlefield awareness • Real-time battlefield intelligence • Security and surveillance • Imagers or video sensors useful in identifying and tracking moving entities

25. Thank You Question And Answer