![Page 1: EE360: Lecture 16 Outline Sensor Network Applications and In-Network Processing](https://reader036.vdocument.in/reader036/viewer/2022070418/568158f6550346895dc631e7/html5/thumbnails/1.jpg)
EE360: Lecture 16 OutlineSensor Network Applications
and In-Network Processing
Announcements2nd summary due today 12am (1 day extension
possible)Project poster session March 15 5:30pm (3rd floor
Packard)Next HW posted by tonight, due March 16Will extend final project deadline
Overview of sensor network applications Technology thrusts Cross-layer design of sensor network protocols Cooperative compression Distributed sensing, communications, and
control
![Page 2: EE360: Lecture 16 Outline Sensor Network Applications and In-Network Processing](https://reader036.vdocument.in/reader036/viewer/2022070418/568158f6550346895dc631e7/html5/thumbnails/2.jpg)
2
Wireless Sensor NetworksData Collection and Distributed Control
•Hard Energy Constraints•Hard Delay Constraints•Hard Rate Requirements
![Page 3: EE360: Lecture 16 Outline Sensor Network Applications and In-Network Processing](https://reader036.vdocument.in/reader036/viewer/2022070418/568158f6550346895dc631e7/html5/thumbnails/3.jpg)
3
Application Domains Home networking: Smart appliances, home security,
smart floors, smart buildings
Automotive: Diagnostics, occupant safety, collision avoidance
Industrial automation: Factory automation, hazardous material control
Traffic management: Flow monitoring, collision avoidance
Security: Building/office security, equipment tagging, homeland security
Environmental monitoring: Habitat monitoring, seismic activity, local/global environmental trends, agricultural
![Page 4: EE360: Lecture 16 Outline Sensor Network Applications and In-Network Processing](https://reader036.vdocument.in/reader036/viewer/2022070418/568158f6550346895dc631e7/html5/thumbnails/4.jpg)
4
Wireless Sensor Networks
Revolutionary technology.
Hard energy, rate, or delay constraints change fundamental design principles
Breakthroughs in devices, circuits, communications, networking, signal processing and crosslayer design needed.
Rich design space for many industrial and commercial applications.
![Page 5: EE360: Lecture 16 Outline Sensor Network Applications and In-Network Processing](https://reader036.vdocument.in/reader036/viewer/2022070418/568158f6550346895dc631e7/html5/thumbnails/5.jpg)
5
Technology Thrusts
Wireless Sensor
Networks
Analog Circuits• Ultra low power• On-chip sensor• Efficient On/Off • MEMS• Miniaturized size• Packaging tech.• Low-cost imaging
Networking• Self-configuration• Scalable • Multi-network comm.• Distributed routing and scheduling
Wireless• Multi-hop routing• Energy-efficiency• Very low duty cycle• Efficient MAC•Cooperative Comm.
Data Processing• Distributed • Sensor array proc.• Collaborative detection/accuracy improvement•Data fusion
System-on-Chip• Integration of sensing, data processing, and communication in a single, portable, disposable device
Applications
![Page 6: EE360: Lecture 16 Outline Sensor Network Applications and In-Network Processing](https://reader036.vdocument.in/reader036/viewer/2022070418/568158f6550346895dc631e7/html5/thumbnails/6.jpg)
Crosslayer Protocol Design
in Sensor Networks
ApplicationNetworkAccessLink
Hardware
Protocols should be tailored to the application requirements and
constraints of the sensor network
![Page 7: EE360: Lecture 16 Outline Sensor Network Applications and In-Network Processing](https://reader036.vdocument.in/reader036/viewer/2022070418/568158f6550346895dc631e7/html5/thumbnails/7.jpg)
Cross-Layer Design with Cooperation
Multihop Routing among Clusters
![Page 8: EE360: Lecture 16 Outline Sensor Network Applications and In-Network Processing](https://reader036.vdocument.in/reader036/viewer/2022070418/568158f6550346895dc631e7/html5/thumbnails/8.jpg)
Double String Topology with Alamouti Cooperation
Alamouti 2x1 diversity coding schemeAt layer j, node i acts as ith antenna
Synchronization required Local information exchange not required
![Page 9: EE360: Lecture 16 Outline Sensor Network Applications and In-Network Processing](https://reader036.vdocument.in/reader036/viewer/2022070418/568158f6550346895dc631e7/html5/thumbnails/9.jpg)
Equivalent Network with Super Nodes
Each super node is a pair of cooperating nodes
We optimize:link layer design (constellation size
bij)MAC (transmission time tij)Routing (which hops to use)
![Page 10: EE360: Lecture 16 Outline Sensor Network Applications and In-Network Processing](https://reader036.vdocument.in/reader036/viewer/2022070418/568158f6550346895dc631e7/html5/thumbnails/10.jpg)
Minimum-energy Routing (cooperative)
![Page 11: EE360: Lecture 16 Outline Sensor Network Applications and In-Network Processing](https://reader036.vdocument.in/reader036/viewer/2022070418/568158f6550346895dc631e7/html5/thumbnails/11.jpg)
Minimum-energy Routing (non-cooperative)
![Page 12: EE360: Lecture 16 Outline Sensor Network Applications and In-Network Processing](https://reader036.vdocument.in/reader036/viewer/2022070418/568158f6550346895dc631e7/html5/thumbnails/12.jpg)
MIMO v.s. SISO(Constellation Optimized)
![Page 13: EE360: Lecture 16 Outline Sensor Network Applications and In-Network Processing](https://reader036.vdocument.in/reader036/viewer/2022070418/568158f6550346895dc631e7/html5/thumbnails/13.jpg)
Delay/Energy Tradeoff
Packet Delay: transmission delay + deterministic queuing delay
Different ordering of tij’s results in different
delay performance
Define the scheduling delay as total time needed for sink node to receive packets from all nodes
There is fundamental tradeoff between the scheduling delay and total energy consumption
![Page 14: EE360: Lecture 16 Outline Sensor Network Applications and In-Network Processing](https://reader036.vdocument.in/reader036/viewer/2022070418/568158f6550346895dc631e7/html5/thumbnails/14.jpg)
Minimum Delay Scheduling
The minimum value for scheduling delay is T (among all the energy-minimizing schedules): T= tij
Sufficient condition for minimum delay: at each node the outgoing links are scheduled after the incoming links
An algorithm to achieve the sufficient condition exists for a loop-free network with a single hub node
An minimum-delay schedule for the example: {2!3, 1!3, 3!4, 4!5, 2!5, 3!5}
1 2
3 4
5
T T
4!5 2!5 3!51!32!3 3!4
![Page 15: EE360: Lecture 16 Outline Sensor Network Applications and In-Network Processing](https://reader036.vdocument.in/reader036/viewer/2022070418/568158f6550346895dc631e7/html5/thumbnails/15.jpg)
Energy-Delay Optimization
Minimize weighted sum of scheduling delay and energy
![Page 16: EE360: Lecture 16 Outline Sensor Network Applications and In-Network Processing](https://reader036.vdocument.in/reader036/viewer/2022070418/568158f6550346895dc631e7/html5/thumbnails/16.jpg)
Transmission Energy vs. Delay
![Page 17: EE360: Lecture 16 Outline Sensor Network Applications and In-Network Processing](https://reader036.vdocument.in/reader036/viewer/2022070418/568158f6550346895dc631e7/html5/thumbnails/17.jpg)
Total Energy vs. Delay
![Page 18: EE360: Lecture 16 Outline Sensor Network Applications and In-Network Processing](https://reader036.vdocument.in/reader036/viewer/2022070418/568158f6550346895dc631e7/html5/thumbnails/18.jpg)
Transmission Energy vs. Delay
(with rate adaptation)
![Page 19: EE360: Lecture 16 Outline Sensor Network Applications and In-Network Processing](https://reader036.vdocument.in/reader036/viewer/2022070418/568158f6550346895dc631e7/html5/thumbnails/19.jpg)
Total Energy vs. Delay(with rate adaptation)
![Page 20: EE360: Lecture 16 Outline Sensor Network Applications and In-Network Processing](https://reader036.vdocument.in/reader036/viewer/2022070418/568158f6550346895dc631e7/html5/thumbnails/20.jpg)
Cooperative Compression
Source data correlated in space and time
Nodes should cooperate in compression as well as communication and routing Joint source/channel/network coding
![Page 21: EE360: Lecture 16 Outline Sensor Network Applications and In-Network Processing](https://reader036.vdocument.in/reader036/viewer/2022070418/568158f6550346895dc631e7/html5/thumbnails/21.jpg)
Cooperative Compression and
Cross-Layer Design
Intelligent local processing can save power and improve centralized processing
Local processing also affects MAC and routing protocols
![Page 22: EE360: Lecture 16 Outline Sensor Network Applications and In-Network Processing](https://reader036.vdocument.in/reader036/viewer/2022070418/568158f6550346895dc631e7/html5/thumbnails/22.jpg)
Energy-efficient estimation
We know little about optimizing this systemAnalog versus digital Analog techniques (compression, multiple access)Should sensors cooperate in compression/transmissionTransmit power optimization
Sensor 1
Sensor 2
Sensor K
Fusion Center
Different channel gains (known)
Different observation quality (known)
1P
2P
KP
)(t
02)ˆ( DE
g1
g2
gK
21
22
2K
![Page 23: EE360: Lecture 16 Outline Sensor Network Applications and In-Network Processing](https://reader036.vdocument.in/reader036/viewer/2022070418/568158f6550346895dc631e7/html5/thumbnails/23.jpg)
Digital vs. Analog
![Page 24: EE360: Lecture 16 Outline Sensor Network Applications and In-Network Processing](https://reader036.vdocument.in/reader036/viewer/2022070418/568158f6550346895dc631e7/html5/thumbnails/24.jpg)
Key Message
Cross-layer design imposes tradeoffs between rate, power/energy, and delay
The tradeoff implications for sensor networks and distributed control is poorly understood
![Page 25: EE360: Lecture 16 Outline Sensor Network Applications and In-Network Processing](https://reader036.vdocument.in/reader036/viewer/2022070418/568158f6550346895dc631e7/html5/thumbnails/25.jpg)
Distributed Sensing, Communications, and
Control
Controller
System
System
Controller
![Page 26: EE360: Lecture 16 Outline Sensor Network Applications and In-Network Processing](https://reader036.vdocument.in/reader036/viewer/2022070418/568158f6550346895dc631e7/html5/thumbnails/26.jpg)
Applications
![Page 27: EE360: Lecture 16 Outline Sensor Network Applications and In-Network Processing](https://reader036.vdocument.in/reader036/viewer/2022070418/568158f6550346895dc631e7/html5/thumbnails/27.jpg)
Joint Design ofControl and
Communications- Generally apply different design principles
Control requires fast, accurate, and reliable feedback. Networks introduce delay and loss for a given rate.
- Sensors must collect data quickly and efficiently
- The controllers must be robust and adaptive to random delays and packet losses.
- Control design today is highly sensitive to loss and delay
- The networks must be designed with control performance as the design objective.
Network design tradeoffs (throughput, delay, loss) become implicit in the control performance index
This complicates network optimization
![Page 28: EE360: Lecture 16 Outline Sensor Network Applications and In-Network Processing](https://reader036.vdocument.in/reader036/viewer/2022070418/568158f6550346895dc631e7/html5/thumbnails/28.jpg)
A Proposed Architecture
OnlineOptimization(RHC, MILP)
Sensing
External Environment
SystemController
StateServer
(KF -> MHE)
Inner Loop(PID, H)
Sensing
OnlineOptimization(RHC, MILP)
Mode andFault
Management
StateServer
(KF -> MHE)
Online Model
OnlineOptimization(RHC, MILP)
Goal Mgmt(MDS)
Attention &Awareness
Memory andLearning
StateServer
(KF, MHE)
![Page 29: EE360: Lecture 16 Outline Sensor Network Applications and In-Network Processing](https://reader036.vdocument.in/reader036/viewer/2022070418/568158f6550346895dc631e7/html5/thumbnails/29.jpg)
Potential Pieces of the Puzzle
Local autonomySubsystems can operate in absence of global data
Estimation, prediction, and planningExploit rich set of existing tools
Command buffering and prefetching Increases tolerance to data latency and loss
Time stamps and delay-adaptive control
Modular designSupervisory control via models, cost functions,
modes
![Page 30: EE360: Lecture 16 Outline Sensor Network Applications and In-Network Processing](https://reader036.vdocument.in/reader036/viewer/2022070418/568158f6550346895dc631e7/html5/thumbnails/30.jpg)
Summary
Cross layer design especially effective in sensor networks.
Node cooperation can include cooperative compressionCooperative gains depend on network
topology and application.
Cross layer design must optimize for applicationRequires interdisciplinary understanding, e.g.
for control
![Page 31: EE360: Lecture 16 Outline Sensor Network Applications and In-Network Processing](https://reader036.vdocument.in/reader036/viewer/2022070418/568158f6550346895dc631e7/html5/thumbnails/31.jpg)
Presentation
An application-specific protocol architecture for wireless microsensor networks
By W. Heinzelman, A. P. Chandrakasan and H. Balakrishnan
Presented by Mainak Chowdhury