Smart Dust
Mohammad Rahimi
CSCI599-Spring2002
1
Back to future!


Colonies of smart ultra small size network
Interaction of advances in technology






VLSI more condense silicon processing
VLSI more silicon speed
VLSI low power short range communication
MEMS Micro-Electromechanical Systems
Networking : Ad hoc wireless Network
Distributed Processing : Operating system, Database
2
Wow!
3
Applications



It is a special class of sensor network
Fine sensing granularity
Applications :






Forest fire warning
Enemy troop monitoring
Large scale Biology or Geology
Smart office spaces
Defense-related sensor networks
Inventory Control
4
5
6
Circulatory Net
Group of sensors inside the body
7
Key Features of these electronic
particles

Power


Survive for extended amount of time
Computation


Sensors


Process Sensor Data and Communicate
To Interface to the environment
Communication

To glue the pieces of information
8
What is really behind the race?
Computer Science!

Ad hoc Networking



Data Fusion



Dynamic Reconfigurable network
Scalability
An efficient semantic to diffuse data in the network
Interpretation of multimodal sensing
Distributed Processing


Network Oriented Operating Systems
Data Aggregation
9
Come Back to Reality!
COTS Dust

Commercial Off-The-Shelf Components Dust

To enable us the research about the algorithms and
semantics
10
Architecture
Low Power
Communication Front End
1-Acoustic
2-RF
3-Optical
Computation Brain
1-MCU
2-FPGA
Sensors
1-Magnetometer
2- Light
3-temperature
4-pressure
5-Hmidity
6-Acceleration
Power
1-Bat
2-Sollar
3-Vibration
4-Acoustic Noise
11
Architecture
12
Power

Power: Lithium Battery




Big Problem
Low capacity per unit of mass and volume
Needs support by sleep mechanism and low power
techniques
Not really so much innovation after Volta!






Solar
Vibration
Acoustic noise
Thermal conversion
Nuclear Reaction
Fuel Cells
13
Computation


Computation: ATMEL91M404000
Micro Controller

Core and variety of different functions
Flash , SRAM , E2PROM
 GPIO , ADC , PWM ,Comparator
 Embedded serial Buses
 Ex: Microcontrollers
Atmel , Microchip, Motorola
Microprocessors Intel Strong-Arms ,Motorola


FPGA

Not Traditional Style Of programming



Hardware Descriptive Languages (VHDL, Verilog)
Faster , Low Power , Reconfigurable
Ex: Xilix,Atmel
14
Sensors

Motion Sensing

Magnetometer


Accelerometer


Study 3 Element of Earth Magnetic field (Compass)
To measure Local vertical (tilt switch) or measure motion vectors
Environmental Sensing(Weather Monitoring)

Pressure




Barometer
Temperature
Light
Humidity
15
Communication



Acoustic
RF radio
Optical


Passive
Active
16
Acoustic Communication






Power Hungry
High Background Noise
Large Size (proportional to harmonics of sound)
Fast Attenuation Curve
Low communication baud rate
Low power receiver

Good for event driven wake up systems
17
RF Communication

New low power techniques



ISM Band
For practical frequencies large ANT Size(~λ)


Not robust (No fancy Spread Spectrum)
It may have a tail!!
Only one RF front end

MCU acts as base band controller

CRC,Encoding(Manchester),DC Balance,Header
18
RF Communication
Train Sequence Flag Byte Number OF Byte
DATA
CRC
•No Addressing
19
Optical Communication

Active



High power laser source
Transmission of modulated laser beam
Passive



MEMS Corner Cube Reflector
Emit modulated ambient light
Extremely low power
20
21
Optical Vs. RF



Both mature technology
Both Electromagnetic waves
light is quiet short wavelength compared to RF


light techniques are extremely directional



Potential of smaller sizes
Peer-Peer
RF techniques can be more omni directional
Directionality Brings the problem of alignment
22
Future Work





Design of multi hop network
Autonomous network configuration
Data Fusion
Network Decision making
Large Scale Distributed Processing
23
Links







http://www.heise.de/tp/english/inhalt/co/5269/1.html
http://robotics.eecs.berkeley.edu/~pister/SmartDust/
http://www.thestandard.com/article/0,1902,27573,00.html
http://www.wired.com/news/technology/0,1282,44101,00.html
http://www-bsac.eecs.berkeley.edu/~warneke/SmartDust/
http://www.ce.berkeley.edu/~glaser/Current%20Civil%20Applications.ppt
http://madmax.me.berkeley.edu/%7Eshadr/overview.html
24
That is all folks!
25
Descargar

Smart Dust - USC Robotics Research Lab