A Whirlwind Tour of Sensor Networks

1
A Whirlwind Tour of Sensor Networks

• CEE 249
• November 26, 2001
• Robert Szewczyk

2
Mote Hardware
3
Current Generation Platform

• 1 x 1.5 motherboard
• ATMEL 4Mhz, 8bit MCU, 512 bytes RAM, 8K pgm flash
• 900Mhz Radio (RF Monolithics) 1-100 ft. range
• ATMEL network pgming assist
• 10 bit A/D converter, 15 ksps
• Radio Signal strength control and sensing
• I2C EPROM (logging) 32 kBytes of storage
• Base-station ready (UART)
• stackable expansion connector
• Analog voltage lines, serial busses, debug pins
• Several sensor boards
• basic protoboard
• tiny weather station (temp,light,hum,prs)
• vibrations (2d acc, temp, light)
• accelerometers, magnetometers,
• current, acoustics

4
Next Generation Mote

• Mica board
• Form factor of a AA battery pack
• 900 MHz radio driven at 50 kbps (RFM chip)
• ATMEL 8-bit CPU, 128 kBytes of pgm flash, 4
  kBytes of RAM
• 512 kBytes of nonvolatile storage
• Same connector as on the Rene board
• Improvements for sensor interfacing analog
  comparator, interrupt lines
• Default Sensor board
• Light, temperature
• 2d accelerometer
• Microphone
• Sounder (fixed frequency speaker)
• 2d magnetometer

5
TinyOS Software Architecture
6
Operating System for Small Devices

• Why create an OS for 8 kByte device
• Define a basic execution model
• provide framework for concurrency and modularity
• Scheduler and graph of components
• constrained two-level scheduling model tasks
  events
• Provides a component based model abstracting
  hardware specifics from application programmer
• Capable of maintaining fine grained concurrency
• Can interchange system components to get
  application specific functionality

7
TinyOS Component Model

• Component has
• Frame (storage)
• Tasks (computation)
• Command and Event Interface
• Constrained Storage Model allows compile time
  memory allocation
• Provides efficient modularity
• Explicit Interfaces help with robustness
• allow appropriate abstractions to emerge, e.g.
  initialization, energy management interfaces,
  etc.
• Concurrency model
• never poll, never block
• Tasks atomic w.r.t. other tasks, interrupted by
  events

Messaging Component
Internal State
Internal Tasks
Commands
Events
8
Appln Graph of Event-Driven Components
Route map
router
sensor appln
application
Active Messages
Radio Packet
Serial Packet
packet
Temp
photo
SW
HW
UART
Radio byte
ADC
byte
Example ad hoc, multi-hop routing of photo
sensor readings
clocks
RFM
bit
9
TinyOS Communication Model

• Physical layer implications
• Single channel radio all communication is a
  local broadcast, need to decide whether message
  should be ignored
• Link-layer properties
• Asynchronous network, CSMA media access
• Unreliable delivery
• Active messages
• Packet specifies the handler which will deal with
  the data in the packet
• The handler on the receiver either performs a
  small amount of processing in the handler, or
  schedules tasks to perform more complex
  processing
• Very simple buffering model fixed size packet
  (36 bytes), in most cases idempotent data units

10
Applications
11
Low-end Digital Oscilloscope

• Capture and display the real-time signal from
  sensors
• Useful in initial development stages of any
  application
• Limiting factors data transmission speed, either
  UART or radio

12
Environmental Monitoring

• Ad-hoc networking making it easy to deploy
  sensors
• Autonomous nodes self assembling into a network
  of sensors
• Sensor information propagated to a few collection
  points
• Intermediate nodes assist distant nodes to reach
  the base station
• Connectivity and error rates used to infer
  distance

13
Data Collection

• Collect the time-stamped data in a relational
  database
• Light, temperature, motion determine room status
• Current sensors determine power consumption

14
Cory Hall and Intel Lab networks

• First Steps
• Monitor power usage and environmental conditions
  at 50 nodes spread around the environment
• Basic administration features network
  programming, node monitoring
• Next
• Lowering power usage how to get the nodes to
  last for a year
• Control Automatically adjust lighting and HVAC
  using nodes as actuators

15
Vehicle detection

• Application scenario
• Deploy sensor network on remote road
• Have nodes automatically determine connectivity
  and synchronize
• Use magnetometers to locally detect vehicles
• Exchange data with neighbors, and use neighboring
  data to compute the vehicle velocity and heading
• Remember the passing vehicles and report history
  to passing airplane

16
Done
17
Take away pieces

• Advantages of intelligent sensors
• Local processing of sensor readings
• Local communication patterns
• Local storage
• Challenges
• Dealing with uncertainty dropped (corrupted)
  packets, reading skew
• Unreliable sensors
• Changing environment
• Limited platform

18
Where to learn more

• Main websites
• http//webs.cs.berkeley.edu/
• http//www.tinyos.net
• How to get started with TinyOS
• http//webs.cs.berkeley.edu/tos/
• Holds instructions, pointers to tools, and code
  release
• Code repository
• http//sourceforge.net/projects/tinyos/
• TinyOS workshops
• TinyOS bootcamp held in October, materials
  available at http//webs.cs.berkeley.edu/tos/ or
  browse the tutorials at http//cvs.sourceforge.net
  /cgi-bin/viewcvs.cgi/tinyos/nest/doc/tutorial/
• Upcoming TinyOS bootcamp in January