Transcript MICA hardware Platform

MICA Node Architecture
WEBS retreat
Jason Hill
1/14/2002
Objective
► To
produce an experimental platform for the
NEST research teams to use explore design
tradeoffs in embedded networked devices.
► Must be ready to deliver nodes in January
2002.
 All parts used must be available in quantity
by Nov. 01
 Development tools must be readily
accessible
Outline
► Previous
designs and factors motivating new
design
► Mica Architecture and Hardware Details
► Architectural Implications (RADIO, CPU
interface)
► Performance characteristics in key areas
► Future directions
Previous Designs
►
MICA is the 4th generation of the Berkeley
Motes.
1.
2.
3.
3.1
COTS dust prototypes, by Seth Hollar
weC Mote
Rene Mote, manufactured by Crossbow
Dot mote, used for IDF
Rene Shortcomings
► Insufficient
memory size
 code size limitation the tightest
 “29 Palms” and “Cory Hall” barely fit
► Radio
performance erratic, slow
 Tied to battery voltage
► No
way to determine battery levels
 Unpredictable operation
► Awkward
attachment to power supply
► No unique Ids
 IDF required manual assignment
Goals
► More
storage
► More communication bandwidth
► Stabilize board voltage
► More capability available to sensor boards
► Retain “cubic inch” form factor and AA/year power
budget with clean package.
► Allow opportunities for flexible communication
protocols
 time synchronization
 other algorithms
The MICA architecture
►
►
►
►
►
Atmel ATMEGA103
 4 Mhz 8-bit CPU
 128KB Instruction Memory
DS2401 Unique ID
Atmega103 Microcontroller
 4KB RAM
4 Mbit flash (AT45DB041B)
Transmission
Hardware
Coprocessor
Power
Control
Accelerators
 SPI interface
 1-4 uj/bit r/w
TR 1000 Radio Transceiver
4Mbit External Flash
RFM TR1000 radio
 50 kb/s – ASK
Power Regulation MAX1678 (3V)
 Focused hardware acceleration
Network programming
Same 51-pin connector
 Analog compare + interrupts
Same tool chain
SPI Bus
►
51-Pin I/O Expansion Connector
8 Programming
Digital I/O 8 Analog I/O
Lines
2xAA form factor Cost-effective
power source
Major features
► 16x
program memory size (128 KB)
► 8x data memory size (4 KB)
► 16x secondary storage (512 KB)
► 5x radio bandwidth (50 Kb/s)
► 6 ADC channels available
► Same processor performance
Major Features (cont.)
► Allows
for external SRAM expansion
► Provides sub microsecond RF
synchronization primitive
► Provides unique serial ID’s
► On-board DC booster
► Remains Compatible with Rene Hardware
and current tool chain
Microcontroller Alternatives
►
Atmega 163
 same pin out as RENE
 2x memory
 Can self-reprogram
►
Not enough memory
ARM Thumb
 lower power consumption, lower voltage
 greater performance
 poor integration  slow radio
►
Peripheral support missing
TI MSP340
 Superior performance
 1/10 power consumption
 Better integration
No GCC, tool chain missing
Radio
► Retained
RFM TR1000 916 Mhz radio
► Able to operate in OOK (10 kb/s) or ASK (115
kb/s) mode
► Design SPI-based circuit to drive radio at full
speed
 full speed on TI MSP, 50 kb/s on ATMEGA
► Improved
Digitally controlled TX strength DS1804
 1 ft to 300 ft transmission range, 100 steps
► Receive
signal strength detector
Network Programming and
Storage
►
►
ATMEGA103 in-circuit, but external reprogramming
 retain secondary co-processor
 AT90LS2343 only small device with internal clock and
in-circuit programming
4 Mbit flash (AT45DB041B)
 Store code images, Sensor Readings and Calibration
tables
 16x increase in prog. mem too large for EEPROM
solution
 forced to use FLASH option
Power
► Incorporated
On-board Voltage Regulation
 Boost Converter provides stable 3V supply
 Stabilizes RF performance
 Allows variety of power sources
 Can run on batteries down to 1.1 V
► Incorporated power supply sensor
 Can measure battery health
 used to adjust wake-up threshold for unregulated
design
Expansion Capabilities
► Backwards
compatible to existing sensor boards
► added two analog compare lines
► added five interrupt lines
► added two PWM lines
► Can connect external SRAM for CPU data memory
(up to 64KB)
 lose most sensor capability
 address lines share with lowest priority devices (LEDS,
Flash ctrl)
 still allows radio, flash, and programming
Why Not Faster/Different Radio?
► RFM
TR1000 is the lowest power RF Transceiver
on the market
► Simplistic design leads to rapid power cycling
► High speed radios usually come with digital
protocol logic forcing users into set communication
regimes
► Raw interface to the RF transmission allows for
exploration of new communication paradigms
(Proximity Mode and Sleep)
Outline
► Previous
designs and factors motivating new
design
► Mica Architecture and Hardware Details
►Architectural
Implications (RADIO,
CPU interface)
► Performance
characteristics in key areas
► Future directions
Wireless Communication Phases
Transmit command
provides data and starts
MAC protocol.
Transmission
Data to be Transmitted
Encode processing
Start Symbol Transmission
MAC Delay
Encoded data to be Transmitted
Transmitting encoded bits
Bit Modulations
Radio Samples
Start Symbol Search
Receiving individual bits
Synchronization
Reception
Start Symbol Detection
…
…
…
…
…
…
…
Encoded data received
Decode processing
Data Received
Radio Interface
► Highly
CPU intensive
► CPU limited, not RF limited in low power systems
► Example implementations
 RENE node:
► 19,200
bps RF capability
► 10,000 bps implementation, 4Mhz Atmel AVR
 Chipcon application note example:
► 9,600
bps RF capability
► Example implementation 1,200bps with 8x over sampling on 16
Mhz Microchip PICmicro (chipcon application note AN008)
Why not use dedicated CPU?
► Dedicated
communications processor could greatly
reduce protocol stack overhead and complexity
► Providing physical parallelism would create a
partition between applications and communication
protocols
► Isolating applications from protocols can prove
costly
Flexibility is Key to success
Node Communication
Architecture Options
Classic Protocol
Processor
Direct Device
Application Controller
Application Controller
Narrow, refined
Chip-to-Chip Interface
Control
Protocol Processor
Raw RF Interface
RF Transceiver
Hybrid Accelerator
RF Transceiver
Serialization Accelerator
Timing Accelerator
Hardware Accelerators
RF Transceiver
Memory I/O
BUS
Application Controller
Accelerator Approach
► Standard
Interrupt based I/O perform start
symbol detection
► Timing accelerator employed to capture
precise transmission timing
 Edge capture performed to +/- 1/4 us
► Timing
information fed into data serializer
 Exact bit timing performed without using data
path
 CPU handles data byte-by-byte
Results from accelerator
approach
► Bit
Clocking Accelerator
 50 Kbps transmission rate
►5x
over Rene implementaiton
 >8x reduction in CPU overhead
► Timing
Accelerator
 Edge captured to +/- ¼ us
►Rene
implementation = +/- 50 us
 CPU data path not involved
Outline
► Previous
designs and factors motivating new
►Future
directions
design
► Mica Architecture and Hardware Details
► Architectural Implications (RADIO, CPU
interface)
► Performance characteristics in key areas
Incremental Improvement
► ATMEGA128
 Self Reprogrammable
 Hardware Multiplier
 JTAG debugging support: real-time, in-system
debugging
► Chipcon
CC1000 radio
 Programmable transmission frequency
 Automatic bit timing
 2.3V operation voltage
The Leap
► MOTE
on a chip
 Integration of radio, CPU, protocol accelerators onto a
single silicon die.
► Investigate
set of hardware accelerators
customized to the needs of sensor networks
► Start
Symbol Detection
► Bit timing
► Bit correlation
► Channel encoding
► Low power channel monitoring
► Optimal communication timing