The Low Power Energy Aware Processing (LEAP) Embedded Networked Sensor System

1
The Low Power Energy Aware Processing (LEAP)
Embedded Networked Sensor System
Dustin McIntire, Bernie Yip, Hing Kei Ho,
Lawrence Au, Timothy Chow, and William J.
Kaiser UCLA Electrical Engineering Department
2
Power Dissipation vs. Energy Efficiency

• High performance components often offer high
  efficiency

• requires new architecture, enables new design
  approach

3
LEAP Hardware Architecture
SPM (Slauson Processor Module)
EMAP (Energy Management and Accounting
Preprocessor)
SDCard/MMC
SPI
I2C
JTAG
GPIO
Ethernet
Shutdown
Address/Data
MOSFET Driver
Vsense
Flash
SDRAM
PCMCIA

-
Current Sensed Supply Outputs SPM Power (2A
Max)
x5 1
Sensor Voltage
Sensor Inputs
x2
180 Pin Interboard Connector
USB Host Controller
USB Host
4
LEAP Design Course Platforms

• LEAP Approach
• Diverse sensor systems
• Fine grained platform instrumentation
• Highly configurable testbeds
• Courses Require
• Usage methods that support many disciplines at
  the undergrad level
• Rapid implementation of complex algorithms
• National Instruments LabVIEW interfaces for LEAP
• New National Instruments program courseware for
  shared development
• Drag-and-drop Actions and Events
• Integration of signal processing, control
  systems, embedded networked sensing

5
LEAP Software Tool Development
6
The New LEAP2 Platform

• Dedicated ASIC performs EMAP functions
• Extensive current sensing and power domain
  controls
• Extendable to future modules through stacking bus

7
LEAP2 Module Block Diagram
GPIO/Serial Busses
Buffers
Address/Data
32MBNOR Flash
512MB NAND Flash
8MB SRAM
64MB SDRAM
Mem Bus
SODIMM 200 Connector
GPS
Ethernet
EMAP ASIC
USB Host
CMOS Imager
Pwr Control
CF
Current Sensors
Low Power Radio Module
LEAP2 Stacking Connector
uWatt FPGA
mPCI WLAN
Current Sensors
LEAP2 Stacking Connector
CC2420
MSP430
Current Sensors
Sensor Power Switching
ADC IF
8
LEAP Research and Deployments

• New design approach based upon a components
  energy efficiency for each sensing task
• Energy efficiency of system components is
  paramount for long lived systems
• Energy profiling is vital for efficient use of
  system resources
• Completed and Planned deployments
• Education (75 users in class project setting)
  USC and UCLA
• Ecosystem (acoustic localization, microclimate,
  multiscale sensing)
• Aquatic (distributed contaminant sensing)
• Seismic (rapidly deployable arrays, GEONet)