A Wireless Sensor Network for Microclimate Monitoring Paul Flikkema Brent West Northern Arizona University April 2002
The Scientific Challenge Understanding microclimates - the long-term weather trends in localized areas. Examples include: the effects of habitat fragmentation and restoration in the Southwest the interaction of insect activity, tree growth, and soil conditions in woodlands of the Colorado Plateau how microclimates influence the growth of redwoods in California Our global environment is a mosaic of microclimates!
Drives the Engineering Challenge Need: Dense array of small sensors to monitor microclimate variables such as temperature and light. Standalone or wired sensor arrays are difficult to deploy and operate. Opportunity: Wireless networking of the sensors dramatically improve coverage and spatial density, and ultimately, our understanding of microclimates... while greatly reducing the total monitoring cost
Requirements Low cost to allow deployment in large numbers Low energy consumption - to realize practical service life in the field Flexible data rate must support low per-sensor, but high aggregate rates Centralized data access - data must be automatically delivered to a central repository Reliability and autonomy - system must be robust to withstand long service intervals
Approach Key requirements can be met using a design which synthesizes: advances in environmental sensing (typified by current sensor/dataloggers) wireless networking infrastructure adapted to the environmental sensing regime
WISARD (Wireless Sensing and Relay Device) Modular sensing, computation, and wireless communication/networking capability Integration of family of technologies---can exploit advances in each Adaptive information rate and transmit power to maximize efficiency Light, Temp., Humidity, etc.. Sensor Package RF Transceiver Microcontroller
Sensor Architecture Microcontroller - 8-bit processor with integrated A/D converter and on-board non-volatile memory Sensor suite - commercially available environmental sensors and required excitation/signal conditioning circuitry Radio transceiver - low-power 900 MHz ISM band transceiver IC; variable transmit power, data rate, and carrier frequency
Digital Processing Fully static CMOS microcontroller with onchip A/D, register file, FLASH program memory, and EEPROM data storage On-board external SRAM with multiplexed data bus WWV time source and FLASH card memory (network controller)
Sensing Functionality Analog Inputs Temperature Light Gen. Purpose Digital I/O Open Drain Output Switched 3.3V Power 1-Wire Bus # of Channels 2 2 1 # of Channels 1 1 1 Sensor Type thermocouple photodiode any analog output device (i.e.-humidity, soil moisture, or pressure sensor) Function general purpose control general purpose switched power source accommodates network of 1-Wire devices (counters, A/D converters, temp. sensors)
Radio Subsystem FSK data radio operating in the 902-928 MHz ISM band DDS based tuner with 230 Hz frequency resolution ~30us hop time for possible FHSS implementation Variable data rate up to ~56 kbps Nominal 4 dbm RF output with 3 step attenuator (10dB increments) RSSI output for measurement by microcontroller
Aggressive Power Management --- provides practical service life with low cost power sources Radio - Data transfer on a predetermined schedule; receiver need not be continuously listening Sensor suite - powered down via program control except during relatively infrequent sampling Processor/memory - sleep mode current drain of 20 µa vs. typical 2.5 ma during operation
Software Reconfiguration Program store may be modified by the microcontroller w/o external device programmer Software updates/bug fixes - modifications may be made simply by uploading new software to the base unit and allowing it to broadcast to all sensor units. Static parameter storage - System parameters may be stored in code space, freeing up on-board non-volatile memory.
Multi-hop Networking Power supply constraints and FCC regulations for unlicensed operation dictate low-power transmitters. Tests with low-power 900 MHz ISM band radio equipment show that point-to-point communication range ~100 m in field conditions. Many applications would span distances greatly exceeding point-to-point range (a linear array of sensors 1 km long would not be uncommon); therefore, capitalize on the high spatial density of units to relay messages from one unit to the next until they reach the base.
WISARDNet Concept Satellite Link Local Network Controller User Community
Network Organization Optimum routing is key to system efficiency since radio is the largest energy consumer. Network organization uses a shortest radio path algorithm with a receive signal strength metric.
Packaging Requirements Weatherproof Allow interfacing with external sensors, batteries, or long-haul comm. equipment Easy to disassemble/reassemble Convenient to deploy in a range of scenarios Low-cost
Integrating Research and Education NAU Communication and Robotics Lab: students learn the fundamentals of embedded, networked devices. The PI s will plan an Environmental Sciences and Informatics Program where students can learn how to develop and use technology for ecosystem monitoring and modeling.
Project Sponsors National Science Foundation Biodiversity and Ecosystem Informatics Program NAU Department of Biological Sciences NAU Merriam-Powell Center for Environmental Research Microchip Technology