fagu2004_seamonster - talus-and

Download Report

Transcript fagu2004_seamonster - talus-and

SouthEast Alaska MOnitoring Network for Scientific TElecommunication & Research:
SEAMONSTER—Application and Challenges of distributed sensor arrays to Southeast Alaska Science
B53A-0979
Edwin Knuth1
1Dept.
ABSTRACT
SEAMONSTER, the SouthEast Alaska Monitoring Network for Scientific
Telecommunication and Research, is a network being developed to provide
infrastructure to benefit research throughout Southeast Alaska. The array will
be composed of ‘bricks,’ which are small Linux-based computers with digitizers
and wireless network communication capabilities. The first use of this network
is for bioacoustic monitoring of bats is Southeast Alaska for species
identification and measurements of population dynamics. Additional projects
during summer 2004 included experiments for testing network link range, brick
durability in adverse conditions, power requirements, and brick capabilities.
We conducted several trial scientific measurements during the 2004 summer
including monitoring of Lemon Glacier and Columbia Glacier.
Matt Heavner1
Eran Hood1
Natural Sciences, University of Alaska Southeast, [email protected], [email protected], [email protected]; 2Vexcel Corporation, Boulder, CO, [email protected]
WHAT IS SEAMONSTER?
SEAMONSTER is a multi-tiered data acquisition network with communication capabilities and some analysis capability at the sensors. The multi-tiered
nature of the network is intended to flexibly accommodate and take advantage of appropriate spatial scale sizes and communication technologies. For
example, monitoring soil chemistry or snow accumulation of a single watershed may require closely spaced sensors communicating over Blue Tooth links.
The data would be fed to a backbone of longer distance communication technologies. A hydrophone network monitoring Glacier Bay would have a much
larger spatial separation and may use 900 MHz modems.
SEAMONSTER nodes include a small, low-power computer.
This computer manages communication, allows for data
acquisition reconfiguration (e.g. adjust sampling rate or
sensitivity), and provides a method for setting alarms if the
data does not pass defined validation tests (if the sensor reports
a value out of a defined range, or if a data value is missing).
This poster is to illustrate the capabilities of SEAMONSTER and seek
collaborators who will benefit from using the network to enhance their ongoing
monitoring of the Glacier Bay region.
Southeast Alaska in general, and Glacier Bay in particular, presents a
challenging environment for both data acquisition and wireless communication.
SEAMONSTER is an ambitious project to address both of these issues to benefit
the scientific community studying Southeast Alaska and Glacier Bay.
Rob Fatland2
Sensor separation can vary from 10 meters to 10 km or more.
This schematic shows multiple telemetry options: in situ
investigator, over-flying plane, satellite. SEAMONSTER will
be linked into the internet and provide data access from the
investigators “desk-top” as well.
BRICK Network Node Capabilities
Existing Prototype
BENEFITS
Long Duration Data Acquisition
One of the main motivations behind the development of SEAMONSTER is to facilitate long-duration
(multi-year) observations of various biological and environmental parameters in a robust manner.
Data Retrieval
Next Generation
(Mini-BRICK, Geo-BRICK)
Long-duration observations are currently possible, but often involve data loggers which much later be
GPS
5 meter accuracy
1 cm differential
retrieved. If the data logger is lost (e.g., a radio collar is lost or destroyed, or a remote site is “visited” by
Digitizer
3000 Hz 4-channel (e.g. seismic)
250 Khz 4-channel (bio-acoustics)
a bear) then the entire data set is lost. SEAMONSTER provides real time data flow from the sensors to
Data Storage
60 MB
1 GB solid state – 80 GB hard drive
overcome this problem.
Power
8 watts: Commercial-off-the-shelf components
½ watt: custom computer, intelligent duty
cycling
Operating System
Linux, Open Source paradigm
Linux, Open Source paradigm
Wireless Range
500 meters
10 km
Telemetry
Fixed station
Internet
Flexible Spatial Scale
Data rate
2 Mbps
TBD
The multi-tier model of SEAMONSTER specifically provides data sampling over diverse spatial scales.
Sample Timing Uncertainty
20 microseconds
5 microseconds
For example, animal radio collar tracking through SEAMONSTER would require receivers at a larger
Interface options
USB / Ethernet / RS232 / Digital I/O Pins /
802.11 wireless link / Parallel port
Same as Existing, also Radio reception (for
radio collar), Ultra-wideband wireless
Operating Temperature
Down to –60C in the laboratory
Down to –60C in the field
Cost
“Parts”
TBD
Real-Time State of Health Information
Real-time access to the data allows for both on-going data validation and monitoring of the instrument
state of health information.
spacing than required for an ongoing soil-chemistry monitoring project.
Existing Systems Can Be Included
Existing data acquisition systems can easily be incorporated to the SEAMONSTER architecture. Please
contact us if you are interested!
NETWORK ELEMENTS
PROTOTYPE EXPERIMENTS
The image at right shows a brick (the silver
box) with a three-axis geophone (orange
cylinder) and an 802.11b wireless bridge. This
is the prototype for the main nodes of
SEAMONSTER. The bricks are designed to
not only acquire and forward data, but also
form an “intelligent” network used for data
storage. If the main link to the internet goes
down for some reason, the data collected is
distributed among the bricks for redundancy
Photo courtesy Tad Pfeffer’s online Columbia Glacier archives.
The photograph at left shows a “waysmall”
computer (the white box with two black dots), a
USB wireless network adaptor, a USB sound
card, (with cable) and a pocket knife to indicate
scale.
The 400 MHz computer provides
significant processing power at very low power,
and is currently being evaluated for use as a
SEAMONSTER node for a location with less
acquisition requirements than a brick.
Other SEAMONSTER elements can include radio
tracking receivers (for relaying radio collar data) or
other existing sensors.
FUTURE PLANS
The wireless communication network is currently being established in the Juneau area to allow easy
access and testing. A robust network will be in place by spring 2005, when we will be expanding the
network to encompass a greater part of Southeast Alaska. Glacier Bay is a high priority for near-term
establishment of the SEAMONSTER network and a link to the internet for real-time data transmission
to investigators.
Locations of bricks with geophones, Columbia Glacier
A large part of why glaciers—like many other environmental and biological
phenomena—remain so mysterious is the difficulty of observations over long periods
of time in fine detail. This is primarily because doing so is prohibitively expensive. We
are taking advantage of rapidly-evolving technology to build better, cheaper, robust,
telemetry-equipped field equipment.
Results
May 2004: Successful air-to-ground telemetry data recovery using a Cessna 172.
June 2004: Columbia Glacier seismic data acquisition: Main trunk centerline serac, 4
geophones, 4 days, 100 samples per second. Additionally several more Bricks were
deployed at the margin near the calving terminus. Recorded seismicity includes
teleseisms and local events; probable ice fracture events, possible calving events.
June 2004: Lemon Creek Glacier (Juneau Icefield) seismic acquisition, 4 geophones, 5
days, 100 samples per second. Seismic record includes the June 28 magnitude 6.7
earthquake on Queen Charlotte Island, 350 km distant from Brick sensor.
The Fish Creek Knob study site is located on the northeast ridge of Mount Ben Stewart
on Douglas Island 8 km southwest of Juneau, Alaska (58° 16' 29" N and 134° 31' 48" W).
The meteorological tower is located in a flat 20 m by 40 m open area just below treeline at
700 meters above sea level and 100 meters below the top of Eagle Crest. The
meteorological tower monitors wind speed and direction, temperature, relative humidity,
and net radiation. Additionally, a sonic snow depth sensor allows observation of real time
snow accumulation during storms.
Last year we added a network link and real time data publishing on the web. The three
plots show representative data (snow depth and temperature) and state-of-health
information (battery voltage).
The Alaska Dept. of Fish & Game has funded a project to build bat detectors based upon the concept of wirelessly networked
small, lower-power computers for long term (approximately a year or longer) monitoring of the bat population in Southeast
Alaska. This project is currently underway and is based on the brick technology and the SEAMONSTER network for data
communication.
Brick Deployed on Lemon Glacier
We are currently developing the bricks to decrease power requirements, further
develop the data communication and networking software, and add additional sensors.
A network of bricks is scheduled to be deployed in Antarctica this fall.
Other sensors, such as video cameras, weather station information, snow depth sensors,
and radio collar animal tracking information is being tested for integration into the
SEAMONSTER network.
The project has successfully recorded bats onto computers and is developing the at-the-sensor software requried to descriminate
the audio signal of bats against background noise (such as rain hitting the microphone). The bat project is a prototype for the
biological monitoring applications possible with a SEAMONSTER network.
CONTACT US
We are very interested in working with researchers who want to investigate the utility
of the SEAMONSTER network. Initially, we would work in parallel with existing
data retrieval techniques.
If you would like to find out more about SEAMONSTER please contact us!