Elkhorn Slough Wireless Project
DRAFT
CSUMB Seafloor Mapping Lab Wireless Project
Brian Wilson
10-June-2003
minor edits and links to pictures 22-Sept-2008
Introduction
The official part of the Wireless Project was to explore the feasibility of using off the shelf wireless network products to enhance the learning experience for ESSP 332 GIS/GPS students. It also happens that while I was managing this project, I was also an ESSP 332 student.
Over the course of working on the Wireless Project I also worked on two other overlapping projects; the set up for a benthic survey of Elkhorn Slough and Henrik Kibak's wireless projects. Equipment for ESSP 332 class
ESSP 332 is a class that provides instruction and lab and field experience using GIS and GPS. "GIS" stands for "Geographic Information Systems"; we learned GIS using ESRI ArcGIS 8.2 and (a little bit of) ArcPad. "GPS" stands for "Geographic Positioning System". The GPS portion of the class focused on using Trimble's Pathfinder Pocket GPS receiver and their Terrasync and Pathfinder Office products.
This picture shows the equipment setup for each student. It consists of a Compaq (now HP) iPaq 3950 equipped with an expansion module. The module holds two PC cards; one is a dual serial card and the other is a wireless network card. The yellow thing is the Pathfinder Pocket GPS receiver. To the left of it is the GPS antenna. the thing that looks like a tin can is a wireless antenna. The setup a fanny pack to carry everything.
The official Trimble baseball cap visible here (and in other photos) has a pocket on the back to hold the GPS antenna. I mention this because getting the GPS antenna up where it can see the satellites is very important; your body can block a substantial amount of signal.
This semester's 332 class took on a project for the Elkhorn Slough Foundation. Eric Van Dyke provided scans of aerial photos of recently acquired properties taken over a 70 year period. The assignment was to georeference, interpret, and ground truth the photos.
Georeferencing and interpretation were done in the lab using ArcGIS. For the ground truthing, students went out to the properties with the GPS equipment to collect field data.
During an orientation session, students went to the Elkhorn Slough visitor center. Students used the wireless facility there to download map data from the CSUMB campus server.
On campus, students can access the network directly from the iPaqs using the 802.11b access points scattered throughout the campus. Though not specifically designed for outdoor access, there are areas near the buildings where the iPaqs can still get connected. Using the coffee can directional antenna with the iPaq increases range.
At the Elkhorn Slough, however, no such wireless access existed, so we built it. We installed a 19' mast on the roof of the ESNERR science building. On this mast, we installed two antennas and a Proxim Orinoco AP2000 access point. We had an SBC Pacbell DSL line installed to connect to the Internet and thence back to the campus servers.
Equipment for benthic survey
Another SFML project at Elkhorn Slough was under way the same time as the wireless project. It consisted of performing a benthic survey of the slough itself.
Basic GPS technology relies on receiving and processing data from a constellation of satellites. Using satellites alone is only good for positional accuracy of 10 meters. This is good enough for finding your way around, but not accurate enough to do the benthic survey.
Accuracy can be improved by using differential correction (DGPS). A fixed base station receives satellite data and records the difference between its known position and the position calculated from the sateillite signals. The base station can feed the differential directly to mobile GPS receivers or it can log it for post processing.
Two forms of DGPS in common use are RTCM and RTK. RTCM corrections provide accuracy in the 2-5 meter range. RTK provide centimeter level accuracy. For our benthic survey, we implemented a base station at the ESNERR science building capable of supporting both RTK and RTCM corrections.
Currently the boat doing the surveys phones a cellular modem at the ESNERR science center. In the picture here, you can see the 8' cell modem mast and antenna, and in the foreground, the GPS receiver antenna on a 3' mast. Part of the area of the slough that was surveyed is visible in the distance.
The GPS antenna connects to a Trimble 4700 receiver. Both the cell modem and the receiver connect to a desktop computer located in an office in the science building. "Trimble Reference Station" feeds data to the cell modem for RTK; it also logs data to disk and copies it to CSUMB so that it can be used for DGPS post-processing for the ESSP 332 class. Collaboration with Henrik Kibak
Another ESSP professor, Henrik Kibak, had his own wireless projects for Elkhorn Slough, so we joined forces. One project involved a remote, solar-powered Web cam located in the Caspian Tern colony out in the slough. Photos from the camera are sent wirelessly back to our tower at the visitor center and copied to a server at CSUMB for archiving distribution on the Web.
Henrik and co-principal investigator Jotham Fisher-Smith did testing with our coffee can antenna and found they can connect at 1 Mbps from Moss Landing, approximately 3 miles away.
Did you say coffee can antenna??
Yes; creme wafer cookie cans work well, too!
The original plans for the ESNERR base station called for one omnidirectional antenna to allow access in the immediate vicinity of the ESNERR visitor center and one unidirectional "Vagi" antenna.
The Vagi points at a barn that is located near the edge of the water. Original plans were to use the barn as a redistribution point for the 802.11b signal. It has not been implemented yet.
When Henrik found out the blind at the tern colony had a line of sight to my antennas at the science building, we removed the omni antenna (which we really did not need) and installed a coffee can antenna.
Our tests show that the cantennas work as well as the $100-150 commercial units. The only problem over the long term might be corrosion. Total cost per unit is around $10, if you include the cookies.
To build the cantennas, I used design information from Greg Rehm's Web page at http://www.turnpoint.net/wireless/cantennahowto.html Details on ESNERR permanent base station The ESNERR base station includes a computer, GPS receiver, and a wireless access point with two antennas. The base station is connected The base station is connected to the Internet via a DSL line What worked and what didn't
PDA What is most strikingly clear here is that these technologies are still in their infancy. Especially the PDAs. The Windows CE / Pocket PC operating system is pretty weak. The PDA is designed to be a personal organizer, not a tiny computer.
Possibly tablet computers would be better, though I have heard that Windows XP on a tablet is even worse.
802.11b Wireless Repeaters We found that the Proxim AP2000 can in fact work as a wireless repeater to extend range. The plan to use the barn as a distribution point is possible with this equipment.
ArcPad and ArcIMS Originally, we wanted to use ESRI products ArcPad (which runs on the iPaqs) and ArcIMS (server based) to update databases from the field. It turned out that ArcPad can't update ArcIMS databases.
Wireless cards There was some delay and confusion over the HP wireless cards. It turns out HP will not be supporting wireless cards for client devices any more. I would have been happier with Orinoco cards, which are well supported.
Portable base station For fieldwork, I built a portable server based on a laptop, a tripod mounted omnidirectional antenna, an AP2000 access point, and powered by an inverter and deep cycle battery. It never got used due to time limitations.
Range I knew from the outset range would be an issue with this project. The technology works quite well. With external antennas, range is good, but the antenna has to be aimed properly. This is difficult to do when walking around with an iPaq in one hand and a stylus in the other. To collect GPS data, you have to walk around so a tripod is not a solution.
When using ArcPad with an ArcIMS server over the wireless network, it's problematic because ArcPad expects a continuous network connection. If conditions change (ie you walk behind a tree) then the connection can fail and ArcPad gets stuck waiting for new map data. So ArcPad does not work well in the field if used to connect to ArcIMS.
What we ended up doing was putting map data into zipped archive files on a Web server so that students could download it in one batch and unzip it on the iPaqs.
DGPS
Pathfinder Pocket GPS receivers can accept a stream of RTCM data to achieve 2-5 meter accuracy in real time. We had a plan to do this by feeding the data over the network link from our base station. There was not enough time allocated to write all the software needed to make this happen. I did get enough running to make me think it's technically possible. The same range limitations with ArcPad apply here too though, so it's probably not a practical application for 802.11b networks.
This project required doing testing, development, and deployment steps in parallel, getting pieces of it ready and operational just in time for use in the projects and in the class. You should NEVER do this! But generally speaking, we made every deadline and nearly everything worked. I am still stunned and amazed at how smoothly it all went.
I was both taking ESSP 332 as a student and teaching my classmates how to use the equipment at the same time! I had enough prior knowledge of computers and networks to pull this off. Overall, I had a blast!
Acknowledgements
Rikk Kvitek, Jorge Goicochea, Pat Iampietro and the whole SFML team Arlene Krebs, Gil Gonzalez, Chip Lenno and the CSUMB IT folk Henrik Kibak and Jotham Fisher-Smith Eric Van Dyke, Kenton Parker, Kim Hayes, Ann Gallenson, Gregg Hoffman at ESNERR Parts suppliers Two of the antennas and the "pigtail" cable are from Fleeman, Anderson, Bird. Hyperlink Technology sold us our best omni antenna. All other electronic components were purchased from Digikey. Stahlin boxes are excellent for mounting our wireless access points outdoors. We got them from of Electrical Distributors of Salinas. The 19' antenna mast at ESNERR is from Radio Shack, where else?
About the author
I am currently wandering around the Pacific Northwest on a bicycle. You can contact me by email. mailto://[email protected]/
