Welcome to the Tracer website. There is nothing here for a non-Testbed user to do. Perhaps you just have to login?

The primary goal of the Tracer website is to help Testbed researchers to do research with GPS data. While there are some data processing tools available, for the most part researchers are expected to do all data processing themselves.

The Tracer website is designed to facilitate the use of GPS devices to collect travel data. At the moment, the GPS devices available are our custom-designed extensible data collection units (EDCUs). These devices are able to collect GPS records as frequently as once per second. The collected data can be transmitted wirelessly over a cellular connection, over a local area wireless (WiFi) connection, using a physical ethernet connection, or by removing a CompactFlash card and manually downloading the data. The amount of data that can be collected is limited only by the size of the CompactFlash card. If one of the wireless data collection options is chosen, the amount of data that can be collected using the Tracer system is unlimited. For example, we have records for a single device that was fielded in a test vehicle for a year, without user intervention.

The Tracer website has evolved over the past few years into a large, complicated system with both administrative and user features. The primary purpose of the Tracer website has always been to manage the use of the EDCUs. This requires tracking the various components, both internal and external, which make up each EDCU, as well as keeping track of which researchers have which EDCUs. The second purpose of the Tracer website, and the reason for its name, is that the website provides a means to collect GPS data wirelessly (or manually) from each EDCU, and a means to redistribute the collected data in a number of formats to various types of authorized researchers or client programs.

Because the EDCUs are modular, and because the embedded code for collecting the GPS points can be rewritten, the Tracer website is a good deal more complicated that one might expect at first blush. Tracer users (non-administrative accounts) have access to the researcher menu, and the operation menu. Tracer admin accounts have access to other menu options which typically are related to adding new EDCUs or reconfiguring existing EDCUs. Administrative accounts are also the only ones who can checkout one or more EDCUs to a researcher. So if you would like to get EDCUs to use in your research, please contact Prof. Mike McNally or Ziggy Bates directly.

The following Javascript application attempts to show exactly what GPS data collection is all about. This application leverages the Google maps API. Some additions and changes have been made to the base API, including the addition of close-up orthoimagery available from the USGS. The map works just like Google maps (because it is Google maps!). For example, double clicking on a point will center the map at that point. Dragging the map will move it. And selecting a different scale will zoom the map in or out. If Javascript is disabled in your browser, this page should show some screen capture images. If Javascript is enabled but Google maps does not support your browser version, then you will see an error message.

As is standard for Google maps, the map scale is controlled by means of a large slider along the left hand side of the map. There are 18 scale levels for the "Map" "Satellite" and "Hybrid" maps is as of this writing. The most zoomed out scale will fit the entire earth into a single 250 by 250 pixel square, which is 1.44 degrees of longitude per pixel. Each scale zooms in by a factor of 2. The most zoomed in scale shows 0.002747 degrees of longitude per map tile, or about 0.00001 degrees per pixel. This translates to approximately 1 meter per pixel.

Since the USGS provides orthoimagery of the LA area at a resolution of 0.1 meter per pixel, the "Tracer" and "USGS" map types are actually zoomed in 3 levels more than the other map types. So zooming to the maximum resolution on the "satellite" map type, and then clicking on the USGS map type, you will see an almost spooky level of detail---a single 250 pixel tile is filled almost entirely by an intersection or a building.

The difference between the Tracer map type and the USGS is simply that the Tracer map is hosted locally, is slightly offset from where it should be, and only contains data for a small portion of Orange County. The purpose of this map type is to speed up map generation by have a local cache of images available. The USGS map type actually hits the USGS image server, and so can be much slower. Currently it only requests images from the Los Angeles area, although future versions may include imagery from other urban areas as well.

Along the right hand side of the map square is a box listing EDCU numbers. For demonstration purposes, data have been made public from EDCU number 2. If you click on that label, you will see a list of dates on which the EDCU was used. If you are logged in and have used the Tracer system to collect GPS data, you old data will be visible rather than the public demo data.

The trips made public were collected in early 2003, using a Garmin 35 GPS unit, with an error of approximately 15 meters RMS. Loading up the GPS points in the Javascript application is rather processor intensive, so you should run in on a fairly modern machine. In a benchmarking run with an 800MHz laptop, the browser threatened to shut down when two of the longer trips were loaded. If you fear your computer isn't up to scratch, click on March 13 or March 11, which are relatively light travel days.

GPS errors and characteristics

The first few dates shown in the sample set---from 2003-02-11 through 2003-03-01---are technically garbage data but demonstrate the range of data quality possible with a GPS. For these points, the GPS antenna was hung up next to my office window while doing some tests on the EDCU. These data points are all colored red, which means that the antenna did not not have fix on the GPS satellites when the measurements were taken. This can happen for many reasons in practice, but in this case, the windows of the offices here at the UC Irvine Institute of Transportation Studies have a rather limited view of the sky, and therefore a small window in which the GPS antenna must find 4 satellites.

The GPS system works by listening to radio signals. Satellites in the sky are broadcasting a radio signal down to the earth. The GPS antennas listen for those signals. Just as you might get a cellphone reception when riding in an elevator, so too you might get a GPS fix when your antenna is indoors or near a window. However, you cannot rely on this, as the GPS signals do not travel very far through solid obstacles (including trees, buildings, and bridges). When collecting travel data, it is equally important to make sure that the GPS antenna has a clear view of the sky. On the other hand, we have successfully collected data using a Garmin 16 with WAAS augmentation while the antenna was (mistakenly) placed in the trunk of the car.

The next two dates---2003-03-03 and 2003-03-04---show short trips in the UCI region. If you click on 2003-03-03, you will see green and yellow points. The yellow points are points at which the GPS fix was two-dimensional. These are not very good data points, but at least the GPS antenna has some idea of where it is as it can see three satellites. When the GPS antenna gets a lock on a fourth satellite, the fix goes to three-dimensional, and the points change from yellow to green.

If you click on March 11, you will see another trip in and around UC Irvine. At the original scale, the points look as if they are one big green blob. If you zoom in to the highest resolution or click on the USGS map type, the points are clearly distinguishable. If you click on any point, a pop up will show the data collected at that point.

Feel free to play with the application, and even hack on the Javascript code that interacts with Google's code. If you check out your own EDCU, you can add your own data and start building your own applications. If you have questions or comments, please email me at .