Review: TamoGraph Site Survey 2.0 - Page 2

By Lisa Phifer

February 14, 2011

Calibration and accuracy

Ultimately, survey accuracy is critical. RF planners and integrated enterprise survey tools from vendors like Ekahau and Motorola can factor in building material attenuation. But a stand-alone survey tool like TamoGraph can only estimate RF propagation based upon specified environment: indoor or outdoor, with high, medium, or low attenuation.

This is done by selecting the most applicable environment, which TamoGraph associates with an adjustable "guess range" (e.g., 10 feet). During walkabouts, clicking any point produces a shaded circle of that diameter. For best results, walk the site until your entire map is shaded. Where doing so is impossible or impractical, TamoGraph offers an "extrapolate beyond guess range" option to fill gaps (available but not recommended).

Accurate results also depend on precise calibration. Calibrating imported floorplan images is easy: just draw a line of known length. However, calibrating maps is a bit trickier. TamoGraph can import saved map images or maps downloaded from OpenStreetMap, Bing, or MapPoint. To download a map, type an address, longitude/latitude, or use GPS-supplied current location. After zooming/panning to isolate the desired area, TamoGraph saves the map image for offline use.

At this point, at least 3 known (preferably distant) coordinates must be marked on the map image. One option: move to 3 locations and take live GPS readings. But our GPS surveys were faster and more reliable when we just typed coordinates copied from an on-line map service. Which begs the question: If TamoGraph can download on-line maps, why can't it auto-calibrate those maps by downloading their coordinates too?

Given good environment selection, decent calibration, and accurate map-clicking, we found that rigorous walkabouts yielded fairly accurate results. Heatmaps from separate walkabouts of the same area were largely consistent; interior AP locations were often predicted within 10 feet. The same cannot be said for edge APs, but that's a common limitation, caused by insufficient data for good triangulation.

In fact, the more we used TamoGraph, the better our results. That says more about proper technique than product, but learning to make best use of TamoGraph didn't take long. One option we'd add: the ability to delete faulty measurement points. This would be especially welcome in continuous mode, where unmarked pauses can require discarding an entire path-walk. To minimize this, break continuous surveys into shorter segments to be combined for analysis. Breaking larger sites into smaller survey zones can also improve map-click accuracy.

Analyzing TamoGraph's results

One TamoGraph project can include many floorplans or maps, each associated with several survey runs. These can be saved for later analysis on the same PC or exported for merged analysis on another (separately-licensed) PC. The latter is great for analyzing a large or geographically-distributed WLAN, or visually comparing old and new surveys of the same site.

TamoGraph analyzes selected survey runs for a single image. For example, we could analyze our first floor perimeter survey results or interior survey results or both at once. But unlike a planner such as RF3D, TamoGraph cannot correlate adjacent floor measurements; heatmaps are all 2D.

During a survey, the left panel displays live AP properties commonly observed by many stumblers: SSID, MAC address, vendor, channel, current signal strength (dBm), encryption type(s), max PHY rate (Mbps), and number of spatial streams. But during offline analysis, the left panel is used to select SSIDs, channels, and APs for analysis (below). APs below a configurable threshold can be eliminated, but this list would be even better if it could be searched or filtered.

Pages: 1 2 3 4


Comment and Contribute
(Maximum characters: 1200). You have
characters left.