Our original plan for displaying the results of our monitoring system was to create a density gradient overlayed on a floor plan of the first floor of the Breakiron Engineering Building, where our experiments took place. The idea was to use Shephard’s Method to determine the appropriate value for the gradient at each pixel in the floor plan. Unfortunately, our best efforts were insufficient to fix our method and produce a meaningful gradient – our initial efforts kept producing gradients like this:
The blue dots here represent our sensors, and we imagine the colored lines were intended to represent the calculated gradient, but clearly something went wrong in our method, and the final image represented not only a flawed gradient, but also an inverted image of the original floor plan (both horizontally and vertically). It became clear to us that we would have to abandon our gradient approach and change tactics.
It was at this point that we decided to try to represent the actual devices we were locating on the map. Given that we already had a wonderfully functional script running a trilateration algorithm, it was relatively simple to redirect our efforts to generate a visual representation of the approximate locations of the devices themselves. We were actually relatively impressed by the accuracy of those locations – we were able to locate exact numbers of residents of the rooms around the floor, despite the fact that the locations within each room were generally all over the place. To our ends, this was a perfectly acceptable result, as it was always our intention to monitor room traffic, not exact device location. In the end, we were very satisfied by the maps automatically being generated by our sensors, which looked like this:
Here, the red dots represent our sensors, and the blue dots represent located devices, which is a clear improvement to the worth of the map, and so we deemed our efforts to produce a meaningful device density map successful.
Recent Comments