In May 2016, we participated at an IPS test bed during the GeoIoT expo in Brussels, Belgium. Several IPS technology providers were invited by the organizer, Grizzly Analytics, to benchmark their systems according to a set of rules and methodology.
After having taken part of the final results, presented in the Grizzly Analytics report, we’d like to highlight some interesting takeaways from the testbed results.
First of all, it is with satisfaction to see that that all of the IPS providers using phones and BLE beacon provide good accuracy. That hasn’t always been the case, and now that the performance accuracy is converging, we take this as a sign that the industry performance is reaching and exceeding the requirements of most use cases. This signals the level of maturity of the technologies and the recent leaps the industry has taken lately.We take this as a sign that the industry performance is reaching and exceeding the requirements of most use casesAs the performance of the providers all fall into an acceptable standard with regards to accuracy, other factors of differentiation becomes more and more prevalent. One of those factors is installation and set up time. The chart below shows the number of beacons that were used by the six solutions that were employing smartphones combined with BLE beacons. Per request by Grizzly Analytics, we’ve masked out some names and numbers but if you are interested in the raw data, I suggest you head over here and purchase the report in its entirety.
As is visible in the chart, the measured accuracy (the solid bars) were all reasonably consistent and well in range for what most smartphone based location services would require, i.e. a few meters. If we take a look at the number of beacons used, there is a larger spread. To cover the roughly 400 sqm (approx. 4300 sqft) testbed area, we used 12 beacons – what turned out to be the lowest number of all participating solutions. This ratio, how many beacons used to cover a set area is commonly referred to as beacon density.
For us indoor positioning has always been about feasibility in commercial enterprise deployments. Ten, twenty or thirty beacons might not sound like a lot, but as the area of a venue grows, the number of beacons scale as well. For a large indoor space, such a factor difference beacon density can easily amount to a large difference in operational costs. One part is the actual cost of hardware but also the effort of installing it and probably most importantly, the cost of maintenance. Recently, when installing our indoor positioning system StepInside® at a large shopping mall, some 580 beacons were used to cover an area of 220 000 square meters, resulting in a beacon density that is even lower than in the GeoIoT testbed. For this particular venue, a higher beacon density could easily have resulted in the range of a thousand beacons.
While accuracy and beacon density are important factors to consider when evaluating and implementing IPS, there are other things to consider as well. Device compatibility (will it work on most smartphones? will it work on both iOS and Android?), latency (does the position lag behind as the user moves about?), boot-up time (how long until the first position is displayed?) and robustness (is the performance consistent over time?), to mention a few.
Have a question or interested in hearing more about recent deployments? Drop me an email and let me know!