Glopos: From Universal Location Positioning to High Accuracy Positioning

A common question people ask about indoor location positioning goes something like this: "Why do you need Wi-Fi, or BLE, or other technologies? Why can’t it be done based on cellular signals?"

The common wisdom, of course, is that cellular signals have so broad a range, so big a coverage area, that they don’t differentiate different locations sufficiently. Many mobile companies have implemented indoor positioning based on cellular signals, using multilateration algorithms on cellular signal strengths and antenna locations, and have ended up with 60 to 100 meter inaccuracies on average, often worse.

Glopos, based in Helsinki, didn’t take "no" for an answer. Instead of relying on cellular signal strength values, Glopos’s technology uses a wider variety of cellular signal parameters, models of cell area and shape, and data from other nearby cells, to build self-learning probabilistic models for estimating positions. Their latest tests have achieved 5-6 meter accuracy based on 3G signals alone. In that same test, the average accuracy of the best 80% of the location estimates was 2.4 meters. (See video below.) This accuracy isn’t as good as some other companies have achieved, but it’s based solely on cellular signals, while most other companies use Wi-Fi or Bluetooth signals.

The aspect of Glopos’s approach that Grizzly Analytics likes the best is that it uses cellular signal parameters that are already being received and processed by every cellphone in the world, as the phones check for incoming calls and messages. So there’s literally no additional radio signal processing used to achieve always-on, always-available, 24×7 location positioning, and virtually no additional battery or CPU usage. In principle their technology can run on non-smartphones as well. This universality can open up the door to a lot of new applications running in the background.

The aspect of Glopos’s approach that Glopos engineers like best, on the other hand, is that the same algorithms can be used to position based on Wi-Fi or other radio data as well as cellular signals. They have tested their technology on Wi-Fi signals in a mall, and achieved 2-3 meter accuracy, without requiring map data for the site or the extensive fingerprinting process that other Wi-Fi based systems require.

But they didn’t stop there. A month ago, when Grizzly Analytics released a report on indoor location positioning systems that achieve sub-meter accuracy, the folks at Glopos felt challenged to show that their technology can also achieve sub-meter accuracy. Their video shows their lab test, using Wi-Fi signals from four Wi-Fi access points around a meter-by-meter grid in their lab. In these conditions, their system achieved 9.4cm accuracy! (See video below)

To be clear, these sub-meter results were in carefully-constructed lab conditions. Their accuracy in real-world conditions is as described above. But even so, this experiment shows the ability of their algorithms to work on a variety of radio signals to deliver a range of results, based on the signals and the application.

Grizzly Analytics believes that universal works-everywhere and always-on location positioning, such as that delivered by Glopos’s technology, will completely revolutionize not only indoor location positioning but location applications in general.


Leave a Reply

Your email address will not be published. Required fields are marked *