Sensor Fusion goes solo with WiFiSLAM

Over a year ago, in our analysis of indoor location positioning technology, we noted a lot of research by major mobile companies in sensor fusion, using a smartphone's sensors to detect the phone's motion and thereby track the phone's location.  At the time we predicted that start-up companies would emerge that take this approach to indoor location positioning.

In the year since, we've seen a number of start-up companies launching sensor fusion technology. In recent months we've blogged about the Senion Labs launch in Singapore, which was the first large-scale commercial launch based on sensor fusion, and Pole Star's launch that included sensor fusion for iPhones, both of which integrated sensor fusion with other technologies. We also blogged about SenionLabs sensor fusion technology when we first saw it at MWC 2012.

The reason that these companies used sensor fusion in conjunction with other technologies is that sensor fusion is very hard to implement with precision.  The reason is clear: If you're tracking the motion of a phone, and you're off by a single degree in measuring the direction you're walking, the error will compound itself very noticeably. The same is true if you are off by a little bit in measuring how far the phone has moved, and might not know which door the user entered. One solution is to implement sensor fusion in hardware, as several chip companies are doing, and another is to use other approaches (map constraints, Wi-Fi) to correct the errors of the sensor fusion.

But now a American start-up called WiFiSLAM has demonstrated sensor fusion technology operating all by itself, tracking location effectively without any use of Wi-Fi or GPS technologies.  Their demo video shows their system tracking the phone's location as it moves around a parking lot, with some deviations but all in all precise enough for any application.

Note that we don't know exactly what it took to get this demo running.  The demo looks like it used map constraints, for example knowing where there are cars and assuming that the person isn't walking through a car, and thereby constraining the sensor fusion. They may also have held the phone very carefully and in a specific orientation.  But it's still a very impressive demonstration of an up-and-coming technology.

WiFiSLAM is a relatively new start-up with Wi-Fi based technology, that's distinguished itself in the past with their self-service location-mapping applications.  This is the first we've seen of their sensor fusion, which is in the process of being integrated into the rest of their solution.

Take a look at their sensor fusion demo below, and read more about it here.



We look forward to WiFiSLAM bringing this new technology to market, in conjunction with their mapping tools and the rest of their solution.

Want to understand all the ins and outs of indoor location positioning technology research and solutions?  Check out the Grizzly Analytics report on Indoor Location Positioning technology.

No comments:

Post a Comment