Dark, rain and Covid. Challenging times but great times for doing research. We got some help from brilliant students and that speeds things up.
1 Route extraction:
We collected a lot of tracks in our 400m2 office area with 15 WiFi APs and 31 BLE beacons. Now let’s see if we can use these tracks for something. We developed a method to combine all these tracks to main paths then extract nodes from the data. As you can see above the “route extractor” works on any positioning data, no matter how noisy. But as always, good positioning data in gives a good result out. Thus using more accurate RTT measurements instead of just RSSI measurements improves the extracted routes. But still amazing that main paths can be extracted automatically this way. Automatic extraction of these paths is a must for being able to do global indoor navigation.
2. RTT studies:
RTT measurements show a strange behaviour. Seems like we get reflections from walls, floor and roof which sometimes gives wrong distance measurements that we would like to detect to remove. Then our researcher at Lund University came up with the idea of let’s try and use these outlier distance measurements to estimate the room and the walls. Could it be done?
As we can calculate the transmitter and receiver location from good measurements, we could estimate the location of the walls by looking at the outlier data received. See picture below. Blue circles shows where a reflection should happen for the outlier measurements. Many blue lines indicates a wall and the best estimate of walls is the green box. Not perfect, but still better than expected!
We have decided to revise the project schedule and prolong the project to end September 2021. So more to come.
Now a well deserved Christmas holiday. Best wishes for the new year and let’s it is much better than 2020.
Stay safe and stay tuned!