Please find another close the same messages or questions on BLE indoor localisation to me as below. I get the question again and again at least once a month. It is time for a block post on this topic.
- What is the recommended way to implement BLE beacons to perform indoor localisation?
- What is the preferred algorithm for iBeacons to perform user localisation? Fingerprint based? Triangulation based?
- How can you triangulate a phone’s exact location in a restaurant with only a few beacons?
- I would like to talk to you about BLE tracking. I have been working on a project for a while and running into walls
- We are a team of 5 people who are interested in the concept of proximity messaging and we are interested in developing a Mobile App in IOS and Android so that Consumers can benefit from the offers which retailers will be offering them by proximity messaging while they are in their stores
- We are working on a project where location Accuracy is critical for our product. I have looked through your articles on Low Energy BT for triangulation. Specifically, we are having problems in Urban Canyon environments. Do you have any articles/ideas that will provide the high level of location accuracy in these environments? Please send me any relevant info.
Triangulation on BLE is impossible because on 2400 MHz you will have a lot of reflections, multipath plus phase shift. If you would like to make a study, then just order a kit with 5 BLE beacons for estimated Euro 50 including free of charge app for iOS and Android. After that, you place the 5 beacon or even more in a room with no people inside. Put your smartphone on a trolley and move it around. The measured field strength you write in an Excel table. After that, you make the same test with 5 beacons and 5 people and put the BLE beacon in their pockets. Let the people turn in steps of 15 degrees and measure it the field strength again. The human body is mainly water and the resonance frequency of water is in the near of 2400 MHz. Remember that your microwave oven is operating in the 2400 MHz frequency range for that reason.
The third test could be to measure the field strength of 5 beacons in front of you at 5 different places. Just walk in steps of one-meter closer to the beacons and write the measured field strength in a table. The result will be that the field strength will maybe decrease even when you are moving closer to the beacon. The reason is multipath and phase shift. Another result will be that the field strength will not increase linearly. This is based on multipath phase shift as well.
Fingerprinting will need a reference. However fingerprinting is not immune against multipath and phase shift as well. If you change the smartphone, the place in your pocket change then the signal loss with change as well. Fingerprinting will help to get a better result, but not accurate enough to generate a position on meter level. The fingerprinting will run better if you place BLE beacons in the room and measure the field strength with another beacon. A beacon to beacon communication will help with more accurate data than a beacon to smartphone communication. There are some further algorithms that you can use for optimising the calculation of the position, but this you can’t do by using smartphones on BLE only.
A detection on room level is possible. If you add Kalman filters and some more sources like the magnetic field of the earth, accelerometers or even ultrasonic sound then the result will be much better than on BLE beacons only.
Proposal for the Euro 50 BLE indoor localisation test set up:
Wall-mounted BLE beacon
Wall mounted bracket
The wall mount bracket gives you the freedom to assemble the Bluetooth Low Energy beacon to the wall, laptop, bicycle, skis or whatever you like.
If you plan to develop an indoor localisation then do not hesitate to email to harald.naumann(at)gsm-modem.de to get an offer on Wi-FI / BLE or any other radio technology. Please feel free to ask for a feasibility study or consulting as well.