WHITE PAPER – Indoor Positioning
Indoor positioning solutions have been around for a long time but yet there is no standardised way to deploy such solutions.
The exact solution to be used is dependent on the customers use-cases and which problems to be solved. The purpose of this white paper is to outline different options and things to consider that will hopefully help you to make their right choices for your needs.
The value of Bluetooth LE
Bluetooth-based tracking solutions, from the tag to the cloud – with meter-level accuracy.
Download White paper for indoor positioning
How to integrate SenseCAP T1000 Tracker
SenseCAP T1000 is a tracker powered by Semtech LR1110, SenseCAP T1000 LoRaWAN® tracker utilizes GNSS/Wi-Fi/Bluetooth for both outdoor and indoor location tracking. With its compact card size, self-geo-adaptive capabilities, local data storage, and battery life lasting months, it’s perfect for various applications for location and status monitoring. Plus, its temperature, light, and motion sensors make it ideal for use in a variety of settings too. Learn more here.
Below you find brief summaries, but please read the rest of the document to get the full picture.
● Login to Traxmate and create your building(s).
● Deploy some more Wi-Fis and/or Bluetooth beacons (if needed).
● Perform an indoor survey to check the provided accuracy of your already installed Wi-Fi and/or Bluetooth infrastructure.
● Make a new indoor survey to validate.
● Either repeat the above steps or start using the indoor positioning solution.
● Connect SenseCAP T1000 LoRaWAN Tracker to TTN and send data to Traxmate.
Deploy the Wi-Fi/Bluetooth Beacons
If you do not already have an installed infrastructure of Wi-Fi access points and/or Bluetooth beacons, you can make a deployment that is optimized for positioning. Most already deployed infrastructures of Wi-Fi access points have been placed out most likely to optimize the range and performance of data connection. When considering positioning as well – think “spread it out and keep it close to corners and walls, and at a higher density in areas where you want higher accuracy”.
TIP Bluetooth beacons are most often battery powered, and can then more easilybe deployed. Wi-Fi access points have a longer range. A dense grid of Bluetooth beacons provides better accuracy than a sparse grid of Wi-Fi access points. Check E5 Bluetooth Location Beacon Deployment for more details.
Add New, then click the map or search an address or a name, click the
Add place popup and submit the details.
The next step is to specify the number of floors of the building and upload floor maps to the floors where you would like to enable indoor positioning. The uploading of floor maps supports PNG and JPEG files. Once you have uploaded the file you use tools to scale, rotate, and place it correctly on the map. There is also a tool to crop the PNG/JPEG file along the shape of the building.
After setting up the building, click
Do the survey
Once the building has been added, upload floor maps and update them.
- For each building, specify the number of floors and upload detailed floor plans (PNG or JPEG) for each floor where you would like to enable indoor positioning.
- On each floor plan in Traxmate Dashboard, you should add the respective level height associated with them in Indicate Level Height. Start from the lowest floor with 0 meters, then add up the respective meters to the higher ones
When you are done with floor maps it is time to perform the survey by using the Combain Indoor Survey APP ( Android version)
- Download the “Combain AR Indoor Survey” app from the Play Store here: Combain Indoor Survey – Apps on Google Play
Install it and start the app.
- Grant all permissions from app prompts, usually, it is 5 prompts; “Take pictures and record video”, “Device location”, “Physical activity”, “Relative position” and “Allow notifications”.
- Log in using your Traxmate account
- Download the “Combain AR Indoor Survey” app from the Play Store here: Combain Indoor Survey – Apps on Google Play
TIP Android normally has a limitation of how often apps can scan for Wi-Fi, it is set to once per 30 second. For the purpose of getting a more accurate survey you are recommended to turn off this throttling. You have to set Android in developer mode to be able to do this. Go to Settings > Developer options > find “Wi-Fi scan throttling” > disable it (or Settings>System>Advanced>Developer options).
Press Survey in the bottom bar on the home screen.
– Turn off WiFi via the prompt or otherwise
– The WiFi will still scan for access points, turning off WiFi improves the scan frequency and thus improves the survey results.
Find your position on the map and place the first reference point on your position
– Keep the phone slightly up, so you have a clear view of the area ahead with both the floor and the area ahead.
Walk for about 10 meters and place the second reference point.
– You should see the path you walked visualized as blue dots.
– If not, walk another 10 meters and place the next reference point. If track does not accurately display how you walked, restart the app and try again.
As a rule of thumb, place a new reference point every 15 seconds or after 10-20 meters traversed. Additionally, place extra manual reference points at corners or when passing through doors.
End the survey by pressing stop ( ⬛ ) in the button right and place the final reference point if prompted by the app.
-If you are happy with the survey, submit it. After a few minutes it should show up on the webpage.
Evaluate the current provided accuracy
Once you have performed the Indoor Surveys, then need to evaluate the results. You log into the Traxmate portal and visit the Positioning tab of your Place/Building. The raw data from the surveys will be processed by the back-end servers and the results will be available as Building Models. The details of the latest published Building Model are shown in the Positioning tab.
Building Models and the Median Error
A building model is an electronic representation of where beacons are placed within a building. The building model is generated when a user (or several) performs either surveys or reference tracks. As soon as these are finalized, they are sent into the system to be calculated into a building mode.
For each building model, a median error is calculated. The median error is based on the difference (the error) between the ground-truth/reference track (generated by the placed reference point during the survey) and the calculated track (based on the indoor positioning provided by the Combain Location API). The system automatically selects to publish the best available building model. The algorithm for “best” is based on a combination of coverage rate and median error.
The automatic process can be overridden by manually publishing a selected building model. The building models can be edited. The users are able to add – or edit beacons to further enhance the positioning effort.
Survey Coverage, RF Coverage, and Median Error Map
When evaluating a specific building model, to see if and how the accuracy could be improved, it can be useful to have a look at the Survey Coverage, RF Coverage, and the Media Error maps.
When reviewing the Median Error map and you find that the provided median error is not aligned with the requirements of your use case, then it is time to look into how the accuracy can be improved.
● Step 1 – More Surveys
Have you performed surveys in all areas of the building where you intend to use the indoor positioning? If not, perform more surveys.
● Step 2 – Increase Wi-Fi and Bluetooth coverage
Do you have Wi-Fi and Bluetooth in all the areas of the building where you intend to use the indoor positioning? If not, deploy more Wi-Fis and Bluetooth beacons in these areas and perform surveys in these areas.
● Step 3 – Increase Wi-Fi and Bluetooth density
You have Wi-Fi and Bluetooth in all the areas of the building, but the accuracy is still not good enough. Review the density of the Wi-Fi and Bluetooth in these areas. Compare with the tables in Appendix 1 and see if you from a theoretical and simulated perspective are recommended to increase the density of Wi-Fis and Bluetooth to meet your requirements. If so, deploy more Wi-Fis and Bluetooth beacons in these areas.
● Step 4 – Increase Survey accuracy
You have Wi-Fi and Bluetooth in all the areas of the building and the density is according to the theoretical levels suggested by Appendix 1 to meet your requirements, BUT the accuracy is still not good enough. Perform more detailed surveys. Make sure you place reference points as often as you can (once per square meter is optimal) and as accurately as you can. It is important that you are precise and correct when placing the reference point on the map. It is quite easy to make mistakes and place them in other places than where you really are located.
Start to Use
The SenseCAP T1000 Tracker gathers the Wi-Fi and Bluetooth scan information, MAC addresses, and signal strength and sends this to The Things Stack, then sends it to the Traxmate portal via the API.
Please check the Connect to TTN first to set up the tracker correctly.
Log in to Traxmate portal, navigate to
Add new ->
Log in The Things Stack, navigate to
Webhooks, and click
Custom Webhook template.
Name your Webhook ID and select the format to
JSON, then copy the Base URL.
TIP – Navigate to
Account, and copy the Service Token.
Enable the following recommended event type, then click
- Uplink message
- Normalized uplink
Check the device data
When the device is successfully connected, we return to the Traxmate portal, you will see the data of the device.
The method and technology we describe in this chapter are suitable for the use cases that require something around 2-10 meters of median error.
All accuracy requirements should be based on the needs of the use cases. Some use cases have higher requirements than others and some use cases have a higher available budget for indoor positioning infrastructure than others. The solution should aim to find a balance between accuracy requirements and the available budget. The rule of thumb is, that the more money and effort put into infrastructure, the better accuracy. But note, that many use-cases can be supported by using an already existing infrastructure.
The following charts show which accuracy you can expect if you have a certain amount of Wi-Fis/Bluetooth for a certain number of square meters.
Estimated Indoor Accuracy Matrix
The tables below show which estimated accuracy you can expect in an indoor environment using the AI (ANN) based indoor positioning. The accuracy will depend on the Area to be covered and the amount of deployed Wi-Fi Access points (APs) or Bluetooth beacons.
The median error accuracy is:
● GREEN from 1 up to 5 meters
● YELLOW 5 up to 10
● Red 10 meters
Table showing the accuracy if you have performed a very detailed survey with one reference point per 1 square meter.
Table showing the accuracy if you have performed a detailed survey with one reference point per 4 square meters.
Table showing the accuracy if you have performed a sparse survey with one reference point per 100 square meters.
How to make an Indoor Survey with the Traxmate app
By performing a survey in your building using the Traxmate app, you capture Wi-Fi and Bluetooth radio data that will be used for providing indoor positioning services. The Traxmate system will aggregate all the surveys performed in a building to create an accurate and complete radio map of that building
You will have a list of decoded devices that is ready to use. Traxmate support 1000+ of devices out-of-the-box but it can be configured to support any type of device.
In just a few clicks, you can connect a device and start consuming live data.
No coding is required!