We have found in testing that some RSSI calculations by Bluetooth chipsets use different scaling for RSSI. Inability to use one distance conversion formula Nevertheless we did discover two interesting aspects of RSSI recording and distribution we shall now detail for background information important to the design of a measure to assess accuracy.

These efforts did not result in the development team being asked to take multiple measurements or perform any other ‘at time of recording’ data processing operations. We ended up rejecting these in favour of a raw RSSI value recorded at a fixed point in time purely because a separate scientific research effort being led by The Alan Turing Institute and the University of Oxford was assessing how RSSI could be processed. Practical observationsĭuring the development of the protocol we noticed some trends and approaches that worked well in testing to make the RSSI value calculated more accurate. Please do contact us with any interesting Distance Estimation discoveries so we canĪdd them to this repository.

Follow the instructions in that file to point to your data and generate analysis output.

Open the r/calibration-analysis.R script in RStudio.

Navigate to the PHONE NAME device in FinderĪfter doing this you can generate analysis output:.

Instructions of how to use the app are within it

Deploy locally on two iPhones and run the app.

Modify the signing authority to be your own.

Currently this supports iOS but Android is also in the works. Herald provides a sample Bluetooth calibration tool. Measuring distance is a case of using a physical protocol based analogue, and later converting this based on physical attributes of the device and environment

Bt rssi