Optical channel model taps LEDs for positioning
A system exploiting the increased use of LEDs for lighting and communication features a functionality that enables it to act as beacons for positioning services.
Epsilon has been the result of collaboration between Microsoft Research and the University of Massachusetts. The system builds on the existing lighting infrastructure, applying trilateration to localise smartphones, tablets, or any device with light sensing capability—with LED lamps as anchors.
Figure 1: Epsilon establishes and experimentally verifies the optical channel model for localisation.
The system takes advantage of LED fixtures being installed for light. This requires little additional cost to provide the beaconing system, unlike the case with radio-frequency beacons or the use of Wi-Fi. In essence, the system's infrastructure constitutes the LED lights and smartphone.
A description on the Microsoft Research website states that Epsilon addresses several practical challenges to using visible LED lights for tri-angulating the position of devices with a light sensor, coming up with localisation accuracies of less than a meter in typical office environments. Guiding consumers to shops and even products are also use cases for the system.
In creating the prototype, the authors adopted binary frequency shift keying (BFSK) and channel hopping to enable reliable location beaconing from multiple, uncoordinated light sources. They added a control board implementing BFSK to an off-the-shelf LED bulb and a light sensor board to a mobile phone that communicated with the phone via the audio jack. Modulation of the optical spectrum was done at 10kHz to 19kHz to produce 30 channels of 300Hz bandwidth.
Figure 2: The system uses the most practical of materials for its infrastructure: LED lights and smartphone.
The system has been implemented and evaluated in a small-scale physical test-bed as well as moderate-scale simulation. The experimental results have achieved the 90th percentile accuracies of 0.4m, 0.7m and 0.8m for three typical office environments. Even in the extreme situation with a single light, the 90th percentile accuracy is 1.1m.
The researchers believe that visible light based localisation is promising to significantly improve the positioning accuracy, despite few open problems in practice.
(With inputs from Peter Clarke)