Unmodulated Visible Light Positioning

Localization based on visible light communication is rapidly attracting more attention as the adoption rate of LED lights increases. The cost of these systems is relatively low due to the integration with existing lighting systems, yet accuracy can be sufficient (< 10cm RMS) by taking advantage of the well-known characteristics of LED lights. However, all current systems rely on modulated light sources. Through this modulation, the fixtures can for example transmit unique codes. A photodiode receives the superposition of the signals from all light sources. Through cross-correlation, the signal can be separated into the components originating from the individual light sources. These components are then converted to distances through a signal strength model. Finally, triangulation is used to determine the position of the receiver. In our work, continuous light sources are used for localization. A receiver mounted on a mobile robot collects light intensity measurements. These measurements are compared to a signal strength model within an Iterated Extended Kalman Filter algorithm. Through data fusion with encoder measurements, a location at every timestep is estimated. This approach, while currently experimental, could enable extremely low cost indoor localization.status: accepte