Indoor Visible Light Tilt-Resilient Positioning Based on RSS Fingerprint Self-Transferability

In this letter, we propose a tilt-resilient positioning method based on received signal strength (RSS) fingerprint self-transfer learning to solve function failure or a reduction in the accuracy of the indoor visible light positioning (VLP) system under tilt receiving and to improve its practicability and reliability in the actual environment. Based on the modified path loss (PL) exponent model, the vertical RSS fingerprints can self-transfer to the tilted RSS fingerprints of arbitrary receiving direction (RD) in the online stage. The constructed tilted RSS fingerprint databsae have the same granularity and dimension as the vertical RSS fingerprint database, making the proposed method have the capability of rapid learning and positioning while completing RSS fingerprint self-transfer. The simulation results show that the proposed method achieves a higher positioning accuracy compared to the other three methods without adding the labor and time costs of RSS measurement caused by the multidimensionality of the RD at each fingerprint point. In the case of dense and sparse light-emitting diode (LED) distributions, when the receiver tilt angle is 31(degrees), the root mean square errors of the proposed method are approximately 4.21cm and 3.52cm in the 10cmx10cm grid size, and the average positioning errors are approximately 3.09cm and 2.43cm, respectively.