Hybrid Positioning Algorithm for Tilted Receiver Using RSS and TDOA with Gaussian Process

In the visible light positioning (VLP) system, received signal strength (RSS) algorithm has a better noise ratio performance than time difference of arrival (TDOA) algorithm, while RSS needs to work under condition that transmitter and receiver are strictly parallel. However, is prone tilt due environmental disturbances, which reduces accuracy algorithm. For tilted receiver, TDOA higher order take full advantage two algorithms, we propose hybrid locate by using Gaussian process (GP). The scheme separately uses distance as inputs GP model estimate position receiver. Then, according proposed selection strategy, more credible estimated in our opinion selected final position. addition, information extracted from signal, allows prevent an increase complexity VLP system. During training testing, normalized meet order-of-magnitude requirements on input data. Simulation results validate based two-dimensional system for When standard deviations azimuth angle elevation 1°, 53.7% artificial neural network, 49.9% GP. localization error 1° angles 20.2% lower 20° angles.