An Indoor Pseudolite Positioning Method Based on Measured and Simulated Fingerprints

At present, the indoor positioning method of pseudolite based on fingerprint is restricted by the range and density of fingerprint collection, and its stability, reliability, and accuracy are still technical bottlenecks. Aiming at the shortcomings of existing indoor fingerprint positioning, combined with the advantages of deterministic ray tracing simulation, a new hybrid method of measurement and simulation is proposed to generate the fingerprint feature input set to drive the position prediction model. The feature parameters generated by the simulation calculation are added to the neural network as part of the input, the original artificial neural network structure is improved, and the sample features of the neural network input data set are expanded. Taking the indoor arrival hall of an airport as an example, the static and dynamic positioning tests are carried out by combining the simulation and measured fingerprint construction methods. Compared with the traditional positioning effect based on fingerprint collection alone, the static positioning error of the proposed method is reduced by 42.2%, and the dynamic positioning error is reduced by 46.9%.