A CSI Fingerprint Method for Indoor Pseudolite Positioning Based on RT-ANN

At present, the interaction mechanism between complex indoor environment and pseudolite signals has not been fundamentally resolved, stability, continuity, accuracy of positioning are still technical bottlenecks. In view shortcomings existing fingerprint methods, this paper proposes a hybrid CSI method for based on Ray Tracing artificial neural network (RT-ANN), which combines advantages actual acquisition, deterministic simulation, network, adds simulation feature parameters generated by modeling to input extending sample features dataset. Taking an airport as example, it is proved that can improve in area where fingerprints have collected, error reduced 54.7% compared with traditional method. It also preliminary be completed without collection.