An Extensible PNT Simulation Verification Platform Based on Deep Learning Algorithm
In this paper, a simulation platform for multi-source PNT (Positioning, Navigation and timing) means is designed. The platform, which adopts a distributed simulation architecture to transmit data through the network, provides a test environment for multi-source PNT users and performs performance evaluation. Through in-depth learning of the output data of various PNT tools collected by different types of users in different scenarios, the platform can achieve parameter settings of multi-source PNT based on user characteristics. Seen from the test results can give the simulation platform, weight value will affect the comprehensive results, deep learning can make the weight value distribution is more close to the human judgment through a large amount of data to judge the artificial experience, so as to realize the autonomous PNT multi-source fusion.
KeywordsPNT Simulation platform Deep learning algorithm Distributed architecture
Our thanks to Beijing Municipal Science and Technology Commission for supporting us to develop the platform.
- 1.Li, N., Zhang, Y., Xi, H.: Some cognition on the new PNT technology under designed by USA. Satell. Appl. 12, 34–37 (2015)Google Scholar
- 2.Zhang, F., Ou, M., Liu, D., Zhen, W.: Future potential PNT review. GNSS World China (2015)Google Scholar
- 4.Nei, Z., Jia, D.: Overview of research on deep learning. Tech. Innov. Rev. 30(2), 224–226 (2015)Google Scholar
- 5.Zhou, G.: Research on Deep Learning Based Image Recognition Application. Beijing University of Technology (2016)Google Scholar
- 6.Zhu, Y.: Research on Very Long Baseline Interferometry Applied to Precise Orbit Determination. University of Electronic Science and Technology of China (2012)Google Scholar