Skip to main content Skip to secondary navigation

Code Design Using Machine Learning for Future Satellite Navigation

Main content start

Research Question: Can We Rethink the Design of Satellite Navigation Codes with the Recent Advances in Machine Learning? 

On January 14, 2020, the first GPS III satellite was marked healthy and available for use, thereby marking the birth of the next-generation GPS constellation. In addition to broadcasting the new L1C signal, the modernized constellation is distinguished by its reprogrammable payload, which allows it to evolve with new technologies. Furthermore, the NTS-3 satellite signal testing platform is planned to be launched in 2022 and will explore new technologies for the future GPS constellations. Additionally, in a 2016 Request for Information, the AFRL has expressed interest in exploring modifications to all layers of the GPS signals. We are indeed entering a new era of satellite navigation. However, the legacy GPS codes are based on linear shift feedback registers, which were designed decades ago before personal computers. It is time to revisit the design methods of the GPS spreading code families. Powered by recent advances in machine learning, we explore new platforms for learning high-quality spreading signal families via genetic algorithms as well as a Natural Evolution Strategy machine learning framework.

Research Team: 

  • Tara Mina
  • Ashwin Kanhere
  • Ridvan Yesiloglu

Related Works: 

  • Tara Mina and Grace Gao, Designing Low-Correlation GPS Spreading Codes with a Natural Evolution Strategy Machine Learning Algorithm, Navigation: Journal of the Institute of Navigation. vol. 69, no. 1, Mar. 2022. doi: 10.33012/navi.506. [paper][video]
  • Tara Mina and Grace X. Gao, Designing Low-Correlation GPS Spreading Codes via a Policy Gradient Reinforcement Learning AlgorithmProceedings of the Institute of Navigation GNSS+ conference (ION GNSS+ 2020), St. Louis, MO, Sep 2020. Best Presentation of the Session Award. [paper] [slides] [video]
  • Tara Mina and Grace X. Gao, Devising High-Performing GPS Pseudo-Random Noise Codes Using Evolutionary Learning AlgorithmsProceedings of the Institute of Navigation GNSS+ conference (ION GNSS+ 2019), Miami, FL, Sep 2019. Best Presentation of the Session Award. [paper] [slides]