A new sensor fusion algorithm for Alternative Position, Navigation and Timing (APNT) is designed with deep Recurrent Neural Network (RNN) using Long Short-Term Memory (LSTM). The proposed deep RNN for APNT estimates the position of the Unmanned Aerial Vehicle (UAV) using the UAV position measurements from the holographic radar and the Radio Positioning System (RPS). For the training dataset generation, flight simulations with multiple episodes are conducted with the measurement models of the holographic radar and the RPS. The testing results of the well-trained deep RNN are provided for verification and validation of the proposed deep RNN. The advantage of the proposed deep RNN over the Extended Kalman Filter (EKF) which is a conventional sensor fusion algorithm is demonstrated by comparing their testing results.
Recurrent Neural Network based Sensor Fusion Algorithm for Alternative Position, Navigation and Timing
03.10.2021
2657131 byte
Aufsatz (Konferenz)
Elektronische Ressource
Englisch
Alternative Position Navigation & Timing (APNT) Based on Existing DME and UAT Ground Signals
British Library Conference Proceedings | 2011
|British Library Online Contents | 2012
|The Case for Alternative Position, Navigation, and Timing (APNT) Development for Aviation
Online Contents | 2011
|Position, Navigation, and Timing for Security
Springer Verlag | 2020
|