Research On Aircraft Localization Method Based On Multi-Source Data Fusion On The Airport Field

In recent years,with the rapid development of China’s civil aviation transportation industry,the number of aircraft takeoffs and landings at the airport,the number of field operations is increasing,causing serious congestion at the airport field and thus making the airport field safety problems increasingly prominent.Timely monitoring of aircraft takeoff,landing,taxiing and other activities on the airport surface is an important technical means to prevent accidents,and is also necessary to ensure the safety of aircraft operations and improve the efficiency of airport operations.In the actual airport field traffic environment,the problem of low accuracy of single sensor positioning and the influence of sensors by environmental noise can occur,which in turn reduces the accuracy and stability of target positioning.Based on Bayesian optimal estimation theory and deep learning theory,it is demonstrated that the deep learning self-encoder with specific loss function constraints can be used to solve the optimal estimation problem of multimodal fusion localization of calibration-free images and GNSS(Global Navigation Satellite System,GNSS),which makes the calibration-free images have the ability to obtain high accuracy spatial localization from the GNSS noisy data.learning to obtain high accuracy spatial localization capability from GNSS noise data.However,the present method of fusing low-dimensional point trace data with high-latitude image data uses noisy GNSS data as sample labels to train the deep learning self-encoder,which greatly limits the localization accuracy and stability of the self-encoder fusion localization method.Therefore,in order to further improve the image/GNSS multimodal fusion positioning capability for aircraft positioning tasks,an image/GNSS multimodal fusion positioning method based on the Unscented Kalman Filter(UKF)preprocessing is proposed in this thesis.Firstly,a highly realistic multimodal digital airport dataset is constructed to provide the data basis for the set of alignment.Then,we use the UKF algorithm to pre-process the GNSS measurement data with noise reduction,which can realize data correction before data fusion.The optimal spatial position of the target is fused by filtering the image and GNSS data.Simulation results show that the method proposed in this thesis effectively improves aircraft positioning accuracy and stability,and the average positioning accuracy improves by53.31% when compared with five classical/frontier methods,indicating that the method in this thesis can lay a good technical foundation for the application of high-precision positioning of aircraft on the field,which effectively contributes to the improvement of field operation safety level and operation efficiency.