Cross-correlation And RFI Excision Research For FAST 19-Beam Data

The wide frequency range and high sensitivity of radio telescope observation make the receiver extremely vulnerable to radio frequency interference(RFI).Because the single-antenna telescope hardly has the ability to distinguish the direction for the received the RFI signal,resulting in the coherent superposition of the astronomical signals and the RFI signal,the single-antenna radio telescope is extremely vulnerable to the RFI.The Five-hundred-meter Aperture Spherical radio Telescope(FAST),as the world’s largest single-antenna radio telescope at present,is also extremely vulnerable to the RFI.The RFI intensity and time domain/spectral density will directly affect the observation results and deteriorate the quality of the observation data.Meanwhile,improve the data processing speed and performance has always been one of the research hotspots in the computer field.The Graphics Processing Unit(GPU)is no longer limited to traditional graphics rendering.Parallel computing architecture is widely used in data processing.The Compute Unified Device Architecture(CUDA)is a high-performance parallel computing platform and programming model launched by NVIDIA.The RFIFIND command in the Pulsa R Exploration and Search Toolkit(PRESTO)determines the RFI information by independently analyzing the FITS file data of a beam.Based on this,the paper explores spatial filtering technology to the data collected by the FAST multi-beam receiver,using CPU + GPU heterogeneous parallel computing to improve the processing efficiency of astronomical data.The main work and innovation of the thesis include the following aspects:(1)A multi-beam cross-correlation and gradient descent algorithm is proposed,and the FITS data of the FAST 19-beam receiver is comprehensively analyzed to obtain the RFI information.The experimental results show that the RFI information obtained by this method is more comprehensive and more accurate.(2)Using gradient descent technology to simulate or locate the RFI distributions of the FAST 19 beams,obtain the intensity distribution of the RFI in the sky,then remove the information of the frequency band where the RFI is located to improve the signal-to-noise ratio of the original data.This research helps to maintain the integrity of observation data and improves the raw data utilization rate,and can greatly reduce the RFI impact during data processing,which is of great significance in improving the processing efficiency of FAST pulsar survey data.(3)Aiming to the problem of low data processing efficiency,a CPU + GPU heterogeneous parallel computing algorithm is designed.With high processing speed,the algorithm can be applied to other signal processing or astronomical data research and has good application value.