The Analysis And Implementation Of Radio Real-Time Direction Finding System Based On GNU Radio

The direction of arrival(DOA)refers to the direction of signals in space relative to a reference element in the array.The research on DOA is always a front research task in the field of array signal processing,of which DOA estimation serves as core technologies in diverse applications,such as communication,radar and exploration.However,with the gradual lack of spectrum resources,security incidents caused by radio interference continue to emerge,in particular,the rise of drone technology has led to risk factors including disrupting public activities,smuggling prohibited items,threatening commercial air traffic,and entering restricted airspace,this makes the need for identification and direction finding of air interface signals increasingly prominent.Since various signals existing in the space are subject to object reflection and diffraction,etc.,multipath effects may be formed,which causes great interference to monitor the direction of incoming signals.Subspace decomposition algorithms with super-resolution capability,such as MUSIC and ESPRIT,can effectively distinguish DOA by using the properties of noise subspace and signal subspace,which is a typical method in array signal processing theory.Shrinkage-MUSIC can effectively improve the DOA estimation performance under the limited number of samples,but in the environment where the coherent signal exists and the signal sources are close together,the algorithm cannot effectively distinguish DOA,and its performance is poor at low signal-to-noise ratio(SNR).SOD-MUSIC is simple to implement and has great resolution performance when the signal sources are close together,and has certain decoherence ability.But in a low SNR environment,its ability to estimate DOA is weak.This paper mainly aims at low SNR,small snapshots,closed signal sources,and proposes a DOA estimation method based on modified MUSIC which is suitable for coherent environment.Then a radio direction finding system is implemented by applying this algorithm.The major contributions of this thesis are as follows.1)Aiming at the problem that DOA estimation is inaccurate when the SNR is low,the number of samples is limited,and signal sources in space are close,a DOA estimation algorithm suitable for uniform linear array is proposed.And the algorithm solves a problem that the performance of DOA estimation is degraded due to the biased estimation of the covariance matrix.The method performs preprocessing on the received signal data matrix to obtain the modified covariance matrix,and then reconstructs the matrix by using the shrinkage estimation method.Thus we can construct the spectral estimation function with eigenvalues and eigenvectors which are obtained by eigenvalue decomposition.Finally,using the secondorder differential method to this function can obtain a result.Simulation results show that the proposed method has a higher performance than the compared methods when the number of samples is limited,and signal sources in space are close.2)Aiming at the problem that the current radio direction finding system is inconvenient and expensive,a radio direction finding system based on GNU Radio is realized,which verifies the proposed algorithm.According to the characteristics of two Twin RX daughter boards build into the USRP X310,the phase correction is required to obtain the phase difference data,and than it can estimate DOA using the proposed algorithm.Experimental results show that this algorithm can effectively estimate DOA,and the system can distinguish the orientation of signals by changing the position of signal sources.It shows that the system meets the requirements of applicability.