Research On Interference Recognition And Individual Recognition Technology For TETRA Signal

Terrestrial Trunked Radio(TETRA)is a unified standard communication system formulated by the European Communication Standards Association.It was first put into use in 1998.Because tetra has many advantages,such as open standards,mature technology and flexible networking,Tetra is widely used,such as some large public utilities,especially some emergency departments.The networking and broadcasting functions provided by tetra can enable these units and departments to complete their work quickly and efficiently.Despite the above advantages,Tetra’s wireless channel space is open,so it is vulnerable to electromagnetic interference inside and outside the system.In the complex electromagnetic environment,if we can effectively judge whether tetra is disturbed and what type of interference it is,then we can take corresponding anti-interference measures to avoid or suppress interference to the greatest extent,which is the goal and significance of interference identification.If the result of tetra signal interference identification is that the signal is not interfered,that is,the non-interference tetra signal,further study on specific emitter identification(SEI)can be carried out to determine which equipment the Tetra signal is transmitted from,so as to realize the accurate identification of specific emitter.Therefore,aiming at the problem of interference signal recognition and individual recognition of specific radiation source,the research contents of this paper are as follows:(1)In the aspect of signal interference identification,three identification algorithms are mainly used to classify and identify communication interference signals.Firstly,based on the Gaussian white noise environment,four typical interferences are superimposed in the undisturbed original tetra signal to complete the mathematical modeling of the interference signal;Secondly,the characteristic parameters of the test signal are extracted from time domain and frequency domain respectively;Finally,the classification and recognition algorithm is used to realize the detection and recognition of interference signals.(2)This paper focuses on the feature extraction method of radiation source fingerprint.Based on the bispectrum theory,a correlation feature of bispectrum is designed,which solves the problem of obvious degradation of recognition performance in the environment of increasing the number of radiation sources or low signal-to-noise ratio.The experimental results show that except for 8PSK modulation signal,the recognition accuracy of other modulation signals is more than 80 %,which shows the effectiveness and generalization of this method.(3)Aiming at the sensitivity of local neighborhood size selection of high-dimensional data in classical manifold learning algorithm,a dimension reduction method of adaptive domain selection is designed,which dynamically selects the domain size of each sample point without changing the parameters artificially.It effectively solves the problem that the domain point selection in the dimensionality reduction algorithm is not adaptive.The experimental results show that the proposed adaptive local linear embedding algorithm not only ensures the recognition accuracy,but also effectively improves the performance of the model.