Based On The Single Pulse Compression Perception Super Resolution Direction Finding Method Research

In modern radar systems, the monopulse technique is widely used to estimate the target direction of arrival (DOA) by computing the output power ratio of the sum and the difference channel, and the angle resolution is about one beamwidth, which means that when multiple targets are closely present in the same azimuth cell, it can not obtain each estimation correctly. With the increasing development of military technology, the need for detecting and estimating certain parameters of multiple targets becomes very common, resulting in more significances of improving the resolution by signal processing methods. Combined with the Compressive Sensing (CS) theory, this thesis focuses on how to achieve the super-resolution angle measurement of multiple targets in single antenna mechanical scanning systems.The main contributions are summarized as follows:1. Analysis the feasibility of the super-resolution angle measurement implemented by the signal from the sum channel only. A single antenna model under mechanical scanning systems for DOA estimation is introduced. Describe the algorithm CML, UML and AML in two cases that the target Doppler frequences are known or unknown. The performance of each algorithm is investigated through simulations.2. Introduce the CS theory into traditional monopulse technique, so that it can be succeed in super-resolution DOA estimation. We propose a CS model for DOA estimation, and make echoes be sparse by dividing the whole space into some equal intervals, as a result, the estimation problem turns into a sparse signal reconstruction problem. Take three kinds of typical reconstruction algorithm BP, OMP and ROMP for example, we verify that these methods based on CS can achieve correct DOA estimations of multiple radar targets present in the mainlobe of a rotating antenna, according to the simulation results.3. The two categories of proposed algorithms are compared from different viewpoints, whose advantages and disadvantages are investigated. The ones based on the ML are able to obtain angle estimated values with small errors compared to ture values, however, their complexities are too high, and there is a requirement for the Doppler frequency difference between targets, that is to say, they are sensitive to the Doppler frequency. The other ones based on the CS cost less time, and they are insensitive to the Doppler frequency of each target.4. To alleviate the effect of inter-atom interference on the performance of greedy algorithms, the concept and design scheme of sensing dictionary are introduced, and applied to a model for DOA estimation. Adding this sensing dictionary into OMP, its effectiveness is showed by simulation results. It can suppress the inter-atom interference successfully with or without noise, which improves the performance of greedy algorithms while maintaining the same complexity.