Research On Algorithms Of Multi-target Passive Localization With Missing Information

Passive localization detects and locates targets by receiving radiated or reflected signals from targets,which is more covert than active localization.Multi-target passive localization has gradually become the mainstream research direction in the complex combat environment.Due to various interference factors in actual scenarios where airborne observation stations locate multiple ground targets,observation stations may not be able to detect all target information as expected,which is referred to as missing information.There are three cases of missing information: First,there are obstacles between the observation station(s)and the targets;Second is that some sensors on the station are damaged;The third case is that some observation stations are invalid.This paper focuses on the problem of multi-target passive localization with missing information.First,Direct Positioning Determination corrects the disadvantage that it is easy to loss information in each processing stage for "two-step" passive localization algorithms.Second,Compressed sensing is a random undersampling technique,which can significantly reduce the amount of measurement data required.Based on these,Direct Positioning Determination is considered to be combined with sparse reconstruction algorithms based on compressed sensing model for multi-target localization with missing information in this paper.For the multi-target passive localization with known target number,consider using several fixed observation stations in the air to observe multiple targets on the ground and the OMP algorithm is adopted.Simulation results show that there are a large number of false points in the inner product vector when locating more than two targets via OMP,resulting in a portion of the targets being estimated.For better positioning performance,an optimized OMP algorithm is proposed by conducting a secondary screening of the selected elements in the index set during iterations,removing false points and zeroing the amplitudes of the selected elements and their adjacent terms to remove duplicate indexes or their adjacent indexes.Simulation results show that compared to OMP,the optimized algorithm has higher positioning accuracy with faster positioning and can estimate more targets correctly under the same conditions.Compared to the improved MUSIC algorithm,it has much better positioning accuracy under low signal-to-noise ratio and can estimate more targets with smaller positioning errors with missing information,resulting in better performance.For the multi-target passive localization with unknown target number,consider using a single aerial motion station to observe multiple far field targets on the ground.The regularized FOCUSS algorithm is adopted.Simulation shows that the algorithm has low positioning speed and there are some interference items in the results.For better positioning performance,a quadratic weighted FOCUSS algorithm with optimized regularization factor is proposed.First,From the point of view of the meaning of regularization factor and formula,updating the regularization factor dynamically instead of taking a fixed value;Second,apply quadratic weighting on the weighting matrix to improve the convergence performance and speed of the sparse estimation vector;Third,a new method of selecting the correct indexes from the sparse estimation vectors was proposed to improve the accuracy of multi-target position estimation results.Simulation results prove that compared to the regularized FOCUSS algorithm,the improved algorithm has better convergence,double the reconstruction speed and higher positioning accuracy,and can estimate two targets when the signal-to-noise ratios is as low as-15 d B;Compared to other algorithm in this paper,it has the best positioning performance with missing information.