Font Size: a A A

Research On Modeling And Filter Algorithm Designing For Underwater Target Passive Tracking

Posted on:2007-08-25Degree:MasterType:Thesis
Country:ChinaCandidate:B WangFull Text:PDF
GTID:2178360185963525Subject:Control theory and control engineering
Abstract/Summary:PDF Full Text Request
Nowadays, there are some disadvantages in underwater target passive tracking field, including the inaccurancy of modeling and low tracking precision, slow convengence speed of filter algorithm. According to these shortcomings, this paper makes deeply analysis. The main research works are as follows:1. The development history of two main problems in passive tracking is surveyed. One is the formulation of the system model, and another is the algorithm of passive tracking.2. The affection of different coordinate system and system state model for target tracking is analyzed, and for underwater passive tracking, "current" statistic model which is under rectangular coordinate and hybrid coordinate is used. The observability of single-observers or double-observers for underwater passive tracking is investigated.3. Contrast to extended kalman filter and pseudo-linear measurement model kalman filter under the same simulation environment which is based on "current" statistic model, the simulation shows that the latter has better performance, but with much more computational complexity. At the same time, the necessary and sufficient condition of astringency for the divergence of EKF is presented.4. The kalman filter algorithm under hybrid coordinate and unscented transformation (UT) algorithm are investigated. By combining the excellent performance of hybrid coordinate with UT algorithm, a new algorithm which is used to this problem is presented. The simulation shows that the new algorithm has more performance in the stabilization, estimate precision and astringency.5. The combination of the kalman linear filter with the UT is discussed after analyzing the scheme of Sigma sampling and the unscented kalman filter (UKF) algorithm is presented. The simulation show that passive tracking of maneuvering targets based on "current" statistic model, UKF has better performance in the estimate precision and astringency.6. This paper introduces the basic principle of particle filter, and analyses importance functions and resampling algorithms of the particle filter. Then , the paper chooses EKF and UKF as importance function, selects bootstrap resampling, systematic resampling, residual resampling as resampling algorithm, and applies PF, EKF-PF, UKF-PF to underwater target passive tracking. The simulation results shows that on the same simulation enverionment, UKF-PF algorithm has better estimating precision and convergence than the formers, but it loses more computation time as expense.
Keywords/Search Tags:Passive Tracking, "Current" Statistical Model, Hybrid Coordinate, Extended Kalman Filter, Unscented Kalman Filter, Particle Filter
PDF Full Text Request
Related items