Research Of Passive Ranging With Cross-Spectra

Active SONAR provides signal for enemy when ranging by transmitting pulse and this is very dangerous for submarine. Therefore, passive SONAR, aiming at universal noise targets, becomes the dominating SONAR of submarine. As we know, noise target ranging technology is one of its main problems. However, the pro passive sonar used middle-small integrate circuit with a huge frame, and it can only range far targets and cannot track move fast targets, such as torpedo. It cannot meet the modern naval battle. With acoustics , signal process theory and electron technique developing, we possible can empolder a more advanced ranging system.The purpose of this thesis is to study and design a digital noise targets passive ranging system which will be used to overcome the imperfectness of conventional SONAR, to track the whole course of a noise target, to follow a fast moving target, as well as to adapt the system to different flank three sub-array types without the limitation of installation place.It first mainly probes into the theory of passive directing and ranging, then analyses some kinds of elements that influence time-delay estimation, finally gets an optimized ranging formula fit both far and near targets in this papers.It tests many ways of the cross-spectra method with simulation, such as chose the weight of least-square estimate, chose the length of correlation window and four frequency domains colligate.It worse to mention about that it will improve greatly the precision of time-delay estimation, if torpedo line spectra exits. Some simulation results is given which the system detecting threshold drops down 3dB and improves both the precision and quality of time-delay estimation, because estimation invariance drops and the mean of estimation approach the true value.A tradition sonar often used Kalman-filter as a post-processor which improve precision with long integral time, however it will bring a serious lag if targets move fast. A parameter-estimator with adapt filter is used in this paper, consisting of two α -β filter and an adapt filter. A post-processor can be got if we choose proper α and β coefficient.