Research On Localization Technologies For Bistatic/Multistatic Sonar

Because of the separate configuration of transmitter and receiver, unlike the mono-static sonar, the bistatic sonar has a lot of different characteristics andlarge advantage in detection range, the probability of target detection, localization accuracy and other aspects. Currently, the research on bistatic/multistatic sonar system has attracted the world-wide attention, and becomes the key point and the hot. Based on the current situation, this paper made a thorough study on the localization technology of bi-static/multi-static sonar system and verified the effectiveness of the algorithm with the theoretical analysis and experimental simulation. Furthermore, the paper studied the multistatic sonar system for multi-target tracking and localization, and studied the theory and performance of the relevancy algorithm of various data. The main contributions of this paper are summarized as follow:1. Four factors that influence the localization accuracy when using the bistatic sonar system locating the target in two-dimensional plane were proposed. And targeting localization formulas under four different measurement subsets were derived. Principle of the algorithm localization was introduced then the corresponding GDOP formulas were derived. The localization error theoretically was analyzed. According to the simulation experiment, the various error factors of the different localization algorithms’ influences were analyzed. And then the GDOP distribution was gotten. The simulation results showed the bear measurement error has much more influence than the time measurement error to the localization accuracy. Both the baseline length and position measurement error have influence on the target-localization error magnitude. The simulation shows the regular patter on errors due to various factors, and it also has certain instructional significance for practical application.2. Correspondingly methods and derived GDOP formula to analyze the error distribution theoretically was proposed in this paper. Through the further simulation to validate the algorithm, the following conclusions can be obtained: the accuracy of distance localization in multistatic sonar system is higher than that of time-only-localization, and using both distance and bear information for the least-squarelocalization or neural network localization could obtain a higher accuracy. Meanwhile, the comparison of simulationverifies different effects on localization accuracy due to different base layouts which has a larger impact on bear-only-localization than on distance-only-localization. BOL is mainly influenced by the angle between the base stations, and either a too large or a too small angel will reduce localization accuracy; while the distance localization mainly based on the distance between base stations.3. Multistatic sonar system, algorithm models, basic principles, and common data association algorithms for target tracking were introduced in this paper.D-S evidence theory, nuclear clustering ideas, nuclear clustering algorithms and combined logical data association algorithm were also introduced.Both algorithms can effectively achieve point track association.Comparingwith the combined logical data association algorithm, the nuclear clustering one could obtain a lower error rate. For the two algorithms above, this paper conducted simulation under four different tracks including single-target linear track, single-target maneuvering track, bi-objectiveparalleled track and bi-objective crossed track. The simulation showsin the case of high noise density. Especially in the case of track intersections the combined logical data association algorithm would fall in error. Because of the proposal of comprehensive application of clutter, the accuracy of localization and speed tracking of nuclear clustering algorithmsis superiorto that of combined logical data association algorithm.