Passive Acoustic Sensor Network Target Probabilistic Localization Algorithm

Passive acoustic sensor detection technology is a target location technology by getting the acoustic signal passively with acoustic sensor. It can cover the shortage of radar which can't detect the low altitude target and it is more anti-jamming and can't be detected by enemies. When the acoustic sensor is working, the nonlinear measurement and time-delay increase the difficulty if target is located. Under the pre-research projects of "11th Five-Year Plan", the paper researches on the single target & multi-target location algorithm in acoustic sensor network. The major research work and achievements are as follows:(1) By the sensor working feature, acoustic time-delay and sensor networking, the research background and status of acoustic sensor system are described. Meanwhile, the system difference between non-linear target tracking and acoustic sensor is analyzed. Some passive localization algorithm is described in the chapter. It contains passive across localization algorithm, passive time-difference localization algorithm and probabilistic localization algorithm.(2) To the characteristic that target signal is measured by acoustic sensor with the character of time-delay and bearing-only, probabilistic localization algorithm and time-delay registration are used to solve the motion target analysis. A passive acoustic sensor network single target probabilistic localization algorithm is present in this chapter. Firstly, we can choose two measures that the intersection angle is close to 90°to get the initial position of the target. Depend on the initial position, time-delay registration is processed, and then we refresh the measure data which the intersection angle is closed to 90°to get the new initial position estimation of the target. Finally, use the new initial position to get the more accurate estimation by probabilistic localization algorithm.(3) To the problem that acoustic sensor network multi-target localization can appear many false location points and the target estimation in different time can't be associated, a passive acoustic sensor network multi-target probabilistic localization algorithm is present based on the real demand of projects. Firstly, by state prediction of the target to eliminate the false location points which are brought by across localization. Secondly, the multi-target location is located by iterative probabilistic localization algorithm combined global nearest neighbour algorithm. Finally, multi-target motion trajectory is implemented by conic fitting algorithm. Simulation results show that by some hypothesis, the passive acoustic sensor network multi-target probabilistic localization algorithm completely applicable the multi-target localization problem in sea surface and low altitude environment.Finally, the whole dissertation is summarized. The algorithm in this paper has been tested in the acoustic detection emulation system, and the results showed that they satisfied the requirement of the forewarning system.