水下对空中声源的运动参数估计

为了解决传统的交互式多模型目标跟踪算法中马尔可夫概率转移矩阵固定不变,造成的模型切换缓慢、跟踪精度不高的问题,提出了一种基于后验信息修正的时变转移概率自适应交互式多模型跟踪算法。算法定义了一种新的修正因子,利用后验信息对概率转移矩阵进行实时修正,提高匹配模型的概率,减小非匹配模型的影响,使得系统模型能够及时、准确地切换到匹配模型。蒙特卡洛仿真实验表明,该自适应交互式多模型算法能够应用于水下目标跟踪中,相比传统交互式多模型算法,模型匹配度更高,滤波效果也更好。

Motion parameters estimation of airborne source from underwater

The spectrum of airborne source includes broadband continuous spectrum and narrowband line spectrum. The energy of line spectrum is higher than that of continuous spectrum. Compared with underwater sound source, the velocity of airborne source is generally higher and the Doppler shift is more obvious, so it can be used to estimate the motion parameters of airborne source from underwater. Firstly, the instantaneous frequency of the received signal is extracted by time-frequency analysis, and then the extracted instantaneous frequency is fitted with the predicted value by using the non-linear least square method. Then, the motion parameters of the sound source are estimated. Both simulation and experiment can accurately estimate the motion parameters of the sound source. Besides, ranging and positioning of the airborne moving source are successfully performed in the experiment, and the ranging error is less than 15.8%. This parameter estimation method has good applicability under the condition of satisfying a certain signal-to-noise ratio and ensuring sufficient Doppler information.