【摘要】 建立了描述目标在三维空间中进行切向与法向机动的非线性状态模型。目标切向与法向机动加速度的幅值表示为修正的瑞利-马尔可夫随机过程;法向加速度的方向角则假定在2π区间内具有均匀的概率密度。在仅有含噪声位置观察数据的情况下,发展了一种推广的卡尔曼滤波和自适应算法,并由此获得一种机动目标切向与法向加速度估值的直接方法。提供了某些计算结果以证实方法的有效性。更多还原

【Abstract】 This paper deals with the problem of target maneuvering acceleration estimation in three-dimensional space.Considering that it is more meaningful to know the acceleration components relative to a moving trace instead of those along longitudinal,lateral and normal axes of targets,a time-varying nonlinear state model describing tangential and normal acceleration maneuvers of targets in three-dimensional space is established.By the use of the concept of "current"statistical model of maneuvering target suggested in[2],the magnitudes of tangential and normal accelerations are both modeled as modified Rayleigh-Markov processes.The directive angle of normal acceleration is assumed to have uniform probability density function with variable mean in the interval of 2π In order to be suitable for Track-While-Scan application it is assumed that only noisy position observation data of target are available.The nonlinear state model is then linearized and discretized in the standard way.Based on these,an adaptive extended Kalman filtering algorithm is developed and therefrom a direct method for estimating tangential and normal accelerations of maneuvering targets is obtained.A flow chart of the adaptive extended Kalman filtering algorithm is given.The analysis and computer simulation confirms that it is possible to estimate tangential and normal accelerations only with noisy position observation data and without knowing the aspect of target.Therefore,the direct method for estimating tangential and normal accelerations proposed in this paper is especially suitable to Track-While-Scan usage.更多还原