基于多智能体导航的高超飞行器信息融合方法

精确导航技术是高超飞行器(hypersonic vehicle, HV)充分发挥威力的关键所在。然而,高马赫数和强机动性致使HV的导航系统误差及其观测噪声无法准确描述,制约着导航信息的精确性和实时性。为及时获取高精度导航信息,设计基于集员框架的卡尔曼滤波算法。一方面采用多智能体分布式协同探测,形成观测椭球交叉集合,提高了观测效率和测量精度;另一方面,通过设计两类噪声模型,求其与状态估值的最小均方误差,实现滤波增益的计算,提高算法对噪声的抗扰动能力,使状态估值达到均方误差最小。通过数字仿真,将设计方法应用到HV导航模型中,并与扩展卡尔曼滤波和集员滤波的状态估值进行比较,结果表明,提出算法在不同噪声影响下具有更高的估计精度。研究成果将为HV实现实时精确导航提供技术支持,并具有重要的理论意义和应用价值。

Information fusion method of hypersonic vehicle based on multi-agent navigation

Precise navigation technology is the key to performing the important function for hypersonic vehicle (HV). However, the high Mach number and strong maneuverability make the navigation system error and make the observation noise of HV hard to be accurately described, so as to constrain the navigation information accuracy and real-time performance. In order to obtain high precision navigation information in time, the Kalman filter algorithm based on the set membership framework is proposed. On the one hand the multi-agent distributed cooperative detection is adopted to get the intersection of observation ellipsoid, which improves the observation efficiency and measurement accuracy. On the other hand, two kinds of error models are designed and used to solve the minimum mean square error together of the state estimation in order to calculate the filtering gain. The anti-disturbance ability of the proposed algorithm to noise is improved and the mean square error of the state estimation can reach the minimum value. Through digital simulation, the proposed method is applied to the navigation model compared with the state estimation of the extended Kalman filter and set-membership filter. The results show that the proposed algorithm has a higher estimation accuracy under different noise influences. It is expected that the research results will provide new technical support for real-time precise navigation of HV, and have important theoretical significance and application value for HV.