自适应滤波理论拓展与改进及其应用研究

Kalman and other filtering theories and algorithms have been studied intensively in the world; however, not very successful in practical GNSS application, until the born of the adaptive filtering theory in 2001. The theory was introduced to the navigational field by Springer original book "GPS" and was honoured by the international expert review as NEW and an innovation progress. The theory and algorithm was first applied in GFZ German Research Centre of Geosciences and Technical University Berlin with great success. Due to the special characteristic of the GNSS survey, this study intends to introduce the Doppler observing information for extension and upgrading the adaptive filter theory. In Kalman filtering, the system descriptions, which up to now without exception using a few empiric system equations, will be upgraded by using Doppler measurement co-determined and more realistic descriptions. Because of the additional velocity information (nearly as much as the positioning information), the much more objective description of the system, and the more reasonable and precise estimation of the error disturbances, the upgraded adaptive filter should be able to be performed more stably and the results could be more accurately. This study intends to base on the kinematic GNSS positioning examples to extend and verify the theory, and studying for application for autonomous orbit determination and manoeuvring.

国际上对kalman滤波及其他滤波理论方法的研究非常多，但在实际GNSS导航定位中的应用并不非常成功，直到自适应滤波理论2001年问世。该理论2003年由Springer原著《GPS》引入导航界，被国际书评誉为是一个非常重要的创新，并在GFZ波茨坦地球科学中心和柏林工大率先应用并获得很大成功。根据GNSS测量的特点，本研究试图通过引入和利用多普勒观测信息对自适应滤波理论方法进行拓展和改进。把自今为止卡尔曼滤波中毫无例外的有限个的先验的系统方程描述，改进为通过多普勒测速信息而动态地确定的更符合实际运动的描述。由于增加了几乎与定位信息等量的测速信息，增强了特别是动态系统描述的客观性，增强了误差干扰统计估值的合理性和准确性，这种获得拓展后的自适应滤波理论，应该能够更稳定精度更高。本研究希望这个研究能通过动态定位为实例予以实现，并开展定轨应用研究，为提高定位定轨精度等做出有意义的理论贡献。

国际上对Kalman滤波及其他滤波理论方法的研究非常多，但在实际GNSS导航定位的应用并不非常成功，直到自适应滤波理论2001年问世。该理论2003年由Springer原著《GPS》引入导航界，并在GFZ波兹坦地球科学中心和柏林工大率先应用并获得很大成功。根据GNSS测量的特点，本研究通过引入和利用多普勒观测信息对自适应滤波理论方法进行拓展和改进。把自今为止卡尔曼滤波中毫无例外的有限个先验的系统方程描述，改进为通过多普勒测速信息而动态的确定的更符合实际运动的描述。目前，完成了该方法的全部理论推导和证明，这种获得拓展后的自适应滤波理论，增强了特别是动态系统描述的客观性，增强了误差干扰统计估计值得合理性和准确性，能够更稳定精度更高。本研究尝试开展我国困难地区的重力测量的应用，利用新疆地区不同航次和航高的实测数据，引入最优自适应比例因子，改进Kalman滤波最优估计过程，最终提高重力扰动的解算精度。还将继续开展定轨应用研究，为提高定位定轨精度等作出有意义的理论贡献。