复杂噪声环境下基于时变渐消广义相关熵的组合导航方法研究

The real-time and accurate acquisition of navigation parameter information of conventional ammunition is the core and key technology of guidance and precision strike. MEMS-INS/GNSS integrated navigation system is considered as a preferred alternative due to its advantages of small size, low cost and high accuracy. Due to the decline of the GNSS measurement accuracy under the complex electromagnetic environment and the influence of the vibration from boost engine in the ballistic extended range, the accuracy and reliability of the existing MEMS-INS/GNSS integrated navigation system are difficult to guarantee. Therefore, this project is aimed at the requirement of high-precision and reliable measurement of navigation parameters for the MEMS-INS/GNSS integrated navigation system under complex noise environment. To solve the bottleneck scientific problems such as the process of uncertain suppression of the integrated navigation system in the trajectory extended range section and the non-Gaussian noise suppression method of integrated navigation in complex electromagnetic environments, this project will focuses on the breakthroughs in key technologies such as intelligent identification of process model uncertain of integrated navigation system, multiple suboptimal fading factor determination, construction of time-varying decreasing generalized maximum correntropy criterion, and adaptive update of Gaussian kernel bandwidth. It is expected to build an adaptive robust integrated navigation theory that is suitable for the flight environment of guided bomb, which provides a strong theoretical and technical support for the accuracy and performance enhancement of integrated navigation system.

常规弹药导航参数信息的实时准确获取是其制导化和精确打击的核心关键技术，MEMS-INS/GNSS组合导航系统以体积小、成本低及精度高等优势，被认为是其首选方案。受复杂高动态弹载环境下GNSS量测信息质量评价准确性下降，微惯性器件受弹道增程段助推发动机振动影响等因素制约，现有弹载MEMS-INS/GNSS组合导航系统的导航精度和可靠性难以保证。因此，本项目面向复杂噪声环境下制导弹用MEMS-INS/GNSS组合导航系统对导航参数的高精度可靠探测需求，针对弹道增程段组合导航系统过程不确定抑制方法、复杂电磁环境下组合导航非高斯噪声抑制方法等瓶颈科学问题，重点突破组合导航系统过程模型不确定智能辨识、多重次优渐消因子决策、时变渐消广义最大相关熵准则构建以及高斯核带宽自适应更新等关键技术，有望构建一套适用于制导炮弹飞行环境的自适应鲁棒组合导航理论体系，以期为弹载组合导航系统的精度性能增强提供强有力技术支撑。

常规弹药导航参数信息的实时准确获取是其制导化和精确打击的核心关键技术，MEMS-INS/GNSS组合导航系统以体积小、成本低及精度高等优势，被认为是其首选方案。受复杂高动态弹载环境下GNSS量测信息质量评价准确性下降，微惯性器件受弹道增程段助推发动机振动影响等因素制约，现有弹载MEMS-INS/GNSS组合导航系统的导航精度和可靠性难以保证。因此，本项目面向复杂噪声环境下制导弹用MEMS-INS/GNSS组合导航系统对导航参数的高精度可靠探测需求，在对制导弹药典型运动特性以及多源复合干扰对组合导航系统影响机理分析的基础上，开展了弹道增程段基于增强型强跟踪容积卡尔曼滤波理论的过程不确定抑制方法研究，提升了弹载组合导航系统在制导弹药增程振动环境下的自适应跟踪能力；开展了复杂电磁环境下基于时变渐消广义最大相关熵准则的非高斯噪声鲁棒性滤波技术研究，确保了弹载组合导航系统在GNSS 测量异常时导航参数测量的可靠性；开展了复杂噪声环境下弹载组合导航方法试验验证研究，在地面半实物仿真条件下验证了 MEMS-INS/GNSS 组合导航系统在弹载复杂噪声环境下导航参数测量的强自适应性、高可靠性以及高精度。研究表明：所提基于假设检验理论的过程模型不确定辨识方法，通过对组合系统残差马氏距离的平方进行实时检测，可有效辨识制导弹药当前飞行状态；在预测协方差阵中引入多重次优渐消因子，可实现组合滤波器对不同通道导航参数状态的自适应跟踪；所提基于时变渐消广义最大相关熵准则的非高斯鲁棒性组合滤波算法，可有效地同步处理系统过程不确定和非高斯量测噪声，导航定位精度相比传统非高斯鲁棒性滤波提高了30%。本项目针对现有鲁棒性滤波方法无法同步处理组合滤波器过程不确定以及非高斯噪声等多源干扰的缺陷，创新性地提出一种新的自适应鲁棒机制。该机制对非高斯噪声更为敏感，在未知组合滤波器噪声分布的情况下，能够有效地同步处理系统过程不确定和非高斯量测噪声，对现有非高斯鲁棒性滤波理论的创新发展具有重要的科学意义。