基于观测序列反馈的多源导航信息融合方法研究

In view of existing integrated navigation method based on Kalman filter and extended Kalman filter algorithms, the number and parameters of navigation sensors are relatively fixed, which cannot meet the requirements of unmanned vehicle under the high threat and strong confrontation conditions when they ask autonomous navigation high precision, high reliability and high anti-interference. Base on the convergence and real-time requirements, this research firstly will continue to analyze the relationship between system parameters and observation sequence, then propose information feedback correction methods based on observation sequence. In order to solve existing multi-source information fusion problem with parameters uncertainty of unmanned systems navigation, we will construct estimating algorithms under Kalman framework based on the direct feedback method of observation information. Based on the previous results, the convergence theory of proposed algorithm will be provided to support that the estimating sequence will be convergent to the real value in different applications. Lastly, we will verify the theory and method under different testing conditions based on the UAV platform simulation system and small flight test platform. In order to improve the precision, reliability and environmental adaptability of unmanned autonomous navigation method under complex environment and tasks, we try our best to improve the theoretical basis of autonomous navigation technology, to provide theoretical.support and method the necessary reference for the popularization and application of unmanned systems in the future.

以Kalman滤波和扩展Kalman滤波算法为基础的多源导航信息融合方法要求传感器参数、类型、数目相对固定，无法满足无人系统对高可靠性和强环境适应性多源导航信息融合方法的需求。本项目从无人系统对滤波融合算法的收敛性和实时性要求着手，挖掘利用观测序列中包含模型参数相关信息，研究提出基于观测序列的反馈方法，并设计实现相对应的滤波估计算法，以期解决存在参数或结构不确定性的多源导航信息融合问题；基于前期预研结果，分析证明算法中参数估计和状态估计序列的收敛性，从而为多源导航信息融合方法提供稳定性理论支撑；基于仿真和小型无人机平台，构建不同的传感器参数、环境和传感器配置测试条件，对所提出的方法及理论结果开展测试验证。本项目的实施可望完善无人系统多源导航信息融合的稳定性相关理论基础，提高无人系统多源导航信息融合方法对复杂任务和环境的适应能力，为无人系统未来的推广应用提供必要的理论支撑和方法参考。

以Kalman滤波和扩展Kalman滤波算法为基础的多源导航信息融合方法要求传感器参数、类型、数目相对固定，无法满足无人系统对高可靠性和强环境适应性多源导航信息融合方法的需求。本项目从无人系统对滤波融合算法的收敛性和实时性要求着手，挖掘利用观测序列中包含的模型参数相关信息，研究提出基于观测序列的反馈方法，并设计实现相对应的滤波估计算法，进一步解决了存在参数或结构不确定性的多源导航信息融合问题；基于前期预研结果，分析证明算法中参数估计和状态估计序列的收敛性，从而为多源导航信息融合方法提供稳定性理论支撑；基于仿真和半实物无人平台，构建不同的传感器参数、环境和传感器配置测试条件，对所提出的方法及理论结果开展测试验证。本项目的实施进一步完善无人系统多源导航信息融合的稳定性相关理论基础，提高无人系统多源导航信息融合方法对复杂任务和环境的适应能力，可以为无人系统未来的推广应用提供必要的理论支撑和方法参考。