基于大数据预测的船用惯性/星敏感器抗干扰组合导航方法研究

Inertial navigation system (INS) has been widely used in aviation field and navigation field as the autonomous navigation system. However, the disadvantage of INS is the navigation error divergence with time. In order to make up this disadvantage, the coupled navigation method based on the inertial navigation system (INS) and the Star Sensor is widely used. Therefore, the INS accuracy can be greatly improved through the Star Sensor. However, when the Star Sensor is installed on the surfaced ship, the ability of star observation is easy to be affected by the rocking of the vehicle and the bad weather, and the accuracy of the Star Sensor is reduced. Hence, the navigation information from the Star Sensor cannot be used as the reference information of the coupled navigation system. To solve this problem, the coupled navigation method of INS/ Star Sensor coupled navigation system based on the large data prediction is proposed in the project request. The navigation information error of the Star Sensor, which is introduced by the various environmental factors, is corrected by the information from INS based on the large data prediction method. In addition, experiment will be designed based on the fiber optical gyro strapdown inertial navigation system and the star sensor in order to confirm the efficiency of the propose method and the technologies. The accuracy of the marine Star Sensor can be improved based on the INS/ Star Sensor coupled system, and the coupled navigation scheme can be better. The navigation accuracy of the surfaced ship INS/ Star Sensor coupled system is improved. In addition, the effectiveness of this approach proposed in this project is demonstrates by experimental study based on the fiber optical gyro INS, which is made by ourselves, and the Star Sensor.

惯性导航系统作为一种全自主导航系统被广泛应用于航海领域，然而导航累积误差发散的缺点又制约了该类系统的长航时可用性。为了弥补该缺点又不失导航自主性，可采用基于惯性/星敏感器的组合导航方式。即以星敏感器信息为基准，对惯导系统发散式误差进行周期性重校。然而，当星敏感器应用到水面舰艇时，应用载体变化和观星高度变化导致星敏感器测量精度下降，使其无法作为惯性误差校准的高精度基准信息。针对这一问题，本项目提出了基于大数据预测的船用惯性/星敏感器抗干扰组合导航方法：从惯性/星敏感器组合导航的角度出发，利用惯性和星敏感器的导航信息，通过引入大数据预测等手段，对环境因素影响下的星敏感器测量误差进行预测与修正，进一步完善惯性/星敏感器组合导航方案，进而提高船用惯性/星敏感器全自主导航系统的导航精度。此外，本项目拟采用自研光纤惯性导航系统与星敏感器搭建组合导航系统，对本项目的研究内容及各项关键技术进行验证。

当星敏感器应用到水面舰艇时，应用载体变化和观星高度变化导致星敏感器测量精度下降，使其无法作为惯性误差校准的高精度基准信息。针对这一问题，本项目提出了基于大数据预测的船用惯性/星敏感器抗干扰组合导航方法。由于水面船舶会由于海面风浪等影响产生摇摆运动，导致星敏感器在星图获取过程中存在拖尾现象，降低星敏感器导航精度，使其无法作为信息融合基准信息。针对这一问题，提出了一种基于自适应差分进化BP神经网络载体姿态预测的惯性/星敏感器信息融合方法，即通过引入ADE-BPNN时间序列预测技术，以消除由于拖尾现象导致的星敏感器导航误差。此外，星敏感器经常会因为受到云层等天气的影响下导致信息融合系统失效，不再能提供精确的载体目标跟踪信息。当星敏感器短期工作失效时，提出了基于粒子群优化算法BP神经网络信息融合方式；当星敏感器长期工作失效状态时，提出了基于自适应联邦滤波算法的惯性传感器/星敏感器/计程仪信息融合算法，以完成星敏感器失效状态下的信息融合过程。最后，本项目采用光纤惯性导航系统与星敏感器搭建组合导航系统，对本课题的研究内容及各项关键技术进行验证。