基于极限学习机的黑障区智能导航算法

在黑障区飞行阶段中，惯性导航系统会因缺少辅助导航系统而持续累积误差，导致飞行器导航系统可靠性下降。针对这一问题，提出了一种新的基于极限学习机的黑障区智能导航算法，通过极限学习机(ELM)对GPS正常工作的导航信息进行学习。在黑障区，利用学习得到的模型对惯性导航系统进行误差补偿，较好地修正了当GPS失锁时惯性导航系统的误差，避免了因误差累积而导致的导航信息发散。仿真结果表明，该算法能够保证在GPS失锁的黑障区中导航系统输出的信息有较好的可靠性和精度，能够为接下来的姿态调整和着陆准备提供良好的基础。

An Intelligent Navigation Algorithm of Blackout Area Based on Extreme Learning Machine

In the flight phase of the blackout area, the inertial navigation system continues to accumulate errors due to the lack of auxiliary navigation system, which leads to the decrease of the reliability of the aircraft navigation system. To solve this problem, a new intelligent navigation algorithm of blackout area based on extreme learning machine(ELM) which can learn the normal navigation information of GPS, is proposed in this paper. In the blackout area, the error compensation of the inertial navigation system is carried out by using the learning model, which can better correct the error of the inertial navigation system when the GPS loses lock, and avoid the navigation information divergence caused by the error accumulation. The simulation results show that this algorithm can ensure the reliability and accuracy of the information output by the navigation system in the blackout area when GPS loses lock, it can provide a good foundation for subsequent attitude adjustment and landing preparation.