首页> 外文会议>International Conference on Computer Vision, Image and Deep Learning >Research on and Application of GNSS Assisted MINS Multiscale Learning Method Based on Chaotic Neural Network
【24h】

Research on and Application of GNSS Assisted MINS Multiscale Learning Method Based on Chaotic Neural Network

机译:基于混沌神经网络的GNSS辅助MINS多尺度学习方法的研究与应用

获取原文

摘要

Applying neural network to MINS/GNSS integrated navigation system is an important method to improve its navigation performance. In this paper, GNSS-assisted MINS neural network learning adopts the strategy of multi-scale comparative learning in continuous time in motion state to obtain the MINS trajectory length error compensation in different time periods, thus reducing the accumulated error of MINS autonomous navigation. At the same time, in order to avoid local minimum when the neural network converges, momentum factor function with chaotic mechanism is introduced. The simulation experiment of the algorithm shows that the method has good effect in inhibiting the divergence speed of MINS navigation error accumulation and improving MINS autonomous navigation capability when GNSS is invalid, and the algorithm is simple and has strong engineering application value.
机译:将神经网络应用于MINS / GNSS集成导航系统是提高其导航性能的重要方法。本文辅助MINS神经网络学习采用多尺度比较学习的策略在运动状态下连续时间,在不同的时间段中获得分钟的轨迹长度误差补偿,从而减少了分钟自主导航的累积误差。同时,为了避免当局部最小值时,当神经网络收敛时,引入了具有混沌机制的动量因子函数。该算法的仿真实验表明,该方法对抑制MINS导航误差累积的发散速度良好的效果,当GNSS无效时,初步的分钟自主导航能力,并且该算法简单,具有强大的工程应用价值。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号