• 主题
高级搜索  >
历史搜索 清空历史
首页> 外文期刊>Measurement Science & Technology >An improved gravity compensation method for high-precision free-INS based on MEC-BP-AdaBoost
【24h】

An improved gravity compensation method for high-precision free-INS based on MEC-BP-AdaBoost

机译：基于MEC-BP-AdaBoost的改进的高精度free-INS重力补偿方法

获取原文
获取原文并翻译 | 示例
           
页面导航
  • 摘要
  • 著录项
  • 相似文献
  • 相关主题

摘要

In recent years, with the rapid improvement of inertial sensors (accelerometers and gyroscopes), gravity compensation has become more important for improving navigation accuracy in inertial navigation systems (INS), especially for high-precision INS. This paper proposes a mind evolutionary computation (MEC) back propagation (BP) AdaBoost algorithm neural-network-based gravity compensation method that estimates the gravity disturbance on the track based on measured gravity data. A MEC-BP-AdaBoost network-based gravity compensation algorithm used in the training process to establish the prediction model takes the carrier position (longitude and latitude) provided by INS as the input data and the gravity disturbance as the output data, and then compensates the obtained gravity disturbance into the INS's error equations to restrain the position error propagation. The MEC-BP-AdaBoost algorithm can not only effectively avoid BP neural networks being trapped in local extrema, but also perfectly solve the nonlinearity between the input and output data that cannot be solved by traditional interpolation methods, such as least-square collocation (LSC) interpolation. The accuracy and feasibility of the proposed interpolation method are verified through numerical tests. A comparison of several other compensation methods applied in field experiments, including LSC interpolation and traditional BP interpolation, highlights the superior performance of the proposed method. The field experiment results show that the maximum value of the position error can reduce by 28% with the proposed gravity compensation method.
机译：近年来，随着惯性传感器（加速度计和陀螺仪）的快速改进，重力补偿对于提高惯性导航系统（INS）的导航精度，尤其是高精度INS变得越来越重要。提出了一种基于神经网络的重力补偿方法，即心进化计算（MEC）反向传播（BP）AdaBoost算法，该方法根据测得的重力数据估算轨道上的重力扰动。在训练过程中使用的基于MEC-BP-AdaBoost网络的重力补偿算法来建立预测模型，以INS提供的载波位置（经度和纬度）作为输入数据，并将重力扰动作为输出数据，然后进行补偿将获得的重力扰动计入INS的误差方程，以抑制位置误差的传播。 MEC-BP-AdaBoost算法不仅可以有效避免BP神经网络陷入局部极值，而且还可以完美解决输入和输出数据之间的非线性问题，这些问题是传统内插方法无法解决的，例如最小二乘配置（LSC） ）插值。通过数值测试验证了所提方法的准确性和可行性。对现场实验中使用的其他几种补偿方法（包括LSC插值和传统BP插值）进行了比较，突出了该方法的优越性能。现场实验结果表明，提出的重力补偿方法可以将位置误差的最大值降低28％。

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