基于鲁棒闭合路径校准的采煤机SINS/OD组合导航系统研究
Shearer SINS/OD integrated navigation system based on robust closed path calibration
-
摘要: 采煤机精确定位是完成综采工作面直线度检测，提升自动化水平的关键技术。基于非完整性约束和闭合路径校准的捷联惯导与里程计的组合导航系统是被广泛应用的采煤机定位方案，其中闭合路径校准利用前次割煤的位置信息结合支架的推移距离获得预测当前位置，以抑制采煤机多次割煤后的误差发散，但是液压支架推移过程中存在较大的执行误差，导致的错误预测位置严重干扰了对于直线度的检测。为解决该问题，本文提出鲁棒闭合路径校准法，采用最大相关熵卡尔曼滤波器取代经典卡尔曼滤波器以拒绝预测位置中的异常值从而提高导航系统的鲁棒性；再此基础上将自适应核带宽算法引入到最大相关熵卡尔曼滤波器中，以提高组合导航系统对于复杂环境的适应性。通过实验表明最大相关熵卡尔曼滤波器可以有效避免预测位置的异常值对直线度检测的干扰，并降低来自首次割煤的累积误差，具有自适应核带宽的最大相关熵卡尔曼滤波器在不预设参数的情况下即可获得良好的性能。
-
关键词:
- 采煤机定位 /
- SINS /
- 液压支架 /
- 闭合路径校准 /
- 最大相关熵卡尔曼滤波器
Abstract: Accurate positioning of shearer is a key technology for completing the straightness detection and improving the level of automation of the fully mechanized mining face. The integrated navigation system of strapdown inertial navigation and odometer based on nonholonomic constraints and closed path calibration is a widely used positioning scheme for shearer. The closed path calibration uses the position information of the previous cutting cycle combined with the displacement distance of support to predict the current position, in order to suppress the error divergence of shearer after multiple cutting cycle. However, there is a large execution error in the hydraulic support displacement process, which seriously interferes with the detection of straightness due to incorrect predicted positions. To solve this problem, a robust closed path calibration method is proposed, which uses the maximum correntropy criterion Kalman filter (MCCKF) instead of the classical KF to reject outliers in the predicted position and improve the robustness; On this basis, the adaptive kernel bandwidth algorithm is introduced into MCCKF to improve the adaptability of the integrated navigation system to complex environments. Experiments have shown that MCCKF can effectively avoid the interference of outliers in predicted positions on straightness detection, and reduce the cumulative error from the first cutting cycle. The MCCKF with adaptive kernel bandwidth can achieve excellent performance without preset parameters. -
计量
- 文章访问数: 42
- HTML全文浏览量: 1
- PDF下载量: 0
下载:
