论文部分内容阅读
在深入研究适用于低成本IMU/GPS组合导航系统非线性滤波的基础上,提出了一种基于交互式多模型自适应鲁棒容积卡尔曼滤波算法.该算法将交互式多模型算法引入H∞滤波容积卡尔曼滤波器,能够有效提升滤波算法的稳定性和质量.根据采集的低成本惯导GPS导航实验数据设计了两组滤波方案,性能分析结果表明,改进的交互式多模型鲁棒容积卡尔曼滤波算法位置估计精度提高了41.4%,速度估计精度提高了36.0%.同时设计的滤波试验结果表明该算法能有效抑制系统噪声取值不准确引起的滤波不稳定,尤其适用于噪声取值偏离最优值较多的情况,也即能够有效抑制野值对滤波结果的影响.
Based on the deep research of non-linear filtering applied to low-cost IMU / GPS integrated navigation system, an adaptive multi-model adaptive robust volume Kalman filter algorithm is proposed, which introduces interactive multi-model algorithm into H∞ The filter volumetric Kalman filter can effectively improve the stability and quality of the filtering algorithm.According to the acquisition of low-cost GPS navigation experimental data, two sets of filtering schemes are designed, the performance analysis results show that the improved interactive multi-model robust volume The accuracy of Kalman filter algorithm is improved by 41.4% and the speed estimation accuracy is improved by 36.0%. The experimental results show that this algorithm can effectively restrain the instability of filter caused by inaccurate system noise, especially for noise Deviation from the optimal value of more cases, that is, can effectively suppress the outliers on the filtering results.