基于模糊自适应卡尔曼滤波的SLAM算法

SLAM algorithm based on fuzzy adaptive Kalman filter

针对同步定位与地图创建(SLAM)问题中难以建立准确的先验噪声模型的问题,提出一种改进的模糊自适应卡尔曼滤波算法.该算法通过在线监测新息的变化,利用模糊逻辑对系统噪声和观测噪声的权重进行实时调整,进而改变系统对观测信息的信赖和利用程度,使滤波器最终趋于稳定.为了保证系统的实时性,提出一种直接将输入和输出进行模糊隶属函数匹配的方法代替模糊推理.将新的滤波算法用于SLAM仿真实验,结果表明该算法能根据噪声变化进行快速调整,滤波精度较高,相比标准EKF对定位和构图精度提升了50%以上.

A improved fuzzy adaptive Kalman filter was proposed to establish accurate priori noise model in the simultaneous localization and mapping （SLAM） problem. Fuzzy logic was used to adjust the importance weights of system noise and observation noise instantaneously through on-line monitoring of innovation, and the reliant and utilization degree of the observation were furthermore modified. To guarantee the real-time performance of system, a direct input-output fuzzy membership function matching approach was proposed to take the place of the fuzzy reasoning. In the simulation, the filter algorithm was applied in the SLAM problem, and the results show that the filter is capable of quickly adjusting according to the variation of noise immediately, and it improves the localization and mapping accuracy by more than 50% compared with the standard extended Kalman filter （EKF）.