Currently most of the existing indoor fingerprint positioning algorithms are based on fingerprint database. The accuracy of the fingerprint database will directly affect the final positioning accuracy. Therefore, through the research of fingerprint data, a method based on skewness-kurtosis normality test and Kalman filter fusion is proposed. In the training phase, the RSSI (Received Signal Strength Indication) samples received on each fingerprint point are tested based on the skewness-kurtosis normality. If the normal distribution model is met, the normal distribution function is used to estimate the probability density of the samples. If not the kernel function will be used. And then the value of the large probability density is taken for Kalman filtering, and finally, the averaged value after filtering is used to establish a high-precision fingerprint database. In the online positioning stage, the weighted KNN (K-Nearest Neighbor) is used to estimate the position, and finally, the positioning point is corrected by the fusion of the Levenberg-Marquardt method and the Kalman filter. The optimization of the three stages can improve the positioning accuracy. The simulation results show that the indoor positioning method proposed in this paper has the least number of iterations and the positioning accuracy is improved by 60% compared with the traditional Kalman filtering method.

当前，大多数现有的室内指纹定位算法都基于指纹数据库。指纹数据库的准确性将直接影响最终的定位准确性。因此，通过指纹数据的研究，提出了一种基于偏度-峰度正态性检验和卡尔曼滤波融合的方法。在训练阶段，根据偏度-峰度正态性测试在每个指纹点接收的RSSI（接收信号强度指示）样本。如果满足正态分布模型，则使用正态分布函数来估计样本的概率密度。否则，将使用内核功能。然后将大概率密度的值用于卡尔曼滤波，最后，滤波后的平均值用于建立高精度指纹数据库。在在线定位阶段，使用加权的KNN（K最近邻）估计位置，最后，通过Levenberg-Marquardt方法和Kalman滤波器的融合来校正定位点。三个阶段的优化可以提高定位精度。仿真结果表明，与传统的卡尔曼滤波方法相比，本文提出的室内定位方法具有最小的迭代次数，定位精度提高了60％。这三个阶段的优化可以提高定位精度。仿真结果表明，与传统的卡尔曼滤波方法相比，本文提出的室内定位方法具有最小的迭代次数，定位精度提高了60％。三个阶段的优化可以提高定位精度。仿真结果表明，与传统的卡尔曼滤波方法相比，本文提出的室内定位方法具有最小的迭代次数，定位精度提高了60％。