一种改进的位置指纹智能手机室内定位算法

为了减小智能手机现有室内定位算法的时间复杂度和空间复杂度,提出一种将确定型算法和概率分布算法融合的智能手机室内定位新方法。利用最近邻算法选出K个最相近的位置点,然后采用贝叶斯算法将K个位置点中匹配概率最大的点作为最终的估计位置。在Android手机上分别采用3种方法进行20组室内对比定位实验,并随机选取10个位置进行定位误差对比实验,结果表明,新方法比贝叶斯算法的复杂度降低了Ο(4n/5),比最近邻算法的定位准确率提高了约4%,且定位误差较小。

Improved fingerprinting algorithm for smart phone indoor positioning

In order to reduce the time complexity and space complexity of smartphone indoor posi- tioning algorithm, a new method of smartphone indoor positioning is proposed in this paper. This method combines the deterministic algorithm and the probability distribution algorithm together. By using the K-Nearest Neighbor algorithm to select K points which are nearest in position and then using the Bayesian algorithm to make the maximum matching probability point in K points, the final position is estimated. Three contrast methods are used on 20 groups in indoor positio- ning experiments respectively on Android mobile phone. Ten positions are then selected randomly for positioning error comparison experiment. The results show that the new method has the low- er algorithm complexity of 0（4n/5） than the Bayesian algorithm, and that the positioning accura- cy has increased 4% than the K-nearest neighbor algorithm. It also has a small position error.