Abstract: In order to improve the shortcomings of indoor location fingerprint positioning technology that offline heavy workload, limited accuracy, poor robustness, a location fingerprint positioning technology method is proposed based on bat algorithm. The method uses the middle point interpolation method and channel attenuation model instead of offline training stage to automatically generate the fingerprint database to realize the function of timing automatic database updates, then combines K nearest neighbor algorithm with bat algorithm using in the stage of matching algorithm, finally realizes the positioning function. Compared with the traditional method of position location technology it reduces the overall positioning of workload and rapidly responds to the effect of the changing environment, at the end, localization performance test is carried out under the given simulation environment. The result shows that the method improves the average precision than other algorithms about 23.14%, the vast most of the blind node position errors range within 1.5 meters, which shows the advantages of positioning accuracy, robustness and adaptation to the changing environment.

Key words: location fingerprint positioning, middle point interpolation, channel attenuation model, bat algorithm

摘要： 为了改进室内位置指纹定位技术存在离线阶段工作量大、定位精度有限、鲁棒性较差的缺点，提出了一种基于蝙蝠算法的位置指纹定位技术的方法。该方法利用中点插值法和信道衰减模型自动生成指纹数据库从而代替了离线训练阶段，实现了定时自动更新数据库的功能，并且将[K]近邻算法和蝙蝠算法结合应用于匹配算法的阶段，最终实现定位功能。该方法与传统位置定位技术相比降低了整体定位的工作量，能快速应对环境变化对定位结果的影响，最后在给定的仿真环境下进行了定位性能的测试，结果显示该方法较其他算法相比平均定位精度大约提高了23.14%，绝大部分的盲节点的定位误差范围在1.5 m以内，显示出在定位精度、鲁棒性和适应环境变化方面的优势。

关键词: 位置指纹定位技术, 中点插值法, 信道衰减模型, 蝙蝠算法