室内空间用户行为的智能手机识别

发布时间：2018-03-16 12:16

本文选题：行为识别　切入点：智能手机　出处：《测绘科学》2017年06期 　论文类型：期刊论文

【摘要】：针对地磁匹配中经常会出现相似点,造成定位偏差较大的问题,该文提出利用智能手机识别用户室内行为方式的方法,为地磁匹配算法提供筛选条件。开发了智能手机传感器数据采集工具,获取用户在室内环境下的行为数据。原始数据首先利用一阶低通滤波和平滑滤波算法进行去噪处理,再经过数据分割和特征提取后,应用于行为识别过程。行为识别模型的建立主要使用两种方法,K最近邻算法和隐式马尔可夫模型,并研究了两种方法的不足以及改进途径。通过针对识别准确度的对比实验,在输入最合适的数据的条件下,隐式马尔可夫模型的准确度略优于K最近邻算法。两种方法的识别准确率均在95%以上,能够有效地提高地磁定位精度。利用室内用户行为数据辅助地磁室内定位,很好地改善了地磁数据单一、定位精度较低的问题。
[Abstract]:In view of the problem that similar points often occur in geomagnetic matching, this paper proposes a method to identify the indoor behavior of users by using smart phones. A smart phone sensor data acquisition tool is developed to obtain user behavior data in indoor environment. The original data is processed by first order low-pass filtering and smoothing filtering algorithm. After data segmentation and feature extraction, it is applied to the process of behavior recognition. Two methods are used to establish the behavior recognition model: K-nearest neighbor algorithm and implicit Markov model. The deficiency of the two methods and the improvement methods are studied. Through the contrast experiment of the recognition accuracy, under the condition of the input of the most suitable data, The accuracy of implicit Markov model is slightly better than that of K-nearest neighbor algorithm. The recognition accuracy of both methods is more than 95%, which can effectively improve the precision of geomagnetic location. The problem of single geomagnetic data and low positioning accuracy is well improved.
【作者单位】：北京建筑大学测绘与城市空间信息学院;现代城市测绘国家测绘地理信息局重点实验室;中铁第五勘察设计院集团有限公司;
【分类号】：P318;TN929.53;TP212

【相似文献】