摘要
针对基于移动终端室内定位的实际应用需求,分析了基于机器学习定位技术运算时间长、单纯的惯性导航累积误差大等问题,提出一种机器学习与惯性导航相结合的联合定位技术。该技术首先利用移动终端收集定位区域内多源数据,通过机器学习算法处理这些数据得到训练模型。然后基于该训练模型和移动终端实时采集的数据周期性进行定位,把定位结果作为惯性导航校正数据,利用惯性导航算法实时计算移动终端位置。实验结果表明,所提技术定位时间快,定位精度高,能够满足室内实时定位追踪及低成本定位系统需求。
Aiming at the practical application requirement of indoor positioning based on mobile terminal,this paper analyses the problems such as the long operation time of machine learning,accumulated error of inertial navigation system etc,an indoor positioning technology based on machine learning and inertial navigation is proposed.A training model is generated by machine learning algorithm based on multi-source data in location area.The model and the current collected data based the machine learning algorithm is performed periodically to present correction data for Inertial navigation.Simulation results show that the proposed scheme has the advantage of real time positioning with high accuracy,and can meet the requirements of indoor real-time location tracking and low cost positioning system.
出处
《电子测量技术》
2016年第8期138-143,共6页
Electronic Measurement Technology
基金
航空电子系统综合技术重点实验和航空科学基金联合资助项目(2013ZC15003)
北京工业大学-青海民族大学合作基础研究基金资助项目
关键词
惯性导航
机器学习
传感器
室内定位
inertial navigation
machine learning
Sensor
Indoor positioning
