【作者】 田婧楠； 丛丽； 秦红磊；

【Author】 Tian Jingnan;Cong Li;Qin Honglei;Beihang University;

【机构】 北京航空航天大学；

【摘要】 在室内定位系统中,行人航位推算(Pedestrian Dead Reckoning,PDR)系统因成本低、自主性强、对环境依赖小的优势被广泛应用,而步长估计是其关键组成部分之一。现有的步长估计方法通常基于加速度或角速度实现,但这些变量易受器件测量噪声以及运动习惯、设备携带方式等的影响,精度与实用性受到限制。针对上述问题,本文将受不同人、携带方式影响较小的无线电信号与加速度进行组合以估计步长。本文首先探究适用于发射塔位于室内/室外环境的多种无线电信号的室内传播模型,在模型基础上进一步建立接收信号指数(Receive Signal Strength Indicator,RSSI)与步长的函数关系,为后续多信号的组合奠定理论基础。最后,应用支持向量机回归(Support vector regression,SVR)组合提取的多信号RSSI特征与加速度特征以实现步长估计。多实验人员及地点的实验结果表明,引入RSSI特征后步长的绝对平均误差减小了18%。所提方法能够有效提升算法对于不同人及携带方式的适用性,实现步长的精确估计。更多还原

【Abstract】 In indoor positioning systems,the Pedestrian Dead Reckoning(PDR) system is widely used due to its advantages of low cost,strong autonomy,and low environmental dependence,among which step length estimation(SLE) is one of its key components.Existing SLE methods are usually implemented based on acceleration or angular velocity.However,these variables are easily affected by device measurement noise,movement habits,and device carrying poses,which limits their accuracy and practicality.To address the above issues,this paper combines radio signals with acceleration to estimate step length.First,this paper explores the indoor propagation model applicable to radio signals transmitted from both outdoor and indoor signal stations.On the basis of the model,a functional relationship between the Received Signal Strength Indicator(RSSI) and step length is further established,laying a theoretical foundation for the subsequent combination of multiple signals.Finally,Support Vector Regression(SVR) is applied to combine the extracted RSSI features of multiple signals and acceleration features to achieve SLE.Experimental results involving multiple experimenters and sites show that the absolute average error of step length decreases by 18% after introducing RSSI features.The proposed method can effectively improve the adaptability to different persons and carrying poses,thus realizing accurate SLE.更多还原