基于胸戴式手机传感器的行人航位推算算法研究

Research on PDR algorithm based on chest-mounted mobile phone sensor

针对基于胸戴式惯性传感器的行人航位推算(PDR)算法中步长模型估计精度低的问题,提出一种多约束的反馈式步长估计模型。首先,用约束型峰值检测方法进行单步划分;在此基础上,通过对身高、步频和单步内加速度计信息进行特征提取,形成不同的权重因子建立步长回归模型;再用前一步步长对当前步进行反馈修正;结合航向补偿算法进行航向修正;最终实现行人位置的解算。基于手机惯性传感器进行了行走实验,结果表明:本文算法可以有效提高步长估计精度,定位误差小于总行程2%。

Aiming at the problem of low estimation precision of step length model in pedestrian dead reckoning(PDR)algorithm based on chest-worn inertial sensor,a multi-constrained feedback step length estimation model is proposed.Firstly,a single-step division is carried out using constrained peak detection method.On this basis,different weighting factors are formed by feature extraction of the height,stride frequency and accelerometer information within a single step to establish step-length regression model.And then,the previous step is used to perform feedback correction at the length of the current step.Combined with the heading compensation algorithm for heading correction.Finally,the calculation of the pedestrian position is realized.A walking experiment is carried out based on the mobile phone inertial sensor.The results show that the algorithm in this paper can effectively improve the precision of the step length estimation,and the positioning error is less than 2%of the total stroke.