基于加速度和速度建模的行人导航位置估计算法研究

Research on Pedestrian Navigation Position Estimation Algorithm Based on Acceleration and Velocity Modeling

位置估计是进行导航解算的最后一步,在基于足绑式惯性测量单元的行人导航中,由于足部的运动加速度变化剧烈,在采样间隔内,直接将运动加速度和速度乘以采样间隔来计算速度变化量和位置变化量的近似误差相对较大,会降低解算精度。为解决这一问题,对运动加速度、速度和位置进行了建模,共得到19种模型,对这些模型在位置估计中的效果进行了比较研究,在行人导航实验中采用矩形路径和8字形路径进行了验证。实验结果表明:对加速度进行一阶至三阶的拟合后,起终点误差和路径航向误差分别平均减小了86.02%和52.39%;进一步对速度进行一阶至三阶的拟合后,精度还可以继续提升。其中误差最小的模型相比于不拟合时的起终点误差和路径航向误差分别平均减小了87.51%和57.74%。

Location estimation is the final step of the navigation. In the pedestrian navigation based on foot binding type inertial measurement unit, due to the acceleration of the foot changes dramatically in the sampling interval, the approximation error of calculating the velocity change and position change by directly multiplying the motion acceleration and velocity by the sampling interval respectively is relatively large, which will reduce the estimation accuracy. In order to solve this problem, this paper carries out modeling with different orders for motion acceleration, velocity and position, and obtains 19 models in total. The effects of these models in position estimation are compared. In the pedestrian navigation experiment, the rectangular path and the figure-8 path are used for verification. Experimental results show that after the first, second or third order fitting of the acceleration, the starting point and ending point error and the course error of the path have been reduced by an average of 86.02% and 52.39%, respectively. After the first, second or third order fitting of the velocity on this basis, the accuracy can be further improved, and the corresponding errors of the model with the minimum error reduces by 87.51% and 57.74% on average respectively.