【作者】 王龙；

【导师】 刘晓文；

【作者基本信息】 中国矿业大学 ， 电子科学与技术， 2015， 硕士

【摘要】 随着国家对矿井安全的重视,煤矿井下人员定位技术也得到了快速发展。研究井下人员定位技术,是为了准确的定位井下人员位置。尤其是发生矿难时,精确的定位信息是快速救援的关键。现阶段井下的人员定位系统多为基于基础设施的定位技术,其需要事先安装系统设施,使用成本较高,而且使用范围有限。本文研究开发了一种基于航迹推算的井下人员定位算法,不需借助外界信息,可实现自主定位,使用方便,成本低廉。该算法主要深入研究了计步算法和航向角解算算法。基于陀螺仪的计步算法容易受到行人身体晃动影响,本文设计了基于加速度计的自适应计步算法,可实现参数的自动设定,精确度高。行人航向角解算引入了无人机姿态解算算法—梯度下降算法和基于Mahony的多轴互补滤波算法。前者利用梯度下降的思想,求得误差函数最小值时对应的旋转四元数,并与陀螺仪计算的旋转四元数相融合,解算出航向角;后者利用加速度计、磁力计的长期有效性和陀螺仪短时间内的精确性,相互补偿来更新姿态角度,解算出行人行走方向。同时本文在多轴互补滤波基础上加以改进,使得算法计算量更小,系统稳定性更高。最后,本文在Matlab环境下对梯度下降算法和改进的多轴互补滤波算法进行仿真分析,同时从算法的估算精度和运算时间两个方面来进行对比,得出改进的互补滤波算法更适合井下个人航向角度解算。更多还原

【Abstract】 With national attention on mine safety, personnel positioning technology of coal mine has also been a rapid development. Underground personnel positioning technology can provide accurate location information of workers and plays a crucial role in determining a plan for helping workers trapped underground quickly and efficiency. Now most of underground personnel positioning system are based on location technology infrastructure, which requires installation of the system facilities in advance, with high cost and limited range.In this paper, personal navigation algorithm based on dead reckoning has been developed, which can enable autonomous positioning without the help of outside information, and is easy to use and cost low. The algorithm is mainly in-depth study of the pedometer algorithm and heading angle solver algorithm. Because pedestrians body sway can affect the gyroscope,so this paper chose an adaptive pedometer algorithm based on accelerometer, whose parameters can be automatically set up. UAV attitude solver algorithm-Gradient descent algorithm and multi-axis complementary filtering algorithm based on Mahony were taken into the pedestrians heading solver. Gradient descent algorithm combined the rotation quaternion obtained by the idea of gradient descent when the error function got the minimum number and the rotation quaternion calculated by gyroscope to get the heading angle. Multi-axis complementary filtering algorithm based on Mahony combined mutual compensation long-term effectiveness of accelerometer and magnetometer and accuracy of gyroscope in a short time to achieve attitude update and get the direction of pedestrians. At the sometimes, this paper made the complementary filter better for a smaller amount of calculation and higher stability.In the end, this paper did a simulation for gradient descent algorithm and improved multi-axis complementary filtering algorithm in Mat lab and gave a comparison from the estimation accuracy of the algorithm and computing time according to the simulation result. Finally the results indicated that improved multi-axis complementary filtering algorithm can do a better result in the underground heading angle solver.更多还原