便携式移动电子设备的步行者航位推算技术研究

发布时间：2018-05-07 17:50

本文选题：MEMS惯性器件 + 步行者导航　；参考：《电子科技大学》2014年硕士论文

【摘要】：便携式移动电子设备的快速发展,给人们的生活带来了很大的改变,其中智能手机和平板电脑已经成为日常生活中必不可少的工具。为了有更加新颖的应用和更绚丽的视觉效果,智能手机和平板电脑中集成了多种微传感器,如三轴MEMS加速度计、三轴MEMS陀螺仪和MEMS磁力计等。本文依据UESTC-NOKIA国际合作项目,旨在研究一种基于MEMS惯性传感器的适用于移动电子设备的新型而有趣的步行者导航方式。因为成本控制,便携式移动电子设备中使用的MEMS惯性传感器有精度偏低、随机误差较大等缺点。针对这些缺点,本文对使用的LIS344ALH加速度传感器和Ex3500A4962A陀螺仪建立了误差模型;使用“六位置法”标定了加速度传感器和陀螺仪的静态误差;针对随机误差设计了低通滤波器和状态扩增卡尔曼滤波器,有效地抑制了传感器信号的漂移,提高了航位推算的精度。惯性测量单元被捆绑于步行者的腰部,其中三轴MEMS加速度计用于测量的步行者的加速度,MEMS陀螺仪用于测量步行者转向时的角速率。通过对步行者一次迈步周期内的动力学模型的分析,提取特征量判别步行者的迈步步态信息:迈步起至点、站立阶段、迈步阶段,进而估计步行者的步频、步数和步长;利用加速度计信号通过姿态算法可间接推算步行者的姿态角。通过角速率可直接计算步行者的姿态矩阵和姿态角。采用互补滤波算法融合两种算法得到的姿态角,提高了姿态角的计算精度,有效地降低了计算姿态角所产生的累积误差对运动轨迹的影响,使之更适用于低精度惯性传感器。最后通过航位推算原理,合成步行者的相对运动轨迹。最后本文对整个系统在Linux系统中进行了算法编程。为验证步行者航位推算系统的精度和对个体差异的适应性,在室内环境中以普通人的正常行走速度进行了现场试验。根据试验结果对步长估计算法和姿态更新算法的精度做了对比分析和评估;并将试验结果与实际参考数据轨迹作对比,分析了步行者航位推算算法的准确性和可行性,且该系统可以满足消费电子领域对实时性的需求。并根据实验所得的结果和数据分析,本文还提出了一些改进意见。
[Abstract]:The rapid development of portable mobile electronic devices has brought great changes to people's lives, in which smart phones and tablets have become an indispensable tool in daily life. For more novel applications and more beautiful visual effects, a variety of microsensors, such as three-axis MEMS accelerometers, three-axis MEMS gyroscopes and MEMS magnetometers, have been integrated into smartphones and tablets. This paper aims to study a new and interesting walker navigation method based on MEMS inertial sensors for mobile electronic devices according to the UESTC-NOKIA international cooperation project. Because of cost control, the MEMS inertial sensor used in portable mobile electronic equipment has some disadvantages, such as low precision and large random error. Aiming at these shortcomings, the error model of LIS344ALH accelerometer and Ex3500A4962A gyroscope is established, and the static error of accelerometer and gyroscope is calibrated by "six-position method". A low-pass filter and a state-amplified Kalman filter are designed for random error, which can effectively suppress the drift of sensor signal and improve the accuracy of dead-reckoning. The inertial measurement unit is tied to the walker's waist and the three-axis MEMS accelerometer is used to measure the walker's acceleration and the angular rate of the walker's turn. Based on the analysis of the dynamic model of the walker in one step cycle, the characteristic quantity is extracted to judge the gait information of the walker: the starting point, standing stage, step stage, and then estimating the walker's frequency, number and length; The attitude angle of the walker can be calculated indirectly by using the accelerometer signal through the attitude algorithm. The attitude matrix and attitude angle of the walker can be calculated directly by angular rate. The attitude angle obtained by the two algorithms is fused by the complementary filtering algorithm, which improves the accuracy of the attitude angle calculation, effectively reduces the influence of the accumulated error caused by the attitude angle calculation on the motion trajectory, and makes it more suitable for the low-precision inertial sensor. Finally, the relative trajectory of the walker is synthesized by the principle of dead-reckoning. Finally, the algorithm programming of the whole system is carried out in Linux system. In order to verify the accuracy and adaptability to individual differences of the pedestrian level reckoning system, a field test was carried out in the indoor environment with the normal walking speed of ordinary people. According to the experimental results, the accuracy of step size estimation algorithm and attitude updating algorithm are compared and evaluated, and the accuracy and feasibility of the algorithm are analyzed by comparing the experimental results with the actual reference data trajectory. The system can meet the real-time demand in consumer electronics field. According to the experimental results and data analysis, this paper also puts forward some suggestions for improvement.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TN96;TP212.9

【参考文献】