基于TDOA和改进粒子滤波算法的智能交通系统运行车辆定位

发布时间：2018-03-07 23:18

本文选题：粒子滤波　切入点：车辆定位　出处：《计算机测量与控制》2014年04期 　论文类型：期刊论文

【摘要】：为了克服已有车辆定位系统依靠GPS或RFID等固定设备进行定位时硬件成本高、同时对运动车辆进行定位时往往精度不高的缺点,提出了一种基于TDOA和改进粒子滤波的运动车辆定位方法;首先,通过TDOA算法和最小二乘法获取多个车辆坐标值估算值并对具有较大误差的坐标进行删除;然后通过RSSI信号强度对剩余的坐标值进行加权获取估计坐标的初始值;为了进一步提高定位的精度,设计了改进的粒子滤波算法以每个时刻的坐标初始值为观测值估算车辆的真实坐标,在算法中通过设计重要性概率密度函数和重采样提高滤波的精确度;取100m×800m的道路区域进行仿真,结果表明:文中方法能较为精确地实现运动车辆的定位,且与其它方法比较,具有硬件成本低和定位误差小的优点,是运动车辆定位的一种可行方法。
[Abstract]:In order to overcome the disadvantage that the hardware cost is high when the existing vehicle positioning system relies on the fixed equipment such as GPS or RFID, and the positioning accuracy of the moving vehicle is often not high, A moving vehicle location method based on TDOA and improved particle filter is proposed. Firstly, the estimation values of multiple vehicle coordinates are obtained by TDOA algorithm and least square method, and the coordinates with large errors are deleted. Then the initial values of the estimated coordinates are obtained by weighting the remaining coordinate values through the RSSI signal strength; in order to further improve the positioning accuracy, The improved particle filter algorithm is designed to estimate the real coordinates of the vehicle with the initial value of the coordinates at each moment as the observation value. In the algorithm the importance probability density function and the resampling are designed to improve the accuracy of the filter. The simulation results of 100m 脳 800m road area show that the proposed method can accurately realize the positioning of moving vehicles, and compared with other methods, it has the advantages of low hardware cost and low positioning error. It is a feasible method for moving vehicle positioning.
【作者单位】：琼州学院电子信息工程学院;琼州学院海南省嵌入式系统重点实验室;
【基金】：海南省自然科学基金(612127)
【分类号】：TP18;U495

【参考文献】