激光导航AGV的马尔可夫定位方法研究

Research on Markov Localization Algorithm for AGV Based on LADAR

针对马尔可夫定位算法由于运算量大而限制了其在激光导航自动导引车中应用的问题,提出了一种降维及栅格信度整体平移计算的简化预测模型计算方法。采用电子罗盘的绝对方向信息直接得到自动导引车的姿态信息,在估计导引车位姿时只考虑求解其平面位置信息代替三维位姿信息;采用傅里叶变换及其逆变换将离散化的栅格信度视为一个整体通过空间域-频域的转换进行平移计算代替对每一个栅格相邻时刻的信度更新。建立了新模型及算法,设计了复合环境下自动导引车的漫游运动仿真,验证了该方法对位姿估计和全局定位的有效性及准确性。

Because the application of the Markov localization algorithm is limited in LADAR navigation AGV,a simplified method for calculating the prediction model by dimensionality reduction and the whole translation of the grid reliability is proposed.By using the absolute direction information of the electronic compass,the orientation information of the AGV is obtained directly.When the AGV position is estimated,only the plane localization information is considered instead of the three-dimensional pose information.The fourier transform and its inverse transformation are used to regard the discrete grid reliability as a whole to be calculated.The reliability update of adjoining times for each grid is instead of whole translation calculation of grid by the spatial domain-frequency domain conversion.The new model and algorithm are established and the simulation of AGV roaming motion is done in a similar composite environment,which are used to verify the validity and accuracy of the method for AGV pose estimation and global positioning.