论文部分内容阅读
针对地图融合时多机器人位姿估计过程中互协方差未知的问题,提出了一种基于协方差交叉点(covariance intersection,CI)因子图的2D地图融合算法.首先通过坐标变换矩阵实现机器人坐标到全局坐标的转换,然后以最小化非线性性能指标为原则求取局部的估计信息权重,通过算法融合各局部估计信息,计算出融合点的位姿和互协方差.最后通过协方差通用公式,计算出融合点到下一级变量节点的概率约束(协方差),进而完成因子图融合.实验结果表明,该算法具有一定的可行性.
Aiming at the problem of unknown mutual covariance during multi-robot pose estimation in map fusion, a 2D map fusion algorithm based on covariance intersection (CI) factor map is proposed.Firstly, the coordinate transformation matrix is used to realize the robot coordinates Global coordinate transformation and then calculate the pose and mutual covariance of fusion point by taking the local estimated information weight as the principle to minimize the nonlinear performance index.At last, through the covariance general formula, The probability constraint (covariance) of the nodes from the fusion point to the next level is calculated, and then the factor graph fusion is completed.The experimental results show that the algorithm is feasible.