cartographer 分析 | 易学教程

原文链接：http://blog.csdn.net/zyh821351004/article/details/52421005

cartographer与karto的比较

1. 两者采取的都是图优化框架。采取的优化库不一致， karto采取的是spa(karto_slam)或g2o(nav2d), cartographer采取的是google的ceres构建problem优化。 karto的前端与后端采取的是单线程进行，cartographer按paper说明，采取的是4线程后端优化，还在进一步确定。

2. 运动预测部分：tracker

karto利用的是odom进行初始位置的预测， cartographer部分利用imu构建预测模型，scanmatcher与odom(可选)构建观测模型，采取UKF进行运动预测， cartographer带有tracker的说法。

// Implementation of a Kalman filter. We follow the nomenclature from  Thrun, S. et al., Probabilistic Robotics, 2006.

// Extended to handle non-additive noise/sensors inspired by Kraft, E., A // Quaternion-based Unscented Kalman Filter for Orientation Tracking.

3. scanMatcher 部分

3.1 karto 采取的的是real-time correlative scan matcher（三维窗口遍历寻优）的方式进行的。采取的是双分辨率的低分辨率和高分辨率的两次搜索。

This is an implementation of the algorithm described in "Real-Time  Correlative Scan Matching" by Olson.
The correlative scan matching algorithm is exhaustively evaluating the scan matching search space. As described by the paper, the basic steps are:
// 1) Evaluate the probability p(z|xi, m) over the entire 3D search window using the low-resolution table.
// 2) Find the best voxel in the low-resolution 3D space that has not already been considered. Denote this value as Li. If Li < Hbest, terminate: Hbest is
 the best scan matching alignment.
// 3) Evaluate the search volume inside voxel i using the high resolution table. Suppose the log-likelihood of this voxel is Hi. Note that Hi <= Li since the
 low-resolution map overestimates the log likelihoods. If Hi > Hbest, set Hbest = Hi.
 This can be made even faster by transforming the scan exactly once over some discretized range.

3.2 cartoGrapher也是采取的双搜索的方式进行的，先用一次real-time correlative scan matcher（三维窗口遍历寻优），再构建优化等式，利用ceres优化求解。（栅格概率， T的偏差，R的偏差）

occupied_space_cost_functor_weight   TranslationDeltaCostFunctor    RotationDeltaCostFunctor

4. submap的说明

4.1 karto没有submap的概念，全部以keyScan的形式存储在sensorManager。无地图缓存，但每次计算地图有计算消耗。

采取的是scan-map的匹配方式，每次keyScan进入主动的依据pose的距离窗口生成localMap进行匹配。 local 与 gloal的loop closure依据graph的结构和sensorManage顺序存储分配的ID信息，选择候选scans，生成localMap，进行匹配，依据score进一步确定闭环。

4.2 . cartographer采用了submap的概念，依据一定数量的scan初始一个submap, 依据窗口大小，插入newScan，更新submap. 有子图缓存，会占用内存。

// An individual submap, which has an initial position 'origin', keeps track of  which laser fans where inserted into it, and sets the  'finished_probability_grid' to be used for
 loop closing once the map no  longer changes.

// Submaps is a sequence of maps to which scans are matched and into which scans are inserted.
// Except during initialization when only a single submap exists, there are
// always two submaps into which scans are inserted: an old submap that is used
// for matching, and a new one, which will be used for matching next, that is
// being initialized.
//
// Once a certain number of scans have been inserted, the new submap is
// considered initialized: the old submap is no longer changed, the "new" submap
// is now the "old" submap and is used for scan-to-map matching. Moreover,
// a "new" submap gets inserted.

5. loopCheck

5.1 karto grapher主要依据pose 和 distance信息创建localMap，scanMatcher（real-time correlative scan matcher）确定。

1) 依据当前的Vertex, 从Graph中找到与之相邻的所有vertex(一定距离范围内).
2) 采取广度优先搜索的方式，将相邻（next）与相连（adjacentVertices）添加进nearLinkedScans.
3) 从sensorManager中取从前到后，依据id序号挑选与当前在一定距离范围内，且不在nearLinkedScans中的candidateScans, 当数量达到一定size，返回。
4）loopScanMatcher进行scanTomap的匹配，当匹配response 和covariance达到一定要求认为闭环检测到。得到调整的correct pose.     
5）Add link to loop :  调整边（全局闭环）
6) 触发correctPose: spa优化

5.2 cartogapher 类似（（real-time correlative scan matcher）），引入了branch and bound的方式，加快了闭环的查找。

依据多分辨率多层的树型结构，单枝生长的方式(branch)，及时剪枝操作（bound），深度优先搜索确定闭环。（Intra-submap Inter-submap ）

添加相应的闭环约束。构建优化问题，利用ceres优化。

// Current optimization problem.

sparse_pose_graph::OptimizationProblem optimization_problem_;
  sparse_pose_graph::ConstraintBuilder constraint_builder_ GUARDED_BY(mutex_);

// This is an implementation of the algorithm described in "Real-Time  Correlative Scan Matching" by Olson.

// It is similar to the RealTimeCorrelativeScanMatcher but has a different  trade-off: Scan matching is faster because more effort is put into the

// precomputation done for a given map. However, this map is immutable after  construction.

[原文：http://blog.csdn.net/zyh821351004/article/details/52421005 ]

安装：

github: https://github.com/googlecartographer/cartographer_ros

　安装指导　https://google-cartographer-ros.readthedocs.io/en/latest/

tip: catkin_make_isolate 编译带非ros repo的　　＝》　可以试试 catkin build 编译

rosbag 数据集出错（bag broke）：校验

shasum ~/Downloads/cartographer_paper_deutsches_museum.bag

2a021fadae9deb0643d73c8ca3acb332fcb3baa2

demo_video:

cartographer 3d http://v.qq.com/x/page/n0334yt1tt1.html

cartographer 2d http://v.qq.com/x/page/z03346p134v.html

https://github.com/googlecartographer/cartographer_ros/issues/41 sensor: two Hokuyo UTM-30LX-EW + 3DM GX4 25

地图保存接口（srv）: 保存位置：～/.ros/cartographer_tb.pgm

rosservice call /finish_trajectory "stem: 'cartographer_tb'"

最早调试遇到的bug，issue贴的图片。

来源：oschina

链接：https://my.oschina.net/u/4323267/blog/4050602

标签

ros

github

spa

cartographer