配置Gmapping

====================

Overview

---

环境：Ubuntu 16.04，ROS：kinetic，需要以下packages：

• urg_node：hokuyo的驱动
• gmapping：gmapping的核心package
• laser_scan_matcher：为gmapping提供odom

laser_scan_matcher

---

详细安装过程参考：Durant35/laser_scan_matcher

为了方便，复制了其中Ubuntu上的配置的部分如下：

• 安装python-wstool和python-rosinstall-generator：

   $ sudo apt-get install python-wstool
   $ sudo apt-get install python-rosinstall-generator
  

• 安装PCL package：

   $ cd <your-catkin-workspace-path>
   $ rosinstall_generator pcl_conversions pcl_msgs pcl_ros ‐‐rosdistro <your-ros-version> ‐‐deps ‐‐wet‐only ‐‐exclude roslisp ‐‐tar > ros_pcl.rosinstall
   $ wstool init src ros_pcl.rosinstall
   # This rosdep command installs all the missing system dependency
   # (must be described in package.xml) in all the packages in your src directory.
   $ rosdep install ‐‐from‐paths src ‐‐ignore‐src ‐‐rosdistro <your-ros-version> ‐y ‐r
   $ catkin_make
  

• 安装csm

   $ sudo apt-get install gsl-bin libgsl0-dev
   $ cd <your-catkin-workspace-path>/src
   $ git clone https://github.com/AndreaCensi/csm.git
   $ cd csm/
   $ ./install_quickstart.sh
  

• 安装laser_scan_matcher

   $ cd <your-catkin-workspace-path>/src
   # clone laser_scan_matcher into catkin workspace's src
   $ git clone https://github.com/Durant35/laser_scan_matcher.git
   # add csm path into your cmake's PKG_CONFIG_PATH
   $ export PKG_CONFIG_PATH=$PKG_CONFIG_PATH:<your-catkin-workspace-path>/src/csm/sm/pkg‐config
   # make targets
   $ cd <your-catkin-workspace-path>
   $ catkin_make laser_scan_matcher
  

• 测试laser_scan_matcher

  先source当前的workspace，然后运行demo.launch：

$ roslaunch laser_scan_matcher demo.launch ```

Gampping

---

参考链接：

• slam_gmapping

• gmapping

• How to Build a Map Using Logged Data

• openslam

具体过程如下：

• 将slam_gmapping和gmapping克隆到workspace

   $ cd <your catkin workspace>/src
   $ git clone https://github.com/ros-perception/slam_gmapping.git
  

• 上面两个package是基于openslam封装的hokuyo devel的ROS package，因此还需要原始的openslam package。因此，在workspace的src下：

   $ git clone https://github.com/ros-perception/openslam_gmapping.git
   # 进行编译
   $ catkin_make
  

• 测试gmapping。详见logged file建图。

  1. 下载一个bag。ROS上How to build a map using logged data提供的bag可能无法得到完整的地图，可以尝试克隆下面的bag。

       $ cd <your ros bag workspace>
       $ git clone https://github.com/RH-Song/slam-bag-files.git
     

  2. 启动一个roscore:

     $ roscore
     

  3. 设置 use_sim_time 为true:

       $ rosparam set use_sim_time true
     

4. 启动slam_gmapping:

   ```
   $ rosrun gmapping slam_gmapping scan:=base_scan
   ```
   

5. 启动rviz，观察建图过程。

   • 在workspace/src下，下载launch文件以及hokuyo驱动：

     ```
     $ git clone https://github.com/RH-Song/gmapping-hokuyo.git
     ```
     

   • 到gmapping-launch目录下启动rviz：

     ```
     $ rosrun rviz rviz -d gmapping.rviz
     ```
     

6. 开始回播bag文件，以上面的包为例子：

   ```
   $ rosbag play --clock gmapping-1.bag
   ```
   

7. 图片保存。等待bag文件播放完成以后，可以将建成的图片保存下来：

   ```
   $ rosrun map_server map_saver -f <map_name>
   ```
   

使用hokuyo建图

---

• 添加hokuyo的驱动urg_node package，上一步已经下载了相关驱动文件，可以进行编译。

   $ catkin_make
  

• 连接hokuyo激光雷达

• 运行gmapping.launch，进行建图。

$ roslaunch gmapping-launch gmapping.launch