# aloha-ros-dev **Repository Path**: linClubs/aloha-ros-dev ## Basic Information - **Project Name**: aloha-ros-dev - **Description**: No description available - **Primary Language**: Unknown - **License**: Not specified - **Default Branch**: master - **Homepage**: None - **GVP Project**: No ## Statistics - **Stars**: 0 - **Forks**: 0 - **Created**: 2024-02-01 - **Last Updated**: 2024-02-02 ## Categories & Tags **Categories**: Uncategorized **Tags**: None ## README # 1 机械臂启动配置 1. 依赖安装 ~~~python sudo apt install libkdl-parser-dev ~~~ 2. 先断电再断can，再接can最后上电 + 13主臂， i进入键盘模式， 24从臂 + `/home/lin/follow/arx5_follow/src/arm_control/src/App/arm_control.cpp`文件中 `modifyLinkMass(model_path, model_path, 0.640)` 质量 3. 编译工程 ~~~python ./tools/build.sh ~~~ 4. 运行 ~~~python # 1 修改配置follow.sh文件 # 1.1 工程目录 workspace=￥{HOME}/test/follow # 1.2 密码 password=1 # 2 运行 ./tools/start.sh ~~~ 5. 删除编译文件 ~~~python ./tools/remove_make_file.sh ~~~ # 2 相机配置 ## 2.1 realsense相机配置 ### 2.1.1 librealsense驱动安装 ~~~python # 1 相机依赖 sudo apt-get install libssl-dev libusb-1.0-0-dev pkg-config libgtk-3-dev libglfw3-dev libgl1-mesa-dev libglu1-mesa-dev libuvc-dev libgoogle-glog-dev # 2. 添加规则库 cd realsense2_camera_ws/thirdparty tar -xzvf librealsense-2.50.0.tar.gz cd librealsense-2.50.0 ./scripts/setup_udev_rules.sh ## 2.1 该步骤可以跳过不执行。 构建和应用补丁内核模块，下面的脚本将下载、修补和构建受 realsense 影响的内核模块（驱动程序）。然后它将尝试插入修补模块而不是活动模块。如果失败，原始 uvc 模块将被恢复。 ./scripts/patch-realsense-ubuntu-lts.sh # 3. 编译安装 mkdir build && cd build && cmake .. sudo make install ~~~ ### 2.1.2 realsense-ros功能包 1. 生成相机的唯一序列号 + 生成相机序列号配置文件, 执行一次即可 + 如果相机更换了， 需要重新操作该步骤，更换对应的相机序列号 ~~~python # 1 依赖安装 sudo apt-get install ros-$ROS_DISTRO-ddynamic-reconfigure ros-$ROS_DISTRO-rgbd-launch # 2 编译源码 cd realsense2_camera_ws catkin_make # 3 生成序列号 # 3.1 生成序列号 ./devel/lib/realsense2_camera/list_devices_device # 将生成的Serial number序列号填入realsense2_camera/launch/目录下rs_multiple_devices.launch文件serial_no_camera*的默认值即可, 这里对rs_multiple_devices.launch复制后重命名为aloha.launch进行操作 # 3.2 自动配置序列号文件 （可以跳过，直接看启动相机） rosrun realsense2_camera list_devices_device && chmod +x ~/.aloha_camera_config.bash # 配置文件打开终端自动生效 echo "source ~/.aloha_camera_config.bash" >> ~/.bashrc # 生效相机序列号 source ~/.aloha_camera_config.bash ~~~ 2. 启动相机 ~~~python # 生效工作空间 source devel/setup.bash # 2 启动相机 roslaunch realsense2_camera aloha.launch ~~~ ## 2.2 ros_astra_camera ~~~python # 1 下载ros_astra_camera git clone https://github.com/orbbec/ros_astra_camera.git # 加代理下载 git clone https://mirror.ghproxy.com/https://github.com/orbbec/ros_astra_camera.git # 本工程自带相机ros驱动源码, 直接进入工作空间, 进行下面操作即可 cd astra_camera_ws # 2 依赖安装 sudo apt install libgflags-dev ros-$ROS_DISTRO-image-geometry ros-$ROS_DISTRO-camera-info-manager ros-$ROS_DISTRO-image-transport ros-$ROS_DISTRO-image-publisher libgoogle-glog-dev libusb-1.0-0-dev libuvc-dev libeigen3-dev # 3 编译相机包 catkin_make # 4 设置规则 source devel/setup.bash && rospack list roscd astra_camera ./scripts/create_udev_rules sudo udevadm control --reload && sudo udevadm trigger # 5 接上相机usb数据线,运行ls /dev/video* ls /dev/video* # 终端出现：/dev/video0 /dev/video1 /dev/video2 /dev/video3 /dev/video4 /dev/video5 # 一个相机会显示2个/dev/video*, 这里显示的6个，即3个相机 # 6 生成相机的唯一序列号 ./devel/lib/astra_camera/list_devices_node # 3个相机会输出3个不同的Serial number,将3个Serial number填入camera_ws/src/ros_astra_camera/launch/multi_dabai.launch中的camera*_serila_number参数即可，这里将multi_dabai.launch复制了一份重命名为aloha.launch # 6 启动相机程序 source devel/setup.bash roslaunch astra_camera aloha.launch # 7 查看相机图像 rqt_image_view # 如果rqt启动报错, 请关闭conda虚拟环境后，重新执行rqt_image_view， 如下 conda deactivate && rqt_image_view ~~~ # 3 采集训练数据 1. 依赖安装 ~~~python pip install h5py dm_env argparse numpy==1.23.4 tqdm==4.66.1 ~~~ 2. 录制bag包 ~~~python # 1 图像压缩格式 rosbag record -o test.bag /camera1/color/image_raw/compressed /camera2/color/image_raw/compressed /camera3/color/image_raw/compressed /master/joint_left /master/joint_right /puppet/joint_left /puppet/joint_right # 2 图像非压缩格式 rosbag record -o test.bag /camera1/color/image_raw /camera2/color/image_raw /camera3/color/image_raw /master/joint_left /master/joint_right /puppet/joint_left /puppet/joint_right ~~~ 3. 录制训练数据 + ros消息多机通讯是通过无线传输可以，图像可以采用CompressedImage压缩格式 + 如果是有线连接通讯建议图像直接用Image格式, 数据转换中耗时更少 + 结束当前回合按`ctrl +c` 即可 ~~~python # 1 不带底盘topic,采集压缩图像格式ros-CompressedImage python scripts/record_data.py --img_compressed --frequency 25 --max_timesteps 500 --episode_idx 0 # 2 带底盘, 图像非压缩格式 python scripts/record_data.py --use_base_robot --frequency 25 --max_timesteps 300 --episode_idx 0 ~~~ 4. 可视化数据 ~~~python python scripts/visualize_episodes.py --episode_idx 0 ~~~