Deep Grasp GPD配置注意事项

启动流程

仿真机器人或者真实机器人

机器人规划和控制器-move group

MTC和GPD节点

每个模块配置注意事项

机器人规划和控制器-move group

1）首先要编译ExecuteTaskSolutionCapability插件，move_group插件。

代码在：moveit_task_constuctor项目里面，capabilities提供这部分能力。

2）同时，在文件里配置插件到move_group里面启动

 <!-- Load ExecuteTaskSolutionCapability so we can execute found solutions in simulation --> <!-- <param name="move_group/capabilities" value="move_group/ExecuteTaskSolutionCapability" /> -->

<!-- MoveGroup capabilities to load -->  <param name="capabilities" value="move_group/MoveGroupCartesianPathService            move_group/MoveGroupExecuteTrajectoryAction            move_group/MoveGroupKinematicsService            move_group/MoveGroupMoveAction            move_group/MoveGroupPickPlaceAction            move_group/MoveGroupPlanService            move_group/MoveGroupQueryPlannersService            move_group/MoveGroupStateValidationService            move_group/MoveGroupGetPlanningSceneService            move_group/ClearOctomapService            move_group/ExecuteTaskSolutionCapability            " />

到此，move_group规划节点已经完成了部署好了接收控制指令的准备。

deep_grasp_task配置

抓取参数配置

 折叠源码

# Total planning attempts max_solutions: 10 # Planning group and link names arm_group_name: "xarm6" eef_name: "xarm_gripper" hand_group_name: "xarm_gripper" hand_frame: "link_eef" # Poses hand_open_pose: "open" hand_close_pose: "close" arm_home_pose: "home" # Scene frames world_frame: "world" table_reference_frame: "world" object_reference_frame: "world" camera_reference_frame: "world" surface_link: "table" # Collision object for picking # CYLINDER object specifications spawn_mesh: false object_name: "object" object_dimensions: [0.25, 0.02] # [height, radius] object_pose: [0.5, -0.25, 0.0, 0, 0, 0] # Camera collision object spawn_camera: false camera_name: "camera" camera_mesh_file: "package://deep_grasp_task/meshes/camera/kinect.dae" camera_pose: [0.5, 0, 0.7, 0, 0, 0] # side camera_pose: [0.5, 0, 0.7, 0, 1.571, 1.571] # over head # Table model spawn_table: false table_name: "table" table_dimensions: [0.4, 0.5, 0.1]  # [length, width, height] table_pose: [0.5, -0.25, 0, 0, 0, 0] # Gripper grasp frame transform [x,y,z,r,p,y] # grasp_frame_transform: [0, 0, 0.1, 1.571, 0.785, 1.571] grasp_frame_transform: [0, 0, 0.172, 0, 0, 0] # Place pose [x,y,z,r,p,y] # place_pose: [0.5, 0.3, 0, 0, 0, 0] place_pose: [0.6, -0.15, 0, 0, 0, 0] place_surface_offset: 0.0001 # place offset from table # Valid distance range when approaching an object for picking approach_object_min_dist: 0.1 approach_object_max_dist: 0.15 # Valid height range when lifting an object after pick lift_object_min_dist: 0.01 lift_object_max_dist: 0.1

其中，相机坐标系参数配置

# camera extrinsics # transformation from camera link to optical link [x y z qw qx qy qz] trans_cam_opt: [0.0, 0.0, 0.0, 7.07072724e-01, -1.89459405e-08,  7.07140837e-01,  1.89477656e-08] # transformation from base link to camera link [x y z qw qx qy qz] # when camera is to the right of the robot (cylinder) trans_base_cam: [0.08, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0]