机器人抓取（一）—— ROS MoveIt! 程序控制真实UR5机器人运动（python）

moveit! ur5 机器人控制

上文在 ubuntu 系统中配置了ur 机器人ros功能包 universal_robot 和驱动 Universal_Robots_ROS_Driver，并实现了用 moveit_rviz 拖动机器人运动。本文旨在学习如何使用 moveit python 用户接口moveit_commander，实现真实的ur5 机器人轨迹规划与控制。

1. 创建功能包
   cd ur5/src # ur5是工作空间
   catkin_create_pkg ur_move_test std_msgs rospy
2. 编写 python 程序
   参考：MoveIt » Tutorials » Move Group Python Interface

在 ur_move_test 文件夹下新建scripts文件夹，用于存放 python 文件。在 scripts 文件夹中创建 ur_move_test_node.py 。此程序仅作为moveit路径规划与执行功能的简单示例。程序如下:

!/usr/bin/env python

use moveit_commander (the Python MoveIt user interfaces )

import sys
import copy
import rospy
import moveit_commander
import moveit_msgs.msg
import geometry_msgs.msg
from math import pi

class MoveGroupInteface(object):
def init(self):
super(MoveGroupInteface, self).init()
######################### setup ############################
moveit_commander.roscpp_initialize(sys.argv)
rospy.init_node('ur_move_test_node', anonymous=True)
self.robot = moveit_commander.RobotCommander()
self.scene = moveit_commander.PlanningSceneInterface() # Not used in this tutorial
group_name = "manipulator" # group_name can be find in ur5_moveit_config/config/ur5.srdf
self.move_group_commander = moveit_commander.MoveGroupCommander(group_name)
self.display_trajectory_publisher = rospy.Publisher('/move_group/display_planned_path',moveit_msgs.msg.DisplayTrajectory,queue_size=20)

    ################ Getting Basic Information ######################
    self.planning_frame = self.move_group_commander.get_planning_frame()
    print "============ Planning frame: %s" % self.planning_frame
    self.eef_link = self.move_group_commander.get_end_effector_link()
    print "============ End effector link: %s" % self.eef_link
    self.group_names = self.robot.get_group_names()
    print "============ Available Planning Groups:", self.robot.get_group_names()
    print "============ Printing robot state:"
    print self.robot.get_current_state()  # get
    print ""

def plan_cartesian_path(self, scale=1):
    waypoints = []
    wpose = self.move_group_commander.get_current_pose().pose
    wpose.position.z -= scale * 0.1  # First move up (z)
    waypoints.append(copy.deepcopy(wpose))
    wpose.position.x += scale * 0.1  # Second move forward/backwards in (x)
    waypoints.append(copy.deepcopy(wpose))
    wpose.position.y += scale * 0.1  # Third move sideways (y)
    waypoints.append(copy.deepcopy(wpose))    

    # We want the Cartesian path to be interpolated at a resolution of 1 cm
    # which is why we will specify 0.01 as the eef_step in Cartesian
    # translation.  We will disable the jump threshold by setting it to 0.0,
    # ignoring the check for infeasible jumps in joint space, which is sufficient
    # for this tutorial.
    (plan, fraction) = self.move_group_commander.compute_cartesian_path(
                                    waypoints,   # waypoints to follow
                                    0.01,      # eef_step  
                                    0.0)         # jump_threshold  

    # Note: We are just planning, not asking move_group to actually move the robot yet:
    pring "=========== Planning completed, Cartesian path is saved============="
    return plan, fraction

def execute_plan(self, plan):
    ## Use execute if you would like the robot to follow
    ## the plan that has already been computed:
    self.move_group_commander.execute(plan, wait=True)

print "----------------------------------------------------------"
print "Welcome to the MoveIt MoveGroup Python Interface Tutorial"
print "----------------------------------------------------------"
print "Press Ctrl-D to exit at any time"
print ""
print "============ Press Enter to begin the tutorial by setting up the moveit_commander ..."
raw_input()
tutorial = MoveGroupInteface()
print "============ Press Enter to plan and display a Cartesian path ..."
raw_input()
cartesian_plan, fraction = tutorial.plan_cartesian_path()
print "============ Press Enter to execute a saved path ..."
raw_input()
tutorial.execute_plan(cartesian_plan)
print "============ Press Enter to go back ..."
raw_input()
cartesian_plan, fraction = tutorial.plan_cartesian_path(scale=-1)
tutorial.execute_plan(cartesian_plan)
注：上述程序中，给定路径点计算轨迹时的返回值 fraction 表示，规划的轨迹在给定路径点列表的覆盖率 [0~1]，如果fraction小于1,说明给定的路径点无法完整规划。

3. 创建 launch文件
   启动机器人控制节点，要在不同终端分别启动 ur5_bringup.launch、ur5_moveit_planning_execution.launch、moveit_rviz.launch（可选）、ur_move_test_node.py 。非常麻烦，因此编写launch文件，一次性启动上述节点。

在ur_move_test 功能包中创建launch文件夹，用以存放launch文件。

ur5机器人的启动文件: ur5_move_test.launch

<launch>

  <!-- robot_ip: IP-address of the robot's socket-messaging server -->
  <arg name="robot_ip" default="192.168.1.2" doc="IP of the controller"/>
  <arg name="kinematics_config" default="$(find ur_description)/config/ur5_calibration.yaml"/>
  <arg name="limited" default="true" doc="If true, limits joint range [-PI, PI] on all joints." />

  <include file="$(find ur_robot_driver)/launch/ur5_bringup.launch">
    <arg name="robot_ip" value="$(arg robot_ip)"/>
    <arg name="kinematics_config" value="$(arg kinematics_config)"/>
  </include>

  <include file="$(find ur5_moveit_config)/launch/ur5_moveit_planning_execution.launch">
    <arg name="limited" value="$(arg limited)"/>
  </include>

  <include file="$(find ur5_moveit_config)/launch/moveit_rviz.launch">
    <arg name="config" value="true"/>
  </include>
  <node name="ur_move_test_node" pkg="ur_move_test" type="ur_move_test_node.py" respawn="true" output="screen" />

</launch>

4.修改 CmakeLists.txt文件, 编译功能包
在ur_move_test 功能包的 CmakeLists.txt文件夹中找到 catkin_install_python函数，解除此部分注释，将代码修改为：

Mark executable scripts (Python etc.) for installation

in contrast to setup.py, you can choose the destination

catkin_install_python(PROGRAMS
scripts/ur_move_test_node.py
DESTINATION ${CATKIN_PACKAGE_BIN_DESTINATION}
)
编译功能包：

cd ur5
catkin_make # 或仅编译单个功能包 catkin_make -DCATKIN_WHITELIST_PACKAGES=“ur_move_test”

5. 运行程序
   运行之前，要给ur_move_test_node.py文件增加可执行权限：

cd ur5/src/ur_move_test/scripts
chmod +x ur_move_test_node.py
运行lunch文件
roslaunch ur_move_test ur5_move_test.launch
注意：不要忘记启动机器人示教器中的external_control.urp程序，详情参考：ubuntu18.04 ros-melodic 安装 ur_robot_driver，驱动真实的 ur5 机器人。

6. 修改 ee_link 固定位置
   ur5_moveit_config/ur5.srdf 文件中可以看到"manipulator" group 的 tip_link为 ee_link。

   <group name="manipulator">
   <chain base_link="base_link" tip_link="ee_link" />
   </group>

   在ur_description/urdf/ur5.urdf.xacro 中，

<joint name="${prefix}ee_fixed_joint" type="fixed">
    <parent link="${prefix}wrist_3_link" />
    <child link = "${prefix}ee_link" />
    <origin xyz="0.0 0.0 0.0" rpy="0.0 ${-pi/2.0} ${pi/2.0}" />
</joint>

本文所用机器人在末端安装了hex-h传感器和rg6机械手，机器人 tcp变为rg6的指尖端点。因此moveit运动规划的 tip_link应修改。
由于机械手的urdf文件添加较为麻烦，此处采用简单粗暴的方法，直接修改ee_fixed_joint的坐标。

启动机器人运行程序，执行指令 rosrun tf tf_echo /wrist_3_link /tool0_controller,可以查看机器人TCP与 wrist_3_link 的坐标系转换关系，以此修改ee_fixed_joint即可。

<joint name="${prefix}ee_fixed_joint" type="fixed">
    <parent link="${prefix}wrist_3_link" />
    <child link = "${prefix}ee_link" />
    <origin xyz="0.0 0.0 0.315" rpy="0.0 0.0 0.0" />
</joint>