Python Library for Controlling UR5 Robot in Realtime

UR5 Realtime client Lightweight python library for controlling UR5 robot from Universal Robot family. Robot kinematics are calculated based on the following paper: tech report ### Requirements * UR Control Box (simulated or real) * Tested on UR Robot software 3.5.1 ### Try it out! python >>> from realtime_client import RTClient >>> import numpy as np # connect to address 127.0.0.1 and port 30003 >>> rtc = RTClient('127.0.0.1', 30003) # move end effector to joint goal [q1, q2, q3, q4, q5, q6] >>> joint_values = [d * (np.pi / 180) for d in [-90, 89, 5]] >>> rtc.move_j(joint_values) # move end effector to pose goal [x,y,z,rx,ry,rz] (3D translation and 3D rotation) >>> pose = np.array([-0.46481069, -0.18235116, 0.13827986, -1.58136603, -2.69628063, -0.01169701]) >>> rtc.move_l(pose) # velocity-based controller, move to pose goal [x,y,z,rx,ry,rz] >>> pose = np.array([0.470, -0.491, 0.430, 0.13, 3.15, -0.00]) >>> rtc.move_v(pose) # kinematic test >>> from kinematics import KinematicsUR5 >>> from math_tools import pose2tf >>> kin = KinematicsUR5() >>> target_ee = np.array([-1.10586325e-01, -4.86899999e-01, 4.31871547e-01, -1.36273738e-01, -3.12118227e+00, 1.18929713e-03]) >>> target_pose = pose2tf(target_ee) >>> solutions = kin.inv_kin(target_pose) >>> closest_solution = kin.get_closest_solution(solutions, rtc.get_feedback('joint_values')) >>> rtc.move_j(closest_solution) >>> assert np.allclose(rtc.get_feedback('tool_pose'), target_ee) >>> rtc.close_connection()