Python, Robotic Manipulation, CoppeliaSim

Overview

Goal: Write a program that automatically plans a trajectory for a KUKA youBot mobile manipulator as it grasps a block and places it in a desired location.

GitHub: https://github.com/henryburon/mobile-manipulation

Process:

1. I wrote three functions to complete this project, which, when combined, generate the final trajectory:
   • TrajectoryGenerator
   • FeedbackControl
   • NextState
2. TrajectoryGenerator
   • Generates the reference trajectory for the end-effector frame {e}.
   • Consists of eight concatenated trajectory segments:
     • A trajectory to move the gripper from its initial configuration to a “standoff” configuration a few cm above the block.
     • A trajectory to move the gripper down to the grasp position
     • Closing of the gripper.
     • A trajectory to move the gripper back up to the “standoff” configuration.
     • A trajectory to move the gripper to a “standoff” configuration above the final configuration.
     • A trajectory to move the gripper to the final configuration of the object.
     • Opening of the gripper.
     • A trajectory to move the gripper back to the “standoff” configuration.
   • Output: A representation of the N configurations of the end-effector along the entire eight-segment reference trajectory.
3. FeedbackControl
   • Calculates the kinematic task-space feedforward plus feedback control law.
   • Output: The commanded end-effector twist V expressed in the end-effector frame {e}.
4. NextState
   • Calculates the robot’s configuration at the next time-step using first-order Euler-step.
   • Output: A 12-vector representing the configuration of the robot time Δt later.
5. Combine Functions
   • Finally, I integrated these three functions, provided the necessary initial and desired final configurations, and generated a .csv file to simulate the KUKA youBot’s movement in CoppeliaSim.
6. Results
   • I finished by generating three trajectories:
     • Best: A well-tuned feedforward-plus-PI controller. Error quickly converged to zero. First clip in the YouTube video.
     • Overshoot: A less-well-tuned feedforward-plus-PI controller. Error takes longer to converge to zero.
     • newTask: Different initial and final configurations of the block. Second clip in the YouTube video.