MIN – Fakultät Fachbereich Informatik The Kinema*cs of End-Effectors in Collabora*ve Robots Soniya Vijayakumar Universität Hamburg Fakultät für Mathematik, Informatik, und Naturwissenschaften Fachbereich InformaKk Technische Aspeckte Mul0modaler Systeme 17. December 2018 1
Outline Introduction | Collaborative Robots | End-Effectors | Kinematics | Hand Guiding | Trajectory | Gripping | Summary Ø Introduction Ø Collaborative Robots Ø End Effector Systems Ø Kinematics in End Effectors Ø Hand Guiding Applications Ø Summary S. Vijayakumar – The Kinematics of End-Effectors in Collaborative Robots 2/18
CO llaborative Ro BO CO BOT s Introduction | Collaborative Robots | End-Effectors | Kinematics | Hand Guiding | Trajectory | Gripping | Summary Cobot - robot with direct physical interaction • a human user • a shared workspace 3 • Invented professors J. Edward Colgate and Michael Peshkin in 1996 • At Northwestern University 2 Sawyer and Baxter cobots from Rethink Robotics. 1 2 http://peshkin.mech.northwestern.edu/cobot/ 1 https://www.roboticsbusinessreview.com/wp- 3 hIps://en.wikipedia.org/wiki/Cobot#cite_note-1 content/uploads/2018/05/Sawyer_and_Baxter-300x229.jpg S. Vijayakumar – The Kinematics of End-Effectors in Collaborative Robots 3/18
End Effectors Introduction | Collaborative Robots | End-Effectors | Kinematics | Hand Guiding | Trajectory | Gripping | Summary Device at end of the arm 2 • Designed to interact with environment • Last Link of the Robot Human-Hand Force- Closure 1 Types Impactive • Ingressive • Astrictive • Contigutive • MagneRc Robot End-Effector 3 2 https://en.wikipedia.org/wiki/Robot_end_effector 1 Richard Greenhill and Hugo Elias (myself) of the Shadow Robot Company 3 https://blog.robotiq.com/bid/65794/Magnetic-Robot-End-Effector-Top-5-Pros-and-Cons S. Vijayakumar – The Kinema3cs of End-Effectors in Collabora3ve Robots 4/18
Importance of End Effectors IntroducAon | CollaboraAve Robots | End-Effectors | KinemaAcs | Hand Guiding | Trajectory | Gripping | Summary https://www.youtube.com/watch?v=1EpJv34gQ88 S. Vijayakumar – The Kinematics of End-Effectors in Collaborative Robots 5/18
Kinema'cs Introduction | Collaborative Robots | End-Effectors | Kinematics | Hand Guiding | Trajectory | Gripping | Summary Forward Kinematics Description of End-Effector Configuration (Position & Orientation) • Function of Joint Coordinates • Reverse Kinematics • Description of Joint Coordinates • Function of End-Effector Configuration Þ What are we learning? Þ Relative pose between frame coordinates S. Vijayakumar – The Kinematics of End-Effectors in Collaborative Robots 6/18
Kinematics Introduction | Collaborative Robots | End-Effectors | Kinematics | Hand Guiding | Trajectory | Gripping | Summary Influencing Factors: [6] Ø End-Effector Weight Degree of Freedom Ø Number of joints Ø Length of the links Ø External force/moment Ø Human force Ø Collisions in dynamic environments Ø Noise Ø Inertial Forces/moments dues to acceleration Ø Friction at joints Ø S. Vijayakumar – The Kinematics of End-Effectors in Collaborative Robots 7/18
Hand Guiding IntroducAon | CollaboraAve Robots | End-Effectors | KinemaAcs | Hand Guiding | Trajectory | Gripping | Summary -> Representative Functionality of Cobots • Teaching Pendant -> Unskilled users interact and program robots Þ Limits intuitiveness Þ What are we teaching them? [4] Þ Position and Orientation https://sites.google.com/site/hoomanleerobot/research S. Vijayakumar – The Kinematics of End-Effectors in Collaborative Robots 8/18
Hand Guiding IntroducAon | CollaboraAve Robots | End-Effectors | KinemaAcs | Hand Guiding | Trajectory | Gripping | Summary Assumption A Force Feedback at the robot end-effector • Three motion groups 1. Position in X,Y,Z Hand Guiding Force -> Linear positioning coordinates -> robot base 2. Orientation in the Cartesian space Hand Guiding Moment -> Angular positioning 3. Rotation around its axis S. Vijayakumar – The Kinematics of End-Effectors in Collaborative Robots 9/18
Self-Learning: Trajectory IntroducAon | CollaboraAve Robots | End-Effectors | KinemaAcs | Hand Guiding | Trajectory | Gripping | Summary Initial State Aim: Learning of a Trajectory Smallest cost State ac8on Trajectory [2] • Controller A length defined by repetition of pattern • Action State – joint position vectors • Controller – optimization function • Dynamics Action – joint target positions • Dynamics – cost function and next state • State Final State Trajectory Pattern [2] S. Vijayakumar – The Kinematics of End-Effectors in Collaborative Robots 10/18
Self-Learning: Trajectory Introduction | Collaborative Robots | End-Effectors | Kinematics | Hand Guiding | Trajectory | Gripping | Summary Aim: Learning of a Trajectory Challenges [3] • Desired future trajectories • Solution • Impedence Control with • Interactive Trajectory Deformation S. Vijayakumar – The Kinematics of End-Effectors in Collaborative Robots 11/18
Self-Learning: Trajectory Introduction | Collaborative Robots | End-Effectors | Kinematics | Hand Guiding | Trajectory | Gripping | Summary 1. Energy Function Maps: trajectory deformations Smooth Trajectories -> cost function 2. Constrainted Optimization Human Force Fixed end point -> Position -> Velocity 3. Variation Field Estimation Desired Trajectory Trajectory Deformation [3] -> Trajectory deformation S. Vijayakumar – The Kinematics of End-Effectors in Collaborative Robots 12/18
Self-Learning: Trajectory Introduc3on | Collabora3ve Robots | End-Effectors | Kinema3cs | Hand Guiding | Trajectory | Gripping | Summary Achievements Ø Compatible with traditional Impedence Control Experiments Ø Reduction in human intervention ü Reduction in torque application ü Improvement in movement quality ü S. Vijayakumar – The Kinema3cs of End-Effectors in Collabora3ve Robots 13/18
Self-Learning: Gripper Introduction | Collaborative Robots | End-Effectors | Kinematics | Hand Guiding | Trajectory | Gripping | Summary Aim: Learning of Gripper Orientation Dextrous Gripper Mechanism 2 • Inputs: Position, force and vision Sensors • Robot joints – Position and Velocity • Fingers and Arms - Force • Controller: Torque Control • Robot joints • Robot hand fingers • Sarcos GRLA Arm 1 2 http://users.cecs.anu.edu.au/~rsl/rsl_dextrous.html 1 h:p://www.cim.mcgill.ca/research/94-95AnnualReport/node99.html S. Vijayakumar – The Kinema3cs of End-Effectors in Collabora3ve Robots 14/18
Self-Learning: Gripper Introduction | Collaborative Robots | End-Effectors | Kinematics | Hand Guiding | Trajectory | Gripping | Summary Aim: Learning of Gripper Orientation Challenges [5] • Correspondence problem • Generalization • Robustness to Disturbances • Solution • Modified Dynamic Movement • Primitive Framework S. Vijayakumar – The Kinema3cs of End-Effectors in Collabora3ve Robots 15/18
Self-Learning: Gripper Introduc3on | Collabora3ve Robots | End-Effectors | Kinema3cs | Hand Guiding | Trajectory | Gripping | Summary 1. Dynamic Movement Primitives 2. Obstacle Avoidance Transformation & Canonical System Coupling Term -> Converges to a Goal -> Rotational Matrix -> Relative angle DMP Control Diagram [5] S. Vijayakumar – The Kinema3cs of End-Effectors in Collabora3ve Robots 16/18
Self-Learning: Gripper Introduction | Collaborative Robots | End-Effectors | Kinematics | Hand Guiding | Trajectory | Gripping | Summary Achievements Ø Improvement to dynamic movement primitives Experiments Ø Adapt movements with changing goals ü Adapt movements with moving obstacles ü Movement library that could be reused ü S. Vijayakumar – The Kinema3cs of End-Effectors in Collabora3ve Robots 17/18
Summary Introduction | Collaborative Robots | End-Effectors | Kinematics | Hand Guiding | Trajectory | Gripping | Summary ü Introduction ü Collaborative Robots ü End Effector Systems ü Kinematics in End Effectors ü Hand Guiding Applications ü Trajectory ü Gripper S. Vijayakumar – The Kinema3cs of End-Effectors in Collabora3ve Robots 18/18
Recommend
More recommend