Card Set Information

2014-01-09 08:39:53

Show Answers:

  1. Teach by Show
    • Teach by show is the process of moving the arm to the desired locations (either using robots own actuators controlled by operator using teach pendent - Drive Through, or by the operator physically moving the arm - Lead Through) and then instructing the system to store the location (by sampling joint sensors).
    • This would be the preferred method since it is unlikely that the required co-ordinates would be known with sufficient accuracy (due to the limitations in the accuracy of building the cell, ie. positioning the robot/conveyors etc).
  2. Why does a robot need 6Dof?
    • An object moving freely in space has 6 DoF; 3DoF of translation along the x,y and z axes plus 3 DoF of rotation about the x,y and z axes. Therefore for a robot to be able to manipulate an object freely in space, it has to possess these 6 DoF at the end effector.
    • A robot generally consists of a serial chain of rigid links connected by single DoF joints and the overall structure generally adopted is to have 3 joints in the main arm that are a mixture of prismatic and revolute joints (giving rise to a range of different configurations: prismatic, cylindrical etc) and together these 3 joints generate the 3 DoF of translation at the end effector.
    • Then at the end of the main arm is a wrist with a further 3 joints, all revolute, where the 3 axes of rotation intersect at a common point and together these joints generate the 3 DoF of rotation at the end effector.
  3. When does a robot not need 6Dof
    • Some process applications such as paint spraying, arc welding, screw driving, nut
    • running do not require all 6 DoF, since they do not require rotation about the process axis either because the process is axi-symmetric (rotationally symmetric) in cases such as paint spraying or arc welding or because the process tool provides the necessary process rotation (screw driving, nut running). In these cases it is possible to dispense with one of the wrist rotations and the robot only has 5 DoF.
    • In assembly, many (approx. 80%) of the assembly operations involve insertion of components from vertically above. In such circumstances, given the parts are normally presented to the robot the right way up, all that is required of the robot is to move freely to any position in space (ie possess the 3 DoF of translation) plus the ability to rotate about the vertical axis (to align the component rotationally). With a suitable main arm configuration (eg. SCARA), only one wrist rotation is required about the vertical axis and the robot only has 4 DoF.
  4. Interpolation
    • Interpolation is process of sending a series of small increments to joint controllers
    • throughout duration of a move at regular intervals of time (eg. 28ms on PUMA), as opposed to sending the full increment at start of move. This minimises any variation in end-effector path due inevitable variation in joint controller response.

    • Joint interpolation involves sending uniform (constant) increments throughout move
    • (in effect ramp input) so that all joints start and finish at same time. Generally this will produce curved end-effector path.

    • World interpolation involves computing intermediate points along the desired
    • end-effector path (typically a straight line) and using inverse kinematic
    • equations to compute the corresponding joints values that are sent to the joint
    • controllers, generally resulting in variable increments. The end-effector
    • follows the desired path but at a much slower speed than with joint
    • interpolation.

    In the task in part a) vertical linear motion is required when picking and placing the components but since in the SCARA robot this motion is produced by the vertical prismatic joint, it is not necessary to use world interpolation.
  5. Discuss the difficulties in obtaining the
    inverse kinematic equations for a general 6 DoF revolute polar arm and explain how this problem is normally solved. If the arm depicted in Figure 3 incorporated a 3 DoF wrist, thus turning it into a 6 DoF arm, explain how the equations derived above would be applicable
    • With a general 6 DoF arm the forward equations will generally be a set of 6 highly non-linear equations which are functions of all 6 joint variables. Being non-linear there is no general method for solving the forward equations to obtain the inverse equations. All that is available is a set of heuristics (guidelines that may work) and nobody yet managed to use heuristics to solve such complex expressions.
    • However if joint successive joints intersect at a single point or are parallel to each other, solutions have been identified. In most robots, the wrist is designed so that the revolute axes intersect at a single point, in the centre of the wrist assembly.
    • The forward equations for position of this point only involve the first three joints (this point lies on the axes of the wrist joints and hence its position is independent of those joints) and would be similar to the above equations for which the solution has been derived.
    • Having solved for first three joints, the 3 orientation equations only involve 3 unknowns (the wrist angles) and again heuristics have yielded a solution to this
  6. Why is Teach by Show often used over textual specification?
    Teach by show is often used in preference to textual specification (ie. entering/typing the required numerical co-ordinates in the program) since very often the required co-ordinates are simply not known with sufficient accuracy – in effect the robot is being used as a digitisation device.

    In the program, textual specification is used purely for the vertical offsets above the pick and place locations where the accuracy is not critical and only nominal offsets are required.
  7. Explain why it is necessary for the robot to go via a point 80mm offset from the pick
    and place locations.
    • Whenever going to/from a pick or place location, robot cannot move directly to/from the location because, depending on where its moving from (ie its previous location) or moving to (its next location), there is a risk of collision either between the gripper and the part or between the gripped part and the fixture or the rest of the assembly.
    • Programming the robot to go via a point vertically above the pick or place location, ensures that the robot arrives or departs along a vertical and hence safe (collision free) path.
  8. Explain the issue that can arise with the gripper speed of response and how this issue is addressed in your program.
    • The issue with the gripper speed of response is that it does not respond instantly to the control signal. It takes a finite time for the air valve to open and for the air pressure to build up in the pipe and the body of the gripper before the fingers start to move, longer still for the fingers to open/close fully.
    • Normally the instruction to open/close the gripper is immediately followed by an instruction for the robot to move away and thus there is a chance that the robot will have moved before the gripper fully opens/closes causing either an incorrect grip position when closing or dropping the part when opening.
    • Thus to give time for the gripper to fully open/close, a delay is explicitly added to the program.
  9. Explain what is meant by the terms “Joint Space” and “World Space” in the context of robot kinematics
    Joint space is a definition of robot location/motion in terms of the controlled joint variables (θ1 - θ6 for a general 6 DoF arm) whereas world space is a definition in terms of the end effector location in space (x,y,z for position and α,β,γ for orientation).
  10. Explain the role ofthe kinematic equations
    The kinematic equations are used by therobot controller to convert between the 2 definitions; forward equations convert from joint to world space whereas inverse equations convert from world to joint space.
  11. Explain why, for a general 6 DoF arm, the wrist
    is designed so that the 3 joint axes intersect at a single point
    • For a general 6 DoF, the forward equations are highly non-linear functions of all 6 joint variables for which no algebraic approach using heuristics (as used above) has yet been derived. With 3 axes intersecting, possible to derive forward equations for the position of the wrist centre (the axis intersection point) which are functions of the first 3 joints (θ1, θ2, θ3) alone - much simpler so that solutions have been derived.
    • The 3 orientation equations are still functions of all 6 joint variables but having solved for (θ1, θ2, θ3), there are only 3 unknowns left (θ4, θ5, θ6) and again heuristic solutions have been developed for these.
  12. Explain what is meant by the terms “Joint Space” and World Space” in the context of robot kinematics and explain the role of the kinematic equations. 

    Illustrate your answer by describing how these equations would be used in the execution of the following VAL instruction:-
                                                                  DEPART 200
    Joint space is definition of location/motion in terms of joint variables (θ1 - θ6), whereas world space is definition of location/motion in terms of end-effector position/orientation (x,y,z,α,β,g)

    Kinematic equations enable robot to convert between joint and world space. Forward Kinematic Equations (FKE) convert from joint space to world space, Inverse Kinematic Equations (IKE) convert from world to joint.

    DEPART 200 - System uses FKE to convert current joint space location to world co-ordinates. The offset of 200mm is added and then the system uses the IKE to convert the offset world location to joint values which are sent to joint controllers.
  13. Describe the difference between passive
    and active compliance
  14. By comparison with a typical analogue camera, explain the advantages of a Charge Coupled Device camera (CCD) and hence
    comment on why it is so suited to industrial tasks requiring component recognition, and the determination of component orientation and location
  15. The image processing tasks of component
    recognition, and determination of object orientation and location, can usually be simplified if the raw images are subject to some form of software-based enhancement.  Describe THREE techniques which are commonly used to enhance images so that the object in question becomes more distinguishable from its background.
  16. Assuming that an object is captured in a binary coded image, describe with the aid of simple sketches, the various software, and hardware-based techniques commonly used for edge detection. and explain, how each method can be achieved. In the case of your chosen hardware-based technique, explain how it could be adapted to Grey Scale images.
  17. Explain the principle of operation of the RCC and compare its performance with a SCARA design of robot
  18. Explain the advantages of active compliance over passive compliance
  19. Describe the construction and principle of operation of the Remote Centre Compliance
    device (RCC), and explain how it solves the problems of misalignment when assembling a round peg into a round hole.
  20. Describe what is meant by hybrid force/velocity control and explain, with regard to software and including any relevant equations, how it might be implemented.

    With specific reference to way in which the order in which the degrees of freedom of the mating parts are constrained
  21. Explain the GUARDI command
    Guardi stops the robot and/or motion once an unexpected event has occurred. It is part of hybrid force control. Not suitable for tracking - too slow, jerky motion. Carries onto the next command
  22. How can you make the threshold more robust?
    Extra terminating conditions, e.g. a distance counter - as soon as you manke contact in the z plane set that as 0, then when you have moved down the specified amount, you know the chamfer cross stage has ended
  23. Gear assumptions
    • Gear B must be prevented from rotating to allow Gear A to mesh
    • Gear teeth are unchamfered - no chamfer crossing stage required for teeth meshing
  24. Kinematics
    • - study of the geometry of motion
    • - generally involves analysis of both the geometry of a mechanism (eg a robot) and the geometry of velocities/accelerations of the mechanism.