aikido

aikido copied to clipboard

VectorField planner that plans a path by following defined joint velocities in configuration space and terminates by defined criteria.

• Use boost::numeric::odeint::integrate_adaptive that integrates joint velocities with adaptive step size

• Refactor implementation to follow the design in prpy; https://github.com/personalrobotics/prpy/blob/master/src/prpy/planning/vectorfield.py

• Add three vector fields

  • [x] [MoveEndEffectorOffset] that moves an end-effector to an offset by a direction; #268
  • [x] [MoveEndEffectorPose] that moves an end-effector to a goal pose; #268
  • [ ] [MoveEndEffectorPoseAlong] that moves an end-effector along a given workspace trajectory.

• In the short-term, VFP returns a linear spline using Spline class. A computeParabolicTime() function is added for timing a linear spline.

• In the long-term, VFP returns a cubic spline using Spline class. A new timing method, e.g. Bobrow's algorithm, is added for timing a spline without altering the geometric path.

@dqyi11 All the issues here resolved by #268? If so, could you please close this?

Not yet, there will be another PR for

[MoveEndEffectorPoseAlong] that moves an end-effector along a given workspace trajectory.

I added a checklist to the description accordingly. Feel free to modify if necessary.