doprakash/traj-gen-and-tracking

Trajectory generation and tracking for Table Tennis

======== INSTRUCTIONS ========

Video containing example trajectories can be found in video/ex_trajectories.mp4

To run a simulation containing solo trials with a ballgun, please do the following:

1. Run tableTennisPractice.m

2. OPT structure includes the following flags:

   DRAW - draw simulation of table tennis if true. PLAN - includes algorithms for generating trajectories and their options. LOOKUP - use lookup table instead of generating trajectories online. TRAIN - train a lookup table using successfully returned balls. VHP - use the Virtual Hitting Plane based method if true. RECORD - record the simulation inside MATLAB. DISTR - initial ball state related options CAMERA - std of the observation noise from the 'cameras'.

3. tt is a Table Tennis class that can be used for table tennis practice:

   tt.practice(q0,N) runs the solo table tennis practice N times from the initial robot posture q0.

Okan Koc, Jan Peters, Guilherme Maeda, Online optimal trajectory generation for robot table tennis, 2017.