The time-based contact planner seems to work well, however it works in unexpected ways.
The code for updating lift-off position and the actual gait phase is currently in the same place the old contact planner was running within the MPC thread. Intuitively it would make more sense to have those two at the end of the MPC iteration, when actual solutions for the newly acquired stance phase are available, for example.
The time-based contact planner seems to work well, however it works in unexpected ways.
The code for updating lift-off position and the actual gait phase is currently in the same place the old contact planner was running within the MPC thread. Intuitively it would make more sense to have those two at the end of the MPC iteration, when actual solutions for the newly acquired stance phase are available, for example.