GitLab

Solved a bug in the heading cost function: If the trajectory had less points than the ones reuired to compute the cost, the node crashed.

Rename stucked with stuck. Add parameters to enable/disable the stuck check and change the vel threshold and max occurrences.

Add bool param to invert cmd_vel steering value. Add bool param to enable/disable inplace steering change before each maneuver

Added a new parameter to generate the steer angle instead of the angular speed for the control.

See merge request !3

Added a new parameter to generate the steer angle instead of the angular speed for the control.

See merge request !2

Heading cost weighted by the number of samples.

Melodic migration

See merge request !1

Added a new input topic with the specific ackermann feedback information (steer angle and velocity and forward speed).
The Odometry information does not expect any additional data.

Improved the indentation.

Steer the car, if necessary, at the first segment.
Added a configurable average of the odometry and cmd_vel info's.

Changed the reference system in which the odometry was computed. Now it is in between the rear wheels.
Added a replanning condition when the final position is farther away than twice the maximum goal tolerance.
Improved the computation of the transverse speed window limits.
Reported some usefull information at each control cycle.

When the car get stucked far from the goal, instead of jumping to the next segment, it forces the gloabal planner to replan.

Substracted a fixed time from the current position time stamp to avoid tf errors (a better solution is needed).

The robot does not get stucked when the goal tolerance distance is not met, it jumps to the next segment.

The robot only starts moving when the error in the steering angle is below a given threshold (improves path tracking).

Solved some bugs of the planner.