This package is for path smoothing. I'm trying to reproduct the conjugate gradient method and the Gauss Process method. The latter considers vehicle kinematics but is unstable now.
the red curve is the result of CG method and the blue curve is the result of GPMP.
- ceres-solver is required for solving Conjugate Gradient method.
- cppad and casadi are two necessary Automatic Differentiation tools.
- grid_map is necessary for obstacle consideration.
- gtsam4.0, gpmp2 are optional. rosparam_handler, dynamic_reconfigure are used for demo.
- test path smoothing without obstacle
rosrun path smoothing without obstacle
cd <path_smoothing-package-dir>/demo
python simple_curve_plot.py
- test path smoothing with obstacle
roslaunch path_smoothing smooth_with_obstacle_demo.launch