Link Search Menu Expand Document
Bench-MR

Post Smoothers

Given a solution from a motion planner, post smoothing, or path improvement algorithms attempt to find a better (e.g., shorter, smoother) path.

A list of smoothers can be provided to the MPB instance which will run a benchmark replicating all the settings across each selected smoother. The user-friendly function is the following that will enable the respective smoothers in under the benchmark.planning group:

MPB.set_smoothers(smoothers: [str])

The smoother names are parsed from the list of strings, showing an error if any entry could not be unified with the available smoothers in Bench-MR. Check the smoother_names dictionary in definitions.py for the mapping of smoother names to be given to the set_smoothers function, and their respective printable titles:

Smoother Name Smoother Title
grips GRIPS
ompl_bspline B-Spline
ompl_shortcut Shortcut
ompl_simplify_max SimplifyMax

Example 

mpb = MPB()
mpb.set_corridor_grid_env(radius = 3)
mpb.set_planners(['rrt'])
mpb.set_steer_functions(['reeds_shepp'])
mpb.set_smoothers(['grips',
                   'ompl_bspline',
                   'ompl_shortcut',
                   'ompl_simplify_max'])
mpb.run(runs=5);

Visualize trajectories (original solution and smoothed paths):

mpb.visualize_trajectories(show_smoother=True);

png

Plot statistics for the 5 runs comparing the original plan against the post-smoothing solutions:

mpb.plot_smoother_stats();

png

View Jupyter Notebook

Table of contents