OMPL (Open Motion Planning Library)#
OMPL, the Open Motion Planning Library, is a generic C++ library for state-of-the-art motion planning algorithms. OMPL is compatible with different robot morphologies and planning spaces which make it widely used in robotics for motion planning problems both in 2D and 3D spaces. (for more details check the official documentation of OMPL)
OMPL with Kompass#
Kompass provides Python bindings (Pybind11 bindings) for OMPL in its navigation core package (kompass_navigation). The bindings enable setting and solving a planning problem using:
‘SE2StateSpace’ which is convient for 2D motion planning. It provides an ‘SE2State’ consisting of a 2D position and rotation in the plane:
SE(2): (x, y, yaw)OMPL geometric planners (listed below)
Built-in ‘StateValidityChecker’ which implements a collision checker (using FCL) to insure finding collision free paths.
Configuring OMPL in Kompass#
The following is an example YAML configuration of OMPL. In the section below we list all the available configuration parameters for each of the planners along with their default values.
ompl:
log_level: 'WARN'
planning_timeout: 10.0 # (secs) Return and consider the planning to fail if solving takes more than planning_timeout
simplification_timeout: 0.01 # # (secs) abort path simplification if it takes more than simplification_timeout
goal_tolerance: 0.01 # (meters) Distance to consider that the robot is at goal point, i.e. no planning is performed
optimization_objective: 'PathLengthOptimizationObjective'
planner_id: 'ompl.geometric.KPIECE1'
Available algorithms from OMPL#
The following geometric planners are supported in Kompass:
Planners test results#
A planning problem is simulated using the Turtlebot3 Gazebo simulation Waffle map. The following table shows the planning results using each of the previous planners. The problem is simulated for 20 repetitions using each planners and the table below shows the average result values. ‘solved’ is set to False if in any repetition the method was unable to find a solution. The solution search timeout is set to 2seconds.
Tip
To check/run the test script and the associated test resources refer to kompass_navigation ompl test
method |
solved |
solution time (s) |
solution length (m) |
simplification time (s) |
Conversion/Publishing ROS2 (s) |
|---|---|---|---|---|---|
ABITstar |
True |
1.071 |
2.948 |
0.0075 |
0.00056 |
BFMT |
True |
0.1128 |
3.487 |
0.0066 |
0.00017 |
BITstar |
True |
1.0732 |
2.962 |
0.0061 |
0.00015 |
BKPIECE1 |
True |
0.0703 |
4.469 |
0.01777 |
0.00019 |
BiEST |
True |
0.06189 |
4.418 |
0.01078 |
0.00014 |
EST |
True |
0.0640 |
4.059 |
0.01074 |
0.00013 |
FMT |
True |
0.1330 |
3.6277 |
0.00630 |
0.00016 |
InformedRRTstar |
True |
1.0679 |
2.962 |
0.00457 |
0.00013 |
KPIECE1 |
True |
0.0676 |
5.439 |
0.0148 |
0.00017 |
LBKPIECE1 |
True |
0.07543 |
5.174 |
0.02 |
0.00017 |
LBTRRT |
True |
1.0696 |
3.221 |
0.005 |
0.00012 |
LazyLBTRRT |
True |
1.0672 |
3.305 |
0.0053 |
0.00013 |
LazyPRM |
False |
1.0811 |
0.0 |
0.0 |
0.0 |
LazyPRMstar |
True |
1.0701 |
3.03 |
0.0063 |
0.00012 |
LazyRRT |
True |
0.0982 |
4.5196 |
0.016 |
0.00013 |
PDST |
True |
0.06781 |
3.836 |
0.009 |
0.00013 |
PRM |
True |
1.0672 |
3.306 |
0.0068 |
0.00012 |
PRMstar |
True |
1.0740 |
3.72 |
0.0085 |
0.00012 |
ProjEST |
True |
0.0683 |
4.19 |
0.0082 |
0.00013 |
RRT |
True |
0.09062 |
4.86 |
0.019 |
0.00014 |
RRTConnect |
True |
0.07535 |
4.78 |
0.014 |
0.00014 |
RRTXstatic |
True |
1.0712 |
3.03 |
0.0041 |
0.00012 |
RRTsharp |
True |
1.0680 |
3.01 |
0.0052 |
0.00013 |
RRTstar |
True |
1.0672 |
2.96 |
0.0042 |
0.00013 |
SBL |
True |
0.0800 |
4.039 |
0.0121 |
0.00015 |
SST |
True |
1.0681 |
2.63 |
0.0012 |
0.00018 |
STRIDE |
True |
0.0679 |
4.12 |
0.0098 |
0.00015 |
TRRT |
True |
0.0798 |
4.11 |
0.01094 |
0.00015 |
Planners Default Parameters#
ABITstar#
ABITstar Default Parameters:
delay_rewiring_to_first_solution: False
drop_unconnected_samples_on_prune: False
find_approximate_solutions: False
inflation_scaling_parameter: 10.0
initial_inflation_factor: 1000000.0
prune_threshold_as_fractional_cost_change: 0.05
rewire_factor: 1.1
samples_per_batch: 100
stop_on_each_solution_improvement: False
truncation_scaling_parameter: 5.0
use_graph_pruning: True
use_just_in_time_sampling: False
use_k_nearest: True
use_strict_queue_ordering: True
AITstar#
AITstar Default Parameters:
find_approximate_solutions: True
rewire_factor: 1.0
samples_per_batch: 100
use_graph_pruning: True
use_k_nearest: True
BFMT#
BFMT Default Parameters:
balanced: False
cache_cc: True
extended_fmt: True
heuristics: True
nearest_k: True
num_samples: 1000
optimality: True
radius_multiplier: 1.0
BITstar#
BITstar Default Parameters:
delay_rewiring_to_first_solution: False
drop_unconnected_samples_on_prune: False
find_approximate_solutions: False
prune_threshold_as_fractional_cost_change: 0.05
rewire_factor: 1.1
samples_per_batch: 100
stop_on_each_solution_improvement: False
use_graph_pruning: True
use_just_in_time_sampling: False
use_k_nearest: True
use_strict_queue_ordering: True
BKPIECE1#
BKPIECE1 Default Parameters:
border_fraction: 0.9
range: 0.0
BiEST#
BiEST Default Parameters:
range: 0.0
EST#
EST Default Parameters:
goal_bias: 0.5
range: 0.0
FMT#
FMT Default Parameters:
cache_cc: True
extended_fmt: True
heuristics: False
num_samples: 1000
radius_multiplier: 1.1
use_k_nearest: True
InformedRRTstar#
InformedRRTstar Default Parameters:
delay_collision_checking: True
goal_bias: 0.05
number_sampling_attempts: 100
ordered_sampling: False
ordering_batch_size: 1
prune_threshold: 0.05
range: 0.0
rewire_factor: 1.1
use_k_nearest: True
KPIECE1#
KPIECE1 Default Parameters:
border_fraction: 0.9
goal_bias: 0.05
range: 0.0
LBKPIECE1#
LBKPIECE1 Default Parameters:
border_fraction: 0.9
range: 0.0
LBTRRT#
LBTRRT Default Parameters:
epsilon: 0.4
goal_bias: 0.05
range: 0.0
LazyLBTRRT#
LazyLBTRRT Default Parameters:
epsilon: 0.4
goal_bias: 0.05
range: 0.0
LazyPRM#
LazyPRM Default Parameters:
max_nearest_neighbors: 8
range: 0.0
LazyPRMstar#
LazyPRMstar Default Parameters:
LazyRRT#
LazyRRT Default Parameters:
goal_bias: 0.05
range: 0.0
PDST#
PDST Default Parameters:
goal_bias: 0.05
PRM#
PRM Default Parameters:
max_nearest_neighbors: 8
PRMstar#
PRMstar Default Parameters:
ProjEST#
ProjEST Default Parameters:
goal_bias: 0.05
range: 0.0
RRT#
RRT Default Parameters:
goal_bias: 0.05
intermediate_states: False
range: 0.0
RRTConnect#
RRTConnect Default Parameters:
intermediate_states: False
range: 0.0
RRTXstatic#
RRTXstatic Default Parameters:
epsilon: 0.0
goal_bias: 0.05
informed_sampling: False
number_sampling_attempts: 100
range: 0.0
rejection_variant: 0
rejection_variant_alpha: 1.0
rewire_factor: 1.1
sample_rejection: False
update_children: True
use_k_nearest: True
RRTsharp#
RRTsharp Default Parameters:
goal_bias: 0.05
informed_sampling: False
number_sampling_attempts: 100
range: 0.0
rejection_variant: 0
rejection_variant_alpha: 1.0
rewire_factor: 1.1
sample_rejection: False
update_children: True
use_k_nearest: True
RRTstar#
RRTstar Default Parameters:
delay_collision_checking: True
focus_search: False
goal_bias: 0.05
informed_sampling: True
new_state_rejection: False
number_sampling_attempts: 100
ordered_sampling: False
ordering_batch_size: 1
prune_threshold: 0.05
pruned_measure: False
range: 0.0
rewire_factor: 1.1
sample_rejection: False
tree_pruning: False
use_admissible_heuristic: True
use_k_nearest: True
SBL#
SBL Default Parameters:
range: 0.0
SST#
SST Default Parameters:
goal_bias: 0.05
pruning_radius: 3.0
range: 5.0
selection_radius: 5.0
STRIDE#
STRIDE Default Parameters:
degree: 16
estimated_dimension: 3.0
goal_bias: 0.05
max_degree: 18
max_pts_per_leaf: 6
min_degree: 12
min_valid_path_fraction: 0.2
range: 0.0
use_projected_distance: False
TRRT#
TRRT Default Parameters:
goal_bias: 0.05
range: 0.0
temp_change_factor: 0.1