Re: [ompl-users] Strange results from RigidBodyPlanningWithODESolverAndControls

[Mark M. <mm...@ri...>]

Hi Andrey,

You may need to explicitly set the propagation step size and maybe the propagation duration as well. The paths well likely never look very optimal, since by default OMPL is simply sampling random controls. You can try the SST planner to get asymptotic optimality. You can also develop your own control sampler that will allow a planner to discover more reasonable trajectories.

Mark



> On Apr 15, 2017, at 12:55 PM, Andrey Nechypurenko <and...@go...> wrote:
> 
>> The second question is regarding the bounds of the state space. My map
>> has dimensions about 1000x1000. So I thought that changing the X and Y
> 
> After playing with bounds and car length, I was able to find the
> configuration where OMPL can find solution. Here is the output of
> printSettings() call:
> 
> OMPL version: 1.3.0
> Settings for the state space 'SE2CompoundSpace0'
>  - state validity check resolution: 3%
>  - valid segment count factor: 1
>  - state space:
> Compound state space 'SE2CompoundSpace0' of dimension 3 (locked) [
> Real vector state space 'RealVectorSpace1' of dimension 2 with bounds:
>  - min: 0 0
>  - max: 1000 1000
> of weight 1
> SO2 state space 'SO2Space2'
> of weight 0.5
> ]
> No registered projections
> 
> Declared parameters:
> longest_valid_segment_fraction = 0.029999999999999999
> valid_segment_count_factor = 1
> Valid state sampler named uniform with parameters:
> nr_attempts = 100
>  - control space:
> Real vector control space 'RealVectorControl[SE2CompoundSpace0]' with bounds:
>  - min: 0 -0.3
>  - max: 50 0.3
>  - can propagate backward: yes
>  - propagation step size: 0
>  - propagation duration: [0, 0]
> Warning: Assuming propagation will always have between 1 and 10 steps
>         at line 56 in
> /home/andrey/projects/piwars/ompl/src/ompl/control/src/SpaceInformation.cpp
> Warning: The propagation step size is assumed to be 42.473531
>         at line 66 in
> /home/andrey/projects/piwars/ompl/src/ompl/control/src/SpaceInformation.cpp
> Info:    No planner specified. Using default.
> Info:    KPIECE1: Attempting to use default projection.
> Info:    KPIECE1: Starting planning with 1 states already in datastructure
> Info:    ProblemDefinition: Adding approximate solution from planner KPIECE1
> Info:    KPIECE1: Created 563 states in 165 cells (55 internal + 110 external)
> Info:    Solution found in 5.002863 seconds
> 
> However, similar to the original demo, the path looks quite weird:
> https://goo.gl/KRfTm1
> 
> Obviously I am doing something wrong and would appreciate if
> experienced OMPL users can provide some hints how to improve my
> results.
> 
> Regards,
> Andrey.
> 
> ------------------------------------------------------------------------------
> Check out the vibrant tech community on one of the world's most
> engaging tech sites, Slashdot.org! http://sdm.link/slashdot
> _______________________________________________
> ompl-users mailing list
> omp...@li...
> https://lists.sourceforge.net/lists/listinfo/ompl-users
>