ompl-users — Users mailing list

[ompl-users] Poor results with basic planning for autonomous car-like robot

[Henry T. <he...@ot...>]

Hi,

I'm getting very poor results while attempting to use OMPL for a basic
autonomous robot with car-like steering. I've installed and used the OMPL
app gui and set up a trivial scenario. Screenshots are linked to below. Is
there anything I can do to improve the quality of the plans? I was
expecting a plan with 2 or 3 turns but instead the robot seems to wander
around aimlessly until it hits the goal. I've also tried using all the
different planners available in the drop-down menu in the GUI with similar
results.

Screenshots here - https://imgur.com/a/KqglH

Any suggestions will be greatly appreciated.

Thanks!

[ompl-users] OMPL 1.3.2 released

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

Hi all,

We are are happy to announce version 1.3.2 of the Open Motion Planning Library (OMPL).


OMPL 1.3.2 contains the following main changes:

• Bug fixes for gcc6, clang5, and new versions of Boost.
• Made PlannerArena work with latest versions of R packages.
• Made MORSE plugin work with latest version MORSE.
• Optionally, use PyPy for Python binding generation. This can significantly speed up the binding generation process (with pypy3 5.9.0 it now takes less than 10 minutes on a modern PC).
• Added support for development version of FCL.
• Misc. bug fixes.

The OMPL web page can be found at http://ompl.kavrakilab.org.

Regards,

Mark Moll

--
Mark Moll
Sr. Research Scientist   |   http://mmoll.rice.edu   |   713-348-5934
Rice University, Dept. of Computer Science — MS 132, PO Box 1892, Houston, TX 77251

Re: [ompl-users] PRM and RRT sample bounds

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

You can create a Goal-derived class where the isSatisfied method always returns false. Use this in combination with the ompl::base::IterationTerminationCondition class to call solve. Next, set the goal you want and call solve again with ompl::base::plannerAlwaysTerminatingCondition(), which makes the planner terminate immediately. The internal data is not cleared between repeated calls to solve, so I think this might work.

Mark



> On Oct 5, 2017, at 11:55 AM, James Lee <JuH...@st...> wrote:
> 
> Thanks for that Mark. Its not exactly what I needed, however.
> 
> What I want to do is find a solution using a roadmap or branch that contains 1000 sample states.
> But when I use the iteration termination condition to and set the value to, say 1000, the algorithm seems to only generate ~480 states.
> I also want to perform benchmark test using set number of sample states, so ideally I don't want to pass the object through Planner::solve().
> 
> Is there another way? or do I need to hard code this?
> 
> James
> 
> From: Mark Moll <mar...@gm... <mailto:mar...@gm...>> on behalf of Mark Moll <mm...@ri... <mailto:mm...@ri...>>
> Sent: Thursday, 5 October 2017 7:39:41 PM
> To: James Lee
> Cc: omp...@li... <mailto:omp...@li...>; Chanyeol Yoo; Robert Fitch
> Subject: Re: [ompl-users] PRM and RRT sample bounds
> 
> Hi James,
> 
> Perhaps this is what you want:
> ompl::base::IterationTerminationCondition
> http://ompl.kavrakilab.org/classompl_1_1base_1_1IterationTerminationCondition.html <http://ompl.kavrakilab.org/classompl_1_1base_1_1IterationTerminationCondition.html>
> ompl::base::IterationTerminationCondition Class Reference <http://ompl.kavrakilab.org/classompl_1_1base_1_1IterationTerminationCondition.html>
> ompl.kavrakilab.org <http://ompl.kavrakilab.org/>
> Public Member Functions IterationTerminationCondition (unsigned int numIterations) Construct a termination condition that can be evaluated numIterations times before ...
> 
> You can pass an object of this type to Planner::solve().
> 
> Mark
> 
> 
> 
>> On Oct 5, 2017, at 5:39 AM, James Lee <JuH...@st... <mailto:JuH...@st...>> wrote:
>> 
>> Hi everyone,
>> 
>> I am trying to implement PRM and RRT (and their respective variants) on OMPL.
>> 
>> Is there a way to add a sample bound to these algorithms? For example, the roadmap or branch is grown until the 1000th sample has been generated, and uses said roadmap or branches to output a trajectory/path.
>> 
>> Regards,
>> 
>> James
> 
> 

Re: [ompl-users] PRM and RRT sample bounds

[James L. <JuH...@st...>]

Thanks for that Mark. Its not exactly what I needed, however.

What I want to do is find a solution using a roadmap or branch that contains 1000 sample states.
But when I use the iteration termination condition to and set the value to, say 1000, the algorithm seems to only generate ~480 states.
I also want to perform benchmark test using set number of sample states, so ideally I don't want to pass the object through Planner::solve().

Is there another way? or do I need to hard code this?

James


________________________________
From: Mark Moll <mar...@gm...> on behalf of Mark Moll <mm...@ri...>
Sent: Thursday, 5 October 2017 7:39:41 PM
To: James Lee
Cc: omp...@li...; Chanyeol Yoo; Robert Fitch
Subject: Re: [ompl-users] PRM and RRT sample bounds

Hi James,

Perhaps this is what you want:
ompl::base::IterationTerminationCondition
http://ompl.kavrakilab.org/classompl_1_1base_1_1IterationTerminationCondition.html
ompl::base::IterationTerminationCondition Class Reference<http://ompl.kavrakilab.org/classompl_1_1base_1_1IterationTerminationCondition.html>
ompl.kavrakilab.org
Public Member Functions IterationTerminationCondition (unsigned int numIterations) Construct a termination condition that can be evaluated numIterations times before ...


You can pass an object of this type to Planner::solve().

Mark



On Oct 5, 2017, at 5:39 AM, James Lee <JuH...@st...<mailto:JuH...@st...>> wrote:

Hi everyone,

I am trying to implement PRM and RRT (and their respective variants) on OMPL.

Is there a way to add a sample bound to these algorithms? For example, the roadmap or branch is grown until the 1000th sample has been generated, and uses said roadmap or branches to output a trajectory/path.

Regards,

James

Re: [ompl-users] PRM and RRT sample bounds

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

Hi James,

Perhaps this is what you want:
ompl::base::IterationTerminationCondition
http://ompl.kavrakilab.org/classompl_1_1base_1_1IterationTerminationCondition.html <http://ompl.kavrakilab.org/classompl_1_1base_1_1IterationTerminationCondition.html>

You can pass an object of this type to Planner::solve().

Mark



> On Oct 5, 2017, at 5:39 AM, James Lee <JuH...@st...> wrote:
> 
> Hi everyone,
> 
> I am trying to implement PRM and RRT (and their respective variants) on OMPL.
> 
> Is there a way to add a sample bound to these algorithms? For example, the roadmap or branch is grown until the 1000th sample has been generated, and uses said roadmap or branches to output a trajectory/path.
> 
> Regards,
> 
> James

[ompl-users] PRM and RRT sample bounds

[James L. <JuH...@st...>]

Hi everyone,

I am trying to implement PRM and RRT (and their respective variants) on OMPL.

Is there a way to add a sample bound to these algorithms? For example, the roadmap or branch is grown until the 1000th sample has been generated, and uses said roadmap or branches to output a trajectory/path.

Regards,

James

[ompl-users] Motion planning of quadruped robot

[Jan K. <squ...@gm...>]

Hi,

Unfortunately, due to my failure, I missed OMPL in research for my
university thesis.

I am trying to write an ROS software for quadruped robot (spider-style
mechanics).  After some basic ideas, I picked sample-based motion
algorithm. I am able to reasonably define the state of the robot, evaluate
simple linear changes of that state and make an efficient search using
Anytime D* Lite. (In theory)

After some implementation in progress. I rechecked actual state of the art
and realized that this is how OMPL works. Which for some reason I missed
previously. (And I consider myself dumb for that)

So, I am in a state when I have mechanics for moving the legs of the robot
based on state search.

Are you aware of anybody trying to do the same? This is not the standard
usage of OMPL and I prefer to ask here, rather than risk that my "search
skill" is going to fail again.

Sincerely,
John Koniarik
squ...@gm...

[ompl-users] Plannerarena error - NAs are not allowed in subscripted assignments

[James L. <JuH...@st...>]

Hi everyone,


I am having difficulties loading the database into the plannerarena.


After I upload the database, no figure appears, and I am met with the following error:
Error: NAs are not allowed in subscripted assignments


Attached is an image of plannerarena with said error, along with the output in the terminal.

I don't know where to go from here. I searched online to see if others had the same problem, but the only relevent information I could find was for the R-code itself.
Could someone point me in the right direction?

Thank you in advance.

Regards,

James Lee

[ompl-users] Starting with Ompl. Using OMPL to park a car robot.

[vanessa f. m. <van...@ms...>]

Hello!


I've done a lot of testing and I do not get a path more or less as I want (a kind of realistic one). I have done several tests with the GUI, you have some of them here: https://www.youtube.com/watch?v=xse7LClDyy8<https://www.youtube.com/watch?v=heUp3urw4a8>

[https://i.ytimg.com/vi/xse7LClDyy8/maxresdefault.jpg]<https://www.youtube.com/watch?v=xse7LClDyy8>

Ompl: Ejemplos de planificación con diferentes tipos de coche<https://www.youtube.com/watch?v=xse7LClDyy8>
www.youtube.com
0:00 Rigid body planning 3D 0:10 Rigid body planning 2D 0:20 Dynamic car 0:47 Kinematic car


My robot is not holonomic, a car like robot. Is the type of robot I am using correct? And the planner? Is there any ideal configuration for my problem?


I've already posted an issue explaining my problem here: https://bitbucket.org/ompl/ompl/issues/347/starting-with-ompl
where you can find some attachments to reproduce my problem if you wish.


Thanks a lot!

Re: [ompl-users] Distance metric and interpolation for RRTs

[Anand <ana...@gm...>]

Thanks Mark, sounds like it's not a very good idea.

Anand

On Wed, Jun 28, 2017 at 9:25 PM, Mark Moll <mm...@ri...> wrote:

> Hi Anand,
>
> You could derive a new class from DubinsStateSpace where you override
> distance. This might lead to some surprising results. The paths will
> probably have many twists and turns that path simplification cannot
> eliminate.
>
> As far as I know there really isn’t an efficient way to search nearest
> neighbors in non-metric spaces.
>
> Best,
>
> Mark
>
>
>
> On Jun 28, 2017, at 5:54 PM, Anand <ana...@gm...> wrote:
>
> In OMPL, the distance metric used (distance()) and the way in which states
> are connected (interpolate()) are encapsulated as part of the state space.
>
> For example, the DubinsStateSpace uses the path length of the Dubins curve
> to compute the distance() and interpolate().
>
> Does it make sense for an RRT to use a different distance metric for
> nearest neighbor computations, than what is used for interpolation?
>
> In my scenario (planning for a car-like robot), I wonder if it makes sense
> to speed up nearest neighbor searches by using a distance metric other than
> Dubins (say Euclidean distance) to find nearest neighbors, but still use
> Dubins to connect states? This will allow me to use nearest neighbor data
> structures like GNAT, which I can't otherwise do if I use the Dubins
> distance metric.
>
> If not, are there any alternative data structures I can use to speed up
> nearest neighbor searches for a Dubins-like non-metric space?
>
> Thanks,
> Anand
>
>
>

Re: [ompl-users] Distance metric and interpolation for RRTs

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

Hi Anand,

You could derive a new class from DubinsStateSpace where you override distance. This might lead to some surprising results. The paths will probably have many twists and turns that path simplification cannot eliminate.

As far as I know there really isn’t an efficient way to search nearest neighbors in non-metric spaces.

Best,

Mark



> On Jun 28, 2017, at 5:54 PM, Anand <ana...@gm...> wrote:
> 
> In OMPL, the distance metric used (distance()) and the way in which states are connected (interpolate()) are encapsulated as part of the state space.
> 
> For example, the DubinsStateSpace uses the path length of the Dubins curve to compute the distance() and interpolate().
> 
> Does it make sense for an RRT to use a different distance metric for nearest neighbor computations, than what is used for interpolation?
> 
> In my scenario (planning for a car-like robot), I wonder if it makes sense to speed up nearest neighbor searches by using a distance metric other than Dubins (say Euclidean distance) to find nearest neighbors, but still use Dubins to connect states? This will allow me to use nearest neighbor data structures like GNAT, which I can't otherwise do if I use the Dubins distance metric.
> 
> If not, are there any alternative data structures I can use to speed up nearest neighbor searches for a Dubins-like non-metric space?
> 
> Thanks,
> Anand

[ompl-users] Distance metric and interpolation for RRTs

[Anand <ana...@gm...>]

In OMPL, the distance metric used (distance()) and the way in which states
are connected (interpolate()) are encapsulated as part of the state space.

For example, the DubinsStateSpace uses the path length of the Dubins curve
to compute the distance() and interpolate().

Does it make sense for an RRT to use a different distance metric for
nearest neighbor computations, than what is used for interpolation?

In my scenario (planning for a car-like robot), I wonder if it makes sense
to speed up nearest neighbor searches by using a distance metric other than
Dubins (say Euclidean distance) to find nearest neighbors, but still use
Dubins to connect states? This will allow me to use nearest neighbor data
structures like GNAT, which I can't otherwise do if I use the Dubins
distance metric.

If not, are there any alternative data structures I can use to speed up
nearest neighbor searches for a Dubins-like non-metric space?

Thanks,
Anand

Re: [ompl-users] Path following while avoiding obstacles.

[Guilherme A. S. P. <gpe...@uf...>]

Hi Adrien,

I think that what you want is the 
VFUpstreamCriterionOptimizationObjective, which is a generalization of 
your function. In fact your function should define the cost to be 
proportional to inner product between the vector defined by the line and 
the derivative of the path:

http://ompl.kavrakilab.org/VFUpstreamCriterionOptimizationObjective_8h_source.html

The VFUpstreamCriterionOptimizationObjective was defined in the context 
of VFRRT. See the paper here:

http://journals.sagepub.com/doi/abs/10.1177/0278364914545812

An example of how to use it with TRRT and RRT* is here:

https://github.com/ompl/ompl/blob/master/demos/VFRRT/VectorFieldConservative.cpp

RRT* is supposed to give you the best results, but you will going to 
spend more time. If time is not a hard constraint, I would go with it.

I hope this can help.

Guilherme



Em 05/06/2017 06:08, Adrien BARRAL escreveu:
> Hello,
>
> I am trying to use OMPL to perform path following while avoiding 
> obstacles.
>
> I guessed that using TRRT algorithm with an optimization function 
> which prune solution far of the path I am following would do the job.
>
> So I create the following optimization function :
>
> classFollowLineObjective:publicob::StateCostIntegralObjective
> {
> public:
> FollowLineObjective(constob::SpaceInformationPtr&si):
> ob::StateCostIntegralObjective(si,true)
> {
> }
> ob::CoststateCost(constob::State*s)const
> {
> Eigen::VectorXdpt=vectorField(s);
> returnob::Cost(hypot(pt[0],pt[1]));
> }
> };
> Where vectorField return the vector from the state to the line I am following.
> I give this optimization objective to my simple setup, and solutions are found.
> The solution is shown here :https://ibb.co/cL5QYF
> As we can see the obstacle (in black) is avoided, and the path I am trying to follow (in red) is followed by the solutuion (in blue).
> But the solution is not realy optimal because when there is no obstacle, the solution is not a "straight line".
> I tryied to "simplify" the solution. But when I simplify it, the optimization seems to be ignored.
> So how can I make the solution better "follow" my red line ?
> Regards,
> Adrien BARRAL.
>
>
> ------------------------------------------------------------------------------
> 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

-- 

Prof. Guilherme A. S. Pereira
Departamento de Engenharia Elétrica
Universidade Federal de Minas Gerais
Tel: +55 31 3409-6687
http://www.ppgee.ufmg.br/~gpereira

[ompl-users] Path following while avoiding obstacles.

[Adrien B. <ab...@ro...>]

Hello,

I am trying to use OMPL to perform path following while avoiding obstacles.

I guessed that using TRRT algorithm with an optimization function which
prune solution far of the path I am following would do the job.

So I create the following optimization function :

class FollowLineObjective : public ob::StateCostIntegralObjective

{

public:

    FollowLineObjective(const ob::SpaceInformationPtr& si) :

        ob::StateCostIntegralObjective(si, true)

    {

    }

    ob::Cost stateCost(const ob::State* s) const

    {

        Eigen::VectorXd pt = vectorField(s);

        return ob::Cost(hypot(pt[0], pt[1]));

    }

};


Where vectorField return the vector from the state to the line I am following.


I give this optimization objective to my simple setup, and solutions are found.


The solution is shown here : https://ibb.co/cL5QYF


As we can see the obstacle (in black) is avoided, and the path I am
trying to follow (in red) is followed by the solutuion (in blue).


But the solution is not realy optimal because when there is no
obstacle, the solution is not a "straight line".


I tryied to "simplify" the solution. But when I simplify it, the
optimization seems to be ignored.


So how can I make the solution better "follow" my red line ?


Regards,


Adrien BARRAL.

Re: [ompl-users] Integrating geometric rrt with another planner

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

Hi Hemang,

> On May 26, 2017, at 12:42 PM, Hemang Chawla <hem...@gm...> wrote:
> I am trying to plan for a non-holonomic wheeled mobile manipulator.
> 
> For planning the path of the base, the control based planners ma be used. However they cannot be extended directly to mobile manipulators due to higher dimensionality.

It *can*, but it will likely be very slow. If the type of problems you want to solve can be solved quasi statically, then the approach you suggest is likely the better way to go.

> 
> On the contrary, the planning with reed shepps planner uses geometric planning. I wish to plan for the base and the arm simultaneously. Is there a way to plan for a robot with base in SE(2) planned with the rest of the joints in R^6 ? Perhaps a new state space will have to be created for the same?

You can create a CompoundStateSpace <http://ompl.kavrakilab.org/classompl_1_1base_1_1CompoundStateSpace.html> with DubinsStateSpace and R^6 as the subspaces. See the tutorial at http://ompl.kavrakilab.org/workingWithStates.html <http://ompl.kavrakilab.org/workingWithStates.html>.


> I recognize the drawback here is the large amount of time it may take as the planning with reed shepps is not optimized. However I wish to plan only in a small area of a couple meters^2.

I’m not sure what you mean with "planning with reed shepps is not optimized.” You should be able to use RRT* or other optimizing planners with DubinsStateSpace (or a CompoundStateSpace with a DubinsStateSpace as a subspace).

Mark

Re: [ompl-users] About the OMPL Web App

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

Hi Bivan,

I can confirm the problem. It’s not just the server; the same problem happens when I try to run the OMPL web app locally. We will investigate.

Thanks,

Mark



> On May 28, 2017, at 8:23 AM, Bivan Alzacky Harmanto <biv...@ka...> wrote:
> 
> Dear Sir/Madam
> Currently I am experimenting on using the OMPL web app in https://omplapp.kavrakilab.org
> 
> However, whenever I tried to use this, it only appears as "Please Wait" and even after I waited for 30 mins, it is still running
> Even for the LazyRRT for the Apartment problem, it is kind of freezing
> 
> I am wondering on what problem is going on in this web app
> I am looking forward for the reply
> Many thanks
> Best regards

[ompl-users] About the OMPL Web App

[Bivan A. H. <biv...@ka...>]

Dear Sir/MadamCurrently I am experimenting on using the OMPL web app in https://omplapp.kavrakilab.orgHowever, whenever I tried to use this, it only appears as "Please Wait" and even after I waited for 30 mins, it is still runningEven for the LazyRRT for the Apartment problem, it is kind of freezingI am wondering on what problem is going on in this web appI am looking forward for the replyMany thanksBest regardsBivan

[ompl-users] Integrating geometric rrt with another planner

[Hemang C. <hem...@gm...>]

Hi,

I am trying to plan for a non-holonomic wheeled mobile manipulator.

For planning the path of the base, the control based planners ma be used.
However they cannot be extended directly to mobile manipulators due to
higher dimensionality.

On the contrary, the planning with reed shepps planner uses geometric
planning. I wish to plan for the base and the arm simultaneously. Is there
a way to plan for a robot with base in SE(2) planned with the rest of the
joints in R^6 ? Perhaps a new state space will have to be created for the
same?

I recognize the drawback here is the large amount of time it may take as
the planning with reed shepps is not optimized. However I wish to plan only
in a small area of a couple meters^2.
-- 
Regards,
Hemang Chawla

Re: [ompl-users] Creating your own optimization objectives

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

Hi Hemang,

Yes, just derive a new class from OptimizationObjective. See this tutorial:
http://ompl.kavrakilab.org/optimizationObjectivesTutorial.html <http://ompl.kavrakilab.org/optimizationObjectivesTutorial.html>

Mark



> On May 17, 2017, at 11:00 AM, Hemang Chawla <hem...@gm...> wrote:
> 
> Hello,
> 
> Some planners allow to use optimization objectives while planning a path. Is there a way to develop your own optimization objective to be included in multi-objective optimization?
> 
> 
> --
> Regards,
> Hemang Chawla

[ompl-users] Creating your own optimization objectives

[Hemang C. <hem...@gm...>]

Hello,

Some planners allow to use optimization objectives while planning a path.
Is there a way to develop your own optimization objective to be included in
multi-objective optimization?


-- 
Regards,
Hemang Chawla

Re: [ompl-users] VFRRT not working

[Guilherme A. S. P. <gpe...@uf...>]

Ok Mark,

I will do that.

Thanks,

Guilherme


Em 12/05/2017 12:11, Mark Moll escreveu:
> Hi Guilherme,
>
> I can’t reproduce this with the latest version of OMPL. The 
> VFRRT-related code hasn’t changed much since 1.2.1 (besides small 
> changes to take advantage of C++11 features). Can you try again with 
> the latest release? If you have the same problem with the latest 
> release, please file a bug report at 
> https://bitbucket.org/ompl/ompl/issues?status=new&status=open. Include 
> OS version, Boost version, compiler version, etc.
>
> Mark
>
>
>
>> On May 12, 2017, at 6:49 AM, Guilherme Pereira <gpe...@uf... 
>> <mailto:gpe...@uf...>> wrote:
>>
>> Hello everyone,
>>
>> I'm trying to use VFRRT but it seems to be not working. It never finds a
>> path. In fact, it does not add any states to the tree.
>>
>> When I run the example in VectorFieldConservative.cpp I get:
>>
>> Info:    VFRRT: Starting planning with 1 states already in datastructure
>> Info:    VFRRT: Created 1 states
>> Info:    No solution found after 10.102015 seconds
>> No solution found.
>> Debug:   TRRT: Planner range detected to be 5.656854
>> Debug:   TRRT: Frontier threshold detected to be 0.282843
>> Info:    TRRT: Starting planning with 1 states already in datastructure
>> Info:    TRRT: Created 66 states
>> Info:    Solution found in 0.001091 seconds
>> Found solution.
>> Total upstream cost: 11.6123
>> Debug:   RRTstar: Planner range detected to be 5.656854
>> Info:    RRTstar: Starting planning with 1 states already in 
>> datastructure
>> Info:    RRTstar: Initial k-nearest value of 4
>> Info:    RRTstar: Found an initial solution with a cost of 7.97 in 21
>> iterations (22 vertices in the graph)
>> Info:    RRTstar: Created 27293 new states. Checked 1141600 rewire
>> options. 4 goal states in tree. Final solution cost 6.766
>> Info:    Solution found in 10.078853 seconds
>> Found solution.
>> Total upstream cost: 6.87743
>>
>> To the VectorFieldNonconservative.cpp example I have this:
>>
>> Info:    VFRRT: Starting planning with 1 states already in datastructure
>> Info:    VFRRT: Created 1 states
>> Info:    No solution found after 10.099492 seconds
>> No solution found.
>>
>> I tried to make some changes without success. Can anyone help me? I'm
>> using OMPL 1.2.1.
>>
>> Thanks,
>> Guilherme
>>
>>
>>
>> ------------------------------------------------------------------------------
>> Check out the vibrant tech community on one of the world's most
>> engaging tech sites, Slashdot.org <http://Slashdot.org>! 
>> http://sdm.link/slashdot
>> _______________________________________________
>> ompl-users mailing list
>> omp...@li... 
>> <mailto:omp...@li...>
>> https://lists.sourceforge.net/lists/listinfo/ompl-users
>>
>

-- 

Prof. Guilherme A. S. Pereira
Departamento de Engenharia Elétrica
Universidade Federal de Minas Gerais
Tel: +55 31 3409-6687
http://www.ppgee.ufmg.br/~gpereira

Re: [ompl-users] VFRRT not working

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

Hi Guilherme,

I can’t reproduce this with the latest version of OMPL. The VFRRT-related code hasn’t changed much since 1.2.1 (besides small changes to take advantage of C++11 features). Can you try again with the latest release? If you have the same problem with the latest release, please file a bug report at https://bitbucket.org/ompl/ompl/issues?status=new&status=open <https://bitbucket.org/ompl/ompl/issues?status=new&status=open>. Include OS version, Boost version, compiler version, etc.

Mark



> On May 12, 2017, at 6:49 AM, Guilherme Pereira <gpe...@uf...> wrote:
> 
> Hello everyone,
> 
> I'm trying to use VFRRT but it seems to be not working. It never finds a
> path. In fact, it does not add any states to the tree.
> 
> When I run the example in VectorFieldConservative.cpp I get:
> 
> Info:    VFRRT: Starting planning with 1 states already in datastructure
> Info:    VFRRT: Created 1 states
> Info:    No solution found after 10.102015 seconds
> No solution found.
> Debug:   TRRT: Planner range detected to be 5.656854
> Debug:   TRRT: Frontier threshold detected to be 0.282843
> Info:    TRRT: Starting planning with 1 states already in datastructure
> Info:    TRRT: Created 66 states
> Info:    Solution found in 0.001091 seconds
> Found solution.
> Total upstream cost: 11.6123
> Debug:   RRTstar: Planner range detected to be 5.656854
> Info:    RRTstar: Starting planning with 1 states already in datastructure
> Info:    RRTstar: Initial k-nearest value of 4
> Info:    RRTstar: Found an initial solution with a cost of 7.97 in 21
> iterations (22 vertices in the graph)
> Info:    RRTstar: Created 27293 new states. Checked 1141600 rewire
> options. 4 goal states in tree. Final solution cost 6.766
> Info:    Solution found in 10.078853 seconds
> Found solution.
> Total upstream cost: 6.87743
> 
> To the VectorFieldNonconservative.cpp example I have this:
> 
> Info:    VFRRT: Starting planning with 1 states already in datastructure
> Info:    VFRRT: Created 1 states
> Info:    No solution found after 10.099492 seconds
> No solution found.
> 
> I tried to make some changes without success. Can anyone help me? I'm
> using OMPL 1.2.1.
> 
> Thanks,
> Guilherme
> 
> 
> 
> ------------------------------------------------------------------------------
> 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
> 

[ompl-users] VFRRT not working

[Guilherme P. <gpe...@uf...>]

Hello everyone,

I'm trying to use VFRRT but it seems to be not working. It never finds a 
path. In fact, it does not add any states to the tree.

When I run the example in VectorFieldConservative.cpp I get:

Info:    VFRRT: Starting planning with 1 states already in datastructure
Info:    VFRRT: Created 1 states
Info:    No solution found after 10.102015 seconds
No solution found.
Debug:   TRRT: Planner range detected to be 5.656854
Debug:   TRRT: Frontier threshold detected to be 0.282843
Info:    TRRT: Starting planning with 1 states already in datastructure
Info:    TRRT: Created 66 states
Info:    Solution found in 0.001091 seconds
Found solution.
Total upstream cost: 11.6123
Debug:   RRTstar: Planner range detected to be 5.656854
Info:    RRTstar: Starting planning with 1 states already in datastructure
Info:    RRTstar: Initial k-nearest value of 4
Info:    RRTstar: Found an initial solution with a cost of 7.97 in 21 
iterations (22 vertices in the graph)
Info:    RRTstar: Created 27293 new states. Checked 1141600 rewire 
options. 4 goal states in tree. Final solution cost 6.766
Info:    Solution found in 10.078853 seconds
Found solution.
Total upstream cost: 6.87743

To the VectorFieldNonconservative.cpp example I have this:

Info:    VFRRT: Starting planning with 1 states already in datastructure
Info:    VFRRT: Created 1 states
Info:    No solution found after 10.099492 seconds
No solution found.

I tried to make some changes without success. Can anyone help me? I'm 
using OMPL 1.2.1.

Thanks,
Guilherme

[ompl-users] Genetic Search for optimal base pose for mobile manipulation

[Hemang C. <hem...@gm...>]

Hello,

I am trying to use the genetic search in OMPL to determine the optimal base
position of a mobile manipulator to reach a EE pose. The OMPL genetic
search is generally used to plan to a valid IK solution near the goal pose.
However, I wish to use the same to solve this optimization problem.

My initial idea is to modify the distanceGoal heuristic to compute the cost
function. And include the validity check within the same.

Could you please give any idea on how to proceed with this using the OMPL
Geneitc search or is that not a good way to approach this?

-- 
Regards,
Hemang Chawla

[ompl-users] Genetic Search for valid base pose

[Hemang C. <hem...@gm...>]

Hello,

I am trying to use the genetic search in OMPL to determine the optimal base
position of a mobile manipulator to reach a pose. The OMPL genetic search
is generally used to plan to a valid IK solution near the goal pose.
However, I wish to use the same to solve my problem.

My initial idea is to modify the distanceGoal heuristic to compute the cost
function. And include the validity check within the same.

Could you please give any idea on how to proceed with this using the OMPL
Geneitc search or is that not a good way to approach this?

-- 
Regards,
Hemang Chawla