From: Luc B. <lb...@pr...> - 2010-08-26 03:21:20
|
Hi everyone, I am using a differential drive robot in Gazebo (svn version), and I am noticing a lot of slip in the wheels that is giving me large errors, particularly in the heading when I am spinning. It is to the point where I visibly watch the robot spin twice before the yaw values indicate that it has spun once. Can someone point me to some setting to play with so that I can tune this? I have tried setting the <mul> of the wheels to infinity without any noticeable change in behaviour. Thanks alot! -- Luc Brunet Co-Founder, Director - Robotic Systems Provectus Robotics Solutions www.provectus-robotics.com |
From: John H. <jo...@wi...> - 2010-08-26 05:38:13
|
Hi Luc, Do you mean <mu1> (and <mu2>)? Also, are you using one of the pioneer models that comes with the svn checkout? John On Wed, Aug 25, 2010 at 8:21 PM, Luc Brunet <lb...@pr...>wrote: > Hi everyone, > > I am using a differential drive robot in Gazebo (svn version), and I am > noticing a lot of slip in the wheels that is giving me large errors, > particularly in the heading when I am spinning. It is to the point > where I visibly watch the robot spin twice before the yaw values > indicate that it has spun once. Can someone point me to some setting to > play with so that I can tune this? I have tried setting the <mul> of > the wheels to infinity without any noticeable change in behaviour. > > Thanks alot! > > -- > Luc Brunet > Co-Founder, Director - Robotic Systems > Provectus Robotics Solutions > www.provectus-robotics.com > > > ------------------------------------------------------------------------------ > Sell apps to millions through the Intel(R) Atom(Tm) Developer Program > Be part of this innovative community and reach millions of netbook users > worldwide. Take advantage of special opportunities to increase revenue and > speed time-to-market. Join now, and jumpstart your future. > http://p.sf.net/sfu/intel-atom-d2d > _______________________________________________ > Playerstage-gazebo mailing list > Pla...@li... > https://lists.sourceforge.net/lists/listinfo/playerstage-gazebo > |
From: Luc B. <lb...@pr...> - 2010-08-26 05:42:04
|
Hi John, thanks for the reply. Yes, I meant <mu1>, although I was unaware of <mu2>, is there a difference? I was originally using my own model, but then verified the same behaviour in the pioneer2dx model that comes as an example. Luc On 08/26/2010 01:37 AM, John Hsu wrote: > Hi Luc, > Do you mean<mu1> (and<mu2>)? Also, are you using one of the pioneer > models that comes with the svn checkout? > John > > > On Wed, Aug 25, 2010 at 8:21 PM, Luc Brunet > <lb...@pr...>wrote: > >> Hi everyone, >> >> I am using a differential drive robot in Gazebo (svn version), and I am >> noticing a lot of slip in the wheels that is giving me large errors, >> particularly in the heading when I am spinning. It is to the point >> where I visibly watch the robot spin twice before the yaw values >> indicate that it has spun once. Can someone point me to some setting to >> play with so that I can tune this? I have tried setting the<mul> of >> the wheels to infinity without any noticeable change in behaviour. >> >> Thanks alot! >> >> -- >> Luc Brunet >> Co-Founder, Director - Robotic Systems >> Provectus Robotics Solutions >> www.provectus-robotics.com >> >> >> ------------------------------------------------------------------------------ >> Sell apps to millions through the Intel(R) Atom(Tm) Developer Program >> Be part of this innovative community and reach millions of netbook users >> worldwide. Take advantage of special opportunities to increase revenue and >> speed time-to-market. Join now, and jumpstart your future. >> http://p.sf.net/sfu/intel-atom-d2d >> _______________________________________________ >> Playerstage-gazebo mailing list >> Pla...@li... >> https://lists.sourceforge.net/lists/listinfo/playerstage-gazebo >> > ------------------------------------------------------------------------------ > Sell apps to millions through the Intel(R) Atom(Tm) Developer Program > Be part of this innovative community and reach millions of netbook users > worldwide. Take advantage of special opportunities to increase revenue and > speed time-to-market. Join now, and jumpstart your future. > http://p.sf.net/sfu/intel-atom-d2d > _______________________________________________ > Playerstage-gazebo mailing list > Pla...@li... > https://lists.sourceforge.net/lists/listinfo/playerstage-gazebo -- Luc Brunet Co-Founder, Director - Robotic Systems Provectus Robotics Solutions www.provectus-robotics.com |
From: Kamil M. <mus...@gm...> - 2010-08-26 08:17:06
|
2010/8/26 Luc Brunet <lb...@pr...>: > Hi John, thanks for the reply. Yes, I meant <mu1>, although I was > unaware of <mu2>, is there a difference? By using <mu1> and <mu2> you can set different friction coefficients along x and y axis of a surface. But if you specify only mu1, mu2=mu1 is used. How about the <mu> parameters for your ground surface? Maybe it is to slippery for your robot :) Regards, Kamil |
From: John H. <jo...@wi...> - 2010-08-26 08:31:13
|
how are you controlling the pioneer? I tried giving it a goal using the example file (attahed position.cc) and was able to achieve target linear and angular velocities. also, can you try the attached patch and see if the slip behavior goes away? John On Wed, Aug 25, 2010 at 10:41 PM, Luc Brunet <lb...@pr... > wrote: > Hi John, thanks for the reply. Yes, I meant <mu1>, although I was > unaware of <mu2>, is there a difference? I was originally using my own > model, but then verified the same behaviour in the pioneer2dx model that > comes as an example. > > Luc > > On 08/26/2010 01:37 AM, John Hsu wrote: > > Hi Luc, > > Do you mean<mu1> (and<mu2>)? Also, are you using one of the pioneer > > models that comes with the svn checkout? > > John > > > > > > On Wed, Aug 25, 2010 at 8:21 PM, Luc Brunet > > <lb...@pr...>wrote: > > > >> Hi everyone, > >> > >> I am using a differential drive robot in Gazebo (svn version), and I am > >> noticing a lot of slip in the wheels that is giving me large errors, > >> particularly in the heading when I am spinning. It is to the point > >> where I visibly watch the robot spin twice before the yaw values > >> indicate that it has spun once. Can someone point me to some setting to > >> play with so that I can tune this? I have tried setting the<mul> of > >> the wheels to infinity without any noticeable change in behaviour. > >> > >> Thanks alot! > >> > >> -- > >> Luc Brunet > >> Co-Founder, Director - Robotic Systems > >> Provectus Robotics Solutions > >> www.provectus-robotics.com > >> > >> > >> > ------------------------------------------------------------------------------ > >> Sell apps to millions through the Intel(R) Atom(Tm) Developer Program > >> Be part of this innovative community and reach millions of netbook users > >> worldwide. Take advantage of special opportunities to increase revenue > and > >> speed time-to-market. Join now, and jumpstart your future. > >> http://p.sf.net/sfu/intel-atom-d2d > >> _______________________________________________ > >> Playerstage-gazebo mailing list > >> Pla...@li... > >> https://lists.sourceforge.net/lists/listinfo/playerstage-gazebo > >> > > > ------------------------------------------------------------------------------ > > Sell apps to millions through the Intel(R) Atom(Tm) Developer Program > > Be part of this innovative community and reach millions of netbook users > > worldwide. Take advantage of special opportunities to increase revenue > and > > speed time-to-market. Join now, and jumpstart your future. > > http://p.sf.net/sfu/intel-atom-d2d > > _______________________________________________ > > Playerstage-gazebo mailing list > > Pla...@li... > > https://lists.sourceforge.net/lists/listinfo/playerstage-gazebo > > -- > Luc Brunet > Co-Founder, Director - Robotic Systems > Provectus Robotics Solutions > www.provectus-robotics.com > > > ------------------------------------------------------------------------------ > Sell apps to millions through the Intel(R) Atom(Tm) Developer Program > Be part of this innovative community and reach millions of netbook users > worldwide. Take advantage of special opportunities to increase revenue and > speed time-to-market. Join now, and jumpstart your future. > http://p.sf.net/sfu/intel-atom-d2d > _______________________________________________ > Playerstage-gazebo mailing list > Pla...@li... > https://lists.sourceforge.net/lists/listinfo/playerstage-gazebo > |
From: Luis M. <lui...@gm...> - 2010-08-26 11:58:53
|
If you're using your own model, you probably didn't set correctly the radius and distances between the wheels: <wheelSeparation>0.34</wheelSeparation> <wheelDiameter>0.16</wheelDiameter> I guess (I didn't actually read the code) that the odometry is computed using this values. Thus, they must match with the actual values of the model. Luis 2010/8/26 John Hsu <jo...@wi...> > how are you controlling the pioneer? I tried giving it a goal using the > example file (attahed position.cc) and was able to achieve target linear > and > angular velocities. > also, can you try the attached patch and see if the slip behavior goes > away? > John > > On Wed, Aug 25, 2010 at 10:41 PM, Luc Brunet < > lb...@pr... > > wrote: > > > Hi John, thanks for the reply. Yes, I meant <mu1>, although I was > > unaware of <mu2>, is there a difference? I was originally using my own > > model, but then verified the same behaviour in the pioneer2dx model that > > comes as an example. > > > > Luc > > > > On 08/26/2010 01:37 AM, John Hsu wrote: > > > Hi Luc, > > > Do you mean<mu1> (and<mu2>)? Also, are you using one of the pioneer > > > models that comes with the svn checkout? > > > John > > > > > > > > > On Wed, Aug 25, 2010 at 8:21 PM, Luc Brunet > > > <lb...@pr...>wrote: > > > > > >> Hi everyone, > > >> > > >> I am using a differential drive robot in Gazebo (svn version), and I > am > > >> noticing a lot of slip in the wheels that is giving me large errors, > > >> particularly in the heading when I am spinning. It is to the point > > >> where I visibly watch the robot spin twice before the yaw values > > >> indicate that it has spun once. Can someone point me to some setting > to > > >> play with so that I can tune this? I have tried setting the<mul> of > > >> the wheels to infinity without any noticeable change in behaviour. > > >> > > >> Thanks alot! > > >> > > >> -- > > >> Luc Brunet > > >> Co-Founder, Director - Robotic Systems > > >> Provectus Robotics Solutions > > >> www.provectus-robotics.com > > >> > > >> > > >> > > > ------------------------------------------------------------------------------ > > >> Sell apps to millions through the Intel(R) Atom(Tm) Developer Program > > >> Be part of this innovative community and reach millions of netbook > users > > >> worldwide. Take advantage of special opportunities to increase revenue > > and > > >> speed time-to-market. Join now, and jumpstart your future. > > >> http://p.sf.net/sfu/intel-atom-d2d > > >> _______________________________________________ > > >> Playerstage-gazebo mailing list > > >> Pla...@li... > > >> https://lists.sourceforge.net/lists/listinfo/playerstage-gazebo > > >> > > > > > > ------------------------------------------------------------------------------ > > > Sell apps to millions through the Intel(R) Atom(Tm) Developer Program > > > Be part of this innovative community and reach millions of netbook > users > > > worldwide. Take advantage of special opportunities to increase revenue > > and > > > speed time-to-market. Join now, and jumpstart your future. > > > http://p.sf.net/sfu/intel-atom-d2d > > > _______________________________________________ > > > Playerstage-gazebo mailing list > > > Pla...@li... > > > https://lists.sourceforge.net/lists/listinfo/playerstage-gazebo > > > > -- > > Luc Brunet > > Co-Founder, Director - Robotic Systems > > Provectus Robotics Solutions > > www.provectus-robotics.com > > > > > > > ------------------------------------------------------------------------------ > > Sell apps to millions through the Intel(R) Atom(Tm) Developer Program > > Be part of this innovative community and reach millions of netbook users > > worldwide. Take advantage of special opportunities to increase revenue > and > > speed time-to-market. Join now, and jumpstart your future. > > http://p.sf.net/sfu/intel-atom-d2d > > _______________________________________________ > > Playerstage-gazebo mailing list > > Pla...@li... > > https://lists.sourceforge.net/lists/listinfo/playerstage-gazebo > > > > > ------------------------------------------------------------------------------ > Sell apps to millions through the Intel(R) Atom(Tm) Developer Program > Be part of this innovative community and reach millions of netbook users > worldwide. Take advantage of special opportunities to increase revenue and > speed time-to-market. Join now, and jumpstart your future. > http://p.sf.net/sfu/intel-atom-d2d > _______________________________________________ > Playerstage-gazebo mailing list > Pla...@li... > https://lists.sourceforge.net/lists/listinfo/playerstage-gazebo > > |
From: Luc B. <lb...@pr...> - 2010-08-26 13:07:26
|
Hi everyone, thanks for the replies. John, my problem isn't achieving the proper angular and linear velocities, it's that the resultant yaw between the observed one in Gazebo, and the one I am reading from the position interface seems off after I have given it a few commands (spin the robot exactly once and the yaw isn't the same as initial). I am guessing it is caused by wheel slip, but I am not sure what the proper mu values should be, and also I don't know if there are other parameters I can tweak. I have attached an example world file that I am using, and the pioneer2dx model from svn, and I am using the position interface libgazebo::PositionIface * _pose; I can set the yaw values to the following: _pose->data->cmdEnableMotors = 1; _pose->data->cmdVelocity.pos.x = 0.0; _pose->data->cmdVelocity.yaw = 1.0; and observe that when the robot is lined up with its initial heading, the value I am reading is not zero, but something way off (2.08 for instance), and changes the more rotations I allow it to do. Any help in this is much appreciated, thanks! Luc On 08/26/2010 07:58 AM, Luis Manso wrote: > If you're using your own model, you probably didn't set correctly the radius > and distances between the wheels: > <wheelSeparation>0.34</wheelSeparation> > <wheelDiameter>0.16</wheelDiameter> > > I guess (I didn't actually read the code) that the odometry is computed > using this values. Thus, they must match with the actual values of the > model. > > Luis > > > 2010/8/26 John Hsu<jo...@wi...> > >> how are you controlling the pioneer? I tried giving it a goal using the >> example file (attahed position.cc) and was able to achieve target linear >> and >> angular velocities. >> also, can you try the attached patch and see if the slip behavior goes >> away? >> John >> >> On Wed, Aug 25, 2010 at 10:41 PM, Luc Brunet< >> lb...@pr... >>> wrote: >> >>> Hi John, thanks for the reply. Yes, I meant<mu1>, although I was >>> unaware of<mu2>, is there a difference? I was originally using my own >>> model, but then verified the same behaviour in the pioneer2dx model that >>> comes as an example. >>> >>> Luc >>> >>> On 08/26/2010 01:37 AM, John Hsu wrote: >>>> Hi Luc, >>>> Do you mean<mu1> (and<mu2>)? Also, are you using one of the pioneer >>>> models that comes with the svn checkout? >>>> John >>>> >>>> >>>> On Wed, Aug 25, 2010 at 8:21 PM, Luc Brunet >>>> <lb...@pr...>wrote: >>>> >>>>> Hi everyone, >>>>> >>>>> I am using a differential drive robot in Gazebo (svn version), and I >> am >>>>> noticing a lot of slip in the wheels that is giving me large errors, >>>>> particularly in the heading when I am spinning. It is to the point >>>>> where I visibly watch the robot spin twice before the yaw values >>>>> indicate that it has spun once. Can someone point me to some setting >> to >>>>> play with so that I can tune this? I have tried setting the<mul> of >>>>> the wheels to infinity without any noticeable change in behaviour. >>>>> >>>>> Thanks alot! >>>>> >>>>> -- >>>>> Luc Brunet >>>>> Co-Founder, Director - Robotic Systems >>>>> Provectus Robotics Solutions >>>>> www.provectus-robotics.com >>>>> >>>>> >>>>> >>> >> ------------------------------------------------------------------------------ >>>>> Sell apps to millions through the Intel(R) Atom(Tm) Developer Program >>>>> Be part of this innovative community and reach millions of netbook >> users >>>>> worldwide. Take advantage of special opportunities to increase revenue >>> and >>>>> speed time-to-market. Join now, and jumpstart your future. >>>>> http://p.sf.net/sfu/intel-atom-d2d >>>>> _______________________________________________ >>>>> Playerstage-gazebo mailing list >>>>> Pla...@li... >>>>> https://lists.sourceforge.net/lists/listinfo/playerstage-gazebo >>>>> >>>> >>> >> ------------------------------------------------------------------------------ >>>> Sell apps to millions through the Intel(R) Atom(Tm) Developer Program >>>> Be part of this innovative community and reach millions of netbook >> users >>>> worldwide. Take advantage of special opportunities to increase revenue >>> and >>>> speed time-to-market. Join now, and jumpstart your future. >>>> http://p.sf.net/sfu/intel-atom-d2d >>>> _______________________________________________ >>>> Playerstage-gazebo mailing list >>>> Pla...@li... >>>> https://lists.sourceforge.net/lists/listinfo/playerstage-gazebo >>> >>> -- >>> Luc Brunet >>> Co-Founder, Director - Robotic Systems >>> Provectus Robotics Solutions >>> www.provectus-robotics.com >>> >>> >>> >> ------------------------------------------------------------------------------ >>> Sell apps to millions through the Intel(R) Atom(Tm) Developer Program >>> Be part of this innovative community and reach millions of netbook users >>> worldwide. Take advantage of special opportunities to increase revenue >> and >>> speed time-to-market. Join now, and jumpstart your future. >>> http://p.sf.net/sfu/intel-atom-d2d >>> _______________________________________________ >>> Playerstage-gazebo mailing list >>> Pla...@li... >>> https://lists.sourceforge.net/lists/listinfo/playerstage-gazebo >>> >> >> >> ------------------------------------------------------------------------------ >> Sell apps to millions through the Intel(R) Atom(Tm) Developer Program >> Be part of this innovative community and reach millions of netbook users >> worldwide. Take advantage of special opportunities to increase revenue and >> speed time-to-market. Join now, and jumpstart your future. >> http://p.sf.net/sfu/intel-atom-d2d >> _______________________________________________ >> Playerstage-gazebo mailing list >> Pla...@li... >> https://lists.sourceforge.net/lists/listinfo/playerstage-gazebo >> >> > ------------------------------------------------------------------------------ > Sell apps to millions through the Intel(R) Atom(Tm) Developer Program > Be part of this innovative community and reach millions of netbook users > worldwide. Take advantage of special opportunities to increase revenue and > speed time-to-market. Join now, and jumpstart your future. > http://p.sf.net/sfu/intel-atom-d2d > _______________________________________________ > Playerstage-gazebo mailing list > Pla...@li... > https://lists.sourceforge.net/lists/listinfo/playerstage-gazebo |
From: John H. <jo...@wi...> - 2010-08-26 22:43:11
|
ah, I see now. I think the error came from bad <wheelSeparation> value in pioneer2dx.model, it is specified as 0.39 but the wheels are actually 0.34 apart. Changing that parameter to 0.34 fixed the orientation odometry for me. John On Thu, Aug 26, 2010 at 6:07 AM, Luc Brunet <lb...@pr...>wrote: > Hi everyone, thanks for the replies. John, my problem isn't achieving the > proper angular and linear velocities, it's that the resultant yaw between > the observed one in Gazebo, and the one I am reading from the position > interface seems off after I have given it a few commands (spin the robot > exactly once and the yaw isn't the same as initial). I am guessing it is > caused by wheel slip, but I am not sure what the proper mu values should be, > and also I don't know if there are other parameters I can tweak. > > I have attached an example world file that I am using, and the pioneer2dx > model from svn, and I am using the position interface > > libgazebo::PositionIface * _pose; > > I can set the yaw values to the following: > > _pose->data->cmdEnableMotors = 1; > _pose->data->cmdVelocity.pos.x = 0.0; > _pose->data->cmdVelocity.yaw = 1.0; > > and observe that when the robot is lined up with its initial heading, the > value I am reading is not zero, but something way off (2.08 for instance), > and changes the more rotations I allow it to do. > > Any help in this is much appreciated, > > thanks! > > Luc > > > On 08/26/2010 07:58 AM, Luis Manso wrote: > >> If you're using your own model, you probably didn't set correctly the >> radius >> and distances between the wheels: >> <wheelSeparation>0.34</wheelSeparation> >> <wheelDiameter>0.16</wheelDiameter> >> >> I guess (I didn't actually read the code) that the odometry is computed >> using this values. Thus, they must match with the actual values of the >> model. >> >> Luis >> >> >> 2010/8/26 John Hsu<jo...@wi...> >> >> how are you controlling the pioneer? I tried giving it a goal using the >>> example file (attahed position.cc) and was able to achieve target linear >>> and >>> angular velocities. >>> also, can you try the attached patch and see if the slip behavior goes >>> away? >>> John >>> >>> On Wed, Aug 25, 2010 at 10:41 PM, Luc Brunet< >>> lb...@pr... >>> >>>> wrote: >>>> >>> >>> Hi John, thanks for the reply. Yes, I meant<mu1>, although I was >>>> unaware of<mu2>, is there a difference? I was originally using my own >>>> model, but then verified the same behaviour in the pioneer2dx model that >>>> comes as an example. >>>> >>>> Luc >>>> >>>> On 08/26/2010 01:37 AM, John Hsu wrote: >>>> >>>>> Hi Luc, >>>>> Do you mean<mu1> (and<mu2>)? Also, are you using one of the pioneer >>>>> models that comes with the svn checkout? >>>>> John >>>>> >>>>> >>>>> On Wed, Aug 25, 2010 at 8:21 PM, Luc Brunet >>>>> <lb...@pr...>wrote: >>>>> >>>>> Hi everyone, >>>>>> >>>>>> I am using a differential drive robot in Gazebo (svn version), and I >>>>>> >>>>> am >>> >>>> noticing a lot of slip in the wheels that is giving me large errors, >>>>>> particularly in the heading when I am spinning. It is to the point >>>>>> where I visibly watch the robot spin twice before the yaw values >>>>>> indicate that it has spun once. Can someone point me to some setting >>>>>> >>>>> to >>> >>>> play with so that I can tune this? I have tried setting the<mul> of >>>>>> the wheels to infinity without any noticeable change in behaviour. >>>>>> >>>>>> Thanks alot! >>>>>> >>>>>> -- >>>>>> Luc Brunet >>>>>> Co-Founder, Director - Robotic Systems >>>>>> Provectus Robotics Solutions >>>>>> www.provectus-robotics.com >>>>>> >>>>>> >>>>>> >>>>>> >>>> ------------------------------------------------------------------------------ >>> >>>> Sell apps to millions through the Intel(R) Atom(Tm) Developer Program >>>>>> Be part of this innovative community and reach millions of netbook >>>>>> >>>>> users >>> >>>> worldwide. Take advantage of special opportunities to increase revenue >>>>>> >>>>> and >>>> >>>>> speed time-to-market. Join now, and jumpstart your future. >>>>>> http://p.sf.net/sfu/intel-atom-d2d >>>>>> _______________________________________________ >>>>>> Playerstage-gazebo mailing list >>>>>> Pla...@li... >>>>>> https://lists.sourceforge.net/lists/listinfo/playerstage-gazebo >>>>>> >>>>>> >>>>> >>>> ------------------------------------------------------------------------------ >>> >>>> Sell apps to millions through the Intel(R) Atom(Tm) Developer Program >>>>> Be part of this innovative community and reach millions of netbook >>>>> >>>> users >>> >>>> worldwide. Take advantage of special opportunities to increase revenue >>>>> >>>> and >>>> >>>>> speed time-to-market. Join now, and jumpstart your future. >>>>> http://p.sf.net/sfu/intel-atom-d2d >>>>> _______________________________________________ >>>>> Playerstage-gazebo mailing list >>>>> Pla...@li... >>>>> https://lists.sourceforge.net/lists/listinfo/playerstage-gazebo >>>>> >>>> >>>> -- >>>> Luc Brunet >>>> Co-Founder, Director - Robotic Systems >>>> Provectus Robotics Solutions >>>> www.provectus-robotics.com >>>> >>>> >>>> >>>> ------------------------------------------------------------------------------ >>> >>>> Sell apps to millions through the Intel(R) Atom(Tm) Developer Program >>>> Be part of this innovative community and reach millions of netbook users >>>> worldwide. Take advantage of special opportunities to increase revenue >>>> >>> and >>> >>>> speed time-to-market. Join now, and jumpstart your future. >>>> http://p.sf.net/sfu/intel-atom-d2d >>>> _______________________________________________ >>>> Playerstage-gazebo mailing list >>>> Pla...@li... >>>> https://lists.sourceforge.net/lists/listinfo/playerstage-gazebo >>>> >>>> >>> >>> >>> ------------------------------------------------------------------------------ >>> Sell apps to millions through the Intel(R) Atom(Tm) Developer Program >>> Be part of this innovative community and reach millions of netbook users >>> worldwide. Take advantage of special opportunities to increase revenue >>> and >>> speed time-to-market. Join now, and jumpstart your future. >>> http://p.sf.net/sfu/intel-atom-d2d >>> _______________________________________________ >>> Playerstage-gazebo mailing list >>> Pla...@li... >>> https://lists.sourceforge.net/lists/listinfo/playerstage-gazebo >>> >>> >>> ------------------------------------------------------------------------------ >> Sell apps to millions through the Intel(R) Atom(Tm) Developer Program >> Be part of this innovative community and reach millions of netbook users >> worldwide. Take advantage of special opportunities to increase revenue and >> speed time-to-market. Join now, and jumpstart your future. >> http://p.sf.net/sfu/intel-atom-d2d >> _______________________________________________ >> Playerstage-gazebo mailing list >> Pla...@li... >> https://lists.sourceforge.net/lists/listinfo/playerstage-gazebo >> > > > ------------------------------------------------------------------------------ > Sell apps to millions through the Intel(R) Atom(Tm) Developer Program > Be part of this innovative community and reach millions of netbook users > worldwide. Take advantage of special opportunities to increase revenue and > speed time-to-market. Join now, and jumpstart your future. > http://p.sf.net/sfu/intel-atom-d2d > _______________________________________________ > Playerstage-gazebo mailing list > Pla...@li... > https://lists.sourceforge.net/lists/listinfo/playerstage-gazebo > > |
From: John H. <jo...@wi...> - 2010-08-26 22:44:04
|
updated is in r8858. On Thu, Aug 26, 2010 at 3:42 PM, John Hsu <jo...@wi...> wrote: > ah, I see now. I think the error came from bad <wheelSeparation> value in > pioneer2dx.model, it is specified as 0.39 but the wheels are actually 0.34 > apart. Changing that parameter to 0.34 fixed the orientation odometry for > me. > John > > > > > On Thu, Aug 26, 2010 at 6:07 AM, Luc Brunet < > lb...@pr...> wrote: > >> Hi everyone, thanks for the replies. John, my problem isn't achieving the >> proper angular and linear velocities, it's that the resultant yaw between >> the observed one in Gazebo, and the one I am reading from the position >> interface seems off after I have given it a few commands (spin the robot >> exactly once and the yaw isn't the same as initial). I am guessing it is >> caused by wheel slip, but I am not sure what the proper mu values should be, >> and also I don't know if there are other parameters I can tweak. >> >> I have attached an example world file that I am using, and the pioneer2dx >> model from svn, and I am using the position interface >> >> libgazebo::PositionIface * _pose; >> >> I can set the yaw values to the following: >> >> _pose->data->cmdEnableMotors = 1; >> _pose->data->cmdVelocity.pos.x = 0.0; >> _pose->data->cmdVelocity.yaw = 1.0; >> >> and observe that when the robot is lined up with its initial heading, the >> value I am reading is not zero, but something way off (2.08 for instance), >> and changes the more rotations I allow it to do. >> >> Any help in this is much appreciated, >> >> thanks! >> >> Luc >> >> >> On 08/26/2010 07:58 AM, Luis Manso wrote: >> >>> If you're using your own model, you probably didn't set correctly the >>> radius >>> and distances between the wheels: >>> <wheelSeparation>0.34</wheelSeparation> >>> <wheelDiameter>0.16</wheelDiameter> >>> >>> I guess (I didn't actually read the code) that the odometry is computed >>> using this values. Thus, they must match with the actual values of the >>> model. >>> >>> Luis >>> >>> >>> 2010/8/26 John Hsu<jo...@wi...> >>> >>> how are you controlling the pioneer? I tried giving it a goal using the >>>> example file (attahed position.cc) and was able to achieve target linear >>>> and >>>> angular velocities. >>>> also, can you try the attached patch and see if the slip behavior goes >>>> away? >>>> John >>>> >>>> On Wed, Aug 25, 2010 at 10:41 PM, Luc Brunet< >>>> lb...@pr... >>>> >>>>> wrote: >>>>> >>>> >>>> Hi John, thanks for the reply. Yes, I meant<mu1>, although I was >>>>> unaware of<mu2>, is there a difference? I was originally using my own >>>>> model, but then verified the same behaviour in the pioneer2dx model >>>>> that >>>>> comes as an example. >>>>> >>>>> Luc >>>>> >>>>> On 08/26/2010 01:37 AM, John Hsu wrote: >>>>> >>>>>> Hi Luc, >>>>>> Do you mean<mu1> (and<mu2>)? Also, are you using one of the pioneer >>>>>> models that comes with the svn checkout? >>>>>> John >>>>>> >>>>>> >>>>>> On Wed, Aug 25, 2010 at 8:21 PM, Luc Brunet >>>>>> <lb...@pr...>wrote: >>>>>> >>>>>> Hi everyone, >>>>>>> >>>>>>> I am using a differential drive robot in Gazebo (svn version), and I >>>>>>> >>>>>> am >>>> >>>>> noticing a lot of slip in the wheels that is giving me large errors, >>>>>>> particularly in the heading when I am spinning. It is to the point >>>>>>> where I visibly watch the robot spin twice before the yaw values >>>>>>> indicate that it has spun once. Can someone point me to some setting >>>>>>> >>>>>> to >>>> >>>>> play with so that I can tune this? I have tried setting the<mul> of >>>>>>> the wheels to infinity without any noticeable change in behaviour. >>>>>>> >>>>>>> Thanks alot! >>>>>>> >>>>>>> -- >>>>>>> Luc Brunet >>>>>>> Co-Founder, Director - Robotic Systems >>>>>>> Provectus Robotics Solutions >>>>>>> www.provectus-robotics.com >>>>>>> >>>>>>> >>>>>>> >>>>>>> >>>>> ------------------------------------------------------------------------------ >>>> >>>>> Sell apps to millions through the Intel(R) Atom(Tm) Developer Program >>>>>>> Be part of this innovative community and reach millions of netbook >>>>>>> >>>>>> users >>>> >>>>> worldwide. Take advantage of special opportunities to increase revenue >>>>>>> >>>>>> and >>>>> >>>>>> speed time-to-market. Join now, and jumpstart your future. >>>>>>> http://p.sf.net/sfu/intel-atom-d2d >>>>>>> _______________________________________________ >>>>>>> Playerstage-gazebo mailing list >>>>>>> Pla...@li... >>>>>>> https://lists.sourceforge.net/lists/listinfo/playerstage-gazebo >>>>>>> >>>>>>> >>>>>> >>>>> ------------------------------------------------------------------------------ >>>> >>>>> Sell apps to millions through the Intel(R) Atom(Tm) Developer Program >>>>>> Be part of this innovative community and reach millions of netbook >>>>>> >>>>> users >>>> >>>>> worldwide. Take advantage of special opportunities to increase revenue >>>>>> >>>>> and >>>>> >>>>>> speed time-to-market. Join now, and jumpstart your future. >>>>>> http://p.sf.net/sfu/intel-atom-d2d >>>>>> _______________________________________________ >>>>>> Playerstage-gazebo mailing list >>>>>> Pla...@li... >>>>>> https://lists.sourceforge.net/lists/listinfo/playerstage-gazebo >>>>>> >>>>> >>>>> -- >>>>> Luc Brunet >>>>> Co-Founder, Director - Robotic Systems >>>>> Provectus Robotics Solutions >>>>> www.provectus-robotics.com >>>>> >>>>> >>>>> >>>>> ------------------------------------------------------------------------------ >>>> >>>>> Sell apps to millions through the Intel(R) Atom(Tm) Developer Program >>>>> Be part of this innovative community and reach millions of netbook >>>>> users >>>>> worldwide. Take advantage of special opportunities to increase revenue >>>>> >>>> and >>>> >>>>> speed time-to-market. Join now, and jumpstart your future. >>>>> http://p.sf.net/sfu/intel-atom-d2d >>>>> _______________________________________________ >>>>> Playerstage-gazebo mailing list >>>>> Pla...@li... >>>>> https://lists.sourceforge.net/lists/listinfo/playerstage-gazebo >>>>> >>>>> >>>> >>>> >>>> ------------------------------------------------------------------------------ >>>> Sell apps to millions through the Intel(R) Atom(Tm) Developer Program >>>> Be part of this innovative community and reach millions of netbook users >>>> worldwide. Take advantage of special opportunities to increase revenue >>>> and >>>> speed time-to-market. Join now, and jumpstart your future. >>>> http://p.sf.net/sfu/intel-atom-d2d >>>> _______________________________________________ >>>> Playerstage-gazebo mailing list >>>> Pla...@li... >>>> https://lists.sourceforge.net/lists/listinfo/playerstage-gazebo >>>> >>>> >>>> ------------------------------------------------------------------------------ >>> Sell apps to millions through the Intel(R) Atom(Tm) Developer Program >>> Be part of this innovative community and reach millions of netbook users >>> worldwide. Take advantage of special opportunities to increase revenue >>> and >>> speed time-to-market. Join now, and jumpstart your future. >>> http://p.sf.net/sfu/intel-atom-d2d >>> _______________________________________________ >>> Playerstage-gazebo mailing list >>> Pla...@li... >>> https://lists.sourceforge.net/lists/listinfo/playerstage-gazebo >>> >> >> >> ------------------------------------------------------------------------------ >> Sell apps to millions through the Intel(R) Atom(Tm) Developer Program >> Be part of this innovative community and reach millions of netbook users >> worldwide. Take advantage of special opportunities to increase revenue and >> speed time-to-market. Join now, and jumpstart your future. >> http://p.sf.net/sfu/intel-atom-d2d >> _______________________________________________ >> Playerstage-gazebo mailing list >> Pla...@li... >> https://lists.sourceforge.net/lists/listinfo/playerstage-gazebo >> >> > |
From: Luc B. <lb...@pr...> - 2010-08-27 00:47:02
|
thanks that did it! On 08/26/2010 06:42 PM, John Hsu wrote: > ah, I see now. I think the error came from bad<wheelSeparation> value in > pioneer2dx.model, it is specified as 0.39 but the wheels are actually 0.34 > apart. Changing that parameter to 0.34 fixed the orientation odometry for > me. > John > > > > On Thu, Aug 26, 2010 at 6:07 AM, Luc Brunet > <lb...@pr...>wrote: > > >> Hi everyone, thanks for the replies. John, my problem isn't achieving the >> proper angular and linear velocities, it's that the resultant yaw between >> the observed one in Gazebo, and the one I am reading from the position >> interface seems off after I have given it a few commands (spin the robot >> exactly once and the yaw isn't the same as initial). I am guessing it is >> caused by wheel slip, but I am not sure what the proper mu values should be, >> and also I don't know if there are other parameters I can tweak. >> >> I have attached an example world file that I am using, and the pioneer2dx >> model from svn, and I am using the position interface >> >> libgazebo::PositionIface * _pose; >> >> I can set the yaw values to the following: >> >> _pose->data->cmdEnableMotors = 1; >> _pose->data->cmdVelocity.pos.x = 0.0; >> _pose->data->cmdVelocity.yaw = 1.0; >> >> and observe that when the robot is lined up with its initial heading, the >> value I am reading is not zero, but something way off (2.08 for instance), >> and changes the more rotations I allow it to do. >> >> Any help in this is much appreciated, >> >> thanks! >> >> Luc >> >> >> On 08/26/2010 07:58 AM, Luis Manso wrote: >> >> >>> If you're using your own model, you probably didn't set correctly the >>> radius >>> and distances between the wheels: >>> <wheelSeparation>0.34</wheelSeparation> >>> <wheelDiameter>0.16</wheelDiameter> >>> >>> I guess (I didn't actually read the code) that the odometry is computed >>> using this values. Thus, they must match with the actual values of the >>> model. >>> >>> Luis >>> >>> >>> 2010/8/26 John Hsu<jo...@wi...> >>> >>> how are you controlling the pioneer? I tried giving it a goal using the >>> >>>> example file (attahed position.cc) and was able to achieve target linear >>>> and >>>> angular velocities. >>>> also, can you try the attached patch and see if the slip behavior goes >>>> away? >>>> John >>>> >>>> On Wed, Aug 25, 2010 at 10:41 PM, Luc Brunet< >>>> lb...@pr... >>>> >>>> >>>>> wrote: >>>>> >>>>> >>>> Hi John, thanks for the reply. Yes, I meant<mu1>, although I was >>>> >>>>> unaware of<mu2>, is there a difference? I was originally using my own >>>>> model, but then verified the same behaviour in the pioneer2dx model that >>>>> comes as an example. >>>>> >>>>> Luc >>>>> >>>>> On 08/26/2010 01:37 AM, John Hsu wrote: >>>>> >>>>> >>>>>> Hi Luc, >>>>>> Do you mean<mu1> (and<mu2>)? Also, are you using one of the pioneer >>>>>> models that comes with the svn checkout? >>>>>> John >>>>>> >>>>>> >>>>>> On Wed, Aug 25, 2010 at 8:21 PM, Luc Brunet >>>>>> <lb...@pr...>wrote: >>>>>> >>>>>> Hi everyone, >>>>>> >>>>>>> I am using a differential drive robot in Gazebo (svn version), and I >>>>>>> >>>>>>> >>>>>> am >>>>>> >>>> >>>>> noticing a lot of slip in the wheels that is giving me large errors, >>>>> >>>>>>> particularly in the heading when I am spinning. It is to the point >>>>>>> where I visibly watch the robot spin twice before the yaw values >>>>>>> indicate that it has spun once. Can someone point me to some setting >>>>>>> >>>>>>> >>>>>> to >>>>>> >>>> >>>>> play with so that I can tune this? I have tried setting the<mul> of >>>>> >>>>>>> the wheels to infinity without any noticeable change in behaviour. >>>>>>> >>>>>>> Thanks alot! >>>>>>> >>>>>>> -- >>>>>>> Luc Brunet >>>>>>> Co-Founder, Director - Robotic Systems >>>>>>> Provectus Robotics Solutions >>>>>>> www.provectus-robotics.com >>>>>>> >>>>>>> >>>>>>> >>>>>>> >>>>>>> >>>>> ------------------------------------------------------------------------------ >>>>> >>>> >>>>> Sell apps to millions through the Intel(R) Atom(Tm) Developer Program >>>>> >>>>>>> Be part of this innovative community and reach millions of netbook >>>>>>> >>>>>>> >>>>>> users >>>>>> >>>> >>>>> worldwide. Take advantage of special opportunities to increase revenue >>>>> >>>>>>> >>>>>> and >>>>>> >>>>> >>>>>> speed time-to-market. Join now, and jumpstart your future. >>>>>> >>>>>>> http://p.sf.net/sfu/intel-atom-d2d >>>>>>> _______________________________________________ >>>>>>> Playerstage-gazebo mailing list >>>>>>> Pla...@li... >>>>>>> https://lists.sourceforge.net/lists/listinfo/playerstage-gazebo >>>>>>> >>>>>>> >>>>>>> >>>>>> >>>>> ------------------------------------------------------------------------------ >>>>> >>>> >>>>> Sell apps to millions through the Intel(R) Atom(Tm) Developer Program >>>>> >>>>>> Be part of this innovative community and reach millions of netbook >>>>>> >>>>>> >>>>> users >>>>> >>>> >>>>> worldwide. Take advantage of special opportunities to increase revenue >>>>> >>>>>> >>>>> and >>>>> >>>>> >>>>>> speed time-to-market. Join now, and jumpstart your future. >>>>>> http://p.sf.net/sfu/intel-atom-d2d >>>>>> _______________________________________________ >>>>>> Playerstage-gazebo mailing list >>>>>> Pla...@li... >>>>>> https://lists.sourceforge.net/lists/listinfo/playerstage-gazebo >>>>>> >>>>>> >>>>> -- >>>>> Luc Brunet >>>>> Co-Founder, Director - Robotic Systems >>>>> Provectus Robotics Solutions >>>>> www.provectus-robotics.com >>>>> >>>>> >>>>> >>>>> ------------------------------------------------------------------------------ >>>>> >>>> >>>>> Sell apps to millions through the Intel(R) Atom(Tm) Developer Program >>>>> Be part of this innovative community and reach millions of netbook users >>>>> worldwide. Take advantage of special opportunities to increase revenue >>>>> >>>>> >>>> and >>>> >>>> >>>>> speed time-to-market. Join now, and jumpstart your future. >>>>> http://p.sf.net/sfu/intel-atom-d2d >>>>> _______________________________________________ >>>>> Playerstage-gazebo mailing list >>>>> Pla...@li... >>>>> https://lists.sourceforge.net/lists/listinfo/playerstage-gazebo >>>>> >>>>> >>>>> >>>> >>>> ------------------------------------------------------------------------------ >>>> Sell apps to millions through the Intel(R) Atom(Tm) Developer Program >>>> Be part of this innovative community and reach millions of netbook users >>>> worldwide. Take advantage of special opportunities to increase revenue >>>> and >>>> speed time-to-market. Join now, and jumpstart your future. >>>> http://p.sf.net/sfu/intel-atom-d2d >>>> _______________________________________________ >>>> Playerstage-gazebo mailing list >>>> Pla...@li... >>>> https://lists.sourceforge.net/lists/listinfo/playerstage-gazebo >>>> >>>> >>>> ------------------------------------------------------------------------------ >>>> >>> Sell apps to millions through the Intel(R) Atom(Tm) Developer Program >>> Be part of this innovative community and reach millions of netbook users >>> worldwide. Take advantage of special opportunities to increase revenue and >>> speed time-to-market. Join now, and jumpstart your future. >>> http://p.sf.net/sfu/intel-atom-d2d >>> _______________________________________________ >>> Playerstage-gazebo mailing list >>> Pla...@li... >>> https://lists.sourceforge.net/lists/listinfo/playerstage-gazebo >>> >>> >> >> ------------------------------------------------------------------------------ >> Sell apps to millions through the Intel(R) Atom(Tm) Developer Program >> Be part of this innovative community and reach millions of netbook users >> worldwide. Take advantage of special opportunities to increase revenue and >> speed time-to-market. Join now, and jumpstart your future. >> http://p.sf.net/sfu/intel-atom-d2d >> _______________________________________________ >> Playerstage-gazebo mailing list >> Pla...@li... >> https://lists.sourceforge.net/lists/listinfo/playerstage-gazebo >> >> >> > ------------------------------------------------------------------------------ > Sell apps to millions through the Intel(R) Atom(Tm) Developer Program > Be part of this innovative community and reach millions of netbook users > worldwide. Take advantage of special opportunities to increase revenue and > speed time-to-market. Join now, and jumpstart your future. > http://p.sf.net/sfu/intel-atom-d2d > _______________________________________________ > Playerstage-gazebo mailing list > Pla...@li... > https://lists.sourceforge.net/lists/listinfo/playerstage-gazebo > -- Luc Brunet Co-Founder, Director - Robotic Systems Provectus Robotics Solutions www.provectus-robotics.com |