ROS Resources: Documentation | Support | Discussion Forum | Index | Service Status | ros @ Robotics Stack Exchange
Ask Your Question
0

RGBDSlam eigen error on Landmark Optimisation

asked 2013-10-21 18:58:46 -0500

NickT gravatar image

updated 2014-01-28 17:18:18 -0500

ngrennan gravatar image

Hi everyone

I have been playing with RGBDSlam to see how it creates 3d maps. I have the package compiled and running on Ubuntu 12.04 with Groovy without too many issues. With the basic kinect+rgbdslam.launch file I can generate a 3d map, save an octomap etc.

However, when I turn on landmark optimisation, after the next node is added to the graph I get the following error.

rgbdslam: /usr/include/eigen3/Eigen/src/Core/PlainObjectBase.h:598: void Eigen::PlainObjectBase<derived>::_init2(Eigen::PlainObjectBase<derived>::Index, Eigen::PlainObjectBase<derived>::Index, typename Eigen::internal::enable_if<(Eigen::PlainObjectBase<derived>::Base:: SizeAtCompileTime != 2), T0>::type*) [with T0 = int, T1 = int, Derived = Eigen::Matrix<double, 6,="" 6,="" 0,="" 6,="" 6="">, Eigen::PlainObjectBase<derived>::Index = int, typename Eigen::internal::enable_if<(Eigen::PlainObjectBase<derived>::Base:: SizeAtCompileTime != 2), T0>::type = int]: Assertion `rows >= 0 && (RowsAtCompileTime == Dynamic || RowsAtCompileTime == rows) && cols >= 0 && (ColsAtCompileTime == Dynamic || ColsAtCompileTime == cols)' failed.

Any advice on how to go about debugging this is appreciated.

edit retag flag offensive close merge delete

1 Answer

Sort by ยป oldest newest most voted
0

answered 2014-03-05 02:43:53 -0500

hm, that's a bug. Could you please try to change the following lines in graph_manager.cpp (right after the #includes):

//typedef g2o::BlockSolver< g2o::BlockSolverTraits<-1, -1> >  SlamBlockSolver;
typedef g2o::BlockSolver< g2o::BlockSolverTraits<6, 3> >  SlamBlockSolver;

to

typedef g2o::BlockSolver< g2o::BlockSolverTraits<-1, -1> >  SlamBlockSolver;
//typedef g2o::BlockSolver< g2o::BlockSolverTraits<6, 3> >  SlamBlockSolver;

I haven't used the landmark optimization in a long time. The gains have turned out to be quite negligible. As it is therefore not regularly tested, you might run into more problems.

edit flag offensive delete link more

Question Tools

1 follower

Stats

Asked: 2013-10-21 18:58:46 -0500

Seen: 266 times

Last updated: Mar 05 '14