Revision history [back]

There are many ways: the one I decided to start with was by using the moveit library and their tutorials. In particular, at first I created a simple node with just a snippet from the whole tutorial code.

You want to move to a joint space configuration (not to a cartesian pose like in the tutorial) so the pose is stored in a vector:

auto panda_home_pose = std::vector<double> {  0.00
                                           , -0.25 * pi
                                           ,  0.00
                                           , -0.75 * pi
                                           ,  0.00
                                           ,  0.50 * pi
                                           ,  0.25 * pi };

and you will want to use setJointValueTarget instead of setPoseTarget like the tutorial does. I recommend also setting very slow velocities and accelerations at first, even more so because of potential reflex errors from the robot due to high jerk and whatnot. For this use setMaxVelocityScalingFactor and setMaxAccelerationScalingFactor. After these you can call plan and execute separately or together as you see fit, but the tutorial seems clear in that regard.

There are many ways: the one I decided to start with was by using the moveit library and their tutorials. In particular, at first I created a simple node with just a snippet from the whole tutorial code.

You want to move to a joint space configuration (not to a cartesian pose like in the tutorial) so the pose is stored in a vector:

auto panda_home_pose = std::vector<double> {  0.00
                                           , -0.25 * pi
                                           ,  0.00
                                           , -0.75 * pi
                                           ,  0.00
                                           ,  0.50 * pi
                                           ,  0.25 * pi };

and you will want to use setJointValueTarget instead of setPoseTarget like the tutorial does. I recommend also setting very slow velocities and accelerations at first, even more so because of potential reflex errors from the robot due to high jerk and whatnot. For this use setMaxVelocityScalingFactor and setMaxAccelerationScalingFactor. After these you can call plan and execute separately or together as you see fit, but the tutorial seems clear in that regard.

There are many ways: the one I decided to start with was by using the moveit library and their tutorials. In particular, at first I created a simple node with just a snippet from the whole tutorial code.

You want to move to a joint space configuration (not to a cartesian pose like in the tutorial) so the pose is stored in a vector:

auto panda_home_pose = std::vector<double> {  0.00
                                           , -0.25 * pi
                                           ,  0.00
                                           , -0.75 * pi
                                           ,  0.00
                                           ,  0.50 * pi
                                           ,  0.25 * pi };

and you will want to use setJointValueTarget instead of setPoseTarget like the tutorial does. I recommend also setting very slow velocities and accelerations at first, even more so because of potential reflex errors from the robot due to high jerk and whatnot. For this use setMaxVelocityScalingFactor and setMaxAccelerationScalingFactor. After these you can call plan and execute separately or together as you see fit, but the tutorial seems clear in that regard.