MoveIt No Motion plan found error.

[Dinesh Madusanke]

Dear all,

I am configuring a 3 link planar arm with moveIt and here is the part of srdf file.

<group name="ahr_arm">
        <link name="down_link" />
        <joint name="arm_1_to_base" />
        <joint name="link1_link2" />
        <joint name="link2_link3" />
        <joint name="link3_down_link" />
 </group>
    <!--END EFFECTOR: Purpose: Represent information about an end effector.-->
    <end_effector name="disk_ef" parent_link="down_link" group="ahr_arm" />

I use move group interface via C++ for motion planning. When I set a random target using  group.setRandomTarget();
and planned it works and I can see the planned result in rviz.

  group.setRandomTarget();
  moveit::planning_interface::MoveGroup::Plan my_plan;
  bool success = group.plan(my_plan);
  group.move();

But When I set a position target using group.setPositionTarget(-0.433892, -0.109893, 0.0000, "down_link");
and planned it gives an error.

[ WARN] [1402937327.375698775]: Fail: ABORTED: No motion plan found. No execution attempted.
and

[ INFO] [1402937327.377013739]: Planning attempt 1 of at most 1
[ INFO] [1402937327.377798236]: No planner specified. Using default.
[ INFO] [1402937327.377905180]: LBKPIECE1: Attempting to use default projection.
[ INFO] [1402937327.378179423]: LBKPIECE1: Starting with 1 states
[ERROR] [1402937332.380049183]: LBKPIECE1: Unable to sample any valid states for goal tree
[ INFO] [1402937332.380161762]: LBKPIECE1: Created 1 (1 start + 0 goal) states in 1 cells (1 start (1 on boundary) + 0 goal (0 on boundary))
[ INFO] [1402937332.380276906]: No solution found after 5.002223 seconds
[ INFO] [1402937332.381313319]: Unable to solve the planning problem


I am new to moveIt and I am stuck with this error. So please give suggestions to solve this.

[Dave Coleman]

Are you sure your target position is kinematically feasible? Your output says that OMPL is unable to sample a valid state of the robot at that x,y,z location. Perhaps your IK is not working properly?

dave

[Dinesh Madusanke]

Dear sir,
Thank you for your reply.
To ensure the position is feasible I tried in two ways.

1. I generated a random pose for end effector using getRandomPose (const std::string &end_effector_link="") method
    and I get the position coordinates from this random pose and set coordinates to target position of end effector.

2. Also I moved the interactive marker of rviz plugin and after planning and executing to that point, I got the end effector pose and that pose position is given as target position.

I thought these two methods will lead to reachable position.
I am not completely sure whether my approach is correct.
Is there any way I can check IK solver using moveit - rviz plugin ?

Thank you very much sir.

[Dave Coleman]

getRandomPose takes the planning group that is associated with the end effector link specified, and generates random joint values for every DOF. So it should usually be creating valid end effector poses, yes, though it appears it does not check for self-collision. So sometimes they may be invalid. 

:: dave | +1.325.261.3283 c

[Dinesh Madusanke]

Thank you for your answer sir, I got to know that KDL IK solver doesn't work well with low degree of freedom arm. Now I planned to write a IK plugin.

[Dave Coleman]

It just happens that I am also testing an IK solver and so wrote a little script to test random poses and display solutions. Perhaps my example code can guide you in your work:

https://github.com/davetcoleman/moveit_hrp2/blob/518a00b2543f9fa38117c15b772e0368cf901c7b/hrp2jsknt_moveit_demos/src/hrp2_demos.cpp#L603

Dave C

[Dinesh Madusanke]

Thank you very much for your help, sir,