[pr2eus_moveit] Add timeout arg for computing IK and increase timeout for test

[708yamaguchi]

In this pull request, I added timeout arg to some functions of robot-moveit.l and set the timeout in the travis test longer.

This pull request aims to prevent the travis test failure of pr2eus_moveit, such as:
https://travis-ci.org/github/jsk-ros-pkg/jsk_pr2eus/builds/674758416
https://travis-ci.org/github/jsk-ros-pkg/jsk_pr2eus/builds/686633237

These tests fail here:

jsk_pr2eus/pr2eus_moveit/test/test-pr2eus-moveit.l

Lines 208 to 214 in 867460f

	(send *ri* :move-end-coords-plan (make-coords :pos #f(700 0 1000)) :move-arm :larm-torso)
	(setq tm-0 (ros::time-now))
	(send *ri* :wait-interpolation)
	(setq tm-1 (ros::time-now))
	(setq tm-diff (send (ros::time- tm-1 tm-0) :to-sec))
	(ros::ros-info "time for duration ~A" tm-diff)
	(assert (> tm-diff 3) (format nil "collsion avoidance motion is too fast ~A" tm-diff))

Error messages are as follows:

  <testcase name="test-move-end-coords-plan" status="run" time="15" classname="test-move-end-coords-plan">
   <failure message="collsion avoidance motion is too fast 0.001" type="AssertionError">
Test:(&gt; tm-diff 3)
Trace:";; ik error at (-31 . NO_IK_SOLUTION)

In detail, I think the reason is that

1. IK calculation in :move-end-coords-plan fails
2. PR2 moves the arm little by :move-end-coords-plan
3. tm-diff becomes too short.

The reason why the IK calculation sometimes fails may be that its timeout is short.
In my environment, the IK calculation sometimes takes ~0.1 [s], but the default timeout is 0.05 [s] in robot-moveit.l.

I checked the IK calculation execution time by this code.
https://github.com/ros-planning/moveit/blob/50e3c21081c03a106ac1157d8a8687710bcc0398/moveit_kinematics/kdl_kinematics_plugin/src/kdl_kinematics_plugin.cpp#L409-L410

The default timeout in robot-moveit.l:

jsk_pr2eus/pr2eus_moveit/euslisp/robot-moveit.l

Line 173 in 867460f

(frame-id default-frame-id) (timeout 0.05) (scene)

I am not confident, but I believe this pull requests improve success rate of the travis test.

By the way, I also tried to use attempts field in moveit_msgs/PositionIKRequest, but this field seems to be unavailable now.
moveit/moveit#1288

[knorth55]

👍 for this PR.
good work!