[Documentation] [TitleIndex] [WordIndex

Description

control-graph.png The control graph (picture above) is created by python script "reach_ball.py". This script is parsed by the interperter embedded in the control framework SoT running on HRP-2 control CPU.

Script "hueblob.py" listen to tf and publishes in topic '/wide/blobs/rose/transform_base_link' the position of the ball in the frame of the robot waist. Note that the position of the ball in the camera reference frame is computed by package hueblob.

"watch_zone.py" implements an actionlib server that tracks the position of the ball and returns when the distance between the ball and the target is less than the tolerance distance.

"supervisor.py" send a request to watch_zone server and triggers the motion when the client returns.

Running the motion on HRP-2 14

Prerequisite: the robot is calibrated.

ssh on hrp2014v and run

roslaunch hrp2_bringup cameras.launch

ssh on hrp2014v and run

roslaunch hrp2_bringup openhrp_bridge.launch
rosservice call restart_controller

switch relay on, ssh on hrp2014v and run

rosservice call switch_servo True all
rosservice call start_dynamic_graph

then put the robot down. ssh -Y on hrp2014v and run

roslaunch ball_reaching robot.launch

Then take the ping ball and move it in front of the robot. When the ball reaches the target position (specified in supervisor.py script) in the waist coordinate frame, up to 5 cm, the robot grabs the ball and moves back to rest position.


2020-08-08 12:28