• cd catkin_ws/src
• git clone https://github.com/bgumodo/bgumodo-dev.git
• git clone https://github.com/robotican/robotican.git
• cd robotican && git checkout temp_branch
• follow this for further robotican installation.
• cd ~/catkin_ws
• catkin_make

• roslaunch robotican_demos armadillo_moveit_and_nav.launch
• rosrun robotican_demos simple_move_group_goals
• rosrun robotican_demos simple_navigation_goals

rostopic pub /move_arm geometry_msgs/PoseStamped "header:
seq: 0
stamp:
secs: 0
nsecs: 0
frame_id: 'base_link'
pose:
position:
x: 0.37
y: 0.0
z: 0.8
orientation:
x: 0.0
y: 0.4
z: 0.0
w: 0.91"

rostopic pub /nav_command geometry_msgs/PoseStamped "header:
seq: 0
stamp:
secs: 0
nsecs: 0
frame_id: 'map' <- (base_link is also an option)
pose:
position:
x: -4.5
y: 8.37
z: 0.0
orientation:
x: 0.0
y: 0.0
z: 0.0
w: 1.0"

rosrun gazebo_ros spawn_model -database coke_can -sdf -model coke_can5 -z 1.1 -y 8.10 -x -8.44
ROS Node API:
char exec[128] = {'\0'};
sprintf(exec, "rosrun gazebo_ros spawn_model -database coke_can -sdf -model coke_can2 -x %f -y %f -z %f",button_x, button_y, button_z);
// bash command FILE *process = popen(exec, "r");
if(process == 0) {
ROS_ERROR("[%s]: can't start the procces shuting down the node :(", ros::this_node::getName().c_str());
ros::shutdown();
}

rosservice call gazebo/delete_model '{model_name: coke_can2}'

• make sure to run "catkin_make"
• run with "roslaunch bgumodo_vision bgumodo_vision.launch"
• to initiate detection - call topic "/detector/observe_cup_cmd" with string "Start"
• notice only one red object should be in the scene. if you get NaN reading check if object is to close. (in gazebo table with objects world I have checked with the Cola can position on the far end of the table.)
• you may need to play around with the kinect's tilt to see the objects better - to do so call use the rqt's message publisher to call the topic "/pan_tilt_controller/command" and in the "data" field enter your desired position in radians ([pan, tilt]), e.g [0.0, 0.3]

Run "roslaunch bgumodo_nav armadillo_moveit_with_nav_alex.launch". To start navigation publish to topic "/navigation/move_cmd" with type=String the following options: (case sensitive)

1. "door_zone_1" : door closed position.
2. "door_zone_2" : door opened position.
3. "room_4" : target room with coke can.
4. "room_2" : final target room for coke can.

Subscribe to topic "/navigation/move_res" type=String for the navigation result. Possible result: 1. "Success" 2. "Failed"

• ROSPlan must be installed (https://github.com/KCL-Planning/ROSPlan)
• Copy our domain.pddl file to rosplan_config/planner in the ROSPlan folder

• kill mongodb service to that clash with rosplan "sudo service mongodb stop"
• Run ROSPlan using "roslaunch rosplan_planning_system planning_system_knowledge.launch"
• (not mandatory) Run "rqt" and open the ROSPlan dispatcher tab to see the plan execution
• Run the simulation using "roslaunch robotican_demos armadillo_office_demo.launch"
• Run the actions node using "rosrun robotican_demos demo_pick_node"
• Run the speech recognition using "roslaunch komodo_speech komodo_speech.launch"
• Run the ROSPlan middleware and dummy action handler (plp dispatchers) using "roslaunch bgumodo_plan plp_dispatcher.launch"
  • The dummy action handler is responsible for the "Observe button" and "Press button" actions which are not currently implemented. In addition, it is responsible for the "Order coffee" action.
• Run the world loader using "rosrun bgumodo_plan world_loader.py"
  • When the world loader is running, it waits 30 seconds for a speech command "komodo get me coffee"/"komodo get me a cup of coffee"/.. and then starts the planning and plan dispatch process

• In order to change the initial state of the world, open the "world_loader.py" script inside "bgumodo_plan". There are three important functions there:
  • createInstances: This functions create the object instances in the world (rooms, zones, doors, etc.)
  • createFacts: Uses the "addFact" generic function to add new grounded predicates (facts) to the world. This function actually builds the initial state.
  • createGoal: The function that activates the speech module and waits for a command. When a command is received, it takes the current position of the robot (loc) and adds the goal: "(coke_at loc)". This function will be changed if another goal/activation process is needed.