This program uses ROS, ROS smach, and YOLO to search for cake in the empty_stage_single_robot floor plan in the rosplan_stage_demo package. Once found the robot will stop the search. Every couple of seconds the robot reports progress (this includes everything it saw up to now).
Prerequisits:
- linux ubuntu 20.04
- ROS, and ROS smach (I installed it using apptainer -- recommended)
- YOLO
- rosplan_stage_demo package
Before you start remember to:
- Change
coco.data
inside the config folder to point to the training set for YOLO - Run the contianer using apptainer:
apptainer run ros-container.sif
- Make the folder available to ROS:
source devel/setup.bash
- Run roscre:
roscore &
To run the program just run it like a normal ROS program:
- In the terminal run the following to start both services: (Note: the package is called
second_coursework
, you may change that if you wish)rosrun second_coursework action_client.py
rosrun second_coursework room_service.py
- You may be using either bags or an actual camera attached to a robot. Please note: that the camera topic in the robot (and therefore, in the ROS bags) on which the
images are published is
/camera/image
.
NOTE: the floor plan in scripts/room_service.py
is taken from empty_stage_single_robot in the rosplan_stage_demo package. Once installed you can use the following to view the floor plan
$ roslaunch rosplan_stage_demo empty_stage_single_robot.launch
$ rosrun rviz rviz -d `rospack find rosplan_stage_demo`/config/rosplan_stage_demo.rviz
The floor plan is meant to look something like this: