This project developed in collaboration with Francesco Zampirollo and Emma Scinti Roger aims to implement an obstacle avoidance control analizing actors attention on Robotnik Kairos simulator.
git clone https://github.com/StePoli-00/Smart-Robotics-Project.git
cd Smart-Robotics-ProjectWarning
for our experience conda and pip environment did not work properly with ROS, we recomand you to install python package in your file system even if is not the best practice. If you are expert with ROS workspace and virtual env, follow this command
conda create --name <env_name>
conda activate <env_name>
pip install -r requirements.txt
conda install python (our version is 3.12)Important
BUILD ALWAYS with catkin_make specify python3
catkin_make -DPYTHON_EXECUTABLE=/usr/bin/python3
cd pedestrian_detectionNote
Before to start the simulation, copy business_man model located at Smart-Robotics-Project/simulation/models/business_man inside <username>/home/.gazebo/models (CTRL+H to see hiding file)
- Just running the simulation, all the required node will be instantiated
roslaunch rbkairos_sim_bringup rbkairos_complete.launchmost of our work is located in catkin_ws/src, inside you can find:
- actor_collision: plugin to make actor model collidable inside gazebo
- pedestrian_detection: rosnode to detect pedestrian pose and send it to robot navigation algorithm
- get_current_pose: get robot position and update it
- mpc algorithm: rosnode to control robot navigation (specifically algorithm_v3.py)
- robot_goal_mover: rosnode that calculate distance from the target
if you want to test different world in catkin_ws/src/rbkairos_sim/rbkairos_gazebo/worlds modify the following line 6-7 of rbkairos_complete.launch located in catkin_ws/src/rbkairos_sim/rbkairos_sim_bringup/launch/
<arg name="launch_rviz" default="true"/>
<arg name="gazebo_world" default="$(find rbkairos_gazebo)/worlds/`filename`.world"/>