The aim of this project is to develop a safe navigation framework for the TIAGo robot moving in a human crowd. Our approach is based on the paper of Vulcano et al., where a sensor-based scheme is presented. This scheme consists of two modules, the Crowd Prediction and Motion Generation modules, which run sequentially during every sampling interval. Our setup is implemented in Python using ROS and to validate our implementation multiple experiments are performed on Gazebo in scenarios of different complexity.
This project was developed for The Autonomous and Mobile Robotics Course 2022/2023 for the Master on Artificial Intelligence and Robotics, Sapienza University of Rome by:
- Antonio Purificato [@antoniopurificato]
- Flavio Galasso [@FlavioGalasso]
- Salvatore Falciglia [@falciglia]
For all the details about the theory or the implementative choices you can refer to the report or the slides;
To make sure everything is working properly make sure you have Ubuntu 18.04 with
ROS Melodic. Install catkin_tools, create a catkin workspace and clone this
repository in the src folder. Make sure you are compiling in Release mode
by properly setting your catkin workspace:
catkin config --cmake-args -DCMAKE_BUILD_TYPE=ReleaseBuild your code by running the following command:
catkin buildInstall the required packages for Python 2.7
pip install -r requirements.txtTo run the Gazebo simulation:
roslaunch crowd_navigation_gazebo tiago_gazebo.launch public_sim:=true end_effector:=pal-gripper world:=WORLDwhere WORLD is one of the worlds in the folder crowd_navigation_gazebo/worlds (e.g. 5_humans or 10_humans).
Remember that if you run regulation experiments (as 5_humans, 10_humans or 15_humans) you have to go to crowd_navigation_core/src/CommonVars.py and change TASK_TYPE=‘REG’ and also modify NUM_ACTORS depending on the number of actors in your simulation.
If you use all the other trajectory tracking experiments you have to change TASK_TYPE=‘8’. In this file you can also change some other parameters as the number of clusters K or you can toggle the printing of debug information from the different modules in your terminal.
To run the crowd prediction module:
roslaunch crowd_navigation_core crowd_prediction.launchTo run the Debugger:
roslaunch crowd_navigation_core debug_tool.launchTo run the motion generation module:
roslaunch crowd_navigation_core motion_generation.launch