A Threat Predictor System of Dynamic Environments by using a Multi-Robot fleet

3D Simulation of a set of robots used to patrol and and secure dynamic environments. This software system is composed of ROS packages, topics, and services. The scenarios are simulated using the Gazebo 3D simulator and the Pedsim pedestrian simulator.

Features

• 3D Simulation the Robotnik Summit-XL mobile robot and all its sensors in Gazebo.
• Coordination of a multi-robot system using Finite State Machines.
• Pedestrian walking simulation using social force models.
• Threats detection and prediction based on cooperative Growing Hidden Markov Models (GHMMs).
• Extensive and detailed visualization using Rviz.

Requirements

• ROS with the Navigation Stack (currently tested on ROS Indigo and Ubuntu 14.04)
• Gazebo 2.2
• C++11 compiler
• Python
• NumPy

Installation

cd [workspace]/src
git clone https://github.com/amattag/mrs_predictor.git
cd ..
catkin_make

Sample usage

roslaunch mrs_monitor mrs_monitor.launch world_file:="$world_file_name" robots_number:=$number_of_robots

Contributors

• Antonio Matta-Gómez
• Antonio Barrientos Cruz, Phd.

This meta-package is a work in progress, currently it is being used for testing in a Phd Thesis at Universidad Politécnica de Madrid. It holds a BSD license.