A multi-robot deep Q-learning framework for priority-based sanitization of railway stations.

Area of Science:

• Robotics and Artificial Intelligence
• Public Health and Epidemiology
• Network Engineering

Background:

• The COVID-19 pandemic highlighted the need for effective sanitization in high-traffic public spaces like railway stations.
• Traditional cleaning methods may not be sufficient for dynamic crowd levels and rapid pathogen spread.
• Efficient and automated disinfection strategies are crucial for maintaining public safety.

Purpose of the Study:

• To develop and evaluate a multi-robot system for automated sanitization of railway stations.
• To leverage WiFi network data for real-time crowd density estimation and disinfection prioritization.
• To implement a distributed Deep Q-Learning approach for cooperative robot navigation and task allocation.

Main Methods:

• A multi-robot framework employing distributed Deep Q-Learning for cooperative sanitization.
• Utilizing anonymous WiFi network data to create dynamic heatmaps of station areas based on crowd density.
• Employing robot-specific convolutional neural networks for effective area sanitization according to priority levels.
• Simulation in a realistic scenario of Roma Termini railway station.

Main Results:

• The proposed multi-robot system demonstrated effective cooperation and prioritized sanitization based on crowd density.
• The framework's scalability was assessed with varying numbers of cleaning robots.
• Performance was validated using real WiFi data from a major railway station.

Conclusions:

• The developed multi-robot sanitization approach is a promising solution for maintaining hygiene in large public transport hubs.
• The integration of WiFi data and AI-driven robot coordination offers an efficient method for dynamic disinfection.
• This technology can significantly contribute to public health resilience in transportation systems.