Tavano, Fabrizio (2023) Multi-Robot Distributed Strategies for Priority-Based Sanitization of Railway Stations. [Tesi di dottorato]

Anteprima

Testo Tavano_Fabrizio_35.pdf Download (14MB) | Anteprima

Abstract

Recent studies have highlighted that shared areas of the railway stations may be locations where there may be contagion by viruses and bacteria during the periods of a pandemic like in the case caused by SARS-CoV-2 diffusion. The propagation of SARS-CoV-2 disease particularly damaged the railway sector because the transportation demand had registered a drastic reduction. People prefer to travel only if strictly necessary, using private cars in alternative to public services such as railway transportation. In this Thesis, answering the Infrastructure Manager Rete Ferroviaria Italiana’s request, we propose sanitization strategies that coordinate a team of robots in a dynamic railway station without interrupting the preexisting human activities such as transportation services, catering services, shopping, and ticketing. Since every important Italian station is equipped with a WiFi Meraki Cisco System Network, our aim is to exploit such infrastructure to monitor the positions of mobile devices (tablets and phones) and to evaluate the most crowded areas of the station to be sanitized. Specifically, our approach is to define a heatmap whose colored zones indicate the presence of contamination (prioritized zones) caused by the aggregation of visitors, which can be used by robots as guidance during sanitization activities. In Chapter 3, we propose a multi-robot approach to sanitize railway stations based on a distributed Deep Q-Learning technique. A team of cleaning robots - each endowed with a robot-specific convolutional neural network - learns how to effectively cooperate and sanitize the station’s areas according to the associated priorities. In Chapter 4 we extend the previous framework allowing to define teams of robots having different sanitizing strategies/capabilities. In Chapter 5 we illustrate a dis- tributed framework, where a centralized server uses the Hierarchical Mixed Integer Linear Programming to coordinate the robots assigning different zones where the cleaning has higher priority; thanks to the MPC-MILP approach, we use historical data about the distribution of people and the knowledge about the transportation service of the station, to predict the future dynamic evolution of the position of people in the environment and the spreading of the contaminants. In Chapter 6, we propose a multi-robot online sanitization strategy that combines the Bioinspired Artificial Cockroach Colony Strategy with the 2-type Fuzzy Logic to coordinate together a team of robot sanitizers. We tested our solution considering real data collected by the WiFi network of the main Italian railway station, Roma Termini shared by Rete Ferroviaria Italiana S.p.A. We compared our results together with other methods conventionally proposed for sanitization, applied to the same scenario. The approaches described in this Thesis may also be applied in every indoor public location as shopping centers, restaurants, and industrial sheds, if WiFi Service is available with visitors’ positioning information, with a correct number of robots selected considering the total surface of the environment.

Downloads

Actions (login required)

Modifica documento