Glasius bio-inspired neural networks based UV-C disinfection path planning improved by preventive deadlock processing algorithm. (January 2023)

• Record Type: Journal Article
• Authors: Rodrigo, Daniel Vicente; Sierra-García, J. Enrique; Santos, Matilde
• Is Part Of: Advances in engineering software. Volume 175(2023)
• Publication Date: 2023-01
• Subjects: Complete coverage path planning -- Mobile robot -- UV-C -- Deadlocks -- Escape routes; Computer-aided engineering -- Periodicals; Engineering -- Computer programs -- Periodicals; Engineering -- Software -- Periodicals; Periodicals; 620.0028553
• DOI: 10.1016/j.advengsoft.2022.103330
• Languages: English
• ISSNs: 0965-9978

Abstract

Abstract: The COVID-19 pandemic made robot manufacturers explore the idea of combining mobile robotics with UV-C light to automate the disinfection processes. But performing this process in an optimum way introduces some challenges: on the one hand, it is necessary to guarantee that all surfaces receive the radiation level to ensure the disinfection; at the same time, it is necessary to minimize the radiation dose to avoid the damage of the environment. In this work, both challenges are addressed with the design of a complete coverage path planning (CCPP) algorithm. To do it, a novel architecture that combines the glasius bio-inspired neural network (GBNN), a motion strategy, an UV-C estimator, a speed controller, and a pure pursuit controller have been designed. One of the main issues in CCPP is the deadlocks. In this application they may cause a loss of the operation, lack of regularity and high peaks in the radiation dose map, and in the worst case, they can make the robot to get stuck and not complete the disinfection process. To tackle this problem, in this work we propose a preventive deadlock processing algorithm (PDPA) and an escape route generator algorithm (ERGA). Simulation results show how the application of PDPA and the ERGA allow to complete complex maps in an efficient way where the application of GBNN is not enough. Indeed, a 58% more of covered surface is observed. Furthermore, two different motion strategies have been compared: boustrophedon and spiral motion, to check its influence on the performance of the robot navigation. Highlights: Model to test Complete Coverage Path Planning (CCPP) methods for UV-C disinfection. Neural Networks, speed controller, and pure pursuit control to enable UV-C CCPP. Development of Preventive Deadlock Processing Algorithm (PDPA) for UV-C CCPP. Development of Escape Route Generator Algorithm (ERGA) for UV-C CCPP. PDPA+ERGA allow to cover complex maps where the application of GBNN is not enough.