RT Journal Article
T1 Glasius bio-inspired neural networks based UV-C disinfection path planning improved by preventive deadlock processing algorithm
A1 Rodrigo-Muñoz, Daniel Vicente
A1 Sierra-García, Jesús Enrique
A1 Santos Peñas, Matilde
AB 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.
PB Elsevier
YR 2023
FD 2023
LK https://hdl.handle.net/20.500.14352/108136
UL https://hdl.handle.net/20.500.14352/108136
LA eng
NO Rodrigo DV, Sierra-García JE, Santos M. Glasius bio-inspired neural networks based UV-C disinfection path planning improved by preventive deadlock processing algorithm. Advances in Engineering Software. 2023 Jan 1;175:103330.
DS Docta Complutense
RD 8 abr 2025