Cooperative strategy to reduce path length in risky environments

Abstract

Objective: Design an artificial intelligence system based on information from the environment that can recommend the shortest path to an individual or vehicle, or robot that moves between two points with the lowest risk of contagion with coronavirus COVID-19. Methodology: The cooperative strategy for path reduction involves a management and monitoring system and two explorer agents. Explorer agents are equipped with path planning algorithms (GBFS and A*) enhanced with incremental heuristics in order to find two different sets of preliminary paths (the first in direction start-goal and the second in the opposite direction). Subsequently, a management and monitoring system estimates a preliminary shortest path for each path planner then obtains a shortest path by comparing the paths attained with the path planners. This research emerges within the field of distributed intelligence in robotics to determine the benefits of teamwork interactions compared to individual work. In this study, 300 tests that involve the cooperative strategy were executed using ten different environments. Results: The results of this paper illustrate that in 79 % of analyzed situations, definitive shortest estimated paths obtained by cooperative strategy outperformed preliminary paths found individually by path planners. Over 20.5 % of tested cases yielded significant path reductions (greater than 100 % in relation to the shortest definitive path). Conclusions: In this work, an artificial intelligence system was designed, whose tests show a good performance. The intelligent system uses Distributed Intelligence implemented in a cooperative team formed by a management and monitoring system and two explorer agents, who, based on information from the environment, recommend the shortest path to an individual or vehicle or robot who wants to travel between two points located in an environment at risk of contagion with coronavirus COVID-19. Financing: This work was supported in part by the Universidad Tecnológica de Pereira through Vicerrectoría de Investigaciones Innovación y Extensión, Project name: Sistema de obtención de rutas más seguras bajo situación de pandemia caso covid-19, Project code: 3-20-11, and in part by the Universidad Nacional de Colombia.

Keywords

• agent-based modeling
• autonomous robots
• collaborative work
• collision avoidance
• cooperative systems
• multiagent systems
• navigation