Proposal of an adaptive system for traffic signalized intersections in the city of Jaén

• Autor: Piedra Tineo, José Luis
• Formato: artículo
• Fecha de Publicación: 2024
• Institución: Universidad Nacional de Jaén
• Repositorio: Pakamuros
• Lenguaje: español
• OAI Identifier: oai:unj:article/347
• Enlace del recurso: http://revistas.unj.edu.pe/index.php/pakamuros/article/view/347
• Nivel de acceso: acceso abierto
• Materia: Sistema adaptativo; intersecciones semaforizadas; algoritmo; programa; demora; Adaptive system; traffic signalized intersections; algorithm; program; delay

The objective of the research was to propose an adaptive system for traffic signalized intersections in the city of Jaén, based on algorithms and program execution. The design of the study was quantitative and experimental. Therefore, field data and primary sources were collected and analyzed to later build algorithms using the Knowlegde Discovery Databases methodology, going on to create an information base, train and validate with prediction algorithms the models that included selected variables, then predict times, which classify to a level of service; and finish with the development of a program using Visual Studio compatible with the generated algorithms, all this bearing in mind the Webster method and Highway Capacity Manual. The intersections were classified in their current situation with delays of 11.41 and 8.32 seconds, and levels of service "B", and "A"; two high precision models using the RandomForest algorithm with R2 of 0.995 and 0.996, estimating delay times and optimal cycle lengths; and an adaptable executable program that calculates and optimizes delays, optimal cycles and levels of service. This system has an R2 of 99.55% with the RandomForest algorithm capable of estimating, improving, and automating the operability of a traffic-light intersection.