RT info:eu-repo/semantics/bachelorThesis
T1 Development of representative driving cycles of the Tenerife metropolitan area through clustering methods
A1 Armas Palmero, Carlos Enrique
K1 Machine learning
K1 Clusters
K1 Driving cycle
AB In recent years, owing to the evolution of science and technology, more efficientmethods of analysing high-dimensional data have been developed. Drawing on thisprogress, data science can be applied to the environmental sector, helping todetermine more accurately the impact of vehicles emissions on the environmentthrough representative driving cycles. This study aims to develop a methodology thathelps to build representative driving cycles from a data set collected in the metropolitanarea of Tenerife. The methodology proposed in this study consisted of the division ofdriving cycles into segments (Microtrips) and subsequently applying various clusteringalgorithms (k-means and Hierarchical clustering), following the application of adimensionality reduction methodology (t-SNE and PCA). The results were split intogroups with similar acceleration-related variables, representing the driving behaviours.. It was found that the highest quality clusters, assessed through silhouettecoefficients, Calinski-harabasz and Davies Bouldin index, resulted from the utilizationof a combination of t-SNE and k-means. The representative Microtrips were thenmerged to obtain the final cycle. This methodology seemed to be unable to satisfy thedesired cycle duration without affecting the data's representativeness. However, whenthe final cycle was compared to the data set, the resulting discrepancies were deemedacceptable.
YR 2021
FD 2021
LK http://riull.ull.es/xmlui/handle/915/24781
UL http://riull.ull.es/xmlui/handle/915/24781
LA es
DS Repositorio institucional de la Universidad de La Laguna
RD 24-abr-2024