Eine Plattform für die Wissenschaft: Bauingenieurwesen, Architektur und Urbanistik
Eine Plattform für die Wissenschaft: Bauingenieurwesen, Architektur und Urbanistik
MFCC Delta–Delta Energy Feature Extraction for Clustering of Road Surface Types
https://orcid.org/http://orcid.org/0000-0002-9982-5218
Identification of road surfaces can help drivers controlling the speed of their vehicles and avoid accidents. Smartphones sensors are widely used to detect road obstacles and alert drivers in advance. However, use of data stream generated by smartphone sensors for road surface detection needs to be studied further. This paper presents a novel way of combination of Mel Frequency Cepstrum Coefficients (MFCC), typically used in audio signal processing, Delta–Delta Energy Feature and dynamic time warping technique for unsupervised learning—clustering to divide data set of accelerometer acceleration values into 4 different types of roads, commonly found in India viz, Unpaved, Bituminous Roads, Concrete Roads, Paver blocks road. Promising results based upon silhouette values in the range 0.6–0.7, Lower Davies–Bouldin index (DBI) (≈ 0.15–0.51) and Higher Dunn’s Index (≈ 0.24–0.54) indicates probability of correct cluster assignment of extracted data points, with reduction of any size of data set to fixed size of 12 Delta–Delta Energy features.
MFCC Delta–Delta Energy Feature Extraction for Clustering of Road Surface Types
https://orcid.org/http://orcid.org/0000-0002-9982-5218
Identification of road surfaces can help drivers controlling the speed of their vehicles and avoid accidents. Smartphones sensors are widely used to detect road obstacles and alert drivers in advance. However, use of data stream generated by smartphone sensors for road surface detection needs to be studied further. This paper presents a novel way of combination of Mel Frequency Cepstrum Coefficients (MFCC), typically used in audio signal processing, Delta–Delta Energy Feature and dynamic time warping technique for unsupervised learning—clustering to divide data set of accelerometer acceleration values into 4 different types of roads, commonly found in India viz, Unpaved, Bituminous Roads, Concrete Roads, Paver blocks road. Promising results based upon silhouette values in the range 0.6–0.7, Lower Davies–Bouldin index (DBI) (≈ 0.15–0.51) and Higher Dunn’s Index (≈ 0.24–0.54) indicates probability of correct cluster assignment of extracted data points, with reduction of any size of data set to fixed size of 12 Delta–Delta Energy features.
MFCC Delta–Delta Energy Feature Extraction for Clustering of Road Surface Types
https://orcid.org/http://orcid.org/0000-0002-9982-5218
Identification of road surfaces can help drivers controlling the speed of their vehicles and avoid accidents. Smartphones sensors are widely used to detect road obstacles and alert drivers in advance. However, use of data stream generated by smartphone sensors for road surface detection needs to be studied further. This paper presents a novel way of combination of Mel Frequency Cepstrum Coefficients (MFCC), typically used in audio signal processing, Delta–Delta Energy Feature and dynamic time warping technique for unsupervised learning—clustering to divide data set of accelerometer acceleration values into 4 different types of roads, commonly found in India viz, Unpaved, Bituminous Roads, Concrete Roads, Paver blocks road. Promising results based upon silhouette values in the range 0.6–0.7, Lower Davies–Bouldin index (DBI) (≈ 0.15–0.51) and Higher Dunn’s Index (≈ 0.24–0.54) indicates probability of correct cluster assignment of extracted data points, with reduction of any size of data set to fixed size of 12 Delta–Delta Energy features.
MFCC Delta–Delta Energy Feature Extraction for Clustering of Road Surface Types
Int. J. Pavement Res. Technol.
Jawale, Anupama (Autor:in) / Magar, Ganesh (Autor:in)
International Journal of Pavement Research and Technology ; 16 ; 631-646
01.05.2023
16 pages
Aufsatz (Zeitschrift)
Elektronische Ressource
Englisch
Online Contents | 1996
Synthesis of Cap delta^0, Cap delta^5 and Cap delta^7 -Sterylglycosides
| British Library Online Contents | 1994
Classification of the Urbanization Types in the Yangtze River Delta
| British Library Conference Proceedings | 2012
Online Contents | 1999