A Fuzzy Logic Approach on the Evaluation of Driving Styles and Investigation of Drivability Calibration Effects


GAZI UNIVERSITY JOURNAL OF SCIENCE, vol.35, pp.668-680, 2022 (Scopus) identifier identifier identifier

  • Publication Type: Article / Article
  • Volume: 35
  • Publication Date: 2022
  • Doi Number: 10.35378/gujs.862867
  • Journal Indexes: Scopus, Academic Search Premier, Aerospace Database, Aquatic Science & Fisheries Abstracts (ASFA), Communication Abstracts, Compendex, Metadex, Civil Engineering Abstracts, TR DİZİN (ULAKBİM)
  • Page Numbers: pp.668-680
  • Keywords: Drivability calibration, Fuzzy logic algorithm, Driver classification, DRIVER BEHAVIOR, CLASSIFICATION
  • Istanbul Technical University Affiliated: Yes


Increased customer expectations lead the automobile manufacturers to develop innovative solutions, such as mode selection functions that provide different performance and comfort settings for the drivers. Almost all brands have different types of driving modes installed on their vehicles, such as sport mode, economy mode, off-road mode, etc. In the current technology, the mode selection is manually done by the driver. Thus, no effort is taken to match the driver style with available driving modes. However, driving mode selection should be done through an intelligent system such as vehicle control unit, in order to optimize customer expectations related to vehicle performance, driving comfort, and fuel consumption. This can be achieved by the analysis of all drivability maneuvers during any driving cycle. Based on the results of these analyses, drivability calibration settings of the vehicle can be adjusted depending on driver behaviors. In addition, fuel consumption can be improved using suitable calibration for each driver type. In this study, an experimental investigation is carried out in which vehicle data is collected for eleven different drivers at three different drivability calibrations. Furthermore, fuzzy logic algorithms are utilized in order to distinguish the driver characteristics. First, data from nine drivers are used in order to train the fuzzy logic approach. Then, the trained fuzzy logic scheme is used to assess the characteristics of two other drivers, who were left out in the training data set. Hence, it is aimed to obtain an intelligent prediction procedure that can estimate the characteristics of a driver based on their driving styles.