An empirical evaluation of importance-based feature selection methods for the driver identification task using OBD data

Priyadharshini, G.; Ukrit, M. Ferni

doi:10.1007/s13198-022-01695-1

Cited by 4 publications

(7 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…On the other hand, there are different feature selection techniques such as Principal Component Analysis (PCA) which are often used to find feature patterns associated with a certain behavior. Priyadharshini et al [ 52 ] conducted research work to identify drivers through data extracted from the On Board Diagnostic II sensor. The main objective of the work is to extract the most important features instead of entering all the features in the machine learning algorithms using PCA, obtaining an accuracy of 75% reducing times, and increasing accuracy.…”

Section: Resultsmentioning

confidence: 99%

Driver Identification Using Statistical Features of Motor Activity and Genetic Algorithms

Espino-Salinas

Luna-García

Celaya-Padilla

et al. 2023

Sensors

View full text Add to dashboard Cite

Driver identification refers to the process whose primary purpose is identifying the person behind the steering wheel using collected information about the driver him/herself. The constant monitoring of drivers through sensors generates great benefits in advanced driver assistance systems (ADAS), to learn more about the behavior of road users. Currently, there are many research works that address the subject in search of creating intelligent models that help to identify vehicle users in an efficient and objective way. However, the different methodologies proposed to create these models are based on data generated from sensors that include different vehicle brands on routes established in real environments, which, although they provide very important information for different purposes, in the case of driver identification, there may be a certain degree of bias due to the different situations in which the route environment may change. The proposed method seeks to intelligently and objectively select the most outstanding statistical features from motor activity generated in the main elements of the vehicle with genetic algorithms for driver identification, this process being newer than those established by the state-of-the-art. The results obtained from the proposal were an accuracy of 90.74% to identify two drivers and 62% for four, using a Random Forest Classifier (RFC). With this, it can be concluded that a comprehensive selection of features can greatly optimize the identification of drivers.

show abstract

Section: Resultsmentioning

confidence: 99%

Driver Identification Using Statistical Features of Motor Activity and Genetic Algorithms

Espino-Salinas

Luna-García

Celaya-Padilla

et al. 2023

Sensors

View full text Add to dashboard Cite

show abstract

“…The pre-processing begins with filtering to remove noise [1,5]. Data cleaning [30,31,32]. Feature scaling [33,32,5].…”

Section: Fleet Managementmentioning

confidence: 99%

“…Data cleaning [30,31,32]. Feature scaling [33,32,5]. Time windowing divides the data into smaller segments using either overlapping or fixed window sizes, enabling the extraction of distinctive features from each segment [34,35].…”

Section: Fleet Managementmentioning

confidence: 99%

“…In feature selection for driver identification, [67] employed Extra Trees and achieved a recognition accuracy rate of 88%. [32] introduced a feature selection approach involved extracting important features through techniques such as Coefficient, tree-based methods, and PCA. The results indicated that the tree-based method, particularly the XGBoost algorithm, produced the greatest results, with a 75% accuracy and reduced training time.…”

Section: Tree-based Feature Importancementioning

confidence: 99%

“…[5], [56], [13], [77], [74], [37], [45], [24], [11], [19] Tree [67], [32], [74], [28], [37], [24], [57] Correlationbased Feature Selection Identifies linear relationships between features and the target. Quick and straightforward to compute.…”

Section: Support Vector Machine (Svm)mentioning

confidence: 99%

See 2 more Smart Citations

A Comprehensive Review: Analysis of Machine Learning, Deep Learning, and Large Language Model Techniques for Revolutionizing Driver Identification

Sohail,

Teh,

Khan

et al. 2024

Preprint

View full text Add to dashboard Cite

Driver identification is crucial for various applications including automotive security, law enforcement, and the ride-sharing industry, as well as for advanced driver assistance systems, fleet management, and usage-based insurance. Machine learning and deep learning techniques have emerged as promising approaches for accurate identification, yet a comprehensive analysis of existing methods remains unattainable. Despite significant research efforts, no in-depth survey has thoroughly evaluated and compared existing driver identification techniques, impeding the development of optimal solutions. This gap restricts our understanding of their strengths, weaknesses, and potential impact. This paper addresses this critical gap by comprehensively reviewing and analyzing existing driver identification techniques. We delve into various methodologies, including preprocessing feature extraction, classification algorithms, and deep learning architectures. We critically evaluate their performance, highlighting unique features, potential advantages, and limitations. Additionally, we proposed the future framework for driver identification with Large Language Model (LLM) and explored the developing potential of LLM in this domain. We employ relevant research articles published in prominent scientific databases. Our analysis reveals diverse driver identification techniques, each offering unique advantages and disadvantages. Traditional methods like Support Vector Machines and Random Forests provide reliable performance, while deep learning architectures achieve higher accuracy but require larger datasets and computational resources. We also identify potential synergies between established techniques and emerging technologies like LLM, suggesting promising avenues for future research. We propose several key research directions based on our findings to further advance driver identification accuracy and robustness. These include exploring hybrid approaches combining traditional and deep learning methods and investigating transfer learning techniques for efficient adaptation to new data sets. This comprehensive review provides a valuable resource for researchers and practitioners interested in driver identification. We highlight the strengths and weaknesses of existing techniques, identify potential for future advancements, and suggest promising research directions to ultimately achieve highly accurate and reliable driver identification systems.

show abstract

A systematic review of the use of in-vehicle telematics in monitoring driving behaviours

Boylan,

Meyer,

Chen

2024

Accident Analysis & Prevention

View full text Add to dashboard Cite

An empirical evaluation of importance-based feature selection methods for the driver identification task using OBD data

Cited by 4 publications

References 12 publications

Driver Identification Using Statistical Features of Motor Activity and Genetic Algorithms

Driver Identification Using Statistical Features of Motor Activity and Genetic Algorithms

A Comprehensive Review: Analysis of Machine Learning, Deep Learning, and Large Language Model Techniques for Revolutionizing Driver Identification

A systematic review of the use of in-vehicle telematics in monitoring driving behaviours

Contact Info

Product

Resources

About