Driver profiling-based anti-theft system.

With the advent and advancement in technology, automobiles have adopted various security features to which they adhere to. This ranges from manual locks to automated security and alerts. According to the results of multiple surveys, it is observed that these systems are often vulnerable to attacks and can get bypassed by notorious elements such as professional hackers and thieves. Since an automobile can always not be protected against these vulnerabilities, it is important to add an additional layer of security which always gives full control to the authentic user. This is where we recognize that the driving pattern of every individual is significantly unique from each other, and it can be used to differentiate a legitimate driver of the automobile from an illegitimate one. Thus, this project aims to analyze the driving pattern of each individual who drives the car and according to the users recognized by the system, it can alert the owner if an unknown person is detected driving the car. The idea is to improve upon the already existing models by experimenting with the different techniques of analysis. We plan to use the CAN Data Extraction and feature preprocessing as a form of data analysis and augmentation to create examples used in the training. This will help in improving the generalization of the model and give us more precise results to detect the user's driving patterns. To identify the user on the basis of his driving habits, algorithms like the Decision Tree, Random Forest and K-nearest neighbors have been used. Multiple models were laid out and trained according to all these different algorithms and the best results were taken in as the result. A simulation program has been used in which a car has been driven around under different circumstances with different obstacle courses to get the CAN data as detailed as possible. By the end of this paper, a more precise identification of different drivers' profiles according to their driving patterns has been established. Some of the driver's profiles have been stored in the system as authentic users, and if the driving pattern deviates from the ones recognized, an alert is sent to the owner of the car. The alert notifies him about the presence of an unknown driver. Therefore, at all times the owner can keep track of the people driving his car. [ABSTRACT FROM AUTHOR]
Copyright of AIP Conference Proceedings is the property of American Institute of Physics and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)