Employing Deep Learning to Improve Driver Acceptance of Automated Features

2022-01-0182

03/29/2022

Authors Abstract
Content
The auto industry is advancing rapidly, offering quite a few features to assist the driver and to make the user experience seamless. However, the response towards such advances in automotive technology differs across drivers. An additional area of focus is improving the driving experience based on driver behavior. Personalization can be a crucial factor in enriching the driving experience. Lane changing is one such driving phenomenon where driver behavior plays a vital role and currently, driver-assist features like blind-spot warning help the driver to make a judgment. As drivers are getting used to driver assist and partial automation technologies, we are progressing towards high automation technology where the vehicle will be able to drive itself smoothly in ideal conditions with the driver behind the wheels, ready to take control when needed. In this paper, we aim to introduce an approach for a level 3 personalized autonomous lane changing model. We are proposing a deep learning-based approach for an autonomous lane change in the ideal conditions. To personalize the lane-changing experience, we are proposing a way to learn and optimize lane-changing operation based on driver actions when the vehicle is not in an autonomous mode. Exploiting the advances in machine learning and computer vision discipline, the model will learn driving patterns specific to a driver based on the parameters such as driver’s gestures prior to lane change, the distance to other vehicles, the frequency of lane change, and the vehicle parameters like steering angle, acceleration profile to gauge the lane changing maneuver more accurately.
Meta Tags
Topics
Affiliated or Co-Author
Details
Citation
Lakhkar, R., and Talty, T., "Employing Deep Learning to Improve Driver Acceptance of Automated Features," SAE Technical Paper 2022-01-0182, 2022, .
Additional Details
Publisher
Published
Mar 29, 2022
Product Code
2022-01-0182
Content Type
Technical Paper
Language
English