Year: 2017
Type: dissertation
Abstract: Driver assistance systems such as Adaptive Cruise Control or Lane Keeping Assist are nowadays part of the driving experience. The research towards these technologies\naims to increase the driver’s safety and comfort. These systems are rapidly becoming\nmore functional with the help of additional sensors and technology. By default, they\nare tuned by test engineers according to fixed requirements, or in other words “one\nsize fits all” model. However, to increase acceptance of these assist functions, drivers\nwant to recognize themselves in the way the car is driven.\nA generic high-level framework for personalized driver-assistance systems is proposed\nin this study. The framework is implemented with a machine learning method known\nas neural networks on two different subtasks, to learn the longitudinal acceleration\nand headway distance preferred by the driver; and to predict the future steering\nwheel inputs provided by the driver. Experiment has been performed on predicting\nthe driver’s steering wheel inputs, using a long short-term memory neural network\nand a multilayer perceptron. The data used in this study has been gathered through\nsimulations with IPG CarMaker.\nThe results show that driver’s steering wheel inputs can be predicted accurately up\nto 10 seconds. Furthermore, the multilayer perceptron was trained additionally by\nsimulating real time learning. This real time learning did not prove to increase the\nperformance significantly.\nKeywords:
Author: Hjörtur Hjartarson
Language: English
Subfield: Automotive Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W2728461071