Human-like driver model for emergency collision avoidance using neural network autoregressive with exogenous inputs

Nurhaffizah, Hassan and Mohd Hatta, Mohammad Ariff and Hairi, Zamzuri and Sarah ‘Atifah, Saruchi and Nurbaiti, Wahid (2024) Human-like driver model for emergency collision avoidance using neural network autoregressive with exogenous inputs. In: Machine Intelligence in Mechanical Engineering. Woodhead Publishing Reviews: Mechanical Engineering Series . Woodhead Publishing, Sawston, United Kingdom, 359 -377. ISBN 978-0-443-18644-8

Preview	Pdf Machine Intelligence in Mechanical Engineering.pdf Download (490kB) | Preview
Preview	Pdf Human-like driver model for emergency collision avoidance_ABST.pdf Download (31kB) | Preview
	Pdf Human-like driver model for emergency collision avoidance.pdf Restricted to Repository staff only Download (1MB) | Request a copy

Abstract

One of the most difficult challenges in providing a reliable and safe autonomous collision avoidance maneuver is developing the driver model that provides the planning system with the risk of an impending collision. Over the last few years, researchers have extensively studied several methodologies to develop such a driver model. The human-like driver model, on the other hand, is a rarely discussed research topic. This work aims to develop a steering maneuver model in emergency collision avoidance that can imitate human emergency intervention. The neural network autoregressive with exogenous inputs (NNARX) is utilized to develop the model and autonomously predict the steering angle response. The work begins by collecting the avoidance maneuver driving data of the expert driver from the automaker company. In the data collection process, a controlled environment target scenario is used to ensure that all drivers encounter real emergencies. To investigate the performance of the developed model, a comparison prediction performance between the developed model and feed-forward neural network (FFNN) is presented. The finding shows that NNARX predicts the steering angle response with a lower prediction error during both training and testing compared to FFNN.

Item Type:	Book Chapter
Additional Information:	Indexed by Scopus
Uncontrolled Keywords:	Artificial neural network; Collision avoidance; Driver model; Human-like driver model; NNARX; Steering maneuver
Subjects:	T Technology > TJ Mechanical engineering and machinery T Technology > TS Manufactures
Faculty/Division:	Faculty of Manufacturing and Mechatronic Engineering Technology
Depositing User:	Mrs Norsaini Abdul Samat
Date Deposited:	03 Jul 2024 03:37
Last Modified:	03 Jul 2024 03:37
URI:	http://umpir.ump.edu.my/id/eprint/41786
Download Statistic:	View Download Statistics

Actions (login required)

View Item