An Artificial Intelligence-Based Vehicular System Simulator

An Artificial Intelligence-Based Vehicular System Simulator

Marvin T. Chan (University of Regina, Canada), Jonathan T. Chan (University of Regina, Canada), Christine Chan (University of Regina, Canada) and Craig Gelowitz (University of Regina, Canada)
DOI: 10.4018/978-1-7998-0951-7.ch034


This paper presents a vehicular system simulator, which enables the human player to race a car against three system-controlled cars in a three-dimensional road system. The objective of the vehicular system simulator is not to support defeating the opponent in a car race, but to provide the player with a challenging and enjoyable racing experience. Therefore, it is important that the system simulates human driving behavior and adopts cognitive computing. The paper discusses development of the vehicular system simulator using the artificial intelligence (AI) techniques that are supported in the game engine of Unity. The design and implementation of the vehicular system simulator are presented. The discussion includes some possible extensions of the current version of the system so that it can be adapted to be a simulation system for education purposes
Chapter Preview

2. Background Literature

Some research studies on cognitive vehicular systems are briefly summarized as follows. Xie & Xu (2015) provided an overview of the algorithms and technologies useful for extracting static properties of vehicles from videos so as to support criminal and traffic violations. With the rapid growth of the video surveillance technology, a large database of video and images has been created. The database can become an important resource for police in performing traffic events analysis: crucial information such as speed and license plate of vehicles can be extracted and examined. Yoshizawa & Iwasaki (2015) examined a type of driving behavior, called “aimless driving”, which is responsible for a large number of fatal traffic accidents in Japan. The study investigates this “aimless driving” behavior by giving subjects non-visual secondary tasks of four difficulty levels while they watch pedestrians and their responses to objects in the road on which they travel are measured. The results indicated that even non-visual tasks influence eye movement and the subjects were not able to react well to objects on the road. Xu et al. (2014) studied the use of semantic technologies for annotating videos so as to enhance comprehensibility of the videos. Their work proposed a video annotation platform, which supports users by providing a search interface of annotated video resources. Some preliminary results indicated that applying semantic technologies for annotation of videos can enhance reusability, scalability, and extensibility of the video resource and support increased adoption of this resource for traffic events analysis.

Complete Chapter List

Search this Book: