This thesis introduces a new integrated algorithm for the detection of lane-level irregular driving. To date, there has been very little improvement in the ability to detect lane level irregular driving styles, mainly due to a lack of high performance positioning techniques and suitable driving pattern recognition algorithms. The algorithm combines data from the Global Positioning System (GPS), Inertial Measurement Unit (IMU) and lane information using advanced filtering methods. The vehicle state within a lane is estimated using a Particle Filter (PF) and an Extended Kalman Filter (EKF). The state information is then used within a novel Fuzzy Inference System (FIS) based algorithm to detect different types of irregular driving. Simulation and field trial results are used to demonstrate the accuracy and reliability of the proposed irregular driving detection method.

Provides an integrated solution for the detection of lane level irregular driving behaviour Presents an extensive literature review to capture the state-of-the-art in the existing irregular driving monitoring algorithms Provides solutions to address basic underpinning issues such as system design, filter choice, vehicle motion model choice and driving pattern detection methods

Inhalt
Table of Contents.- Acknowledgements.- Declaration of Contribution.- Copyright Declaration.- Abstract.-Chapter 1 Introduction.- Chapter 2 Road Safety and Intelligent Transport Systems.- Chapter 3 State-of-the-art in Irregular Driving Detection.- Chapter 4 A New System for Lane Level Irregular Driving Detection.-Chapter 5 Testing, Analysis and Performance Validation.- Chapter 6 Conclusion and Recommendations for Future Work.- Publications Related to This Thesis.- Reference.- APPENDIX 1. Field Test Risk Assessment.