No description available.

No description available.

No description available.

No description available.

No description available.

Sleep related vehicle accidents have been under publicised but remains as one of the main causes of road traffic accidents, as much as drink driving. This research aims to reduce this worldwide problem by developing a system to monitor fatigue driving. The thesis describes the research into the application of chromatic data processing techniques to detect early physiological and physical indicators of fatigue. Physiological factors that influence drivers are based on the duration of the drive, how much rest they have throughout the journey and the quality of sleep they had prior to the drive. The physiological indicator algorithm of the system is developed to take account of these factors and calculates the tiredness level. The chromatic technique is then used to analyse the results to establish trends and signatures of early fatigue situations where a warning system can be introduced. The chromatic signatures of fatigue have been established using results from 20 road tests conducted by professional drivers. Physical indicators such as early drowsy driving are detected by monitoring the behaviour of the vehicle. Micro sleep (e.g. head nodding, slow eye-blinking) can lead to lane drifting and vehicle swerving. These events are being regarded as early physical signs of sleepy driving. The main sensor for detecting the lateral yaw motion of the vehicle is a miniaturised gyroscope. Chromatic analysis is applied to the gyroscope output to identify and differentiate fatigue related events (e.g. swerves and lane drifting) from normal driving (e.g. left and right turning, roundabouts and bumpy roads) Combining the extracted information of the physiological and physical indicators, a Chromatic Fatigue Driving System can be developed as a tail safe system which monitors and alerts driver during critical fatigue conditions.

This thesis reports five studies that investigate the impact of trust on driver response to forward collision warning systems (FCWS). The experiments, while self-contained were conceived to relate together in a cohesive way. The first three studies investigated the relationship between alarm timing and driver performance in collision situations in a broad range of driving conditions. These studies also established trust models describing changes in driver subjective ratings of trust in response to alarm timing. It was found that an early alarm timing led to quick braking reaction times, resulting from prompt accelerator release. A middle alarm timing induced more consistent braking response than a control condition in which no alarms were presented. A late alarm timing had the potential to delay braking response when driving with long time headways. With respect to trust, early alarm timings induced higher levels of trust than late or middle alarm timings. Moreover the results suggest that the conflict between driver expectation of alarm performance and actual alarm timing results in decreased trust.

No description available.

Over the recent years, a considerable growth in the number of vehicles on the roads has been observed. This increases importance of vehicle safety and of minimization of fuel consumption, subsequently prompting manufacturers of cars to equip their products, with more advanced features such as Adaptive Cruise Control (ACC) and Collision Avoidance and Collision Warning System (CWS). In this thesis we concentrate on new methods for ACC. This work will include: Design of the simulation models suitable for this application, Investigation and design of suitable hybrid control algorithm by using classical and advanced control algorithm's consisting of the gain scheduling PI and Linear Quadratic (LQ) controllers, Design of the Nonlinear Model Predictive Control (NMPC) and the nonlinear Balance-Based Adaptive Control (B-BAC), Real-time implementation and tests of the algorithms by using NI Lab View Starter Kit Robot from National Instruments, Implementation and tests of the models and the controllers in MA TLAB/Simulink(R). The applications of the different control methods in the ACC are tested and compared against different traffic scenarios considering both velocity tracking (CC) and distance tracking (ACC) modes. Judging about the performance of ACC by utilizing the two advanced control methods; B-BAC and NMPC includes trade-offs between tracking-distance and velocity and the vehicle acceleration. However, both the B-BAC and the NMPC has demonstrated significantly smoother responses in controlling the throttle and the brake compared to PI control and linear MPC which in tum could improve the vehicle acceleration and fuel efficiency. The methods in order of producing better performance in terms of the values of control errors and their influences on fuel saving; NMPC, B-BAC, linear MPC and PI control. Improvement of fuel efficiency is investigated in this thesis through two approaches; first, by calculating the optimal control actions corresponding to the throttle and the brake signals through utilising the advanced control methods, second, by reducing the engine speed to idle speed during coast phase of the vehicle which causes the engine friction to be reduced. The engine speed can be reduced through transition between locked and unlocked states of the torque convertor. Possibility of achieving fuel efficiency through coasting in the vehicle has been investigated in the simulation and it has been demonstrated that longer coasting duration could be achieved i.e. more distance can be covered, and the fuel efficiency could be improved.