Road safety has become a major concern due to the increased rate of deaths caused by road accidents. For this purpose, intelligent transportation systems are being developed to reduce the number of fatalities on the road. A plethora of work has been undertaken on the detection of different styles of behaviour such as fatigue and drunken behaviour of the drivers; however, owing to complexity of human behaviour, a lot has yet to be explored in this field to assess different styles of the abnormal behaviour to make roads safer for travelling. This research focuses on detection of a very complex driver’s behaviours: ‘tafheet’, reckless and aggressive by proposing and building a driver’s behaviour detection model in the context-aware system in the VANET environment. Tafheet behaviour is very complex behaviour shown by young drivers in the Middle East, Japan and the USA. It is characterised by driving at dangerously high speeds (beyond those commonly known in aggressive behaviour) coupled with the drifting and angular movements of the wheels of the vehicle, which is similarly aggressive and reckless driving behaviour. Thus, the dynamic Bayesian Network (DBN) framework was applied to perform reasoning relating to the uncertainty associated with driver’s behaviour and to deduce the possible combinations of the driver’s behaviour based on the information gathered by the system about the foregoing factors. Based on the concept of context-awareness, a novel Tafheet driver’s behaviour detection architecture had been built in this thesis, which had been separated into three phases: sensing phase, processing and thinking phase and the acting phase. The proposed system elaborated the interactions of various components of the architecture with each other in order to detect the required outcomes from it. The implementation of this proposed system was executed using GeNIe 2.0 software, resulting in the construction of DBN model. The DBN model was evaluated by using experimental set of data in order to substantiate its functionality and accuracy in terms of detection of tafheet, reckless and aggressive behaviours in the real time manner. It was shown that the proposed system was able to detect the selected abnormal behaviours of the driver based on the contextual data collected. The novelty of this system was that it could detect the reckless, aggressive and tafheet behaviour in sequential manner, based on the intensity of the driver’s behaviour itself. In contrast to previous detection model, this research work suggested the On Board Unit architecture for the arrangement of sensors and data processing and decision making of the proposed system, which can be used to pre-infer the complex behaviour like tafheet. Thus it has the potential to prevent the road accidents from happening due to tafheet behaviour.

The objective of this thesis is to analyse driving behaviour and to characterize the effectsof an efficient way of driving, termed eco-driving, that enables the driver to reduce fuelconsumption and CO2emissions.The approach used to assess driving style is a collection of data from a CAN bus of acar equipped with OBD-II (on-board diagnostic) system. The driving experiment wasperformed for nine drivers who drove in a normal way or regular driving style and onedriver was an eco-driver who drove in an economical driving style. The driving routewas approximately 18.7 kms (which took between 25 to 30 minutes) in Halmstad city,Sweden.The drivers are compared using a statistical analysis of the driving parameters such as,speed, accelerator (gas pedal) and brake pressure, which are obtained from CAN busdata. A hierarchical clustering algorithm also used to classify the drivers based on theaverage result of the signals.In the results, a driving difference between the eco-driver and the normal drivers is visi-ble, most of the normal drivers have more or less similar behaviour. The average speed ofthe eco-driver lower than the normal drivers and the accelerator (gas pedal) result is alsoshown less usage by the eco-driver than the normal drivers. On the other hand, the eco-driver has braked more often than the normal drivers, but gently. Nevertheless, differenttraffic conditions during the experiment obstructs comparisons between the drivers.