A number of advanced driver assistance systems have recently been developed and implemented to improve road safety and driving experience. However, most of them are designed based on generic driver behaviours, where driving characteristics and preferences of a specific driver are neglected. Therefore, it is not surprising that conventional driving status recognition systems may lead to degraded performance for a particular driver. To this end, driving status recognition systems should be personalised to enhance system performance. This thesis aims to develop driving status recognition systems of high accuracy in order to ensure driving safety.