Conceptual design of a driving habit recognition framework

All drivers operate vehicles differently and demonstrate varying habits behind the wheel. Some drivers may execute vehicle maneuvers more cautiously than others, and some drivers may operate the vehicle with extreme inefficiencies. The habits developed by drivers can be viewed as a sequence or pattern of events that uniquely define the habitual behavior of the vehicle operator. In this paper, a conceptual design of a recognition system is discussed to classify sequences or patterns in vehicle data extracted from the Engine Control Unit in order to provide information about the vehicle operator's driving habits. Through an application of accepted pattern recognition techniques, Fuzzy Adaptive Resonance Theory, and Modern Control System Theory, a conceptual system framework was realized. To complement the conceptual design relationships between certain vehicle data parameters and certain human behaviors, models were developed to demonstrate these relationships created by this conceptual framework. These relationships were categorized and simulated in terms of vehicle safety and efficiency. Variables or factors were chosen to develop driving habit behavior models, such as wheel slippage, vehicle braking, fuel efficiency, and base or vehicle efficiency. The new conceptual framework was successfully validated through MATLAB simulations, consisting of 4 behavior models with a range of 11 variants. Evaluations of these behaviors provided the necessary feedback, via direct mapping of vehicle data points to a continuum of behavior types, to improve the vehicle operator's decision making.