This article proposes a new method on recognition of driving style, Roadway Type, and level of congestion
based on some of information available in electronic control unit (ECU) of vehicles. Vehicle speed,
engine speed, indicatory torque, acceleration pedal position, brake activity, and clutch pedal are used to
achieve this goal. Driving style is the driver's behavior that can be variable according to driver's personal
characteristics and level of congestion in the road. This paper is part of the research on «Driving Pattern
Recognition and Traffic Identification in Tehran for Control of Hybrid Vehicles» that is supported by
Irankhodro Powertrain Corporation (IPCO). In the proposed method, Neural Networks are used to classify
the Driving Style into three categories: calm, normal, and aggressive based on the features extracted from
ECU information.
Data collected in this research in calm, normal, and aggressive classes in collaboration with IPCO in
the real traffic conditions. Results show Driving Style can be recognized with neural network with high
performance, although Roadway Type and level of congestion didn't recognized well. Correct classification
rate that reached are 70% for Roadway Type and level of congestion, and above 90% for Driving Style. The
results attained in this research have many profits and can be used for control of vehicles (especially hybrid
electric vehicles) in future works.