Rapid Driving Style Recognition in Car-Following Using Machine Learning and Vehicle Trajectory Data

Table 1

Comparison of different driving behavior data collection approaches.

Collection Approaches	Advantages	Disadvantages
In-vehicle camera [29–33],; sensors, hardware [10, 24]	Real-world driving data; high-accuracy data; Access to driver’s personal data and vehicle control data	Expensive and time consuming; Lack of data in extreme and dangerous driving condition
Driving simulator [24–26]	Collect drivers’ behavior in designed and controlled driving scenarios	Driving behavior observed in the simulator may not always correspond to real-world driving
Traffic video [35–37]	Low expense; Easy to collect enormous vehicle data in a short time; Observe vehicle interaction in real traffic flow	Video extraction is challenging; No access to driver’s personal data and vehicle control data
Smartphone-based method [52–55]	Real-world driving data; Low expense in the smartphone	Data accuracy is critical