Driver evaluation based on classification of rapid decelerating patterns
暂无分享,去创建一个
This paper presents a novel method to evaluate risk levels of driving behaviors based on acceleration patterns while braking. Acceleration patterns were recorded with drive recorders mounted on such vehicles as taxis and trucks to detect "events," i.e., remarkable scenes while driving such as rapid acceleration and deceleration. They also captured video images of the vehicles in front of the drivers. The event data and video resources are used to analyze the causes of crashes and near-miss incidents and to evaluate the risk levels of the driving behaviors of individual drivers. Conventional driver evaluation methods generally use the frequencies of event occurrences detected with a certain acceleration threshold. Yet it remains unclear how the driver depresses and releases the brake pedal at each event or how dangerous each event is. To make it apparent, we introduce a method that characterizes braking patterns based on the time series of the acceleration signals around the events. Events of rapid deceleration were classified into four typical types of braking patterns based on how the brake pedal was depressed and released. Driver risk levels of braking actions were evaluated based on these four braking patterns from different points of view. We show results applying our proposed method to evaluate the driving behaviors of 35 drivers.
[1] Brian N. Fildes,et al. Young Novice Driver Subtypes: Relationship to High-Risk Behavior, Traffic Accident Record, and Simulator Driving Performance , 1999, Hum. Factors.
[2] Robert M. Gray,et al. An Algorithm for Vector Quantizer Design , 1980, IEEE Trans. Commun..
[3] C. Miyajima,et al. Abrupt Steering Detection Based on the Road Construction Ordinance and Vehicle Acceleration Captured with Drive Recorders , 2008, 2008 3rd International Conference on Innovative Computing Information and Control.