Research On Dangerous Situation Identification Of Driving Behavior

Driver is the most important factor in traffic accidents because vehicle is the most important transportaion.Now,the Driving Safety Assistance Systems mainly use a more fixed risk decision-making program,its auxiliary strategy and control parameters do not consider the driver’s current driving behavior and specific environmental conditions;The Intelligent Driving Safety Assistance Systems need to be able to identify the driver’s current driving behavior,and combine with the current vehicle operating conditions,giving the auxiliary control strategy close to the driver and the current situation of the vehicle.If the Driving Safety Assistance Systems’ control and risk decision-making can be combined with the driver’s current driving behavior,and not interfere with driver’s normal driving maneuvers,the Driving Safety Assistance Systems’ intelligence level will be greatly improved.One of the key technologies to achieve this goal is to study the identification method of dangerous behavior in driving behavior.Based on the development of above issues,the main contents of this paper are as follows:(1)By collecting the driving behavior manipulation data from the driving simulator and using the data-mechanism combination method,the characteristic of the driver’s manipulation action can be got and be drawn into the driving manipulation mode hierarchy diagram in the way of the directed graph.(2)According to the characteristics of the data of driving manipulation,the Adaptive Resonance Theory neural network algorithm is adaptively changed,and the corresponding driving behavior identification method is studied.(3)According to the classification of driving behavior whether to keep the lane or not,the method of trajectory prediction of lane keeping behavior based on the state of the vehicle and the method of trajectory prediction of lane-change behavior integrating the state of the vehicle and the driver’s intention are studied in this paper.(4)Based on the theory of minimum safety distance and the consideraion of the random distribution of predicted trajectories,the method of calculating the risk coefficient using probability density function is studied,and the relationship between driving manipulation action and hazard coefficient can be analyzed by means of the principle of the driving manipulation mode hierarchy diagram.Research indicates:(1)The driving manipulation mode based on the characteristics of driving behavior manipulation can intuitively reflect the sequence relationship of the driving behavior.(2)The driving behavior recognition method based on the Adaptive Resonance Theory variant algorithm can make a good recognition effect result.(3)Both types of the trajectory prediction methods have good prediction effects in their driving behaviors.(4)The hazard identification method based on trajectory prediction can calculate the hazard coefficient and its distribution in advance,and can analyze the inherent relationship between the driving manipulation action and the hazard coefficient by using the hazard driving manipulation mode hierarchy diagram,it can provide more diversified information for the Driving Safety Assistance Systems.