Research On Braking Intention Recognition Based On The Modeling Of The Dilemma Zone Of Signal Intersection

In recent years, the occurrence of more traffic accidents at intersections and accident will inevitably lead to regional traffic blocking lag, order confusion, waste of resources and other conditions, therefore, reduce the intersection traffic accident has become a hot research topic, and it is of great significance. On signalized intersection at the driver of the vehicle braking behavior research is one of the active vehicle safety research important content, rate of one of the important methods to reduce the occurrence of intersection accidents. According to the research of traditional automobile braking intention recognition is still confined to the operation of the driver part, automobile driving itself is a closed loop system, composed of man car road, driving environment and driving itself is composed of a driver vehicle road closed-loop system, while the traditional method is unable to determine the driver decision making complex make detailed and accurate quantification in this paper.In this paper, a simulation driving simulation platform consisting of Pre Scan and G27 is constructed. On the basis of summarizing the research results of driver’s braking behavior and intention recognition of signal intersection at home and abroad, divided into two cases, the innovative dilemma zone value, influence and the two factors of the stop line distance on the braking behavior, principal component dimension reduction processing for multiple input parameters, and analysis to determine the membership as the input interval fuzzy reasoning using k-means clustering, the establishment of braking intention recognition model. The Ministry of transport information technology research project(project number:2013-364-836-900) and the National Natural Science Fund Project(project number:61573171) to rely on carry out related research. The specific contents are as follows:⑴ Aiming at the traffic accident happened at the intersection of the signal intersection, the paper further analyzes the causes of the accident.⑵ The according to the vehicle braking characteristics of signalized intersection, respectively from the angle of vehicle driving angle and driver proposed the definition of two kind of dilemma zone, combined with the signalized intersection, the driver’s braking behavior, get the intersection dilemma zone braking intention prediction of key factors, these factors by mechanical components and driving environment, mechanical components of the factors including brake pedal opening degree and the brake pedal opening rate of change and driving factors including intersection dilemma zone, and signalized intersection stop the distance between the lines, the parameters of the car, in front of the vehicle parameters, the front distance, collision between reciprocal.⑶ To establish the model of the braking behavior in the dilemma zone of the signal intersection. Research on intersection vehicle braking is divided into two scenarios. The first scenario of vehicle in front of the car, the second scenario research vehicle comprises a front vehicle. In the second scenario, BP network input more by principal component analysis(PCA) for dimensionality reduction of BP simulation experimental results confirm. BP neural network model can accurately predict the vehicle deceleration.⑷ Research on the identification of braking intention based on the braking behavior model of signal intersection. First, choose and determine the input parameters for the fuzzy reasoning, k-means clustering method is used to determine the membership degree interval range of the input parameters, the fuzzy rule base is established. Simulation results show that the established based of fuzzy inference of braking intention model can effectively identify the driver’s braking intention, compared with the BP network prediction model, the recognition accuracy rate is 91.39%.