Research On Collision Avoidance Early Warning Model Based On RBF Neural Network

With the rapid development of Chinese industrialization process,Chinese per capita car ownership has grown rapidly.At the same time,during the 13th Five-Year Plan period,Chinese highway construction has developed rapidly and the mileage of traffic has increased dramatically.Faced with increasingly serious traffic safety issues.This paper summarizes the research status of vehicle collision avoidance early warning system and braking intent recognition algorithm in the context of vehicle-to-vehicle communication or vehicle-road collaboration at home and abroad.The safety distance model based on RBF braking intent recognition network in vehicle-vehicle communication environment And in-depth study of the vehicle longitudinal collision avoidance warning system.First,the RBF braking intent recognition network that uses the throttle valve opening degree,brake pedal opening degree,and brake pedal opening degree change rate as input variables and the braking intent as output to determine the driver's braking intent,and According to the braking intention of the vehicle in front,it is divided into three working conditions:emergency braking mode,normal braking mode and non-braking mode.Secondly,in order to achieve accurate estimation of the distance between the front and rear vehicles,this paper calculates the relative running distance of the front and rear vehicles based on the positioning data of the front and rear vehicles and combines the multi-source data complementary filter algorithm to estimate the workshop distance of the front and rear vehicles,and estimates the workshop distance of the front and rear vehicles Under the distance and vehicle-to-vehicle communication safety distance model based on the braking intention identification algorithm,a secondary warning model with the minimum braking deceleration of the vehicle as the boundary threshold is established.In order to verify the safety distance model based on the RBF braking intent identification network and the vehicle longitudinal collision avoidance warning system built in this paper,a vehicle braking parameter collection and monitoring system based on LABVIEW was developed to collect the relevant system of the vehicle during the driving process.Dynamic parameters and operating conditions data,and transfer the information of operating conditions,longitudinal acceleration,vehicle speed,etc.of the preceding vehicle to the vehicle,prompting the vehicle to make corresponding braking measures.Finally,the system was simulated and tested,and the actual vehicle was tested.The accuracy and timeliness of the minimum safety distance model and collision avoidance warning system constructed in this paper were verified under different operating conditions and driving conditions of the preceding vehicle.The experimental results show that the safety distance model based on the RBF braking intention identification network is 28.12%lower than the safety distance model under the traditional vehicle-road collaborative environment.The workshop safety distance is significantly reduced.Under the four conditions of emergency braking mode,normal braking mode,uniform speed driving and uniform acceleration driving in front of the vehicle,the average false alarm rate of the collision avoidance warning model based on RBF neural network is 2%,the performance is reliable,and it can be used for driving Provide early warning of collision avoidance.