Vehicle Following Collision Is Based On Fuzzy Neural Network Control System Research

In the recent years,with the surging of car’ number in our country, China’straffic safety face huge challenge. In order to increasing traffic safety, manyresearch institutions and experts around the world have carried out a lot ofeffort.The active safety of vehicles in the field of anti-collision technologies hasbeen becoming a popular research focus.Car-following anti-collision system combines advanced informationtechnology,and sensor technology with control technology,that make up thedeviation of relative distance which is caused by the fuzziness,randomness anduncertainty of drivers’ perception,which can control vehicles more precisely andreduce the chance of rear-end accident to improve the safety of driving.In this paper,we analyze the experience and defect of vehicles followinganti-collision system of domestic and foreign. According fuzziness、randomnessand uncertainty of drivers’ behaviour inadequacies vehicle using fuzzy controlsystem constructed according to the ambiguity and uncertainty of drivingbehavior, and nonlinear of vehicles’ movement, the vehicle followinganti-collision model is structured by fuzzy control system and optimized byfuzzy basis function.The summary of this paper is represented as follows:First,we study the previous researches and analyze the results and problemsof them.Aim to vagueness,randomness and uncertainty of driver behavior thedistracted driving behavior on the reaction time is anylized to calculate the safedistance for different speeds.Second,the vehicle following anti-collision system we designed use radarand speed sensor as a road information collection, the relative distance and therelative speed of car on same lane as two inputs and expected acceleration ordeceleration to design an fuzzy control model of vehicle following anti-collisionsystem and related algorithms. The Simulation results show that the model andalgorithm can control the vehicle to generate an appropriate response for the relevant stimulus to maintain a safe distance,and improve the safety of thecar-following.Third, aim to the problems of the general fuzzy neural network, fuzzy basisfunction has been optimized to the control model, the establishment of a vehiclebased on fuzzy basis function following anti-collision control model, usingrecursive least squares method and the reference signal can be online updatesystem control parameters, thus speeding up the convergence rate of the system,making the system control process more smooth; optimization of membershipfunctions and fuzzy rules for fuzzy control model and solve the final relativedistance and relative velocity shocks.In this paper, the design of fuzzy neural network car-followinganti-collision control two vehicles relative distance to the time required toachieve safe distance between vehicles is about6-7s, much faster than othermethods. More importantly, this approach compensate for the shortcomings ofconventional empirically determined membership function, and thus can providea more scientific and reliable anti-collision system for the vehicle active safetycontrol model.