Threat-Estimation-Based Vehicle Collision Avoid Method

anti-collision system of Intelligent vehicle can effectively avoid traffic accidents,and improve the safety of vehicles driving.Anti-collision system uses anti-collision algorithm and action device to remind the drivers to pay attention to driving safety or active control of the vehicle to avoid collisions.The existing anti-collision algorithms are not fully considered,on the one hand,the influence of sensor measurement error is not considered,on the other hand,the influence of traffic environment factors and target characteristic factors is not fully considered.Therefore,it's great practical significance to study the anti-collision algorithm and its improvement measures,which can effectively improve the warning performance of anti-collision system and ensure the safety of drivers.Based on the time-to-collision algorithm in anti-collision algorithm,this thesis improves the insufficiency by improving the calculation method of time-to-collision's threshold.Aiming at the influence of sensor measurement error,this thesis established a time-to-collision probability model.Firstly,the relative distance between the main vehicle and the target,the speed of the main vehicle and the target speed are considered as stochastic processes in the time-to-collision algorithm,and the distribution function is established.Then the time-to-collision probability model is obtained by converting the formula.Finally,the time-to-collision's threshold is calculated by combining the collision probability index.Aiming at the influence of traffic environment factors and target characteristic factors,this thesis constructs the threat estimation correction model.Using the fuzzy reasoning method in threat estimation,the risk coefficient considering traffic environment factors and target characteristic factors are calculated,and the time-to-collision's threshold is modified by using the risk coefficient.The anti-collision algorithm includes forward,rear and lateral scenes in practical application,designs the warning method according to each scene,and then compares the time-to-collision's threshold with the real-time time-to-collision,and realizes the warning to the target with the warning method.In this thesis,the forward,rear and lateral scenes are established through the PreScan,and the method of this thesis is simulated to verify the feasibility and effectiveness of this method,and then the vehicle anti-collision system is designed and developed on the basis of the intelligent vehicle experimental platform.Finally,under the actual environment,the real vehicle experiments are carried out on the forward,rear and lateral scenes respectively to verify the effectiveness of this method.