Skip to Main Content
The paper proposes a new vehicle crash-avoiding method using the fuzzy reasoning system and neural net work. The method used neural net work to calculate collision risk instead of fuzzy inference. A vehicle crash-avoiding adaptive network fuzzy interference system model is proposed. The hybrid learning algorithm is proposed to improve rapidity of convergence. For some linear parameters such as consequent parameters, recursive least square algorithm is used to update it. For other nonlinear parameters such as premise parameters, steepest descent method are used to identity it. By comparing the simulation result and experiment data, it shows that the membership function and fuzzy rules for fuzzy control model is optimized effectively by using adaptive network fuzzy inference system. It has a good and self-adaptive performance for vehicle auto-control under the dangerous condition.