| A great of progress has been made in Intelligent Transportation System(ITS) in recent years. Safety Driving Assistance(SDA), which is an important parts in ITS, is mainly applied in many areas,such as speed advice and overspeed alarming, driveway aberrancy alarming and defense, collision alarming and defense, tiredness drive watching etc.Safe-distance control of the vehicle is mainly used in the alarming and defense of collision. At present, there are some progress have been made in the field of theory research and practical application of SDA, in these method, generalized forecast control, PID control, classical fuzzy control were adopted, but there are some limits in them,for the parameters of the system are preset, While the exterior environment has changed, the control effect isn't satisfied.Instead of constructing the accurate mathematics, fuzzy logic control uses linguistic variables to describe the under-control system; The neural network has many advantages such as nonlinear map, self-learning, parallel processing; B-Spline functions, because of its recursion, local positive support, iteration and its differential characteristic, can satisfy both fuzzy system and neural-network system very well. So by combination of fuzzy control and neural-networks with B-Spline functions, can advance both advantages to control the system. Then, by analying the robustness of the system, we proposed the improved B-Spline fuzzy neural-network controller. In this paper, by analyzing the driving action first, we construct the safe-distance driving model by using the Adaptive Neural-Fuzzy Inference System(ANFIS) to control the following car's acceleration across the speed and distance gap changing on the leading and the following car, and optimize the system on-line by gradient descent method.Using Matlab to simulate the method. The result shows that the car can keeping safe-distance and prevent collision, which also proved the correctness of the method proposed in this paper. |
