Multi-Senso Fusion Technology In Obstacle Avoidance Of The Car

As China's rapid economic development, the tenure of cars has been increasing. Vehicle security has received increasing attention, the obstacle avoidance system of the car as an active safety technology will have a broader space. The obstacle avoidance system of the car is a complex nonlinear systems, has a great impact by the external environment, therefore, the establishment of the exact mathematical model has been very difficult. the theory of multi-sensor data fusion for the rapid development to address this problem has brought vitality.Based on the analysis of the most current vehicle obstacle avoidance system and various common obstacle avoidance sensor, for driving with unstructured environment and the characteristics of uncertainty,in order to solve a single sensor was the one-sidedness of the problem of access to information. Based on the driving characteristics of the environment , and combine the characteristics of fuzzy control and neural network, propose FNN multi-sensor fusion algorithm. By data of vehicle speed, vehicle location in the road, car relative distance barrier and obstacles relative to the location of the car fusing, adjusting the speed and traveling direction of the car, the issue of obstacle avoidance systems obtaining one-sided information can be satisfactorily resolved, in order to the pilots can safely avoid obstructions. I had simulated by this fuzzy neural network by software of Matlab ,simulation results show that neural networks used by the fuzzy multi-sensor data fusion algorithm can fast convergence, and suitable for the automotive obstacle avoidance system.