The Multi-sensor Information Fusion Technology In The Intelligent Vehicle Obstacle Avoidance

To improve driving safety, handling stability, flexibility of obstacle avoidance and human-computer interaction of vehicles now is the focus of intelligent vehicle research, covering Sensor Technology, motor drive and control, Vehicle Engineering, data processing and integration and other areas. This paper mainly studies driving safety and flexibility of obstacle avoidance of intelligent vehicles, including software and hardware design for motion control system, multi-sensor information fusion technology, obstacle avoidance algorithm based on fuzzy neural network(FNN) in the S-T Model, controller design and other related technologies.In this paper, Multi-information fusion technology of multi-sensors is applied to obstacle avoidance control of intelligent vehicles. Firstly, this paper gives a brief overview about fundamentals, structures and fusion levels of multi-information fusion technology, analyzes and compares common information fusion methods, on which basis the method used in this paper is chosen.Subsequently, in order to verify the reliability and validity of obstacle avoidance algorithm, an intelligent model vehicle equipped with ultrasonic sensor and infrared sensor is designed as the experimental carrier, a detailed introduction is given in this part about its main function modules and its controller hardware circuit is also projected. Then, packet processing and information fusion of data captured by the sensor are discussed according to the characteristics and installation locations of ultrasonic sensor and infrared sensor equipped in the intelligent vehicle, and we propose obstacle avoidance algorithm based on fuzzy neural network in the S-T Model and design obstacle avoidance controller on this basis. We testify each initial membership function of fuzzy neural network by offline training and obtain T-S model FNN control algorithm of obstacle avoidance combining advantages of neural networks control and fuzzy logic control through simulation experiment. This algorithm has high precision approximation to nonlinear system and then the smart vehicle can perform obstacle avoidance control safely and effectively.In the third part, the whole program design and process of intelligent model vehicle movement control system is brought up and each main function module of controller is program-developed, debugged and tested. According to operating features and differences of ultrasonic sensor and infrared sensor, we conduct corresponding program design and illustration and also come up with matters that need attention in practical experiments. Finally, obstacle avoidance control experiment is taken on the model smart-car, which testifies the validity and reliability of obstacle avoidance algorithm. At the end of this paper, we give a detailed explanation about the problems and shortages in the experiment.