Research Of Decision System For Smart Car Based On Camera

With the development of the automotive electronics, people need the car with security, energy conservation, environmental protection, intelligence and information more and more. Smart car system is an important component of the Intelligent Traffic System. The self-driving technology of smart car has a great perspective in road management and traffic transport. In this paper, the system of smart is designed on the background of "Freescale" National Smart Car Tournament and the decision making system for self-driving and control algorithm is researched based on the mechanism and hardware of the model.First, the design and improvement of the mechanical structure of the smart car model is discussed. Caster angle, toe-angle and the parameter of steerable wheel and the gap between the chassis and the ground are rectified and optimized. Then, the paper introduced the basic principle and application of CMOS sensor. Further more, the blue print of camera based on CMOS sensor OV6620 is stated. Third, the circuitry of the system has been discussed including MCU model, power model, motor drive model and wireless communication module. Finally, the available information is extracted by digital image processing algorithm by which the controlling algorithm is discussed. The traditional controlling algorithm cannot settle the problem of parameters mismatching because of the complexity of the path and the vehicle dynamic system's strong nonlinearity. This paper introduced a decision system by fractional order PID controller combining with conventional PID to control the smart car.Experiments prove that the decision system has the characteristics of real-time and adaptability and can control the smart car module to identify the path accurately and self-driving stably.