Research On The Vision-based Local Path Planning Method And Its Embedded Realization

Intelligent Vehicles have extensive application prospects in both military and civilian fields, which are still in the exploratory stage at present, where the method of local path planning is an important research topic in mobile robots, intelligent vehicles and other automated guided devices. This paper is mainly on local path planning method based on visual navigation in intelligent vehicle. It's focus on the driving control module and the local path planning method on the basis of model vehicle which is designed independently by our selves. The simulation experiment has been carried on and the result shows that there is great practical value in related fields.First, according to the theme of this research, it extensively retrieves the relevant fields of researches and summarizes the available research results so that we can grasp the research status and development tendency in domestic and foreign.Second, it introduces the relevant design and production process of foundation platform in intelligent vehicle model. Using proportional scaling model vehicle as the research platform it completes the test of the controller. It introduces the design details of the hardware platform at length, including the hardware sensor module, decision-making module and the controlling module. We build the system software platform based on real-time operating systemμC-II software platform and design hierarchical programs structure, and to the greatest extent to ensure the instantaneity in operational process and the inheritable in development process.Third, it proposes the path planning method based on "S" type and uses the concept of a virtual path, making the algorithm can be considered the relationship between the lateral vehicle dynamic and the curvature of reference lane. It combines the preview theory and the vehicle kinematics and establishes the kinematics model of vehicle trajectory tracking. It designs a single neuron adaptive PID correction links and a rudder stability loop; analyses the two degrees of freedom linear model on vehicle. It also establishes the lateral controller based on correction link, stability loop and vehicle model.Finally, it realizes all the hardware circuit designed in this paper and conducts a large number of tests. The results show that the design of the vehicle control systems and the sensors measuring system can work steadily in a long-term. It proposes the control law, using Matlab for simulation of experiment. The result shows that the local path planning method which is proposed in this paper can effectively achieve the trajectory tracking task.