Research And Implementation Of Information Forecasting In Parking Guidance System

The contravention of vehicle parking between supply and demand becomes more and more conspicuous and serious in some major cities of China during these years, while Parking Guidance System (PGS) offers an effective solution to solve this issue. By deeply analysing the deficiency and the current of PGS, the main work of this thesis is to design and realize a prediction subsystem, which can enhance the efficiency of the guidance .By adopting the neural network theory and Time-series theory, the prediction subsystem offers a efficient solution for parking prediction by designing a forecasting model of Parking Occupancy Rate which is based on BP(Error Back Propagation). Meanwhile, it also has done some good expansion in application scope of neural network. The subsystem becomes more flexibility and efficiency with a strong adaptability by adding the dynamic redeployment function.And through a lot of experiments, the optimum prediction model is finalized. The application of the subsystem can forecast a short term Parking Occupancy Rate. And done a good development for solving the parking contravention between supply and demand.