Research On Intelligent Parking Lot Parking Prediction And Guidance Method

With the sharp rise in the ownership of urban residents’private cars, the difficulty of parking is becoming more and more prominent. Parking guidance system is mainly to solve supply and demand balance between the parking space of parking lots and the car of the drivers. But, the existing parking guidance technology is mainly committed to outsider area parking guidance, as for the parking guidance of parking lots is mechanism singling. Generally, without considering the favor factors of drivers selecting parking space, parking lots parking guidance system is only give a random shortest path to parking space. In addition, free parking space prediction technology can effectively helping outsider area parking guidance system if joined.This paper analyzes mainly time series forecasting techniques at first, on this basis, using the BP(Back Propagation) neural network algorithm to realize predication of parking lots free paking spaces. Making detailed research about the principle of standard BP neural network and determining BP algorithm network structure for the problem of parking lots free parking spaces predication. Improving standard BP neural network by using BP momentum method combined with adjusting learning rate to solve shortcoming such as easy volatiling, falling into local minimum and slowing convergence speed in the process of BP neural network training. Transforming the paking lot free parking spaces data through the data preprocessing technology to prevent large values caused network paralysis when simulating the experimental data. Through simulating results, the improved BP neural network is effective predicting of parking lots free parking spaces. This predicting results provide help for outersider area parking guidance.As for the parking lot part, this paper using MADM(Multiple Attribute Decision Making) method for optimal routing to realize the parking guidance of the parking lot. By analysising the drivers’main consideration about how to choose a parking space, the decision attribute matrix is identified by driving distance which is deduced through Dijkstra algorithm, walking distance which is deduced through Euclidean distance and parking space environment value which is deduced through the entropy of triangular fuzzy number. Finally, using the grey correlation entropy method of MADM to sorting the effective free parking spaces, the optimal attributes of free parking spaces is the optimal free parking spaces, the corresponding route is optimal route.