Research On Short Term Prediction Of Unoccupied Parking Space Number And Path Guidance In Large Parking Lots

In recent years,with the accelerating of urbanization process,the per capita car ownership is increasing,the problem of parking difficulties is also increasingly prominent.As an important part of intelligent transportation system,the urban parking guidance system plays a vital role in improving the efficiency of existing parking facilities and solving the problem of urban parking difficulties.This thesis chooses two key links in urban parking guidance system as the object of study: one is the predication of parking spaces in short period time,the other is optimal parking location selection and arrival path planning in parking lots.In view of the first research,using the method of grey neural network which corrected by modified Markov chain to predict the short-time empty parking spots.In view of the second research,A two-way search Dijkstra algorithm is used to optimize the path to the optimal parking space.The main work of this thesis is as follows:Because of Randomness,complexity and correlation of the short time parking change and based on analyzing the advantages and disadvantages of traditional intelligent prediction methods and the flaw of single object forecasting,the grey neural network formulate prediction method has been designed.And it also can improve the prediction accuracy and timeliness by correcting the error that produced be using the Markov chain system.Based on this,a short time idle parking space prediction model of "Grey Neural Network-Markov chain" is established.The temporal and spatial variation of the actual parking space data is analyzed,the input of the prediction model is determined,and the simulation verification is carried out.The test results show that the prediction model accords with the change rules of actual parking lot data,and it can provide high accuracy short time prediction of idle parking space.The short time prediction of idle parking space is mainly for guiding the choice of parking lots.This thesis also analyzes the optimal parking space selection and the optimal path in the parking lot.According to the main factors considered by the driver to choose the parking space,the optimal model of the optimal parking space is established,and a two-way search Dijkstra algorithm is proposed to optimize the path to reach the best parking space.The simulation experiment is carried out on a certain large parking lot at a certain time.The result shows that the method not only improves the search efficiency,but also makes the optimal decision result more reasonable.