Design And Implementation Of Intelligent Service Platform For Urban Parking

In recent years,the number of car ownership has been increasing year by year,but the corresponding parking resources have grown slowly,and the existing parking resources have not been fully utilized,which makes the problem of parking difficulties more serious.Therefore,the difficulty of parking can be alleviated by improving the utilization of parking resources.This paper designs and implements a service platform that shares parking resource information,and provides reference information for drivers to park in order to improve the utilization of parking resources.The platform adopts a front-end and back-end separation architecture.It uses Django and DRF as the back-end development framework to realize the various business logic of the platform’s back-end functional services;uses Vue and uni-app as the front-end development framework to realize the development of front-end applications for various users.Then analyze the required data information of each function,design the database table in detail,realize the My SQL database cluster through PXC,and store the data information of the platform persistently.Finally,Redis is used to cache frequently read data information into the memory of the server to improve the response speed of the platform in processing requests.In order to ensure the timeliness of parking information,the platform needs to provide forecast information on the number of vacant parking spaces.This function is mainly to predict the number of vacant parking spaces at a certain time in the future through the historical record information of the parking lot.The recorded information of the parking lot is the sequence data indexed by time.The methods for predicting the time sequence data include exponential smoothing,autoregressive moving average model,and neural network-based prediction model.In view of the limitations and insufficient accuracy of these methods in the prediction of vacant parking spaces,an improved algorithm(A-RNN)based on the combination of Attention Mechanism and Recurrent Neural Network(RNN)is proposed for vacant parking.Make predictions.In this algorithm,the vacant parking space record data is processed through the recurrent neural network to extract the hidden feature information;then the attention mechanism is used to calculate the degree of influence of each feature information on the predicted target,and then according to the degree of influence The size of the feature information is weighted and summed to obtain new feature information;finally,the number of vacant parking spaces is predicted based on the feature information processed by the attention mechanism.Compared with the traditional algorithm,the experimental results show that the prediction accuracy of the improved algorithm has been significantly improved.After testing,the various functions of the parking platform have reached the expected goals,can realize the information sharing of parking resources,and provide a reference for the driver to park.