Long-time prediction method for number of idle parking stalls in parking lot

Abstract

The invention proposes a long-time prediction method for the number idle parking stalls in a parking lot. The method comprises: preprocessing the acquired car parking data and corresponding weather data to obtain the number characteristics data of idle parking stalls; normalizing the idle parking stall characteristics data; cutting the characteristics data into a training set and a testing set; constructing an LSTM neural network; determining the idle parking stall characteristics data at a previous time of a to-be-predicted first time period used as the input neural elements of the LSTM neural network to obtain the number of idle parking stalls of the first time period; creating a long-time prediction model for the number of the idle parking stalls; performing repeated iterative predictions and reverse normalization to obtain the prediction result. According to the invention, based on the parking behavior characteristics of users, a long-time prediction method based on LSTM neural network for the number of idle parking stalls in a parking lot is proposed to support the number information of idle parking stalls in a future designed period and present it to users so that the users could select a best parking lot. These functions make the method of significant market value.

Description

technical field [0001] The invention belongs to the technical field of parking lot prediction applications, in particular to a long-term prediction method for the number of idle parking spaces in a parking lot. Background technique [0002] In recent years, my country has only started to use and study the intelligent parking management system. However, with the rapid development of my country's economy, the number of motor vehicles is also increasing year by year. Demand intelligent parking guidance management system. However, due to the relatively backward development concept of urban traffic planning in my country, it cannot meet the rapidly changing traffic demand. At the same time, the theoretical research on the number of free parking lots and the prediction of parking time is still in its infancy and lacks practical application. [0003] The above theories often only focus on finding the nearest parking lot in the city and the parking space induction work inside the pa...

AI Technical Summary

Problems solved by technology

However, for the prediction technology of parking spaces, the prediction models used before, such as classic RNN and ARIMA, all have a situation where the prediction time is too long and the accuracy drops significantly.

Images