A metadata-based smart parking system using ensemble learning mechanisms for intelligent transportation

In today’s world, urban planning requires intelligent parking management for dealing with congestion and providing convenience to users. Efficient allocation of parking slots is one of the key challenges faced by planners, especially when there is rapid growth in population. Most of the existing systems rely upon database and historical data for predicting future demands. However, the existing systems cannot account for changes in user behavior and real-world situations. This paper provides an automatic parking allocation system using ensemble learning and boosting methods. The system uses historical data as well as current data like location of users and availability of slots. Using an adaptive learning process, the system accurately predicts future demands and allocates slots accordingly. The system also eliminates the problem of wastage due to cancellation of reservation requests. With intelligent prediction and allocation capability, this parking allocation system provides efficient management of parking lots within digital urban landscapes. Performance assessment for the suggested framework is carried out through the use of standard performance measures, which illustrate the efficiency of the ensemble boosting algorithm. It can be observed from the experimental analysis that the combination of XGBoost and random forest results in a success rate of 98.8%, surpassing traditional techniques. In addition, it is ensured that there is optimal use of resources, considering efficient allocation of the parking slots and related performance factors.