Research On Parking Scheduling Technology Of Office Building Based On Federated Learning And Blockchain

In recent years,the construction speed of parking space can’t catch up with the growth rate of motor vehicles,’ parking difficulty ’ has become the main problem of urban parking management.Especially in the parking lot of office buildings in densely populated commercial areas,parking difficulty is more prominent.In order to solve the parking problem,the office building usually adopts the hybrid parking strategy of monthly rental parking space and temporary parking space.Through the digital management of parking space,it can help the driver quickly find the available parking space,but the unreasonable allocation strategy causes the waste of parking resources.How to scientifically schedule the parking space is of great significance to the improvement of the utilization efficiency of parking resources.At present,the methods of parking reservation and prediction are mainly used to solve the problem of parking difficulty.By using reservation parking spaces and parking periods for vehicles with reservation parking spaces,the parking time and flow of vehicles with non-reservation parking spaces are predicted and the available parking spaces are matched.This method has achieved good results in the case of vehicle behavior.However,in practice,many vehicles with non-reservation parking spaces lack historical parking data and are difficult to predict effectively,resulting in low utilization rate of parking spaces.At the same time,overtime parking of vehicles with non-reservation parking spaces will affect the parking of vehicles with subsequent reservation parking spaces.In addition,with the increasing emphasis on data privacy and security,parking data is often difficult to obtain and utilize in practice due to the relevant privacy information of users,and it is unable to effectively predict the parking flow of vehicles.In view of the problems existing in the existing office building parking solutions,such as parking conflicts,low utilization rate of parking spaces,and unable to effectively predict the parking flow under the centralized parking data,this paper studies the parking scheduling technology of office buildings,and puts forward the parking scheduling method of office buildings based on federated learning and blockchain.It can effectively predict the parking flow without centralized parking data,dynamically optimize the parking space,and timely allocate it to the corresponding vehicles,so as to avoid the parking conflicts of different types of vehicles.In addition,while ensuring the quality of parking service of monthly rental vehicles,it also meets the parking demand of more temporary vehicles and maximizes the utilization rate of office parking.The main work and innovations of this paper are as follows:1.Aiming at the problem that parking data is difficult to be analyzed and utilized due to privacy and security,a parking flow prediction method based on federated learning is proposed to effectively train the parking flow prediction model of office building parking lot without centralized data,which can meet the actual requirements.2.Aiming at the problem of low efficiency of model training in the parking flow prediction method based on federated learning,a parking flow prediction method based on federated learning and blockchain is proposed.Based on blockchain and IPFS,the parameters of the parking flow prediction model are stored on and off the chain,and the incentive mechanism based on the contribution of data volume and the contribution of model performance improvement is designed.3.Aiming at the problem of low utilization rate of parking spaces in the existing optimization methods of office building parking lots,a parking scheduling method based on federated learning and blockchain parking flow prediction is proposed.The parking flow prediction method based on federated learning and blockchain trains the monthly rental vehicle parking flow prediction model.According to the predicted monthly rental vehicle parking flow in each period,two parking scheduling algorithms are designed.The type of available parking spaces is dynamically planned,and the utilization rate of office building parking spaces is improved.