Research On Edge-Computing-and-HP2P-Protocol-Based Smart Parking System And Path Planning Algorithm

With the rapid development of economy,the urbanization process is accelerating,and the number of cars has increased sharply,and the problem of "parking difficulty" is becoming more and more prominent.Based on the continuous progress of Internet of things and artificial intelligence technology,smart parking systems have sprung up.The current smart parking systems have reached the current level of intelligent technology in all kinds of situations,but because these systems focus on highlighting intelligent services,they ignore some more basic problems.The current smart parking systems are mostly centralized systems,and the stabilities are low.In order to deal with the exponential growth of the huge data and the ultra-high calculation amount of intelligent services,the storage resources and computing resources are accumulated in the central server constantly,which is too expensive.In order to solve these problems,this thesis proposes a smart parking solution based on edge computing and HP2P network.It designs and develops a smart parking system,which provides efficient and convenient intelligent parking services for users and intelligent parking management services for parking lots.The smart parking system has three parts: cloud end,parking lot end and mobile end.The data transmission network is built based on HP2 P protocol.The cloud end is used as the guide node of HP2 P network.The parking lot end and mobile end are guided to join the appropriate cluster by the guidance strategy based on the latitude and longitude of adjacent regions,and the computing resources and storage resources of the cloud provide cloud computing services for the system.The parking end,as a super node in HP2 P network,is responsible for managing other nodes in the cluster.Meanwhile,the parking end,as an edge server,filters massive information and complex data for the cloud,and provides intelligent parking services for the mobile end.The mobile end,as an ordinary node in HP2 P network,collects massive traffic information data for the system,and is responsible for the interaction with users,providing convenient and fast intelligent parking services for users.The design and development of the smart parking system makes the research of HP2 P protocol and edge computing enter the practical application stage.Due to the distributed characteristics of HP2 P protocol and the idea of edge computing to lower computing to the edge,the parking lot end,as an edge server,greatly shares the computing pressure,storage pressure and network congestion pressure of the cloud,which provides a novel and efficient mode for the internet of things and the big data.This thesis proposes a double-layer hybrid D*Lite algorithm,which is a D*Lite path planning optimization algorithm.The strategy is modeling the environment information by low-resolution map to reduce the state space,planning global path by D*Lite algorithm.According to the global path node sequence,curve fitting is carried out in high-resolution map,and the intersection of fitting curve and local area of high-resolution map is taken as the starting point and target point of local area path planning,so that the path planning of each local area is not constrained by the sequence,and the path planning of each local area can be paralleled and greatly reduced computing time.If the information in the map changes dynamically,it only needs to re-plan the path of the local area,so as to reduce the number of re planning.In this thesis,the contrast experiment is simulated by Matlab.The experimental results show that the efficiency of double-layer hybrid D*Lite algorithm is better than D*Lite algorithm in large complex map.