Research And Application Of Intelligent Parking Navigation Method

With the rapid growth of private car ownership,the behavior of disorderly parking and occupation of parking spaces in urban centers and commercial areas is increasing.In the process of parking,people lack an information management platform to obtain parking resources to provide users with convenient parking navigation.Therefore,based on the internet of things,mobile applications,artificial intelligence and other computer and internet key technologies,the parking lot information in the city is effectively integrated to realize real-time query and effective navigation of parking space information to avoid disorder of vehicles mobility,so as to improve the utilization rate of parking lot resources and meet user needs,which has become a hot spot in the research of intelligent parking navigation platform construction.Based on the in-depth analysis and research on the development status and limitations of the intelligent parking navigation system at home and abroad,this paper proposes a set of mobile agent-based parking navigation cloud platform design.The main work completed in this article is as follows:(1)Aiming at the indoor parking navigation module,combined with reinforcement learning,the generation of adversarial network algorithm in the field of path planning and perception and decision-making ability,a parking navigation algorithm based on reinforcement learning and generative adversarial network is designed for indoor path planning of the platform.Starting from the reinforcement learning algorithm,combining the respective advantages of reinforcement learning and the generation of the confrontation network algorithm,the state space suitable for reinforcement learning training is generated by generating the confrontation network,and the network structure of the confrontation network is optimized using the Inception module.Through experimental comparison,the results show that the improved algorithm can improve the convergence stability of the algorithm,have better learning efficiency,and can accurately find the optimal parking space path of the indoor parking lot.(2)In view of the problems of heterogeneous parking resources,weak intelligent computing capabilities,third-party access difficulties,and low security levels in traditional multi-parking resource unified guidance system,this paper proposes a mobile agent-based parking guidance cloud platform.The integration of mobile agent technology solves the problems of network load,processing capacity bottleneck and poor coordination in traditional distributed computing,and improves the efficiency of network communication.The mobile agent collaboration model,migration model,sharing model and security of this platform are described.Based on the model,the experiment is compared with the traditional SNMP-based network management method,and the results verify the superiority of the platform.(3)The development and implementation of each functional module,based on the design of the overall framework of the platform,the functional flowchart is analyzed and designed,and the parking navigation algorithm based on reinforcement learning and generated confrontation network is integrated into the indoor parking path For planning issues,it is used for the design of indoor navigation routes.The system has completed the functions of information acquisition and navigation,parking space reservation,parking space rental,indoor parking path navigation and other functions in the surrounding parking lot,and realized the integration of inquiry,reservation,navigation and parking at the wrong time.(4)A functional test was carried out in response to the performance requirements of the parking navigation cloud platform,which verified the effectiveness and practicability of the platform.At present,the system has been put into use in a company in Suzhou city,to a certain extent to meet the needs of the company and employees for parking spaces.