节点文献

车位共享系统设计及服务端开发

Design of Shared Parking System and Development of the Service-side

【作者】 陆斌

【导师】 张赤斌;

【作者基本信息】 东南大学 , 机械工程, 2017, 硕士

【摘要】 随着我国经济的快速发展,城市汽车保有量激增,带来了日趋严峻的停车难问题。本文调查研究发现,城市混合用地的停车资源在时间上存在互补性。这为缓解我国目前停车难问题带来了契机。首先,本文通过对共享车位理论的研究,提出了物业管理服务下基于云平台的车位共享系统。该系统有效解决了混合用地的停车位资源信息孤岛问题,提高了车位使用效率。其次,根据车位共享系统的需求分析,本文分解系统功能树为登录注册、车位认证、车位发布、车位租用、反向寻车、用户缴费、日志记录共七个模块。本文采用MySQL数据库系统,并进行数据库概念设计和逻辑设计,完成数据库开发;采用Java开发平台,选用MVC模式,借助SSM开发框架开发了高内聚低耦合的系统服务端。然后,本文针对反向寻车实现的关键难题设计对应解决方案,并进行实验验证。本文分别提出基于BLE技术的指纹定位方案以及基于直线优化Dijkstra算法的路径规划方案。定位方案中,本文在离线训练阶段采用了高斯滤波、邻域均值滤波方法保证指纹的准确性,在在线定位阶段采用Kalman滤波和加权KNN算法保证定位的实时性,并结合实际环境提出优化设计。路径规划方案中,本文主要研究了 Dijkstra算法和直线优化Dijkstra算法,并通过实际停车场环境建模验证了优化算法的优异性能。最后,本文针对各功能模块,设计测试用例并完成各模块的功能测试,经测试整个系统功能基本实现。

【Abstract】 With the rapid development of economic and the surge of vehicle’s possessive quantity,parking problem is coming out.According to the latest survey,parking lots of residential land and others are strongly complementary mutually in the time.It brings an opportunity for relieve the acute problem of parking.Firstly,the shared parking system which is based on cloud platform and managed by the property is designed through the study of theories about shared parking.This system solves the problem of "Isolated Information Islands"in parking effectively,and improves the efficiency of parking spaces.Secondly,according to the demand analysis of shared parking system,system function is decomposed into seven modules:login with registration,input parking information,release parking lots,rent parking lots,vehicles search,payments and history records.The MySQL database management system is selected to complete the development of database with the database concept design and the logic design.Java development platform,MVC pattern together with SSM framework are choosed to develop the service side of this system.Thirdly,the key problems of reverse vehicle search and corresponding solutions are proposed.A fingerprint localization scheme based on BLE technology is proposed to slove indoor location.Gauss filter and neighborhood mean filter are used to ensure the accuracy of fingerprints in the offline training.Kalman filter toghther with weighted KNN algorithm are used to ensure real-time performance in the online positioning.A routh planning scheme based on beeline optimizing Dijkstra algorithm is proposed to solve indoor path planning.The dijkstra algorithm and the straight line optimization dijkstra are studied by the modeling of actual parking lots in this paper.Finally,test cases for function modules are designed to test the function of system’s service side.These tests shows the whole system function is realized.

  • 【网络出版投稿人】 东南大学
  • 【网络出版年期】2018年 04期
  • 【分类号】TP311.52
  • 【被引频次】6
  • 【下载频次】566
节点文献中: 

本文链接的文献网络图示:

本文的引文网络