Construction Services And Performance Optimization On The Positioning Platform Based On LBS

Motion Location Service is called Location Based Service (LBS) too. It uses the network of telecom mobile operators (like GSM and CDMA) to access the location information of the motion end-user. With the help of Electronic mapping platform, it provides the services of value-added business. Now in China, there are a few companies who provide LBS with modern communications technology, computer technology, control technology, location technology, GIS technology and etc. They can achieve real-time tracking, location, navigation and etc, and provide users more travelling information to meet the rapid development of modern city traffic.This paper based on LBS design and implementation of a positioning platform according to the LSP system architecture. This system is mainly to achieve precise map-matching, a comprehensive traffic navigation services (location of graphical information, navigation systems and the best walking route), security and tracking services, location of group information and interactive functions. This paper aimed at optimizing the performance of the entire system, making the system faster and better able to provide users with convenient services.This text can be mainly divided into three parts, which are on the overall system architecture of the service building, the second position to optimize the performance of middleware, Web Pictures manner to optimize client performance, server performance optimization.The system's goal is to realize a platform based on LBS on the J2EE platform and to provide accurate LAC CI, latitude and longitude coordinates, photos and vector data of the positioning information in four ways interface for service providers. We used free API to develop the lightweight HTTP server FGI. In order to enhance the system concurrency and the overload, it removed some redundant functions which in the HTTP servers which this system does not need. We used Mapinfo Corporation's mapxtreme for java API to process map essential factor, and build the map data to the FGI server, analysis and withdraws the path node information for the services of twice localizations,optimal remote search,navigation and etc. The picture image's production uses the phantom pre-production mechanism which likes the service stations Google Map, Baidu Map, Virtual Earth use, with the high speed of map response. The picture retrieval's way selects four fork tree methods, which is fast and convenient. Client using Java Applet technology, set up on the Web with an intuitive, highly interactive user interface and a stronger ability to deal with the graphical client, and Java archive files using a tool to overcome the jar in the Java Applet security mechanisms exist bound .From the view of the overall system, we use the current framework of the more advanced open source, with mapxtreme for java technology, and less use of the other company's products, we developed the lightweight picture servers, map elements server, and high security user server.This study focused on the performance of the entire system are optimized, the original system had an in-depth analysis and research, in view of the inadequacy of the original system to improve, making the system faster and better able to provide users with convenient services,and simple to achieve the shortest path tracking and positioning services.In the system structure, the original LBS system does not meet LSP system architecture, according to the needs of the mobile companies to re-build the positioning platform,and provide accurate LAC CI, latitude and longitude coordinates, photos and vector data of the positioning information in four ways interface for service providers.At the second position, since the original Dynamic limit the search region algorithm exist incomplete extraction of road, GIS buffer of the SAFDGIS-MM algorithm has incomplete path information to be matched. It can not provide accurate positioning information for the second match. So the original system only used a distance parameter for the second match and did not add the matching of direction angle. In the second position, when GIS buffer has the more roads to be matched, the more important direction angle to be matched. We used Mapinfo Corporation's mapxtreme for java API to optimize and implement the Dynamic limit the search region algorithm, for making the second position more accurate.Display on the client, the implementation of the system has two types of mobile phone and Web Client. Image segmentation uses the Level Matrix Image Segmentation Strategy. Image segmentation segments the map into many smaller pictures because of the actual situation that the mobile phone screen is smaller, so that Web clients can not make a good show. At the time of positioning and tracking, anchor out of this picture can show the next picture. However, in navigation, the shortest path search, the client which is the way of picture exist the problem that can not compute the shortest path. To address this problem, System uses a seamless drag picture strategy to solve the problem of the jump, and pre-pictures strategy to reduce the network access delay, implementation of more friendly, more convenient client program.Scheduling at the server, the original system on the assumption that all the maps store on a map server deploy of a number of such a map server to provide services, finally using of polling to the load balancing algorithm to provide users with service. The actual situation in all of the maps will be to picture the form of documents stored in a server on this map is very difficult to achieve, even if the implementation will be made to reduce the performance of the server map. Load Balancing Algorithm poll question is: do not consider the specificity of each server, but simply to carry out the distribution, and there should be no basis for the status of the Service switch. For the server for the special circumstances of the performance of servers designed to further optimize the deployment of three types of server strategy, the introduction of the Advanced Dynamic Feedback and Random Dispatch algorithms, and in accordance with the special circumstances of the system have been adjusted so that more ADFRD algorithm effective implementation of load balancing the system.Finally to finish the work in this paper provides a brief summary. Discussed the limitations of the system and the existence of defects, and the further development of this system, put forward some views and expectations.