Research And Implementation Of Mobile Location Platform Based On OpenLS And Map Matching Algorithm

Location Based Service (LBS) is a new value-added service which determines the actual location of mobile users through the mobile terminal and wireless network and provides users with information tracing, vehicle navigation, Mobile Yellow Pages in the support of GIS systems. With the development of society and the accelerating pace of life, time saving and efficiency has become a very common pursuit, the need for location information is more and more pressing. Under the background of growing demand of LBS, on the premise of combination of mobile communication technology and vehicle navigation technology, this paper implements LBS platform based on OpenLS and conducts a study on the map matching.This paper first introduces the LBS technology, including communication technology of mobile networks, Location Based Service, Open Location Service, mobile positioning technology and other key technologies, such as map matching. The establishment of a complete LBS will be a huge systematic project, which integrates a large number of the latest technology in the field of wireless network and space information. Its research and development will directly benefit from the development and update of the embedded systems, mobile communications technology, geographic information systems, and interoperability technology.The design and implementation of LBS platform based on OpenLS is the main focus of this paper. This paper elaborates on LBS framework of China Mobile and core services, information model and framework of OpenLS. On the foundation of the thorough understanding of the relevant international and domestic norms, we propose the LBS architecture. The entire system is built and deployed on the J2EE platform and can be divided into application layer, service layer and data layer from the logical view. A majority of server programs have been implemented, including map server, application server and location server. Map server provides the abstract interface and concrete implementation of accessing a variety of map data. It unifies, massively caches map data and provides map data for the client and location server. Application server deals with the client request, gets the client necessary location information through location server, calculates the client demand map information link and sends them to the client. Location server is mainly responsible for the provision of location information which gets through the GSM base station. It provides map matching algorithm and optimal route navigation algorithm. Two forms of the client are implemented, J2ME client and Applet client. J2ME client uses MVC design pattern. Browser-based client uses Applet technology. In business function, we have implemented essential location services.Map matching is an important component of the system. Its matching rate, accuracy and matching speed will directly affect the performance of LBS. Therefore, to find a reasonable and accurate map matching algorithm is of great significance. The basic idea of map matching is to compare the vehicle location track from location device with road network in the map database to find the most similar route on the map, and to allude the actual location data to the visual digital map. The current study involves many map matching algorithms. The commonly used map matching algorithms are direct projection algorithm, relevance algorithm, semi-deterministic algorithm, probability and statistics algorithms, fuzzy logic algorithm and knowledge-based algorithm for computational geometry. The research of map matching algorithm is another focus of this paper.This paper makes an in-depth study and comparison about the existing map matching algorithm. We select the suitable map matching algorithm for point positioning and real time positioning respectively, which could enhance the positioning accuracy. In terms of point positioning, centroid algorithm, center algorithm and maximum likelihood algorithm are introduced. To suit the point positioning, the algorithm must have a high computing efficiency, occupy less storage space, and be programmed easily. Through a comprehensive comparison of three algorithms, the system finally adopts centroid algorithm based on the principle of practicality.In terms of real time positioning, taking the continuity, real time and accuracy of the map matching algorithm into account, the fuzzy map matching algorithm with confidence feedback is introduced. A complete map matching algorithm includes three main processes, that is, determining the error region, selecting the matching location, amending the location result. With the support of MapXtreme Search function, the region of searching is limited, and the real time of algorithm is improved. In this paper, the candidate road is the road in the circle with a center of receiving the vehicle location coordinate from GPS and a radius of 100 meters. The Algorithm determines the final matching road through the functionη=αφd +βφθ+γφc, taking into account the distance, direction and connectivity for a fuzzy comprehensive evaluation. In the matching comprehensive decision, we firstly calculate the possible matching path set according to the maximum subjection degree. If there are multiple elements of the set, and one of them meets the consistency condition, the road is the matching road; otherwise any road is taken as the result. The matching comprehensive decision determines the final matching road from the candidate matching road set according to the decision criteria, but the result is not fully credible because of errors. So we define the matching confidence to measure the confidence of decision result, for short, matching degree. In the calculation of matching degree, two main factors are considered: The own comprehensive subjection degree of the matching road, it reflects the actual matching degree of the road and the vehicle location information. The higher the matching degree is, the higher the confidence of the matching result will be; the comprehensive subjection degree difference of the matching road and other candidate matching road, it reflects the relation of roads. The closer the comprehensive subjection degree of the matching road and other candidate matching road is, the lower the confidence of the matching result will be. Then we make a series of improvements to the algorithm: The confidence is used as feedback for analyzing the connectivity in map matching considering various factors. It is more effective to use the connectivity to improve the map matching result; the error region is determined by limiting the search scope. It reduces the computing complexity to adapt the requirement of the real time navigation system.Under the environment of Eclipse with MapXtreme, we implement the map matching algorithm, simulate GPS receiving process and the actual traffic situation, then we give rendering pictures of map matching. After the experimental verification, the algorithm is considered feasible and has better effect on reducing the matching error.Finally, this paper summarizes the research work of LBS platform and map matching, and describes the prospect for improvement in the future.