Application And Research Of Mobile Positioning Technology In Traffic Flow Prediction

With the rapid development of wireless communication systems and the widespread use, wireless location technology has aroused interest and concern in a number of research workers, and thus it has also became a research focus of related fields.In their tireless efforts, wireless location technology has been greatly developed,and widely applied in other related industries.In short, it is also becoming more and more important in the role of national security, mobile navigation, intelligent transportation,and monitoring of vulnerable groups.This paper mainly studies wireless location technology which is based on time difference estimate and its application in the traffic flow induction system, to achieve short-term traffic flow prediction.First of all, it is mainly studied and discussed about the implementation of wireless location technology and the factors of affect the positioning accuracy and stability,such as the improvements of localization algorithm, NOLS transmission error suppression, etc.; Secondly, it is focused on the matching process that the user's mobile trajectory which deviates from the true road is matched to the most similar road from the vicinity of roads,and applying RBF neural network that is a way of vague identification to complete the matching process; Finally, obtaining the traffic density according to the number of users which is matched to relevant road,meanwhiles,getting average speed according to the multiple users'speed to achieve the purpose of traffic flow prediction by the two characteristic parameters which reflect the situation of traffic flow.As the GSM network still is currently most major wireless communications network, this research work is mainly devoted to GSM communications network.This main contents of article are as follows:1)This discusses a variety of wireless location technologys which is based on cellular networks and the various indicators which is used for evaluating positioning accuracy. After these positioning technologys was comprehensively compared, finally, determining the E-OTD location technology belonging to mobile phone location technology as research basis of this paper,and determining the mean square error (MSE) as main evaluation indicators of positioning accuracy.2)This paper introduces the mathematical models of E-OTD location technology and the basic principles of two types of classical location algorithms—Taylor series unwrapping algorithm and Leastsquare algorithm, and then analyze their applicable precondition , difficultness and easiness level,and other advantages and shortcomings. Then to discuss the basic ideas of weighted residuals localization algorithm proposed by the Pi-Chun Chen[1],the algorithm can play a positive role in inhibiting or reducing the impact of NOLS,but the algorithm itself is extremely complex,and its performance about speed is poor. According to Pi-Chun Chen's research ideas,the article proposed a method that it can identify the non-line of sight transmissiont—interval recognition judgments method,and a weighted positioning algorithm which is based on the interval recognition judgments method;The former can effectively identify the base station of NOLS transmission,the latter can play a certain inhibition of NOLS, and the algorithm is also relatively simple, the speed performance can be ensured.Finally, the simulation results verified its validity and rationality, and show that it has some theoretical value and practical significance.3)Finally,the paper discusses the development status of map-matching algorithm at home and abroad and their own basic principles.Then briefly introducing the basic concepts and principles of RBF neural network, and training network by OLS learning method, making the deviation point located as the training samples of network.after the network is tested,and it shows that:the network can meet the basic matching purposes when the deviation point located is processed by the network.Then three features parameters of traffic flow and the relationship between them are introduced, meanwhile the purpose of traffic flow prediction also is reached by simulation experiments, and the accuracy of prediction results is further analysed and evaluated.Firstly,this paper intensively studys the related algorithms of both mobile positioning technology and digital map-matching technology; secondly, and collects in real time the traffic flow data in the form of simulation through through the two techniques,according to the data collected,then calculates the travel speed value and the traffic density values that can reflect the traffic flow conditions very well;finally,it is to achieve the purpose of traffic flow prediction through the comparison with the priori values.