Research On The Location Algorithm Based On Fingerprint Matching For NB-IoT Terminal

With the rapid development of the Internet of Things industry,it is closely integrated with traditional industries,and the ecosystem of the Internet of Everything has been built with the help of technologies such as cloud computing and big data.Due to its advantages of low power consumption,low cost,large connection,and wide coverage,Narrow Band Internet of Things(NB-IoT)communication technology has become one of the main transmission technologies of the Internet of Things,and it is widely used in business scenarios such as shared bicycle,remote meter reading,and logistics tracking.Terminal positioning and location based services are the main business characteristics of the Internet of Things.Therefore,research on high-precision and low-cost NB-IoT terminal positioning technology solutions is particularly important for promoting the development of the NB-IoT industry.Aiming at studying the NB-IoT positioning technology and improving its positioning accuracy,the specific work of this thesis includes the following aspects:Firstly,this thesis analyzes the key technologies of terminal location of NB-IoT network,including NB-IoT network architecture,physical layer frame structure and positioning standard protocol principle,etc.Then focus on the principle and algorithm of positioning technology based on NB-IoT network.Combined with the positioning of business scenarios and performance requirements,this thesis chooses the location technology of scene analysis as the key technology for NB-IoT terminal location,and uses this technology to carry out algorithm research and positioning scheme design.Secondly,aiming at the problem that the search time of the fingerprint matching algorithm increases linearly with the increase of the data volume of the fingerprint database in the positioning technology of scene analysis,this thesis proposes a fast positioning algorithm solution.The main idea of the algorithm is to introduce a clustering algorithm based on fuzzy c-means for the problem of large classification error of subpartition boundary points.The sub-partition can be quickly located by the maximum membership value.Aiming at the problem that Euclidean distance can not effectively measure the mapping relationship between signal distance and actual distance,a partition metric learning based on large interval nearest neighbor is proposed to improve fingerprint matching accuracy.Then,aiming at the problem that the weighted average algorithm based on signal distance for position estimation cannot accurately locate the terminal position,an improved K-nearest neighbor location algorithm based on the actual distance of signal distance mapping is proposed.The algorithm learns the mapping relationship between the signal distance and the actual distance of the K-nearest neighbors by polynomial fitting method,and constructs the actual distance between the terminal to be located and the Knearest neighbor,and combines the least squares algorithm for final positioning.For the application scenario of NB-IoT terminal suitable for low-speed mobile,this thesis designs and compares the trajectory fusion positioning algorithm.Finally,the MATLAB tool is used to build the positioning algorithm verification platform of NB-IoT terminal,and the positioning scheme proposed in this thesis is simulated,verified and optimized.The simulation results show that the proposed NB-IoT terminal positioning scheme based on fingerprint matching has certain improvement in fingerprint search speed and positioning accuracy.