Research On Indoor Three-dimensional Positioning Technology Based On The Location Fingerprint

With the continuous development of computer technology and network communication technology, location-based information and services people demand more and more, especially GPS is not complex indoor environments targeting. Currently known indoor wireless location technology is mainly a problem for two-dimensional positioning, the positioning studies for three-dimensional space is little, but in many scenarios, such as mines, disaster relief, underground cable tunnels, indoor irregular structures stadiums, etc., often need to achieve localization of persons and goods within the three-dimensional space, and thus the three-dimensional positioning indoor issues of more practical value.In the conventional indoor positioning technique, a location fingerprinting based on the received signal strength indicator (Received Signal Strength Indication, RSSI) of (Location Fingerprint, LF) positioning method because of its high precision, the indoor environmental factors that affect the small, flexible and become wireless indoor positioning technology in the mainstream approach has been more widely used. In this paper, after a three-dimensional position of the fingerprint indoor location positioning method for the sampling stage in the offline and online positioning stage problems in depth analysis, combined with Latin Hypercube Sampling guidance Kriging interpolation position fingerprint database construction algorithm, spectral clustering LANDMARK coarse positioning combined with precise positioning of online localization algorithm, the formation of a new deterministic method of three-dimensional positioning. The method in the efficiency and positioning accuracy than the traditional method of three-dimensional positioning. The main research work carried out specifically as follows:Offline phase location fingerprinting positioning method need to establish accurately reflect the value of the radio signal according to a different reference point within the targeted area received intensity of each position of the feature position fingerprint database, which is the basis for the effective realization of indoor location. Conventional position fingerprint database construction method with full main mining method for three-dimensional spatial position of the fingerprint information collection, the work of a tedious, huge workload, fingerprint library to build low efficiency, it greatly reduced practicality. Therefore, we propose to build Kriging interpolation algorithm based on efficient fingerprint database location, just off a small collection of fingerprint data location reference point, you can build a complete fingerprint database location by interpolation calculation. While introducing pull Dingchao Li SAMPLING algorithm to guide the selection of a known reference point, to ensure uniform distribution of the sampling points within the location area as possible, in order to make better use of the position data of the global characteristics of the fingerprint. The above method, to ensure the position of the fingerprint library to build on the basis of accuracy, effectively reducing the workload off the sampling phase. Lay the foundation for the precise three-dimensional positioning.Online positioning stage is the core location fingerprinting positioning method, there are many theories can be used for localization algorithm design, the fundamental method is mainly based on the position vector fingerprint matching to achieve positioning. But for three-dimensional positioning problem, or traditional nearest neighbor K-nearest neighbor method, you need to position the data one by one matching fingerprint library, the calculation is very large, and difficult to ensure accuracy. Therefore, we propose spectral clustering binding LANDMARK two-phase localization algorithm. First, the use of advanced spectral clustering algorithm to locate the position of the fingerprint database data within the region are classified to obtain each cluster center line positioning when the target position to be fingerprint data with a variety of classes were matched to determine where the target class, area of determining the target to complete the rough location; and then the binding site localization algorithm mature LANDMARK goals in the identified class area estimate that fine positioning. The method can ensure the positioning accuracy in the premise, greatly improve the positioning efficiency, and ensure real-time positioning.For the above algorithm, theoretical analysis, to carry out computer simulation and experimental study of a combination of research, select the campus a particular interior scenes, build ZigBee-based wireless indoor positioning experiment platform, the proposed algorithm has been fully experimental verification, further evidence of the theoretical value and practical significance of this algorithm, but also for the practical application of the algorithm provides a reference.