Research Of Position Technology Based On WFI Networks In Three-Dimensional Space

The requirement for location-based services, promote the research of indoor positioning technology. In accordance with the different wireless carrier, positioning technology can be divided into ultrasonic, infrared, UWB, and RFID positioning technology. WiFi positioning technology is a kind of RFID positioning technology, wide-range and low-cost made it attract much attention. And now it becomes mature under2D-space environment. Based on these previous studies, this paper aims to study the WiFi position technology under3D space environment, and hope to enhance the3D-Positioning accuracy and stability by exploring the new algorithm.Basic wireless location technology, generally include time of arrival based or time difference of arrival based positioning (TOA/TDOA), angle of arrival based positioning (AOA), received signal strength indicator based positioning (RSSI). The paper compared the performance of these various positioning technology, and the result indicated that for WiFi networks, using the RSSI location technology is better. Thus, in order to improve the3D-Positioning accuracy and stability, we can optimize the indoor propagation model, RSSI measurement, RSSI filtering, or positioning algorithm. Here, we mainly discussed from the RSSI filtering and the positioning algorithm.Firstly, the paper discussed the RSSI filter. By statistical analysis of the RSSI signal values, got using the Logarithmic distance ranging model was more accurate. But, due to more impact factors at indoor wireless transmission channel, the measured value of RSSI has a fluctuant larger range, and it would result in a high error when ranging. So, the RSSI filter is necessary. In this article, the author compared and analyzed the performance of mean filter, Gaussian filter, weighted modification filter and Kalman filter, and the simulation results stated the Kalman filter was the best one for RSSI signal processing. Because, the measurement and estimates process of RSSI both met the Gaussian white noise process, moreover, to this kind of Linear stochastic differential system, the Kalman filter is the most excellent and efficient processor. It can not only guarantee filtering performance, but also implement simply and operation fast.Besides, the paper discussed3D-Positioning algorithm either. Firstly, the author studied the traditional2D-positioning algorithm, mainly including Triangular-centroid algorithm, OLS, and MLE. In addition, make the simulation to these algorithms under the experimental environment, and got their performance. Secondly, based on these theories, the author discussed the3D-localization algorithm further. By comparison the positioning accuracy and the time complexity of2D and3D positioning, could obtain that applying the extended2D-positioning algorithm to3D had more problems. It would result in a high calculation cost and a declined positioning accuracy. Therefore, we need to explore a new3D-positioning algorithm, and provide an effective positioning scheme. In fact, some scholars have proposed the COLA (Complexity-reduced trilateration approach) method to solve the3D-position problems. In the COLA scheme, the super node instead of traditional AP in the networks, this kind of architecture could make the3D-positioning translate into2D-positioning. Thus, made the3D-positioning have a high performance as2D. But, the COLA method requires the WiFi networks have a special architecture, it’s not universal to traditional WiFi network. Against this background, the author proposed an improved positioning algorithm FCPSO (Feature-points Constraint-PSO), through studying the PSO (Particle Swarm Optimization) optimization theory and the CPSO (Constraint-PSO) traditional location algorithm. The FCPSO algorithm constrains the solution space by feature points, and then uses the PSO algorithm to find the optimal solution. It can not only avoid the solution into a local optimum as PSO, but also have a faster convergence feature. In a word, the improved FCPSO algorithm has a higher positioning accuracy and less time complexity than traditional CPSO algorithm. Finally, the author uses the Matlab simulation platform building an analog positioning Simulation System, and the RSSI filter and3D-positioning algorithm both are applied into the system. Then, the author compared the positioning performance of each3D algorithm, beyond that, also discussed the affection of node density, node coverage range and range error to positioning results.