| As the development of the Global Positioning System(GPS)and cellular wireless location technology,outdoor mobile positioning technology has been widely used in our daily life,providing the convenience for traveling.However,in the Shopping Mall,airports,underground parking and some other indoor places,GPS can not provide accurately positioning since the obstacles blocking and the complex indoor environment.Typical indoor positioning technology include Bluetooth positioning,RFID positioning,WiFi positioning.Since the WiFi has been widely covered in public facilities and the cost is small,WiFi indoor positioning system become more and more popular.However,because of signal be blocked and reflected,there are still existing some problems such as the low positioning accuracy and the instability positioning results.So study the WiFi indoor positioning algorithm which has higher positioning accuracy is valuable both in theoretical and actual use.First of all,the paper studied the basic algorithms of indoor wireless location,including the positioning method based on distance,the positioning method based on Time Difference of Arrival(TDOA),the positioning method based on Angle of Arrival(AOA),and the positioning method based on Received Signal Strength(RSS).Drew a conclusion that the positioning method based on RSS technology was more suitable for the indoor environment.Then studied and compared two different positioning methods based on WiFi — Propagation Model method and the positioning method based on signal fingerprint.Explained the advantage of selecting signal fingerprint for indoor positioning.Secondly,the paper deeply studied the complexity of the WiFi signal propagation environment,analyzed the factors that influence the signal propagation.Then studied the WiFi location fingerprint features in the positioning method based on signal fingerprint,explained the mapping relation between the received signal strength and position,and explained the received signal strength volatility caused by environmental factors.On this basis,compared and analyzed several classic fingerprint localization algorithm,deeply studied KNN fingerprint localization algorithm.First optimized the parameter selection.Then improved the algorithm according to the characteristics of signal propagation volatile in the indoor environment.Improved algorithm adds a weighted average window before generating positioning result,using a weighted average value of the last three times as the final positioning result.Also proposed to improve the positioning algorithm by dynamic forecasting the node location based on KNN algorithm.Filtering the RPs from radio map,which don’t have similar RSS vector on tags,to find the nearest neighbor,in order to reduce the time complexity and the computational complexity of KNN algorithm and increase positioning accuracy at the same time.The experimental results showed that positioning accuracy of the improved algorithm had been greatly improved.Finally,the paper designed and achieved a WiFi indoor positioning system based on improved KNN algorithm,the system includes the offline fingerprint collection module and online positioning module.The paper analyzed and designed the offline maps and fingerprint database.In the positioning stage,used the improved KNN algorithm to realize positioning through comparing with fingerprint database,and the results had been verified in some application scenarios. |
