Research And Implementation Of Indoor Positioning Algorithm Based On WIFI

With the rapid development of smart phone and mobile Internet technology,there is an increasingly urgent need for precise location services such as personnel,goods and devices.The global navigation and positioning system(GPS)can meet the positioning requirements of the meter or decimeter level in most outdoor scenes.However,due to the signal attenuation of the penetrating wall in the complex indoor environment,the positioning is inaccurate or fails,which makes the indoor positioning technology become a research hotspot.In indoor positioning technology,indoor positioning technology based on Wi Fi(Wireless Fidelity)has the advantages of high positioning accuracy,easy laying,strong anti-interference ability and easy realization,so Wi Fi indoor positioning technology has become the key research direction of indoor positioning technology.The existing indoor positioning algorithm based on Wi Fi is unstable,and there is still room for improvement in positioning accuracy and efficiency.In this paper,based on the ease of operation of Android and MATLAB platforms,the construction of fingerprint database and matching positioning algorithm based on location fingerprint method are discussed and studied.The main work of this paper is as follows:(1)Firstly,Set up the software development environment and develop the signal acquisition software by using Android Studio tools.Then,the software was used to analyze the influence of human body shielding,indoor environment,multipath effect and other factors on the propagation of Wi Fi signal,and the time-varying characteristics of Wi Fi received signal intensity were statistically analyzed,and the conclusion of its close to normal distribution was obtained.(2)The direct construction of the fingerprint database based on the noise of the original Wi Fi signal will affect the positioning accuracy.The dual processing method of gaussian filtering and mean filtering of Wi Fi signal for multiple measurements is adopted to make the constructed fingerprint data more representative.In fingerprint database classification processing,the traditional K-means algorithm in the initial clustering center and clustering number choice of randomness in clustering effect is poorer,therefore put forward the initial clustering center to select parameters and comprehensive environmental factors determining K-means algorithm the initial clustering center and clustering number,the method of comparing the traditional Kmeans algorithm to classify the original fingerprint database training at the same time,get the fingerprint classification result,positioning results validate the rationality of the method.(3)An adaptive dynamic k-weighted k-nearest neighbor method is proposed to determine the nearest neighbor K by using the median of test point density.On the MATLAB platform,the simulated location of the original fingerprint database with different K values,the simulated location of the fingerprint database with different K values trained by the traditional k-means algorithm and the simulated location of the adaptive dynamic K values of the fingerprint database trained by the improved k-means algorithm were realized respectively.The experimental results show that the improved positioning scheme proposed in this paper has reduced the positioning time consumption,improved the positioning efficiency and improved the positioning accuracy compared with the traditional positioning method.In terms of positioning system,this paper designs and develops Wi Fi signal acquisition software on Android platform to complete the construction of offline fingerprint database.The simulation of real-time matching positioning process on MATLAB platform,through the analysis of the results,proved the effectiveness of the simulation system positioning.