Research On Indoor Localization Algorithm Based On RSSI And CSI Fusion Under WiFi

In the past few decades,the explosive growth in the demand for indoor location services has provided fertile soil and foundation for the research of indoor positioning technology.Since GPS signals cannot penetrate the wall in an indoor environment,various wireless positioning methods based on Wi Fi,Bluetooth,radio frequency identification,ultra wideband and electromagnetic have been proposed successively.Among these approaches,the widespread deployment of Wi Fi access points and mobile devices has drawn great interest in Wi Fi-based indoor positioning.In general,the indoor environment is complex and dynamic.Many existing methods of indoor positioning based on Wi Fi ignore the environmental change factors that can easily lead to the failure of positioning.Based on the research and analysis of a large number of Wi Fi positioning algorithms,this paper proposes a new Wi Fi positioning algorithm to reduce the impact of environmental factors on positioning accuracy and improves the accuracy of positioning services by introducing knowledge of feedback correction,Voronoi diagram area division and error filtering.The main work is as follows:(1)Based on the signal attenuation model,the indoor received signal strength indicator is ineffective because of the change of the received signal.Therefore,an indoor location optimization algorithm based on RSSI feedback correction is proposed.In order to reduce the ranging error caused by environmental factors,the algorithm first determines the minimum positioning region by fitting experiment before positioning,and estimates the path loss parameters by using the geometric relationships among nodes in this sub-region.At the same time,the concept of vector mixed product is introduced to filter the selected beacon nodes,so that the anchor nodes can be fully used,the number of initial anchor nodes is also reduced.Finally,the position of the target point is corrected by calculating the positioning error of the beacon point to further reduce the positioning error.(2)In order to solve the problem of low accuracy and high computational complexity in the existing wireless network location fingerprint feature location method,a clustering density weighted algorithm of k-nearest neighbors fingerprint location based on Voronoi diagram is proposed.First,the seed points are selected by using uniform design method within the location region,then the location region is divided based on seed points and Voronoi diagram.At the same time,aim at estimating the location area accurately,the Dixon's test is employed to filter the gross errors.Finally,considering the problem of low accuracy for traditional k NN method,it combines the clustering algorithm and k NN method,proposes a new positioning algorithm with density weighted to obtain the final results.(3)Considering that in the uncertain indoor environment,RSS will limit its positioning accuracy due to multipath effect and other factors,an improved indoor positioning algorithm based on channel state information is proposed.The collected CSI signal is used as fingerprint information to establish a fingerprint map for positioning,which greatly slows down the impact of multipath effects on the positioning results.