Algorithm Design And Implementation Of Rssi Indoor Position Based On Unscented Kalman Filter

With the rapid development and maturation of wireless indoor sensor positioning technology,the positioning technology based on the Received Signal Strength Indicator(RSSI)has been caught concern by the domestic and international scholars because of its low power consumption and simple equipment,and it has become the major topic of discussion in the current indoor positioning research.Although the indoor positioning technology based on Bluetooth RSSI has an huge advantage over other indoor positioning technologies,due to the complex indoor environment,such as noise interference,multipath effects and other factors,the value of received RSSI is easily caused a large deviation which lead to the low accuracy of positioning.Therefore,the ability to smooth the random errors and reduce the interference of external factors to the value of received RSSI is the key to improve the accuracy of the indoor positioning algorithm based on RSSI.The purpose of this thesis is to study and improve the indoor positioning algorithm of RSSI based on Unscented Kalman Filter.In general,the Sigma point in the UKF algorithm move farther from the central sampling mean point with the increase of the dimension of the state vector,which can easily produce the non-local effects of sampling point and lead to filtering divergence.In view of this limitation,a modified Unscented Kalman Filter based on the ratio correction and the improved filter initial value has been taken.The modified algorithm is divided into two parts:the sampling strategy of proportion correction in the unscented transform by using covariance weight and mean value weight;the improvement of the Unscented Kalman filtering initial value by combination of quadrilateral measurement method and centroid algorithm.Finally,the improved algorithm will be brought into the Kalman Filter framework to conduct the filtering operation and follow-up positioning calculation.The results of MATLAB simulation and test in the actual indoor environment show that in the functional test,the improved algorithm can realize the indoor positioning and track the trajectory,and in the performance index,during the all test points,the positioning error of 80%test points is under 1.2 meters,compared with the unimproved UKF algorithm,the accuracy of improved indoor localization algorithm increased by 10%,it can be considered that the improved algorithm has reached the expected requirement.In this thesis,the indoor positioning algorithm of RSSI based on the improved unscented Kalman Filter algorithm in both function and performance meet the design requirements and indicators,it can realize the object to be positioned in the indoor and dynamic tracking.The research results have a certain practical engineering value for the application in the positioning technology of RSSI.