| At present, the positioning system based on the satellite signal is widely used in outdoor location tracking, due to the satellite signal is not easy to penetrate building, so the positioning system based on the satellite signal cannot apply to indoor positioning tracking. In recent years, with the rapid development and mature of Internet, some tech giants such as Microsoft, Google, Apple, Baidu and many world famous university are working on indoor location tracking technology, people on indoor environment for its own position information of increasing demand, making indoor positioning tracking technology has become research hot spot, this paper finishes the following work on the hot.For sampling RSSI signals contain a lot of outliers in the interference noisy environment, on the basis of Kalman Filtering, an adaptive KF algorithm against outliers is proposed, which can adaptively control prediction and gain according to the innovation, real-timely detect and eliminate outliers in the dynamic observation data.For using Euclidean distance to measure the similarity of two RSSI signal vector, to a certain extent, it amplifies the RSSI signal vector of the larger components in the role of distance measure problem, this article using the Dice coefficient to measure the similarity of two RSSI signal vector, a WKNN location algorithm based on Dice coefficient is proposed.A performance evaluation index is defined, which can real-timely reflect the tracking performance of KF, EKF, UKF or PF algorithm.Defining a forward switching threshold condition between KF and UKF algorithm, thus a KF+UKF target tracking algorithm is proposed, which is dominated by KF algorithm. When the tracking performance of KF algorithm falls to a certain degree and reaches the switching threshold, switching the KF algorithm into the higher complexity of UKF algorithm is not necessary, thus the proposed algorithm can keep the system real-time performanc of KF algorithm and keep the fault tolerance of UKF algorithm. |
PDF Full Text Request