Research On Indoor Position Estimation Technology With High Precision Based On Multi-source Data Fusion

With the popularization of smartphones and mobile internet,people's demand for location-based services and indoor positioning is ever increasing.Among the existing indoor positioning technologies such as Bluetooth,WiF,RFID and ultrasound,there are some shortcomings,which makes them difficult to have universality.As a result,the combination of multiple positioning technologies is one of the solution to this problem.On the background,positioning technologies based on Ultra-Wide Band(UWB)and inertial measurement unit(IMU)are studied in this paper.After making analyses on their advantages and disadvantages,the combination of UWB and IMU is put forward in this paper to achieve higher positioning precision.The main contributions of this dissertation are as follows.Firstly,for the problem of non-linear ranging method and suffering from multipath effect and non-lineof-sight(NLOS)error,a method of calibrating the Anchor is proposed to decrease the standard deviation between Anchor and Tag.In order to reduce the influence of abnormal ranging data on positioning accuracy,extended Kalman filter(EKF)and particle filter(PF)are used to calculate position,as well as Taylor-series expansion method.Experiments of different motion paths such as loop-shaped,S-shaped and straight line and different motion speeds such as slow,normal and fast speed are carried out.The results show that PF achieves the highest accuracy in UWB positioning experiments,which is 7.74% higher than EKF and 15.60% higher than Taylor-series,but PF takes the longest time to complete the tracking process.Secondly,for the problem of IMU positioning error accumulation in the process,a joint detection method of zero-speed interval based on acceleration amplitude,acceleration sliding variance and angular velocity amplitude is proposed,and the pseudo-zero-speed point and pseudo-step point are eliminated by using time threshold judgment method.Kalman filter(KF)is used to update the state information of zerospeed time.The experimental results show that Zero-Velocity update method works in reducing the cumulative error.Thirdly,after analyzing the characteristics and shortcomings of UWB and IMU positioning technologies,the loose combination model based on KF and tight combination model based on EKF between UWB and IMU are put forward.The constructions of loose and tight combinations and data fusion methods are introduced in detail.The results show that when compared with the accuracy of UWB,it is increased by 60.61% with tight combination and 31.47% with loose combination.