Novel double-Kalman filtering method

Abstract

The invention discloses a novel double-Kalman filtering method. The method comprises the following steps: determining a current NLOS error range; adjusting the state error covariance and the noise error covariance of the classical Kalman filter at the current moment according to the NLOS error range; inputting the adjusted state error covariance and noise error covariance into a classical Kalman filter to carry out state estimation at the current moment; the state estimation value and the state error variance output by the classical Kalman filter are input into an extended Kalman filter, and the extended Kalman filter carries out prediction by taking the state estimation value and the state error variance output by the classical Kalman filter as a measurement value and a measurement noise variance of the extended Kalman filter respectively. According to the method, fast convergence can be achieved, large system error fluctuation does not exist, and short-time or long-time continuous NLOS errors and outliers can be restrained.

Description

technical field [0001] The present invention relates to the technical field of filtering methods, in particular to a novel dual Kalman filtering method. Background technique [0002] In 1960, R. E. Kalman published a work for solving linear discrete systems based on recursive algorithms. In this work, Kalman filtering got rid of the shortcomings of Wiener filtering that was widely used. Since then, with the With the invention of the computer, the computing power has been greatly improved, and the Kalman filter has been widely used in scientific research, military and civilian fields. The advantage of Kalman filtering is that it is not only used to estimate the current state of the signal, but also to predict the future state of the signal without knowing the exact model of the system, so Kalman filtering is widely used in radar and navigation systems, Follow the established goals. Its basic idea can be expressed as: obtain the accurate value and predict the estimated value...

AI Technical Summary

Problems solved by technology

[0011] 1. The classic Kalman filter algorithm, extended Kalman filter algorithm, unscented Kalman filter algorithm, and volumetric Kalman filter algorithm cannot realize the adjustment of Q and R, but in the actual target tracking process, due to various objective factors such as the distance of the target Influenced by factors, the observation noise changes at any time. In the above filtering algorithms, Q and R are constant values, and need to be set in advance based on empirical values. The noise variance is inconvenient to determine and may cause unsatisfactory tracking results.

[0012] 2. When the classic Kalman filter algorithm, extended Kalman filter algorithm, unscented Kalman filter algorithm, and volumetric Kalman filter algorithm have a large NLOS distance error in the measured value, short-term or continuous tracking failure will occur. The positioning effect is poor, and it is difficult to cope with the relatively large error impact on the system caused by the short or long time delay caused by non-line-of-sight in complex indoor environments

[0013] 3. The classic Kalman filter algorithm is only suitable for linear systems and not for nonlinear systems. The extended Kalman filter, unscented Kalman filter, and volumetric Kalman filter algorithms are suitable for nonlinear systems, and their operation speed is faster than that of classical Kalman filters. The algorithm is slower, the adaptive Kalman filter algorithm is suitable for linear systems, but can also be improved to be suitable for nonlinear systems

[0014] 4. The adaptive Kalman filter algorithm can adjust the noise parameters in real time during the iterative process, but Q and R cannot be adjusted at the same time. Generally, only the R value is adjusted, and when there is a large NLOS distance error in the measured value, Adaptive Kalman filter may diverge and be difficult to converge

Therefore, the accuracy of the dual Kalman filter method cannot meet the requirements of practical applications

In addition, when there are outliers in the measured values, due to the large step size of one filter in the double Kalman filter method, the system delay will produce a large tracking error, and the convergence speed of the state posterior estimation will also be reduced

Images