A Combined UWB/IMU Localization Method with Improved CKF

Pengfei Ji,Zhongxing Duan,Weisheng Xu

DOI: https://doi.org/10.3390/s24103165

IF: 3.9

2024-05-17

Sensors

Abstract:Aiming at the problem that ultra-wide band (UWB) cannot be accurately localized in environments with large noise variations and unknown statistical properties, a combinatorial localization method based on improved cubature (CKF) is proposed. First, in order to overcome the problem of inaccurate local approximation or even the inability to converge due to the initial value not being set near the optimal solution in the process of solving the UWB position by the least-squares method, the Levenberg–Marquardt algorithm (L–M) is adopted to optimally solve the UWB position. Secondly, because UWB and IMU information are centrally fused, an adaptive factor is introduced to update the measurement noise covariance matrix in real time to update the observation noise, and the fading factor is added to suppress the filtering divergence to achieve an improvement for the traditional CKF algorithm. Finally, the performance of the proposed combined localization method is verified by field experiments in line-of-sight (LOS) and non-line-of-sight (NLOS) scenarios, respectively. The results show that the proposed method can maintain high localization accuracy in both LOS and NLOS scenarios. Compared with the Extended Kalman filter (EKF), unbiased Kalman filter (UKF), and CKF algorithms, the localization accuracies of the proposed method in NLOS scenarios are improved by 25.2%, 18.3%, and 11.3%, respectively.

engineering, electrical & electronic,instruments & instrumentation,chemistry, analytical

What problem does this paper attempt to address?

The problem that this paper attempts to solve is: in an environment with large noise variations and unknown statistical characteristics, the problem that ultra - wideband (UWB) cannot achieve accurate positioning. Specifically, the paper proposes a combined UWB/IMU positioning method based on an improved cubature Kalman filter (CKF) to address the following challenges:

1. **Initial value limitation**: When the traditional UWB positioning method uses the least - squares method for solution, the initial value needs to be close to the optimal solution; otherwise, it may lead to inaccurate local approximation or even non - convergence. For this reason, the paper adopts the Levenberg - Marquardt (L - M) algorithm to optimize the solution of the UWB position.

2. **Influence of measurement noise and non - line - of - sight (NLOS) noise**: These noises will reduce the positioning accuracy. In order to improve the positioning accuracy, the paper introduces an adaptive factor to update the measurement noise covariance matrix in real - time and adds a forgetting factor to suppress filter divergence, thereby improving the traditional CKF algorithm.

3. **Influence of time - varying noise**: Existing navigation algorithms combining UWB and IMU usually process UWB outliers based on known noise statistical characteristics and rarely consider the influence of time - varying noise on the positioning system. Therefore, the paper designs a new fusion method that can better handle time - varying noise.

Through the above improvements, the paper aims to improve the positioning accuracy of robots in complex indoor environments, especially in non - line - of - sight (NLOS) scenarios. The experimental results show that compared with the extended Kalman filter (EKF), the unbiased Kalman filter (UKF) and the traditional CKF algorithm, the proposed method improves the positioning accuracy in NLOS scenarios by 25.2%, 18.3% and 11.3% respectively.

### Key formulas

1. **UWB ranging model**:
\[
R_i = d_i+\eta_i + b_i=c\tau_i,\quad i = 1,2,\ldots,N
\]
where
\[
d_i=\sqrt{(x - x_i)^2+(y - y_i)^2}
\]
\(R_i\) is the measured distance between the tag and the base station, \(d_i\) is the true distance, \(b_i\) is the measurement noise (assumed to be zero - mean Gaussian distribution), and \(\eta_i\) is the positive distance deviation caused by the NLOS effect.

2. **Objective function of the L - M algorithm**:
\[
\varepsilon(x,y)=\min\left(\frac{1}{2N}\sum_{i = 1}^N\beta_i\left[\sqrt{(x - x_i)^2+(y - y_i)^2}-R_i\right]^2\right)
\]

3. **State equation (considering process noise)**:
\[
\begin{cases}
x_k=x_{k - 1}+\Delta t\cdot\dot{x}_{k - 1}+\frac{1}{2}\Delta t^2a_{x,k - 1}+w_{1,k}\\
y_k=y_{k - 1}+\Delta t\cdot\dot{y}_{k - 1}+\frac{1}{2}\Delta t^2a_{y,k - 1}+w_{2,k}\\
\dot{x}_k=\dot{x}_{k - 1}+\Delta t\cdot a_{x,k - 1}+w_{3,k}\\
\dot{y}_k=\dot{y}_{k - 1}+\Delta t\cdot a_{y,k - 1}+w_{4,k}\\
\theta_k=\theta_{k - 1}+\Delta t\cdot w_{5,k}
\end{cases}
\]

4. **State transition equation of the improved CKF algorithm**:
\[
X_k = F_{k - 1}X_{k - 1}+B U