In a non-line-of-sight (NLOS) environment, high accuracy ultra-wideband (UWB) positioning has been one of the hot topics in studying indoor positioning. Aiming at the UWB and inertial measurement unit (IMU) fusion vehicle positioning, a constraint robust iterate extended Kalman filter (CRIEKF) algorithm has been proposed in this paper. It has overcome the innate defect of the extended Kalman filter against non-Gaussian noise and the shortcoming of the robust extended Kalman filter algorithm, which has just processed the non-Gaussian noise solely based on the prior information. Our algorithm can update the observation covariance based on the posteriori estimate of the system in each iteration, and then update the posteriori distribution of the system based on the obtained covariance to significantly reduce the influence of non-Gaussian noise on positioning accuracy. Also, with the introduction of motion constraints, such as zero velocity, pseudo velocity and plane constraints, it can achieve a smoother positioning result. The experimental result proves that through the CRIEKF-based UWB/IMU fusion robot positioning method, a mean positioning accuracy of around 0.21 m can be achieved in NLOS environments.

中文翻译：

---

基于运动约束的移动车辆UWB / IMU融合定位研究

在非视距（NLOS）环境中，高精度超宽带（UWB）定位已成为研究室内定位的热门话题之一。针对UWB与惯性测量单元（IMU）融合车辆的定位问题，提出了一种约束鲁棒的迭代扩展卡尔曼滤波（CRIEKF）算法。它克服了扩展卡尔曼滤波器针对非高斯噪声的先天缺陷以及健壮的扩展卡尔曼滤波器算法的缺点，后者仅根据先验信息就已经处理了非高斯噪声。我们的算法可以基于每次迭代中系统的后验估计来更新观测协方差，然后根据获得的协方差更新系统的后验分布，以显着降低非高斯噪声对定位精度的影响。另外，通过引入运动约束，例如零速度，伪速度和平面约束，可以实现更平滑的定位结果。实验结果证明，通过基于CRIEKF的UWB / IMU融合机器人定位方法，在非视距环境下平均定位精度可以达到0.21 m左右。