UWB

Ultra-wideband (UWB) has attracted much attention in indoor positioning due to its high accuracy, low power consumption, and excellent anti-jamming capability. However, due to the complexity of the usage environment, UWB signals are often obstructed by objects such as metals and walls, leading to non-line-of-sight (NLOS) conditions and a decrease in positioning accuracy. To fully consider the effect of ranging errors caused by different occlusions on UWB signals, firstly, this paper analyzes the attenuation characteristics of UWB in different NLOS scenarios. By combining the channel impulse correspondence (CIR) and Markov Transition Field (MTF), an efficient classification method for NLOS identification is proposed. Further, for the signal attenuation characteristics in different scenarios, this paper proposes a polynomial fitting based on the Negative Log-Likelihood function (NLLPF), which greatly mitigates the ranging errors caused by different occlusions. Additionally, the dynamic scene adaptation of UWB localization is enhanced by fusing the Inertial Measurement Unit (IMU) using a confidence-considering EKF (CEKF). Finally, experimental results show that in indoor scenarios, the proposed method achieves a mean absolute error (MAE) of 5.60 cm, while in corridor scenarios, the MAE is 16.36 cm, significantly lower than the 12.42 cm and 54.88 cm of the original methods. 

超宽带（Ultra-wideband, UWB）凭借其高精度、低功耗与优异的抗干扰能力，在室内定位领域受到广泛关注。然而，由于使用环境复杂多变，UWB信号常受到金属、墙体等物体遮挡，进而引发非视距（Non-Line-of-Sight, NLOS）场景并导致定位精度下降。为充分考量不同遮挡对UWB信号产生的测距误差影响，本文首先分析了UWB在不同NLOS场景下的信号衰减特性；结合信道冲激对应（Channel Impulse Correspondence, CIR）与马尔可夫转移场（Markov Transition Field, MTF），提出了一种高效的NLOS识别分类方法。进一步针对不同场景下的信号衰减特性，本文提出了一种基于负对数似然函数的多项式拟合方法（Negative Log-Likelihood function based Polynomial Fitting, NLLPF），可有效缓解各类遮挡引发的测距误差。此外，通过采用考虑置信度的扩展卡尔曼滤波（confidence-considering EKF, CEKF）融合惯性测量单元（Inertial Measurement Unit, IMU）数据，本文进一步提升了UWB定位的动态场景适配能力。最后，实验结果表明，在室内场景下，本文所提方法的平均绝对误差（Mean Absolute Error, MAE）为5.60厘米；在走廊场景下，MAE为16.36厘米，显著低于原有方法的12.42厘米与54.88厘米。