Ultra-wideband (UWB) sensor technology is known to achieve high-precision indoor localization accuracy in line-of-sight (LOS) environments, but its localization accuracy and stability suffer detrimentally in non-line-of-sight (NLOS) conditions. Current NLOS/LOS identification based on channel impulse response’s (CIR) characteristic parameters (CCP) improves location accuracy, but most CIR-based identification approaches did not sufficiently exploit the CIR information and are environment specific. This paper derives three new CCPs and proposes a novel two-step identification/classification methodology with dynamic threshold comparison (DTC) and the fuzzy credibility-based support vector machine (FC-SVM). The proposed SVM based classification methodology leverages on the derived CCPs obtained from the waveform and its channel analysis, which are more robust to environment and obstacles dynamic. This is achieved in two-step with a coarse-grained NLOS/LOS identification with the DTC strategy followed by FC-SVM to give the fine-grained result. Finally, based on the obtained identification results, a real-time ranging error mitigation strategy is then designed to improve the ranging and localization accuracy. Extensive experimental campaigns are conducted in different LOS/NLOS scenarios to evaluate the proposed methodology. The results show that the mean LOS/NLOS identification accuracy in various testing scenarios is 93.27 %, and the LOS and NLOS recalls are 94.27 % and 92.57 %, respectively. The ranging errors in LOS(NLOS) conditions are reduced from 0.106 m(1.442 m) to 0.065 m(0.739 m), demonstrating an improvement of 38.85 %(48.74 %) with 0.041 m(0.703 m) error reduction. On the other hand, the average positioning accuracy is also reduced from 0.250 m to 0.091 m with an improvement of 63.49 %(0.159 m), which outperforms the state-of-the-art approaches of the Least-squares support vector machine (LS-SVM) and K-Nearest Neighbor (KNN) algorithms.

Implementing line-of-sight (LOS) and none-line-of-sight (NLOS) identification in ultra-wideband (UWB) systems is crucial. Convolutional neural network (CNN) based identification methods can extract higher-level features automatically, but they are based on channel impulse response (CIR)-turned image ingested features that impose calculation complexity and do not make use of manual features due to the data inundation risk. In this letter, we propose a novel multilayer perceptron (MLP)-based LOS/NLOS identification algorithm that can utilize both manually extracted features and feature from CNN based on raw CIR inputs with only 7.39% calculation complexity as compared to the traditional image-based CNN. Three experiments, at a teaching building, an office, and an underground mine, were conducted to verify the proposed method’s performance. Our proposed features are conducive to LOS/NLOS identification, especially the proposed raw CIR-based feature from the CNN, achieving 26.9% improvement over existing manual features. Furthermore, the proposed method outperformed the traditional image-based CNN with an improvement of 44.16%.

Indoor positioning using the geomagnetic field has become a popular technique because of the infrastructure-free characteristic and the ubiquitous magnetic signals in indoor environments. Geomagnetic intensity will generate fluctuations due to the factors such as device heterogeneity, dates, and electronic facilities. Stable geomagnetic data are essential for high-precision positioning. To find a robust and low time-consuming geomagnetic positioning model (GPM), the extreme learning machine (ELM) optimized by the enhanced genetic algorithm (EGA) is constructed. The EGA with three optimization strategies is designed to find the best initial parameters solutions for the ELM. Magnetic patterns collected by different participants using different mobile phones on different dates are extracted characteristics and divided into segments for ELM training. Extensive experiments are conducted to evaluate the proposed model's performance. The experimental results demonstrate that the EGA-based ELM model can achieve meter-level location accuracy and has good robustness under different testing conditions. The model construction is much faster than that of the popular learning algorithms, such as the convolutional neural networks (CNNs) and backpropagation (BP) networks.

Ultra-wideband (UWB) is a high-precision positioning and navigation technology, it faces significant challenges due to the abundance of non-line-of-sight (NLOS) conditions in complex indoor environments. In this study, we introduce the bidirectional encoder representations from transformers (BERT) to identify and mitigate the impact of NLOS paths using the channel impulse response (CIR). We derive three new CIR features that comprise both the time and energy characteristics of CIR sequences. These proposed features are fused with fuzzy probabilities into BERT (F-BERT), in order to identify the NLOS paths. Based on the NLOS identification results from F-BERT, a ranging classification and mitigation strategy with another BERT is further designed to enhance the ranging and positioning accuracy. The experimental results indicate that F-BERT outperforms state-of-the-art algorithms such as least-squares support vector machine (LS-SVM), convolutional neural network (CNN), and CNN with long short-term memory (CNN-LSTM) by 12.5%, 13.9%, and 14.9%, respectively, in terms of NLOS identification accuracy with LOS and NLOS recall. The proposed BERT also outperforms the existing algorithms by 36.2% in ranging error reduction in an NLOS environment. Furthermore, our proposed algorithms similarly outperform existing algorithms in mean positioning accuracy by 37.9%. Finally, our BERT algorithms achieve generality as, although they were trained in one environment, they are shown to still work well in another unknown environment.

In the rapidly evolving Internet of Things (IoT) landscape, accurate indoor positioning is increasingly vital. The proposed algorithm synergizes Ultra-Wideband (UWB) sensor with Inertial measurement unit (IMU) and Artificial Intelligence to obtain precise positioning in Non-Line-of-Sight (NLOS) scenarios. In the proposed UWB module, Large language model (LLM) such as Bidirectional Encoder Representations from Transformers (BERT) algorithm is designed to utilize the Channel Impulse Response (CIR) for effective NLOS identification and UWB ranging trustworthiness evaluation. Concurrently, the IMU module is also designed with BERT to recognize various pedestrian activity states, thereby optimizing positioning. BERT’s self attention mechanism and deep learning bidirectional training efficiently extract essential features from sequential data, capturing both local and global information. The integration of both UWB and IMU through a proposed tightly coupled algorithm significantly boosts positioning performance. Experiment campaigns demonstrate an average NLOS identification accuracy, Line-of-Sight (LOS) and F2 of 98.8%, 99.4% and 0.9926, respectively. These performance surpass the state-of-the-art Least Squares Support Vector Machine (LS-SVM), Convolutional Neural Network (CNN), CNN with Long Short-Term Memory (CNN-LSTM) up to 17.66% in NLOS identification. In terms of pedestrian activity recognition using BERT, the BERT algorithm achieves a precision(recall) of 99.3%(99.4%), notably outperforming CNN and CNN-LSTM by 17.9%(16.2%) and 11.9%(10.9%), respectively. Finally, the UWB-IMU algorithm significantly enhances positioning accuracy by 80.5%, outperforming Kalman, LSTM-EKF, and Particle filter methods by 68.1%, 48.3%, and 45.0%, respectively. The proposed approach presents a robust solution for indoor positioning for IoT applications, particularly in challenging NLOS environments.

The effective identification and mitigation of non-line-of-sight (NLOS) ranging errors are essential for achieving high-precision positioning and navigation with ultra-wideband (UWB) technology in harsh indoor environments. In this paper, an efficient UWB ranging-error mitigation strategy that uses novel channel impulse response parameters based on the results of a two-step NLOS identification, composed of a decision tree and feedforward neural network, is proposed to realize indoor locations. NLOS ranging errors are classified into three types, and corresponding mitigation strategies and recall mechanisms are developed, which are also extended to partial line-of-sight (LOS) errors. Extensive experiments involving three obstacles (humans, walls, and glass) and two sites show an average NLOS identification accuracy of 95.05%, with LOS/NLOS recall rates of 95.72%/94.15%. The mitigated LOS errors are reduced by 50.4%, while the average improvement in the accuracy of the three types of NLOS ranging errors is 61.8%, reaching up to 76.84%. Overall, this method achieves a reduction in LOS and NLOS ranging errors of 25.19% and 69.85%, respectively, resulting in a 54.46% enhancement in positioning accuracy. This performance surpasses that of state-of-the-art techniques, such as the convolutional neural network (CNN), long short-term memory–extended Kalman filter (LSTM-EKF), least-squares–support vector machine (LS-SVM), and k-nearest neighbor (K-NN) algorithms.