Ultra-Wide Band (UWB) Non-Line-of-Sight (NLOS) identification is a crucial task in wireless localization systems. Various Deep Learning (DL) solutions have demonstrated promising outcomes in UWB NLOS identification by utilizing Channel Impulse Response (CIR) and channel characteristics. However, effective and robust UWB NLOS identification on resource-constrained edge devices remains a challenge. Hence, this paper presents a self-attention-assisted Tiny Machine Learning (TinyML) solution that offers an effective representation for UWB NLOS identification. To overcome computational limitations, a feature selection method is devised for the proposed data-driven DL-based approach. By leveraging feature selection, the self-attention mechanism enhances the representation capability of a pre-trained model for UWB NLOS identification. The proposed method is evaluated on both personal computer (PC) and edge platforms, and compared against multiple baselines. The evaluation demonstrates its effective representation and optimal performance on both PC and edge platforms, as indicated by various metrics. Thanks to the effective representation, the proposed method also enables the quantized model to achieve State-of-The-Art (SOTA) in UWB NLOS identification, while significantly accelerating inference efficiency at the edge.

Ultra wide-band (UWB) brings many benefits to wireless sensing. Theoretically, UWB can achieve centimeter-level ranging accuracy. However, in practical applications, multipath fading channel (MPF) problems and antenna delay effects will adversely affect the ranging accuracy of UWB. The sample data of UWB contain rich channel characteristics, which can be extracted to characterize the implicit relation of UWB ranging error. Undoubtedly, the UWB channel characteristics will show different importance in characterizing UWB ranging errors in different environments. Therefore, this letter proposes an attention-assisted UWB ranging error compensation algorithm. Using the attention mechanism, the significance of the extracted UWB channel characteristics in various environments can be re-evaluated to improve the performance of the deep neural network (DNN) model. The experimental results prove that using the proposed algorithm, the 75 % error lines of the original indoor and outdoor ranging errors are reduced from 13.32 cm and 19.41 cm to 5.74 cm and 5.05 cm, and median errors are compensated from 7.00 cm and 15.42 cm to 2.78 cm and 2.69 cm, respectively.

Radio Frequency Identification technology (RFID) is the key technology to realize the Internet of things. Currently, using RFID to locate objects indoor is a hot topic in the research on the application of RFID. In this paper, we propose an RFID system based on RFID technology and Wi-Fi wireless communication technology. And we design a corresponding handheld device client software to realize a visual book search and management system. This system can not only greatly improve the efficiency of book search and management but also save the manpower and material resources to a large extent, which has a practical application value.

A lightweight and effective algorithm is benefit for real-world applications and edge devices. This letter presents a lightweight algorithm for Non-Line-of-Sight (NLOS) detection in Ultra-Wideband (UWB) systems, using the referenced Channel Impulse Response (CIR) similarity metrics. It retains essential UWB channel characteristics while incorporating a Radar Cross-Section (RCS)-inspired feature engineering method to capture the effective NLOS representation. It bears novel features to significantly reduce the input data dimensions, model parameters, and Floating-point operations (Flops) by 64.9%, 78.5%, 76.0%, respectively, with comparable performance to the latest State-Of-The-Art (SOTA) solutions. We demonstrate its effectiveness through actual measurements and comparative experiments in mixed complex Line-Of-Sight (LOS)/NLOS dataset, offering a computationally efficient solution for UWB NLOS detection with effective representation.

The ability to autonomously navigate and explore complex 3D environments in a purposeful manner, while integrating visual perception with natural language interaction in a human-like way, represents a longstanding research objective in Artificial Intelligence (AI) and embodied cognition. Vision-Language Navigation (VLN) has evolved from geometry-driven to semantics-driven and, more recently, knowledge-driven approaches. With the introduction of Large Language Models (LLMs) and Vision-Language Models (VLMs), recent methods have achieved substantial improvements in instruction interpretation, cross-modal alignment, and reasoning-based planning. However, existing surveys primarily focus on traditional VLN settings and offer limited coverage of LLM-based VLN, particularly in relation to Sim2Real transfer and edge-oriented deployment. This paper presents a structured review of LLM-enabled VLN, covering four core components: instruction understanding, environment perception, high-level planning, and low-level control. Edge deployment and implementation requirements, datasets, and evaluation protocols are summarized, along with an analysis of task evolution from path-following to goal-oriented and demand-driven navigation. Key challenges, including reasoning complexity, spatial cognition, real-time efficiency, robustness, and Sim2Real adaptation, are examined. Future research directions, such as knowledge-enhanced navigation, multimodal integration, and world-model-based frameworks, are discussed. Overall, LLM-driven VLN is progressing toward deeper cognitive integration, supporting the development of more explainable, generalizable, and deployable embodied navigation systems.

Developing more intelligent instrument via artificial intelligence (AI) is highly valuable. Ultrawide band (UWB), with rich radio frequency (RF) and channel impulse response (CIR) features, is widely favored in geospatial applications with AI-based enhancement. This article presents an AI-enhanced UWB solution using a dynamic neural network (DNN) approach, for the improvement of ranging and positioning performance. This dynamic AI-enabled UWB (DAI UWB) solution, employs a pretrained channel estimation model as the policy network to adjust a changeable error fitting network structure for efficient ranging error compensation. Tested in various complex indoor-outdoor environments (up to about 110 m), DAI UWB’s performance was evaluated using mixed line-of-sight (LOS)/nonline-of-sight (NLOS) datasets. Comparative results indicate that DAI UWB outperforms the state-of-the-art baselines in ranging and positioning stability. Specifically, in ranging tasks, DAI UWB performs outstanding ranging error compensation while reducing computational waste through dynamic inference routing. Moreover, DAI UWB surpasses all baselines in providing the most stable and accurate positioning results, with the most robust statistical significance. Additionally, DAI UWB achieved a 92.8% accuracy and 93.5% accuracy of channel estimation in real-world and the public benchmark tests, playing a positive role to boost positioning system performance. Furthermore, we believe that DAI UWB’s “teaching according to students’ aptitude” design philosophy through dynamic AI methods is instructive for developing other AI-enabled advanced instruments.