The attitude of an unmanned aerial vehicle (UAV) is highly related to its motion status, such as velocity and acceleration, and thus needs to be taken into consideration in UAV localization and tracking. In this paper, we study a short-range UAV localization and tracking system, where a UAV flies in the near-field region of an extremely large antenna array (ELAA). The ELAA is arranged to track the UAV by continuously estimating its position and attitude. To accomplish this task, we leverage an array partitioning approach to establish the signal model between the UAV and the ELAA based on the subarray-wise far-field assumption. Then, we characterize the relationship between UAV’s attitude and motion status based on force analysis. Building on the analysis, we formulate a probabilistic UAV tracking problem that jointly estimates the UAV position and attitude in an online fashion. A new message-passing-based algorithm is proposed to solve this problem, which combines attitude and motion status information to enhance tracking performance. We also derive the Bayesian Cramér Rao bound (BCRB) of the problem as a performance benchmark. Numerical results show that the proposed algorithm outperforms other alternatives, and demonstrate that the information fusion of the UAV attitude and motion status can effectively improve the accuracy of the UAV localization and tracking.

This paper studies a localization system, where an extremely large-scale antenna array (ELAA) is deployed at the base station (BS) to locate a user equipment (UE) residing in the near-field (Fresnel) region. We propose a novel algorithm, named array partitioning-based location estimation (APLE), for scalable near-field localization. The APLE algorithm is developed based on the basic assumption that, by partitioning the ELAA into multiple subarrays, the UE can be approximated as in the far-field region of each subarray. We establish a Bayeian inference framework based on the geometric constraints between the UE location and the angles of arrivals (AoAs) at different subarrays. Then, the APLE algorithm is designed based on the message-passing principle for the localization of the UE. APLE exhibits linear computational complexity with the number of BS antennas, leading to a significant reduction in complexity compared to existing methods. We further propose an enhanced APLE (E-APLE) algorithm that refines the location estimate obtained from APLE by following the maximum likelihood principle. The E-APLE algorithm achieves superior localization accuracy compared to APLE while maintaining a linear complexity with the number of BS antennas. Numerical results demonstrate that the proposed APLE and E-APLE algorithms outperform the existing baselines in terms of both localization accuracy and computational complexity.

Existing near-field localization algorithms generally face a scalability issue when the number of antennas at the sensor array goes large. To address this issue, this paper studies a passive localization system, where an extremely large-scale antenna array (ELAA) is deployed at the base station (BS) to locate a user that transmits signals. The user is considered to be in the near-field (Fresnel) region of the BS array. We propose a novel algorithm, named array partitioning based location estimation (APLE), for scalable near-field localization. The APLE algorithm is developed based on the basic assumption that, by partitioning the ELAA into multiple subarrays, the user can be approximated as in the far-field region of each subarray. The APLE algorithm determines the user's location by exploiting the differences in the angles of arrival (AoAs) of the sub arrays. Specifically, we establish a probability model of the received signal based on the geometric constraints of the user's location and the observed AoAs. Then, a message-passing algorithm, i.e., the proposed APLE algorithm, is designed for user localization. APLE exhibits linear computational complexity with the number of BS antennas, leading to a significant reduction in complexity compared to the existing methods. Besides, numerical results demonstrate that the proposed APLE algorithm outperforms the existing baselines in terms of localization accuracy.

Extracting entities and relations from text is a significant task of information extraction. Existing extraction models often straightforwardly produce their confident prediction results without any reconsideration or double-checking, resulting in avoidable mistakes and sub-optimal performance. In this paper, we propose a novel coarse-to-fine extraction framework, which first extracts high-potential relations as well as entities via knowledge distillation, and then rechecks the predictions via handcrafted natural language inference (NLI) task in a fine-grained manner. Specifically, based on the knowledge distillation mechanism, we train multiple teacher models iteratively through an adaptive loss function for making one teacher concentrate more on the data that others are incompetent for. Then, these complementary teacher models are utilized to provide valuable soft-label information for training a considerate student model, enabling it to generate reliable preliminary predictions. Further, these generated potential relations and entities are formulated as hypotheses, together with the original sentences as premises, serving as the input for an NLI model. Considering the linguistic diversity of relational expression, we automatically generate various semantic templates for hypotheses through an$\mathcal{N}$-gram mining strategy. Moreover, due to the existence of multi-fact sentences, a relation-guided Gaussian attention is designed to reduce the gap between the single-relation hypothesis and the multi-relation premise. To implement efficient training, we also develop several ways to generate high-quality negative samples, which help the NLI model learn to identify errors. Experimental results show that the proposed method is effective and outperforms other strong baselines on public benchmarks.

In this paper, we build up a new intelligent reflecting surface (IRS) aided multiple-input multiple-output (MIMO) channel model, named the cascaded LoS MIMO channel, that is applicable to both near-field and far-field scenarios. The proposed channel model consists of a transmitter (Tx) and a receiver (Rx) both equipped with uniform linear arrays (ULAs), and an IRS is used to enable communications between the transmitter and the receiver through the line-of-sight (LoS) links seen by the IRS. When modeling the reflection of electromagnetic waves at the IRS, we take into account the curvature of the wavefront on different reflecting elements. Based on the established channel model, we show that an IRS-assisted MIMO channel is able to support spatial multiplexing solely by the cascaded LoS links. We generalize the notion of Rayleigh distance originally coined for the single-hop MIMO channel to full multiplexing region (FMR) for the cascaded LoS MIMO channel, where the FMR is the union of all the Tx-IRS and IRS-Rx distance pairs that enable full multiplexing communication. We derive an inner bound of the FMR under a special passive beamforming (PB) strategy named reflective focusing, and provide the corresponding orientation settings of the antenna arrays that achieve full multiplexing.

In this paper, we study the performance limit of the cascaded line-of-sight (LoS) multiple-input-multiple-output (MIMO) system consisting of a transmitter (Tx) and a receiver (Rx) both equipped with uniform linear arrays (ULAs), and an intelligent reflecting surface (IRS) that enables communications between the Tx and the Rx through the cascaded LoS Tx-IRS-Rx link. We investigate the potential gain of the cascaded LoS MIMO system achieved by the optimization of the array orientations, especially when the Tx and the Rx are in the near-field of the IRS. Under a recently established near-field channel model, we formulate the problem of maximizing the input-output mutual information (MI) of the cascaded LoS MIMO system over active and passive beamforming as well as Tx and Rx array orientations. We give analytical solutions to the problem under asymptotic conditions, such as in the high/low signal-to-noise ratio (SNR) regime or with sufficiently large Tx-IRS and IRS-Rx distances. For non-asymptotic cases, we propose an alternating optimization method to solve the problem. We show that compared with only optimizing active and passive beamforming, the cascaded LoS MIMO system can harvest a significant MI gain from the additional optimization over the array orientations.

Low earth orbit (LEO) satellite has been considered as a potential supplement for the terrestrial Internet of Things (IoT). In this paper, we consider grant-free non-orthogonal random access (GF-NORA) in the orthogonal frequency division multiplexing (OFDM) system to increase access capacity and reduce access latency for LEO satellite-IoT. We focus on the joint device activity detection (DAD) and channel estimation (CE) problem at the satellite access point. The delay and the Doppler effect of the LEO satellite channel are assumed to be partially compensated. We propose an OFDM-symbol repetition technique to better distinguish the residual Doppler frequency shifts, and present a grid-based parametric probability model to characterize channel sparsity in the delay-Doppler-user domain, as well as to characterize the relationship between the channel states and the device activity. Based on that, we develop a robust Bayesian message-passing algorithm named modified variance state propagation (MVSP) for joint DAD and CE. Moreover, to tackle the mismatch between the real channel and its on-grid representation, an expectation–maximization (EM) framework is proposed to learn the grid parameters. Simulation results demonstrate that our proposed algorithms significantly outperform the existing approaches in both activity detection probability and channel estimation accuracy.

In this paper, we consider the cascaded line-of-sight (LoS) multiple-input-multiple-output (MIMO) system, in which an intelligent reflecting surface (IRS) is employed to enable the communication between a multi-antenna transmitter (Tx) and a multi-antenna receiver (Rx) by connecting the line-of-sight (LoS) links seen by the IRS. We formulate an optimization problem to maximize the input-output mutual information (MI) of the system over the passive beamforming and the Tx/Rx array orientations. Analytical solutions to this problem are provided under asymptotic conditions, such as high or low signal-to-noise ratio (SNR) regimes or sufficiently large Tx-IRS and IRS-Rx distances. For general cases, we propose an alternating optimization method to solve the problem. Numerical results validate the effectiveness of the proposed optimization method and our analyses.

In this paper, we build up a new intelligent reflecting surface (IRS) aided multiple-input multiple-output (MIMO) channel model, named the cascaded line-of-sight (LoS) MIMO channel. The proposed channel model consists of a transmitter (Tx) and a receiver (Rx) both equipped with uniform linear arrays (ULAs), and an IRS used to enable communications between the Tx and the Rx through the LoS links seen by the IRS. To model the reflection of electromagnetic waves at the IRS, we take into account the curvature of the wavefront on different reflecting elements (REs), which is distinct from most existing works with the plane-wave assumption. Based on the established model, we study the spatial multiplexing capability of the cascaded LoS MIMO system. We come up with the notion of full multiplexing region (FMR), which is the union of Tx-IRS and IRS-Rx distance pairs that enable full multiplexing communication over the cascaded LoS MIMO system. We also derive an inner bound of the FMR under a specific passive beamforming strategy named reflective focusing.

With the advent and applications of sensor networks, pervasive traffic information can be gathered rapidly and collaboratively, which can upgrade traditional traffic monitoring system and enable many unprecedented services for travelers. In this paper, three-tiered sensor network architecture is proposed to approach the novel traffic information service system that would dramatically improve the ways and qualities of traffic information collection. A brief analysis of the important requirements of system, as well as the key issues that are faced in the design process, is described. The detailed strategies at each level, from the network architecture, to the sensor unit configuration and software deployment, to the operation of system, are explained. The implementation of a pilot platform is discussed in the end.