Massive heterogeneous IoT networks encounter significant random access (RA) challenges due to diverse Quality of Service (QoS) requirements and resource constraints. To address these issues, we first propose a fuzzy logic-assisted multi-criterion access priority ranking (FL-MCAPR) scheme to prioritize RA for IoT devices, integrating delay, channel interference, and energy factors. The resulting suitability values enable adaptive and fine-grained backoff adjustments in large-scale IoT deployments. Next, hybrid RA control schemes with a deployability-descending double-queue (D3Q) structure and access priority-backoff window model optimize preamble and backoff allocation. In addition, preamble parallelization and early-stage collision detection enhance RA throughput by expanding resources and reducing collisions. Using D3Q, the analytical RA throughput is derived, informing an optimization problem to determine Access Class Barring (ACB) factors balancing low delay and energy efficiency, considering all RA resources. Building on these results, the hybrid fuzzy logic-guided RA management (HFL-RAM) scheme is developed for comprehensive RA management, systematically evaluated in terms of delay and throughput. Finally, a lightweight pseudo-Bayesian estimation method is applied, which relies solely on two observable quantities to estimate the contending MTCD traffic. Simulation results demonstrate that the proposed HFL-RAM scheme consistently outperforms conventional approaches, effectively managing traffic heterogeneity across a wide range of traffic loads.

In this paper, we consider a practical integrated sensing and communication (ISAC) scenario in industrial Internet of Things (IIoT). In this scenario, a robot acted as a mobile base station (BS) performing sensing to locate a logistics transport robot (LTR) while also communicating with multiple production lines (PL). We predict the motion parameters of LTR in each time slot and derive the Cramér-Rao bound (CRB) of angle and distance estimation. Afterward, we formulate a joint CRB minimization problem by optimizing the transmit beamforming for communication and sensing. We convert the formulated problem into a two-tier alternating optimization approach by constructing precise surrogates for the non-convex objective functions and constraints. A closed-form expression is derived for solving the outer layer problem. In addition, we employ a convex framework to address the inner layer problem. Numerical results validate the superiority of the proposed algorithm, especially when the BS is far away from the sensing target.

In this paper, a comprehensive and detailed uplink performance analysis is provided for cell-free massive multipleinput multiple-output orthogonal frequency division multiplexing (CF m-MIMO OFDM) systems, which consider the impact of multiple user equipment (UE) heterogeneous factors. This is the first performance analysis work on CF m-MIMO OFDM systems that simultaneously accounts for the heterogeneous mobility speed, activation probability and serving priority. Considering that UE's serving priority determines the amount of its allocated time-frequency resources, a novel closed-form expression of uplink spectral efficiency (SE) is derived by weighting each UE's SE based on its allocated time-frequency resources. The derived SE expression can quantify the impact of multiple UE heterogeneous factors on the uplink performance. Additionally, the SE performance analysis of local processing and fully centralized processing is also included for comparison. Simulation results show that CF m-MIMO OFDM systems under multiple UE heterogeneous factors outperform existing CF m-MIMO systems in terms of the 90 %-likely uplink SE, and allow a trade-off among fronthaul overhead, complexity and SE performance.

Recent advancements in large language models (LLMs) have significantly improved text-to-SQL systems. However, most datasets and LLM-based methods tend to focus narrowly on SQL generation, often neglecting the complexities inherent in real-world conversational queries. This oversight can result in unreliable responses, particularly for ambiguous questions that cannot be directly addressed with SQL. To address this gap, we propose MMSQL, a comprehensive test suite designed to evaluate LLMs’ question classification and SQL generation capabilities by simulating real-world scenarios with diverse question types and multi-turn Q&A interactions. Utilizing MMSQL, we assessed the performance of popular LLMs, including both open-source and closed-source models, and identified key factors influencing their performance in these contexts. Furthermore, we introduce an LLM-based multi-agent framework that employs specialized agents to identify question types and determine appropriate answering strategies. Experimental results demonstrate that this method effectively enhances baseline models’ ability to handle diverse question types in conversational scenarios. Our approach simultaneously considers multiple question types and multi-turn interactions, providing a new, realistic perspective, and offering a valuable advancement toward more reliable and versatile text-to-SQL systems. Our dataset and code are publicly available at https://mcxiaoxiao.github.io/MMSQL.

The integration of Internet-of-Things (IoT) and the fifth-generation (5G) networks presents challenges due to low access efficiency caused by massive random access (RA) requests. To this end, both preamble parallelization (PP) and colliding preambles reuse (CPR) modes are proposed as critical RA control methods to enhance access performance. In this paper, we aim to maximize the random access efficiency (RAE) in a smart city scenario to determine the optimal control mode selection between the PP and CPR over the device heterogeneity with limited RA resources. We establish an access order-backoff window (AOBW) mapping model, where RA requirements are mapped onto the backoff time. It offers greater flexibility of backoff window size than previous work to guarantee diverse application and service requirements. Thanks to the derived closed-form expressions of the actual RAE, an RAE maximization algorithm is developed, which optimizes performance across both PP and CPR modes, achieving optimal performance in access delay and access throughput.

In industrial Internet of Things (IIoT), short-packet communication requires higher reliability and lower latency, enabling Grant-free (GF) access in massive machine-type communication (mMTC) to gain a tremendous research interest. However, in practical applications, the accurate channel state information (CSI) required by coherent GF (C-GF) is challenging to acquire. Concurrently, the spectral efficiency and reliability of the non-coherent (NC-GF) scheme are areas necessitating enhancement. Inspired by sparse vector coding (SVC), this paper introduces an innovative SVC-GF scheme. This scheme redefines joint user activity detection (UAD) and data decoding as a bifurcated sparse recovery issue. A novel block refinement orthogonal matching pursuit and multipath matching pursuit (BROMP-MMP) algorithm is designed to resolve this intricate problem. Simulation results demonstrate that the proposed SVC-GF scheme outperforms traditional C-GF and NC-GF schemes in terms of probability of successful detection and average block error rate (BLER) performance, heralding a paradigm shift for massive GF access.

Enabling Internet-of-Things (IoT) in fifth generation (5G) networks is challenging due to the low access efficiency in the presence of massive random access (RA) requests. To tackle this, we investigate multi-criterion RA ranking and random access efficiency (RAE) maximization for massive IoT networks to deal with devices' heterogeneity and limited RA resources. A fuzzy logic-guided suitability ranking (FL-SR) scheme is proposed, where multiple criteria are considered such as RA delay, movement speed, and battery capacity to ensure that various service and application requirements are met. With normalized suitability and deployability from the FL-SR scheme, the backoff window size gets more flexible than previous work. A fuzzy logic guided random access efficiency optimization (FL-RAEO) algorithm is proposed to maximize the RAE. Thanks to the derived closed-form expressions for the optimal RAE, the FL-RAEO algorithm achieves optimal performance in average access delay, access throughput. and successful access rate.

In the context of the industrial Internet of Things (IIoT), accommodating massive connectivity presents significant challenges for random access (RA) networks, primarily due to scalability and diverse quality-of-service (QoS) requirements, resulting in severe preamble collisions. We propose a novel Preamble Parallelization Based Random Access with Colliding Preamble Reuse (PP-RACPR) scheme applicable to the RA procedure. Our method enhances the RA procedure by allowing machine-type communication devices (MTCDs) to transmit multiple preambles in parallel during the initial step of the RA procedure, increasing the successful access rate. Additionally, MTCDs are empowered to reuse part of colliding preambles by identifying them earlier in the process, thereby improving the preamble utilization ratio (PAUR) for the RA network. Finally, we conduct a comprehensive mathematical analysis of the proposed scheme, focusing on the system PAUR, and corroborate our analytical framework through extensive simulations, demonstrating its feasibility and efficacy in supporting massive MTCDs and mitigating preamble collisions.

Multimodal summarization (MS) aims to generate a summary from multimodal input. Previous works mainly focus on textual semantic coverage metrics such as ROUGE, which considers the visual content as supplemental data. Therefore, the summary is ineffective to cover the semantics of different modalities. This paper proposes a multi-task cross-modality learning framework (CISum) to improve multimodal semantic coverage by learning the cross-modality interaction in the multimodal article. To obtain the visual semantics, we translate images into visual descriptions based on the correlation with text content. Then, the visual description and text content are fused to generate the textual summary to capture the semantics of the multimodal content, and the most relevant image is selected as the visual summary. Furthermore, we design an automatic multimodal semantics coverage metric to evaluate the performance. Experimental results show that CISum outperforms baselines in multimodal semantics coverage metrics while maintaining the excellent performance of ROUGE and BLEU.

In industrial Internet of Things (IIoT), the existing access class barring (ACB) random access (RA) strategy suffers severe performance degradation with massive contention devices, due to high probability of access collision. In this paper, we propose a collision-aware (CA) ACB RA scheme by reusing the colliding preambles, to enhance the resource utilization. The proposed scheme employs dynamic adjustment of the ACB factor and the preamble resources for delay-sensitive and -non-sensitive devices, respectively. A joint optimization problem is formulated and solved to maximize the preamble utilization ratio (PAUR) subject to the delay constraints and the available preambles. The system performance is evaluated by a Markov Chain based analytical model. Simulation results verify the correctness of our analysis and also show that the proposed CA-ACB RA scheme significantly outperforms the existing ACB RA schemes in terms of PAUR, network throughput, and average access delay.