The evolution of Artificial Intelligence (AI) has reached a critical point, where agentic AI systems demonstrate strong capabilities in goal formulation and planning but remain difficult to deploy in real-world settings due to their limited grounding in physical environments. These limitations arise from the challenges of partial observability, actuation uncertainty, and strict resource constraints that characterize the physical world. This survey argues that the Artificial Intelligence of Things (AIoT) provides the necessary foundation to embed agentic intelligence into such environments by enabling continuous interaction between sensing, reasoning, and action. We analyze the synergy between goal-driven agentic AI and distributed AIoT infrastructures and present a unified taxonomy of AIoT-enabled agentic architectures, highlighting trade-offs across centralized, edge-native, and hybrid deployment models. The survey further examines key enabling technologies, including edge intelligence, semantic communication, digital twins, and trust mechanisms, and discusses how they integrate into cognitive control loops. Through representative applications in smart cities, industrial automation, healthcare, and energy systems, we show how this convergence moves automation beyond rule-based behavior toward context-aware autonomy. Finally, we identify open challenges related to long-horizon safety, resource-aware intelligence, and ethical governance, and outline research directions toward robust, trustworthy, and socially embedded autonomous systems.

Fraud detection in Internet-of-Things (IoT) applications remains a pressing challenge. Adversaries exploit injection, eavesdropping, and man-in-the-middle attacks that often evade conventional detection pipelines. Existing blockchain and Machine Learning (ML) based solutions improve accuracy but lack verifiability, auditability, and resilience against quantum-era threats. We proposeQuanFraud, a protocol that integrates Greenberger–Horne–Zeilinger (GHZ)–$\theta$quantum state verification, Decentralized Identifiers (DID), and a Quantum Support Vector Classifier (QSVC) within an auditable blockchain framework. The scheme ensures that fraud detection outcomes are not only data-driven but also cryptographically verifiable and resistant to identity-correlation and replay attacks. We evaluateQuanFraudon a financial dataset of 20,000 records (117 features), using Principal Component Analysis (PCA) and the Synthetic Minority Oversampling Technique (SMOTE) under 10- fold cross-validation. Results show that classical baselines such as Random Forest and XGBoost achieve balanced accuracy above 77%, while QSVC alone yields 42.1$\pm$2.8%. This gap indicates that the contribution ofQuanFraudis not accuracy leadership but a verifiable, auditable fraud-detection protocol under Noisy Intermediate-Scale Quantum (NISQ) constraints, where QSVC provides kernel-level privacy, quantum state verification, and on-chain checks that classical models do not offer. We further discuss complexity and scalability, highlighting the scheme's suitability for deployment in resource-constrained IoT environments.

Emerging sixth-generation (6G) edge networks demand intelligent, scalable, and privacy-preserving learning systems that support real-time decision-making and natural language-driven control. In addition to training, these systems must also support federated unlearning (FU), the ability to selectively remove user data without full model retraining. However, existing federated learning (FL) and FU frameworks lack adaptability, require manual hyperparameter tuning, and are ill-suited for dynamic, resource-constrained environments. To address these challenges, we propose SLM-FARL, a hierarchical multi-agent deep reinforcement learning (MARL) framework that integrates small language models (SLMs) with FL and FU processes for autonomous, privacy-compliant learning aligned with user-level data removal demands. We implement a customized MAPPO algorithm to enable stable and adaptive policy updates across distributed SLM agents, orchestrated by a central LLM controller that supports human-in-the-loop interaction. To ensure real-time responsiveness and deployment efficiency, we incorporate SLM optimization techniques such as quantization and knowledge distillation, reducing model size and latency while maintaining performance. The proposed framework is evaluated on the UCI Adult dataset using 120 clients and demonstrates up to 15.78% higher FL-FU accuracy compared to baseline methods. The MAPPO Loss decreased by 90.12% indicates highly effective MARL convergence and Policy Entropy drop by 84.31% shows it confident policy decisions. The KD demonstrated an overall 17.56% improved performance over other model compression techniques. Thus, these metrics affirms the robustness and adapt-ability of our proposed SLM-FARL framework.

Tomatoes are essential fruits in numerous nations for their vast demand. It is very important to maintain the freshness of tomatoes. One of the primary challenges in the recent culinary landscape is accurately identifying healthy tomatoes while effectively eliminating damaged or rejected ones. Existing approaches employ various strategies for categorizing tomato fruit, but they often suffer from inaccuracies, slow detection, and suboptimal performance. Thus, motivated by this gap, in this paper, we propose a novel machine learning (ML) framework, ViT-SENet-Tom , which is a hybrid vision transformer (ViT) model with squeeze and excitation (SENet) block network for fast, accurate, and efficient tomato fruit classification. The framework works on three tomato classes, respectively, the ripe, unripe, and reject. In developing the proposed model, we utilized advanced and newly designed layers and functions. This integration created a more complex and sophisticated neural network, significantly enhancing efficiency and contributing to the model’s novelty. Our chosen dataset was small initially, but we implemented augmentation techniques to increase its size. This approach made our system more reliable, efficient, and effective. The hybrid ViT-SENet framework employs encoders and self-attention networks with squeeze and excitation channel functions to allow precise, robust, fast, and efficient tomato classification. In simulation, the framework achieves a training accuracy of 99.87% and validation accuracy of 93.87%, indicating the precise classification of tomatoes. Besides, this work tests accuracy using fivefold cross-validation. The highest accuracy seen at fold-5 is 99.90%. These testing results demonstrate the efficacy of the proposed framework in real-deployment scenarios. The implementation has the potential to provide enhanced and more sustainable food security and safety in future.

In smart communities, Electric vehicles (EVs) have grown in popularity as a key component of the energy ecosystem where the focus has turned to the generation of clean, sustainable energy. The integration of EVs, charging stations (CS), and smart grids (SG), however, poses significant challenges in terms of energy trading (ET) optimization and profit maximization. Next, trust is another challenge in the ET ecosystem among the communicating entities (EVs, CS, and SG) to buy and sell energy. Recent studies have overlooked the fact of ET among CS and SG, and mostly have focused on ET by EVs. However, at peak loads, SG may experience bottlenecks in energy dissipation, and thus excess energy collected by CS from EVs might be traded to SG to manage loads during peak times. So, we propose a framework, StackGrid, that leverages the capabilities of Vehicle-to-Grid (V2G) systems over a blockchain network. We design a Stackelberg game between CS and SG for profit maximization of both parties and to obtain optimal payoff equilibria. The framework is powered over the 5G ultra-reliable low latency communications (uRLLC) service for real-time ET response and data exchange. To address blockchain scalability concerns, we incorporate Interplanetary File Systems (IPFS) as local off-chain ledgers, where only meta-information is stored on-chain to handle blockchain scaling issue. The framework is evaluated on metrics like 5G service latency, optimal payoff scenario, attack probability, and node throughput. The obtained results indicate StackGrid viability in real ET setups, with benefits for sustainable and efficient energy management.

The emergence of massive automation has transformed Industrial Internet-of-Things (IIoT) to become adaptive, self-healing, and autonomous. In IloT, the increased volume of data traffic has raised questions about the privacy and security of shared sensor data, resource management, the accuracy of trained models, and the authenticity of network traffic in operation. Thus, conventional security paradigms and centralized learning models are outdated to support the IloT operational space. Modern solutions like federated learning (FL) and onion routing (OR) are integrated into IloT to secure and optimize link communication and improve the computational requirements of central model training. Thus, the paper integrates FL and OR in IloT, and presents a framework FedOnion, where federated classifiers are proposed at intermediate OR circuits, which preserves anonymity and privacy of data sharing among nodes in IloT. In this frame-work, an FL-assisted network traffic classification approach is presented for malicious or non-malicious data requests forwarded to the OR network. Malicious requests are discarded at the next onion router, and it prevents the shared key from getting compromised, as the hash is computed at each hop to signify that data is not tampered with. FL-classifiers divide the overall dataset into small segments, which alleviates the computational burden on OR links and improves the detection rate of malicious data requests. The proposed framework's effectiveness is demonstrated on real-world IloT datasets, based on security and computational parameters. The obtained results indicate the practical viability of the scheme for critical industrial setups which paves the way towards a robust and secured industrial future. The proposed framework is assessed using different performance parameters, such as FL MSE$(10^{-9}$at 300 epochs), onion circuit compromisation rate (16%), and 5G modulation scheme.

The paper examines the complex structure of quan-tum computing and outlines its key elements, including qubits, quantum gates, superposition, and entanglement. Our inves-tigation goes beyond standard analysis, providing a singular synthesis of quantum theoretical foundations with cutting-edge applications like Brain Computer Interface (BCI) technology. We highlight fresh connections between emerging disciplines like quantum machine learning and quantum image processing. Through rigorous examination and innovative methodology, this paper contributes not only an advanced understanding of the fundamentals of quantum computing but also serves as a pioneering compass, directing future research in aligning quantum computational prowess with contemporary technological challenges.

Summary User mobility represents the movement of either individual or group. In smart cities, detection and prediction of mobility patterns are required for numerous applications like resource distribution, traffic management, and user behavioral analysis. With the increase in the number of smart vehicles, urban mobility detection and prediction have become a critical problem for study. Bike‐sharing ecosystems (BSS) form an integral part of such ecosystems, as it supports the green revolution, ease of access, and solves traffic problems. However, recent schemes have suggested that BSS are challenged by issues of high density, mobility complexity of bikes (stations), large commute cost, uneven distribution, and route imbalances. To address the critical issues, the article proposes a hybrid scheme that combines rebalancing using clustering that addresses the mobility complexity. Once rebalancing is done, we address the uneven distribution among clusters using prediction models. This article is presented a comparative analysis of algorithms like fuzzy C‐means clustering, linear regression, decision tree, and random forest classifiers for predictive analysis performed on weather data and nonweather data. The presented results indicate the viability of the proposed model in real‐world scenarios.

Abstract The next wave in smart transportation is directed towards the design of renewable energy sources that can fuel automobile sector to shift towards the autonomous electric vehicles (AEVs). AEVs are sensor‐driven and driverless that uses artificial intelligence (AI)‐based interactions in Internet‐of‐vehicles (IoV) ecosystems. AEVs can reduce carbon footprints and trade energy with peer AEVs, smart grids (SG), and roadside units (RSUs). It supports green transportation vision. However, the sensor information, energy units, and user data are exchanged through open channels, and thus, are susceptible to various security and privacy attacks. Thus, AEVs can be remotely operated and directed by malicious entities that can propagate false updates to the peer nodes in IoV environment. This can cause the failure of components, congestion, as well as the entire disruption of IoV network. Globally researchers and security analysts have addressed solutions that pertain to specific security requirements, but still, the detection and classification of malicious AEVs is a widely studied topic. Malicious AEVs exhibit an anomaly behavior that differentiates them from normal AEVs, and thereby, the detection of anomalous AEVs and classification of anomaly type is required. Motivated from the aforementioned facts, the survey presents a systematic outlook of AI techniques in anomaly detection of AEVs. A solution taxonomy is proposed based on research gaps in the existing surveys, and the evaluation metrics for AI‐based anomaly detection are discussed. The open challenges and issues in AI deployments are discussed and a case study is presented on anomaly classification through a weighted ensemble technique. Thus, the proposed survey is designed to guide the manufacturing industry, AI practitioners, and researchers worldwide to formulate and design accurate and precise mechanisms to detect anomalies.

Abstract There is a drastic increase in urbanization over the past few years, which requires energy‐efficient and optimized solutions for transportation, governance, quality of life in a smart city among all the citizens. The Internet‐of‐Energy (IoE) ecosystem offers many sophisticated and ubiquitous applications for smart cities. The energy demand of IoE applications is increased while IoE devices continue to grow. Therefore, smart city solutions must have the ability to utilize energy and handle the associated challenges efficiently. Moreover, energy Optimization (EO) techniques can be used to reduce energy consumption to meet the sustainability goals in IoE. Different techniques have been proposed for EO in various fields by researchers worldwide. Computing systems also need energy optimization. The energy consumption in the data center, clouds, and blockchain (BC)‐based architectures are a point of concern at the current time. Due to the enormous energy demands of these systems, we cannot take advantage of the latest technologies to their fullest. Due to the emergence of new technologies and some limitations of proposed techniques, we can still not optimize energy usage more than some extent. There is minimal exploration done in energy optimization in BC‐based systems. In this paper, we have proposed a survey on the energy optimization techniques in various systems, including the optimization techniques in BC‐based systems. We have proposed a taxonomy that classifies energy optimization techniques. We have also proposed an energy‐efficient consensus mechanism, Proof‐of‐High Performance optimization (named as PoHPo), for High‐Performance Computing (HPC) based ecosystems. The open issues and challenges are then discussed in EO. The survey intends to propose future directions for industry professionals, green‐energy stakeholders, and researchers worldwide to explore this topic further.

This article proposes a SaTya scheme that leverages a blockchain (BC)-based deep learning (DL)-assisted classifier model that forms a trusted chronology in fake news classification. The news collected from newspapers, social handles, and e-mails are web-scrapped, prepossessed, and sent to a proposed Q-global vector for word representations (Q-GloVe) model that captures the fine-grained linguistic semantics in the data. Based on the Q-GloVe output, the data are trained through a proposed bi-directional long short-term memory (Bi-LSTM) model, and the news is classified as real-or-fake news. This reduces the vanishing gradient problem, which optimizes the weights of the model and reduces bias. Once the news is classified, it is stored as a transaction, and the news stakeholders can execute smart contracts (SCs) and trace the news origin. However, only verified trusted news sources are added to the BC network, ensuring credibility in the system. For security evaluation, we propose the associated cost of the Bi-LSTM classifier and propose vulnerability analysis through the smart check tool for potential vulnerabilities. The scheme is compared against discourse-structure analysis, linguistic natural language framework, and entity-based recognition for different performance metrics. The scheme achieves an accuracy of 99.55% compared to 93.62% against discourse structure analysis. Also, it shows an average improvement of 18.76% against other approaches, which indicates its viability against fake-classifier-based models.

Healthcare big data (HBD) allows medical stakeholders to analyze, access, retrieve personal and electronic health records (EHR) of patients. Mostly, the records are stored on healthcare cloud and application (HCA) servers, and thus, are subjected to end-user latency, extensive computations, single-point failures, and security and privacy risks. A joint solution is required to address the issues of responsive analytics, coupled with high data ingestion in HBD and secure EHR access. Motivated from the research gaps, the paper proposes a scheme, that integrates blockchain (BC)-based confidentiality-privacy (CP) preserving scheme, CP-BDHCA, that operates in two phases. In the first phase, elliptic curve cryptographic (ECC)-based digital signature framework, HCA-ECC is proposed to establish a session key for secure communication among different healthcare entities. Then, in the second phase, a two-step authentication framework is proposed that integrates Rivest-Shamir-Adleman (RSA) and advanced encryption standard (AES), named as HCA-RSAE that safeguards the ecosystem against possible attack vectors. CP-BDAHCA is compared against existing HCA cloud applications in terms of parameters like response time, average delay, transaction and signing costs, signing and verifying of mined blocks, and resistance to DoS and DDoS attacks. We consider 10 BC nodes and create a real-world customized dataset to be used with SEER dataset. The dataset has 30,000 patient profiles, with 1000 clinical accounts. Based on the combined dataset the proposed scheme outperforms traditional schemes like AI4SAFE, TEE, Secret, and IIoTEED, with a lower response time. For example, the scheme has a very less response time of 300 ms in DDoS. The average signing cost of mined BC transactions is 3,34 seconds, and for 205 transactions, has a signing delay of 1405 ms, with improved accuracy of ≈ 12% than conventional state-of-the-art approaches.

This paper proposes a generic scheme VaCoChain, that fuses blockchain (BC) and unmanned aerial vehicles (UAVs) underlying fifth-generation (5G) communication services for timely vaccine distribution during novel coronavirus (COVID-19) and future pandemics. The scheme offers 5G-tactile internet (5G-TI) based services for UAV communication networks (UAVCN) monitored through ground controller stations (GCs). 5G-TI enabled UAVCN supports real-time dense connectivity at ultra-low round-trip time (RTT) latency of [Formula: see text] and high availability of 99.99999%. Thus, it can support resilient vaccine distributions in a phased manner at government-designated nodal centers (NCs) with reduced round trip delays from vaccine production warehouses (VPW). Further, UAVCNs ensure minimizes human intervention and controls vaccine health conditions due to shorter trip times. Once vaccines are supplied at NCs warehouses, then the BC ensures timestamped documentation of vaccinated persons with chronology, auditability, and transparency of supply-chain checkpoints from VPW to NCs. Through smart contracts (SCs), priority groups can be formed for vaccination based on age, healthcare workers, and general commodities. In the simulation, for UAV efficacy, we have compared the scheme against fourth-generation (4G)-assisted long term evolution-advanced (LTE-A), orthogonal frequency division multiplexing (OFDM) channels, and traditional logistics for round-trip time (RTT) latency, logistics, and communication costs. In the BC setup, we have compared the scheme against the existing 5G-TI delivery scheme (Gupta et al.) for processing latency, packet losses, and transaction time. For example, in communication costs, the proposed scheme achieves an average improvement of 9.13 for block meta-information. For 4000 transactions, the proposed scheme has a communication latency of 16 s compared to 36 s. The packet loss is significantly reduced to 2.5% using 5G-TI compared to 16% in 4G-LTE-A. The proposed scheme has a computation cost of 1.6 ms and a communication cost of 157 bytes, which indicates the scheme efficacy against conventional approaches.

In this paper, we propose a dynamic spectrum allocation (DSA) scheme DeepBlocks at the backdrop of sixth-generation (6G) communication networks that address the challenges of fixed spectrum allocations (FSA). The scheme exploits the advantages of deep-Q-network (DQN) and minimizes the search state explosion through a reward-penalty framework. A dynamic allocation of unallocated resource blocks (RBs) to mobile units (MUs) is carried out and once the allocation of RBs is complete, we integrate blockchain (BC) to record the transactional ledgers. The resource usage of MUs is recorded through smart contracts (SCs). We model the proposed scheme as a convex optimization problem, and subproblems are decomposed into a Pareto-optimal solution via Techebyecheff decomposition. In the simulation, we compare our scheme against FSA, and fifth-generation (5G) based DSA schemes like reinforcement learning (RL), deep neural networks (DNN)-based, and duelling DQN based schemes. The comparative analysis of 6G-DQN is modeled in terms of reward formulation, scalability of 6G-DQN-assisted DSA, and profit scenarios of BC-based allocation through intelligent channel control. The scheme proposes significant findings, with the best fit learning rate of 0.0001, and takes 500 episodes to converge to 60 total resource blocks. The servicing latency of the scheme is 272.4 ms, compared to 2010 ms in the duelling DQN approach. In spectrum allocation, an improvement of 26.32% is observed against non-DQN approaches, and 13.57% in the fairness parameter for spectrum allocation due to BC inclusion. The findings present the scheme efficacy for DSA over the aforementioned conventional approaches.

The 6G-based spectrum bands allocation to telecom providers would guarantee ultra peak rates, high availability, and extremely low-latency for various user applications. However, the spectrum allocation still suffers from the limitations of fair allocation process, delays in auction process, and collusive bidding due to inherent centralization. Thus, this paper proposes a scheme, Block6Tel, that integrates blockchain (BC) in 6G-envisioned spectrum allocation to ensure secure and trusted band allocation among telecom providers, and ensure transparency among telecom stakeholders. The scheme operates in two phases. First, a 6G-based protocol stack model is proposed that leverages a cell-free communication infrastructure. Then, in the second phase, a BC-based auction algorithm is proposed for inter-operator spectrum allocation, and resource allocations among service providers are finalized. Finally, smart contracts (SC) are executed among telecom providers as bidders, and government authorities (GA) as auctioneers. Through extensive simulations, we prove the superiority of Block6Tel compared with traditional static allocation approaches, in terms of parameters like- resource utilization, requests overhead, and allocation fairness. The results demonstrate that the proposed scheme outperforms the traditional schemes using various parameters.

The paper proposes a sixth-generation (6G)-enabled cellular vehicle-to-anything (C-V2X)-based scheme, Res6Edge, that supports high-data ingestion rate through artificial intelligence (AI) models at edge nodes, or Edge-AI. Through Edge-AI in 6G supported C-V2X, we address the research gaps of earlier schemes based on fifth-generation (5G) resource orchestration. 6G improves decision analytics and real-time resource sharing among C-V2X ecosystems. The scheme operates in three phases. In the first phase, a layered network model is proposed for V2X communication based on 6G-aggregator and core units. Then, based on the proposed stack, in the second phase, 6G resource allocation is proposed through macro base station (MBS) units. MBS ensures channel gain and reduces energy loss dissipation. Finally, in the third phase, an intelligent edge-AI scheme is formulated based on deep-reinforcement learning (DRL) to support responsive edge-cache and improved learning. The proposed scheme is compared to 5G baseline services in terms of parameters like- throughput, latency, and DRL scheme is compared to random allocation approaches. Through simulations, Res6Edge obtains a V2X user throughput of 43.24 Mbps, compared to 0.7 Mbps for 4 x 108connected ACV sensors. The reduced latency is ≈ 13.84 times of 5G. DRL learning algorithm achieves a satisfaction probability of 0.5 for 500 vehicles, compared to 0.35 using conventional schemes. The obtained results indicate the viability of the proposed scheme.

In modern decentralized Internet-of-Things (IoT)-based sensor communications, pseudonoise-diffusion oracles are heavily investigated as random oracles for data exchange among peer nodes. As these oracles are generated through algorithmic processes, they pass the standard random tests for finite and bounded intervals only. This ensures a false sense of privacy and confidentiality in exchange through open protocol IoT-stacks in public channels, i.e., Internet. Recently, blockchain (BC)-envisioned random sequences as input oracles are proposed about financial applications, and windfall games like roulette, poker, and lottery. These random inputs exhibit fairness, and nondeterminism in SC executions termed as probabilistic smart contracts (PSCs). However, the IoT-enabled PSC process might be controlled and forged through humans, machines, and bot-nodes through physical and computational methods. Moreover, dishonest entities like contract owners, players, and miners can co-ordinate together to form collusion attacks during consensus to propagate false updates, which ensures forged block additions by miners in BC. Motivated by these facts, in this article, we propose a BC-envisioned IoT-enabled PSC scheme,SaNkhyA, which is executed in three phases. In the first phase, the scheme eliminates colluding dishonest miners through the proposed miner selection algorithm. Then, in the second phase, the elected miners agree through the proposed consensus protocol to generate a stream of random bits. In the third phase, the generated random bit-stream is split through random splitters and fed as input oracles to the proposed PSC among participating entities. In simulation, the scheme ensures a trust probability of 0.38 even at 85% collusion among miners and has an average block processing delay of 1.3 s compared to serial approaches, where the block processing delay is 5.6 s, thereby exhibiting improved scalability. The overall computation and communication cost is 28.48 ms, and 101 bytes, respectively, that indicates the efficacy of the proposed scheme compared to the traditional schemes.