Forthcoming Articles

International Journal of Autonomous and Adaptive Communications Systems

International Journal of Autonomous and Adaptive Communications Systems (IJAACS)

Forthcoming articles have been peer-reviewed and accepted for publication but are pending final changes, are not yet published and may not appear here in their final order of publication until they are assigned to issues. Therefore, the content conforms to our standards but the presentation (e.g. typesetting and proof-reading) is not necessarily up to the Inderscience standard. Additionally, titles, authors, abstracts and keywords may change before publication. Articles will not be published until the final proofs are validated by their authors.

Forthcoming articles must be purchased for the purposes of research, teaching and private study only. These articles can be cited using the expression "in press". For example: Smith, J. (in press). Article Title. Journal Title.

Articles marked with this shopping trolley icon are available for purchase - click on the icon to send an email request to purchase.

Online First articles are also listed here. Online First articles are fully citeable, complete with a DOI. They can be cited, read, and downloaded. Online First articles are published as Open Access (OA) articles to make the latest research available as early as possible.

Open AccessArticles marked with this Open Access icon are Online First articles. They are freely available and openly accessible to all without any restriction except the ones stated in their respective CC licenses.

Register for our alerting service, which notifies you by email when new issues are published online.

International Journal of Autonomous and Adaptive Communications Systems (6 papers in press)

Regular Issues

  • An Optimisation Approach for Path Prediction and Detection in Autonomous Vehicles   Order a copy of this article
    by Swati Jaiswal, Renu Kachhoria, Rupali Chopade, Trupti Pawase, Neha Joshi, Spandan Surdas 
    Abstract: Intelligent transportation systems are rapidly evolving with the rise of autonomous vehicles, emphasising travel safety. However, challenges such as road damage, occlusion, low illumination, shadows, and complex road conditions often result in unclear images, making accurate vehicle decisions difficult. This study proposes an ensemble framework for lane line and traffic sign detection, integrating metaheuristic optimisation with a modified Hough-enabled Lane Generative Adversarial Network (GAN) designed to process fuzzy road images. To strengthen decision-making in dynamic urban crossings, the model incorporates the features of Grey Wolf and Starling bird optimization, enhancing classifier performance. A hybrid ensemble classifier combining CNN and BiLSTM improves traffic sign recognition, while synthetic data generation significantly reduces lane detection misclassifications. The proposed system demonstrates improved accuracy, sensitivity, and specificity, with the lane detection model showing a lower error rate than existing methods. This approach advances autonomous driving by offering a more reliable, adaptive, and robust decision-making framework.
    Keywords: Optimization; decision-making; autonomous vehicles; generative adversial network; convolutional neural network.
    DOI: 10.1504/IJAACS.2026.10080285
     
  • PSO-Optimized Compact Two-Port Planar MIMO Antenna with Double U-Shaped DGS for 3.5 GHz 5G Applications   Order a copy of this article
    by Riski Ramadani, Rohim Aminullah Firdaus, Afiyah Nikmah, Nisaul Fadhilah, Hanan Zaki Alhusni 
    Abstract: 5G communication systems require antennas that support high-speed data transmission, stable impedance matching, and compact wireless-device integration. This study proposes a compact two-port planar multiple-input multiple-output (MIMO) antenna for 3.5 GHz applications. The design integrates a double U-shaped defected ground structure (DGS), inset feeds, symmetrical parasitic elements, and particle swarm optimisation (PSO)-based dimensional tuning. Its novelty lies in jointly optimising these techniques to enhance impedance matching, bandwidth, envelope correlation coefficient (ECC), and diversity performance, rather than applying each technique individually. Measurements show a return loss of 45.60 dB, voltage standing wave ratio (VSWR) of 1.01, and matched resonance at 3.5 GHz. The antenna also achieves a gain of 3.41 dBi, bandwidth of 362 MHz, diversity gain of 9.99, and ECC of 3.34 x 105. Despite its moderate gain, the compact design demonstrates competitive impedance-matching, bandwidth, and diversity characteristics for practical integration into compact 5G wireless communication devices and related applications.
    Keywords: 5G communication; planar MIMO antenna; double U-shaped DGS; inset feed; parasitic elements; Particle Swarm Optimization; impedance matching.
    DOI: 10.1504/IJAACS.2026.10080628
     
  • Topology-Aware Federated MARL: a Knowledge-Graph Approach to DAG-Structured Edge Orchestration   Order a copy of this article
    by Amin Mohajer, Abbas Mirzaei, Babak Noori Moghaddam, Ali Ghaffari, Xavier Fernando 
    Abstract: DAG-structured edge applications comprise interdependent microservices whose performance depends on workflow precedence, network topology, container availability, and heterogeneous regional resources. Existing federated orchestration methods often model these factors separately, leaving workflows vulnerable to critical-path stalls, link variations, cold starts, migration failures, and non-IID workloads. This paper proposes TopoKG-FMARL, a topology-aware federated multi-agent reinforcement-learning framework for knowledge-assisted DAG orchestration. It builds a dynamic cross-layer knowledge graph over regions, links, queues, containers, microservices, deadlines, dependencies, and execution outcomes. A temporal relational encoder derives workflow infrastructure embeddings, while policy memory retrieves successful and failed traces under workload drift. Regional agents use a conditional hierarchical actor critic to coordinate admission, placement, migration, and communication computation resources, with feasibility masking and projection. Semi-asynchronous federation weights updates by sample support, staleness, topology divergence, and reliability without sharing raw data. Simulations show improved deadline satisfaction, latency, robustness, and convergence over representative baselines under heterogeneous workloads and unreliable edge-network operating conditions.
    Keywords: DAG-Structured microservices; edge orchestration; topology-aware learning; semi-asynchronous federation.
    DOI: 10.1504/IJAACS.2027.10080978
     
  • Imbalance-Robust Multi-Task Learning for NILM Signals with Convolution and Attention   Order a copy of this article
    by Kaijie Chen, Honghuai Feng, Yulin Luo, Shiliang Zhang, Yannan Guo, Wuxing Chen 
    Abstract: Non-Intrusive Load Monitoring (NILM) is crucial for energy efficiency, as it provides users with detailed electricity consumption information to enable more efficient power management and reduce energy losses. However, most state-of-the-art deep learning-based methods fail to establish a complete NILM framework that encompasses all necessary steps. To address this gap, this study proposes an integrated approach named IRMTLCA, which is based on the publicly available LIT-Dataset and combines the detection and classification tasks of high-frequency NILM signals. In the detection phase, the proposed method achieves performance exceeding 90% in most scenarios, whereas state-of-the-art techniques remain below 70% for eight loads. Furthermore, this study incorporates a multi-label strategy to bypass the disaggregation stage, directly identifying loads operating at a specific time and thereby enhancing the recognition capability of simultaneous loads.
    Keywords: Deep learning; cross-channel attention; multi-label; NILM.
    DOI: 10.1504/IJAACS.2026.10081017
     
  • Static Software Security Vulnerability Identification Integrating GNN and Causal Inference   Order a copy of this article
    by Lin Zhou, Xiaoli Jia 
    Abstract: With the continuous expansion of software systems, the hidden security risks in source code are increasingly showing diversity and concealment characteristics. Therefore, conducting research on joint GNN causal inference for static software security vulnerability identification. Firstly, by constructing a causal diagram structure of code semantic units, the dependency relationship between program control flow and data flow is transformed into directed edges. Secondly, we innovatively adopt an attention based GNN model for causal inference, separating causal and confusion graphs through deconvolution techniques to eliminate non causal interference factors. Finally, by integrating the node embeddings of Node2Vec and the causal features of GNN through structural enhancement modules, a high-dimensional vector representation of the source code is generated using global aggregation and max pooling operations. The test results show that the vulnerability identification accuracy of the proposed method reaches 94.8%-97.4%; At the same time, the false alarm rate is controlled at around 2%.
    Keywords: GNN; Causal inference; Static software; Security vulnerability identification.
    DOI: 10.1504/IJAACS.2026.10081226
     
  • State Estimation Model Development and Simulation Study for Active Distribution Networks   Order a copy of this article
    by Pratik Patel, Ravi Kumar, Sheshang Degadwala, Feng Liu 
    Abstract: The increasing integration of distributed energy resources, photovoltaic generation, and bidirectional power flow complicates active distribution network state estimation (ADNSE), necessitating accurate, robust, and computationally efficient monitoring. This study proposes an integrated framework combining bad-data preprocessing, long short-term memory (LSTM)-based dynamic state estimation, structured weight pruning, and digital twin synchronisation. Experimental data were generated using the IEEE 34-bus active distribution system under diverse operating conditions, including load variations, renewable generation, measurement noise, missing measurements, and gross errors. The proposed LSTM model was evaluated against the conventional weighted least squares (WLS) estimator. Results demonstrate improved estimation accuracy, reduced computation time, and robust performance under incomplete and noisy measurement conditions. Weight pruning reduces model complexity and memory requirements, supporting deployment in resource-constrained environments. Digital twin synchronisation enables real-time network monitoring, enhanced situational awareness, and informed operational decision-making. The proposed framework offers an efficient and resilient solution for next-generation active distribution network state estimation.
    Keywords: State Estimation (SE); Active Distribution Networks (ADN); Bad Data Processing; Neural Network-Based Estimation; Dynamic State Estimation; Digital Twin (DT).
    DOI: 10.1504/IJAACS.2027.10082211