Omm Prakash Sahoo

The integration of Artificial Intelligence (AI) into wireless communication has enabled adaptive, efficient, robust, and scalable system designs. Reconfigurable Intelligent Surfaces (RIS) offer a promising paradigm for AI-driven control by dynamically adjusting the wireless environment through Phase Configuration Sequence (PCS) bits sent from the Base Station (BS). By steering signals intelligently, RIS can improve coverage and direct transmissions to specific locations. However, the increasing size of RIS and rapidly changing IoT device positions in dynamic environments impose significant feedback bandwidth constraints due to frequent PCS bits updates. To address this, we propose a novel Attention-Guided Encoder-Decoder with Normalization Enhancement (AGENE) framework that compresses PCS bits at the BS using a lightweight encoder and reconstructs them at the RIS controller using a decoder. Our design incorporates a custom attention mechanism and Generalized Divisive Normalization (GDN)/Inverse GDN layers to enhance feature extraction and normalization. We also evaluate our method under different noise models, including Gaussian and Rician noise, to test robustness in practical scenarios. Finally, to evaluate the effectiveness of our proposed method (AGENE), we compared its performance with existing methods across different compression ratios and noise conditions, focusing on loss reduction and Normalized Mean Square Error (NMSE). In Additive White Gaussian Noise (AWGN) conditions, AGENE achieved a loss reduction of 28.12% compared to PSC-DN, 43.9% compared to DL-CsiNet, and 58.18% compared to CsiNet at a compression ratio of 2/9. Under Rician noise, AGENE showed a reduction of 19.39% compared to PSC-DN, 37.01% compared to DL-CsiNet, and 45.21% compared to CsiNet, again at the same compression ratio. These results consistently demonstrate that AGENE outperforms existing methods by a significant margin across both metrics and noise conditions.

Pixel failures on Reconfigurable Intelligent Surfaces (RISs) cause major challenges in wireless communication systems. These failures are especially problematic for Internet of Things (IoT) networks, where reliability and Quality-of-Service (QoS) are critical parameters. Failed RIS pixels can significantly impact system efficiency. These failures disrupt signal quality, reduce Signal-to-Interference-Plus-Noise-Ratio (SINR), lower the sum rate, and affect energy efficiency. As a result, maintaining stable communication links becomes challenging. The primary function of RIS is to improve the connectivity and signal strength between the IoT devices and the Base Station (BS). However, when a large number of pixels fail, this function is compromised. Motivated by the aforementioned challenges, this paper focuses on severe RIS pixel failures. The method uses a Deep Neural Network (DNN) to detect failed RIS pixels and a mathematical model to optimize the configuration of the affected regions, including the failed pixels and their neighboring healthy elements. To achieve this, a dynamically adapting mask is employed, ensuring that the total failed pixels within the mask always remain at 25 % or less. The remaining functional elements retain their standard configuration, allowing for minimal and targeted adjustments. This selective reconfiguration focuses on improving the SINR and sum rate at the BS, while minimizing adjustments to ensure optimal performance. The effectiveness of the proposed method is validated through comprehensive experimental evaluations in various scenarios and time instances. The results consistently demonstrate that the proposed Pixel and Configuration (P&C) RIS method improves the SINR and sum-rate performance at the BS, outperforming conventional RIS systems. This adaptive approach offers a resilient and scalable solution to address the challenges posed by faulty RIS elements, ensuring high-quality communication in dynamic and real-world IoT applications.

Node localization in time-varying Internet of Things (IoT) networks is an essential problem due to increased delay and poor Signal-to-Interference plus Noise Ratio (SINR) at the Base Station (BS). To improve the received signal strength at the BS, Reconfigurable Intelligent Surface (RIS) has recently been used between transmitter and receiver. Additionally, novel phase prediction methods and optimal weight assignment frameworks have been proposed over RIS and BSs, respectively. Nevertheless, these methods suffer from poor performance due to their heuristic approach, resulting in more time consumption and poor SINR. Motivated by the aforementioned challenges, we propose a novel node localization method over a RIS-assisted time-varying IoT network using Machine Learning (ML) frameworks in this work. Firstly, the method computes the optimal phase configuration over the RIS corresponding to each element using coeff2phaseNN, which has been trained on channel coefficients among the transmitter, receiver, and RIS. Subsequently, the weight of the individual antenna element at the BS is optimized using the proposed VectorSync model. The results confirm that the coeff2phaseNN method demonstrates a reduction of 89.79% in total MSE loss compared to the Artificial Neural Network-RIS (ANN-RIS) method. Additionally, it demonstrates a 71.04% reduction in the absolute RIS phase prediction deviation from the optimal phase compared to the ANN-RIS method. Moreover, the proposed VectorSync method attains a 79.28% and 92.29% reduction in time required for optimal weight assignment compared to the Bartlett and Capon methods, respectively. Finally, the Localization Error(LR) using the proposed method is compared to conventional methods in a time-varying experimental scenario and found to be the minimum, i.e., 6.156%.

Cholangiocarcinoma is one of the rarest yet most aggressive cancers that has a low 5-year survival rate (2% - 24%) and thus often requires an accurate and timely diagnosis. Hyperspectral Imaging (HSI) is a recently developed, promising spectroscopic-based non-invasive bioimaging technique that records a spatial image (x, y) together with wide spectral (λ) information. In this work, for the first time we propose to use a three-dimensional (3D)U-Net architecture for Hyperspectral microscopic imaging-based cholangiocarcinoma detection and classification. In addition to this architecture, we opted for a few preprocessing steps to achieve higher classification accuracy (CA) with minimal computational cost. Our results are compared with several standard unsupervised and supervised learning approaches to prove the efficacy of the proposed network and the preprocessing steps. For instance, we compared our results with state-of-the-art architectures, such as the Important-Aware Network (IANet), the Context Pyramid Fusion Network (CPFNet), and the semantic pixel-wise segmentation network (SegNet). We showed that our proposed architecture achieves an increased CA of 1.29% with the standard preprocessing step i.e., flat-field correction, and of 4.29% with our opted preprocessing steps.