High-precision indoor positioning remains sensitive to non-line-of-sight (NLOS) propagation and multipath interference. This study investigates cross-environment adaptation of Encoder-only Transformer (ET) models for raw channel impulse response (CIR) sequences. The adaptation configuration combines optional Adapter insertion after self-attention with selective updates of existing transformer modules and is selected separately for NLOS classification and ranging-error regression under limited target-domain supervision. TPE and hyperband are used for configuration search and early pruning. Experiments on a public dataset collected in four indoor environments show that, under the evaluated settings, ET obtains an average classification accuracy of 84.94
The superheterodyne digital receiver (SDR) is the main equipment for frequency guidance in modern digital radio frequency memory (DRFM) jamming devices. Low-interception signals can significantly reduce the frequency guidance effectiveness of SDR. Nevertheless, at present, the design principle of anti-SDR low-interception signals is still unclear, which severely restricts the quality of the design of low-interception signals. This paper reveals the mechanism of SDR signal interception and frequency scanning failure. Then, the calculation formula for the critical bandwidth required to achieve low interception of signals by the radio detector is derived. Finally, the design principle of multi-frequency group virtual broadband low-interception signals is proposed. Aiming at the limited frequency band and computing resources of a radio detector, based on the proposed principles, in typical cases (i.e., α =0.8, β =0.2, B_total=3GHz, B_J=1GHz ), the total bandwidth is divided into four frequency groups and evenly spaced to synthesize virtual broadband. The synthetic virtual broadband can cause the failure of SDR frequency scanning. The simulation results show that the interception probability of the signal designed by the signal design principle proposed in this paper is less than 20
Alzheimer’s disease (AD) is a progressive brain disorder that is caused by the nerve cells and results in memory loss and cognitive decline, which highly impacts daily function. In the present times, several artificial intelligence (AI)-based models, such as deep learning strategies, are utilized to automatically analyse the MRI and detect AD. Various researchers integrate explainable AI (XAI) into deep learning to advance the detection process and to understand the reasons behind the outcome to support clinical settings. These XAI-based frameworks are highly efficient and able to evaluate the subtle changes and complex patterns in the MRI. In addition, these types of advanced models improve the diagnosis process and provide early diagnosis results to enhance the patient’s prognosis. Thus, this work develops an enhanced AD detection model by using advanced XAI approaches. Initially, the MRIs are gathered from the standard datasets and fed into Gabor Convolution with Channel Self-Attention-based Adaptive Explainable ShuffleNetV2 (GC-CSA-AXAI) to detect the presence and absence of AD. It uses the Gabor convolution, which aids in capturing the fine-grained textures, along with the edges that are present in the given images, while the channel self-attention allows the developed GC-CSA-AXAI to provide more intense concentration only on the essential brain regions. Moreover, the ShuffleNetV2 is the core network, which provides optimal results by ensuring that the network is rapid and suitable for clinical applications, while the explainable nature of the framework helps to build crucial trust and transparency in the automated process. Finally, to enhance the performance, an adaptive component is used, which applies the Reformulated Arbitrary Term of Puma Optimizer (RATPO) to fine-tune the critical parameters, resulting in quicker and more accurate results. Finally, the developed model offers the detected outcome, and then, the efficiency is validated by differentiating the introduced approach from several earlier models and showing its improved functionality over others.
The power system is a core infrastructure for the national economy and energy security. As key assets in the grid, transmission-line components must be accurately inspected from UAV images despite small targets, large scale variation, complex backgrounds, and partial image degradation. This paper proposes KED-GNN, a graph neural network enhanced by space-to-depth convolution, Kolmogorov–Arnold Network-based nonlinear transformation, and efficient multi-scale attention, for UAV-based transmission-line component and fault detection. Graph-based image representation methods, such as Vision GNN, provide a way to represent an image as a graph by treating image patches as nodes and their relationships as edges, which has potential for UAV transmission-line inspection. However, direct use of this representation still faces small-target information loss, multi-scale feature variation, and complex background interference. Therefore, KED-GNN introduces an SPD-Conv-based detail-preserving sampling algorithm, a KAN-based nonlinear feature transformation mechanism, and an EMA-based efficient multi-scale attention mechanism. Under the same Faster R-CNN detection framework, KED-GNN achieves a final mAP @ 0.5 of 0.86 and precision of 0.85, improving mAP @ 0.5 by 3.61
Channel estimation errors significantly degrade signal detection performance in massive multiple-input multiple-output (massive MIMO) systems, limiting the effectiveness of conventional linear, near-optimal, and learning-based detectors under practical channel uncertainty. To address this challenge, this paper proposes a channel state information (CSI)-aware deep neural network-assisted detector (DNN-ML) that explicitly incorporates channel estimation error variance into the detection process. The proposed framework employs CSI feature extraction, dimensionality reduction, error-adaptive feature fusion, and reliability-aware decision refinement to jointly exploit the received signal, estimated CSI, and CSI uncertainty. The detector is evaluated for a 256×256 massive MIMO system employing 256-QAM modulation over Rayleigh fading channels with 10
