FIGURE 5 - available via license: Creative Commons Attribution 4.0 International
Content may be subject to copyright.
Source publication
The performance of existing signal detection methods depend heavily on the amount of prior information acquired by the sensor of interest. Therefore, to improve cognitive radio-based detection in low-SNR environments, we propose a deep learning method based passive signal detection. A convolution neural network (CNN) and long short-term memory (LST...
Context in source publication
Context 1
... used blind detection algorithm. The ED method assumes the signal follows a zero-mean Gaussian distribution with covariance matrix 2 s I . The noise is also characterized by a zero-mean Gaussian distribution with covariance matrix 2 I . The dataset was collected from a real world communication station. The experiment equipment is shown in Fig. 5. It comprises PCs, Anykey AKDS700 radios, a digital receiver, and an oscilloscope. Two radios are linked to two computers respectively to form a transmitter and a receiver. Then, real time wireless communication is performed between the transmitting side and the receiving side. Following this, a digital oscilloscope is utilized to ...
Similar publications
Glaucoma, a leading cause of blindness, is a multifaceted disease with several patho-physiological features manifesting in single fundus images (e.g., optic nerve cupping) as well as fundus videos (e.g., vascular pulsatility index). Current convolutional neural networks (CNNs) developed to detect glaucoma are all based on spatial features embedded...
Citations
... With accurate signal classification and modulation recognition, secondary users (SUs) and the primary users (PUs) can efficiently share the spectrum, thereby significantly improving the spectrum utilization. Given the potential of CR, the related technologies are attracting more and more attention of researchers [3]. As the wireless channels, signal waveforms and the modulation formats become increasing complex and dynamic, the accurate signal parameter estimation and modulation classification become more and more challenging. ...
Automatic modulation classification (AMC) technology is one of the cutting-edge technologies in cognitive radio communications. AMC based on deep learning has recently attracted much attention due to its superior performances in classification accuracy and robustness. In this paper, we propose a novel, high resolution and multi-scale feature fusion convolutional neural network model with a squeeze-excitation block, referred to as HRSENet, to classify different kinds of modulation signals. The proposed model establishes a parallel computing mechanism of multi-resolution feature maps through the multi-layer convolution operation, which effectively reduces the information loss caused by downsampling convolution. Moreover, through dense skip-connecting at the same resolution and up-sampling or down-sampling connection at different resolutions, the low resolution representation of the deep feature maps and the high resolution representation of the shallow feature maps are simultaneously extracted and fully integrated, which is benificial to mine signal multilevel features. Finally, the feature squeeze and excitation module embedded in the decoder is used to adjust the response weights between channels, further improving classification accuracy of proposed model. The proposed HRSENet significantly outperforms existing methods in terms of classification accuracy on the public dataset “Over the Air” in signal-to-noise (SNR) ranging from -2dB to 20dB. The classification accuracy in the proposed model achieves 85.36% and 97.30% at 4dB and 10dB, respectively, with the improvement by 9.71% and 5.82% compared to LWNet. Furthermore, the model also has a moderate computation complexity compared with several state-of-the-art methods.
... Considering the single antenna SU scenarios, authors of [13] and [14] have developed an SS method employing a hybrid structure of CNN-LSTM to extract the spatio-temporal information from observations of single sensing intervals. A combination of 1-D CNN and LSTM network has been employed to learn the time and frequency domain features, to find the presence of PU specifically in low signal-to-noise ratio (SNR) environments [15]. In order to enhance the SS performance further, xie et. ...
... Further, max pooling operation is performed over the feature of all SUs to get output representation provided in equation (15). Instead of spatial pooling (like typical max pooling in CNNs, which pools over spatial dimensions), this operation pools over the dimension representing different SU and combines their predictions by selecting the maximum value for each position across SU dimension. ...
In this paper, we develop a novel mobility-aware transformer-driven tiered structure (MASSFormer) based cooperative spectrum sensing method that effectively models the spatio-temporal dynamics of user movements. Unlike existing methods, our method considers a dynamic scenario involving mobile primary users (PUs) and secondary users (SUs)and addresses the complexities introduced by user mobility. The transformer architecture utilizes an attention mechanism, enabling the proposed method to adeptly model the temporal dynamics of user mobility by effectively capturing long-range dependencies within the input data. The proposed method first computes tokens from the sequence of covariance matrices (CMs) for each SU and processes them in parallel using the SUtransformer network to learn the spatio-temporal features at SUlevel. Subsequently, the collaborative transformer network learns the group-level PU state from all SU-level feature representations. The attention-based sequence pooling method followed by the transformer encoder adjusts the contributions of all tokens. The main goal of predicting the PU states at each SU-level and group-level is to improve detection performance even more. We conducted a sufficient amount of simulations and compared the detection performance of different SS methods. The proposed method is tested under imperfect reporting channel scenarios to show robustness. The efficacy of our method is validated with the simulation results demonstrating its higher performance compared with existing methods in terms of detection probability, sensing error, and classification accuracy.
... In [25], the authors have proposed a CNN-based approach that fully combines historical and current sensing data to recognize the state of primary users. To enhance detection performance, several algorithms combining CNN and LSTM have been proposed to capture both spatial and temporal features from sensing data [26], [27]. In [26], the authors have designed a spectrum sensing architecture based on CNN and LSTM, where the CNN model captures energy correlation features and the LSTM model captures temporal features from multiple sensing periods based on the output of the CNN. ...
... In [26], the authors have designed a spectrum sensing architecture based on CNN and LSTM, where the CNN model captures energy correlation features and the LSTM model captures temporal features from multiple sensing periods based on the output of the CNN. Ke et al. [27] have employed a mixed model that combines CNN with LSTM to extract time-frequency features from received data in low SNR. In [28], the authors have designed the DS2MA model, which constructs autocorrelation and cross-correlation matrices for multi-antenna spectrum sensing. ...
Spectrum Sensing plays a crucial role in cognitive radio and serves as a fundamental requirement for achieving dynamic spectrum access. This work investigates a novel multi-antenna spectrum sensing framework based on graph neural networks to accurately identify the state of primary users. Specifically, the work proposes a graph spectral convolution-based spectrum sensing scheme (SCM-GNN), which employs stacked graph convolutions to capture the dependencies contained in test statistics. To further enhance the detection performance of SCM-GNN, the work introduces a covariance matrix with smooth factor as the test statistic. The covariance matrix includes more discriminative information and assists the SCM-GNN in achieving state-of-the-art detection performance. Simulation results demonstrate that the proposed algorithm outperforms existing works in terms of detection performance under the influence of various non-ideal factors, such as general Gaussian noise, channel fading, large-scale fading, real-world scenario, and imperfect reporting channel.
... The rapid advancement in deep learning has injected new vitality into signal detection and recognition [29][30][31]. Ke et al. [32] employ convolutional long short-term deep neural networks (CLDNNs) cascaded with convolutional neural networks (CNNs) and long short-term memory (LSTM) to extract time-domain and frequency-domain features from input signal sequences. which gives better performance than traditional energy detection algorithms. ...
In complex electromagnetic environments, satellite telemetry, tracking, and command (TT&C) signals often become submerged in background noise. Traditional TT&C signal detection algorithms suffer a significant performance degradation or can even be difficult to execute when phase information is absent. Currently, deep-learning-based detection algorithms often rely on expert-experience-driven post-processing steps, failing to achieve end-to-end signal detection. To address the aforementioned limitations of existing algorithms, we propose an intelligent satellite TT&C signal detection method based on triplet attention and Transformer (TATR). TATR introduces the residual triplet attention (ResTA) backbone network, which effectively combines spectral feature channels, frequency, and amplitude dimensions almost without introducing additional parameters. In signal detection, TATR employs a multi-head self-attention mechanism to effectively address the long-range dependency issue in spectral information. Moreover, the prediction-box-matching module based on the Hungarian algorithm eliminates the need for non-maximum suppression (NMS) post-processing steps, transforming the signal detection problem into a set prediction problem and enabling parallel output of the detection results. TATR combines the global attention capability of ResTA with the local self-attention capability of Transformer. Experimental results demonstrate that utilizing only the signal spectrum amplitude information, TATR achieves accurate detection of weak TT&C signals with signal-to-noise ratios (SNRs) of −15 dB and above (mAP@0.5 > 90%), with parameter estimation errors below 3%, which outperforms typical target detection methods.
... Our proposed scheme does not require prior knowledge of the waveform of the signals to be sensed. "Energy detection (ED) is the most widely used spectrum sensing method because of its low complexity and also because no prior information is required" [33]. Therefore, we compare the performance of our proposed scheme with that of ED in terms of spectrum sensing. ...
This paper introduces an innovative methodology for spectrum sensing and signal classification, leveraging generative artificial intelligence and incorporating out-of-distribution (OOD) detection mechanisms. Our proposed approach demonstrates remarkable resilience in challenging, low signal-to-noise ratio (SNR) environments. Through comprehensive simulations, our method demonstrates remarkable results even under challenging conditions with low SNR scenarios. For example, our proposed method achieves 100% accuracy in spectrum sensing and a notable 95.8% accuracy in signal classification at -14dB SNR, which is a fairly low SNR. What sets our approach apart is its capability to identify and classify novel signal classes not encountered during the training phase. This exceptional feature significantly enhances system robustness and adaptability, particularly in dynamic, real-world applications such as the broader landscape of cognitive radio technologies, aiming to enhance adaptability and responsiveness in dynamic spectrum environments. The ability to successfully operate in diverse and unanticipated signal scenarios positions our methodology as a valuable contribution to the advancement of spectrum sensing and signal classification technologies, with promising implications for future wireless communication systems.
... This approach enhances the model's robustness to minor variations in the input. The inclusion of LSTM in the CLDNN architecture is instrumental for capturing long-term dependencies in sequential or time series data [27]. LSTM's proficiency in learning evolving patterns over time makes it particularly well-suited for tasks involving temporal sequences. ...
Nowadays, wireless communication plays a pivotal role in our daily lives, encompassing technologies such as wireless fidelity (Wi-Fi) and the internet of things (IoT). The backbone of the wireless communication is modulation, which involves various techniques with its own unique characteristics. As modulation techniques evolve in intricacy and diversity, the need for modulation recognition becomes apparent. Traditional modulation recognition relies on human intervention to classify modulation types in received signals, a time-consuming and laborious process prone to human error and inefficiency. Consequently, automatic modulation recognition (AMR) is introduced to autonomously classify modulation types without human interventions. In the current era, artificial intelligence (AI), specifically deep learning (DL) has gained prominence, providing numerous advantages across various domains, including AMR. While many DL-based AMR models have been developed, their efficacy reduces at low signal-to-noise ratio (SNR). Consequently, we propose a hybrid DL model for AMR, named the in-phase and quadrature - temporal graph convolutional network (IQ-TGCN) to enhance the recognition performance at low SNR. Integrating graph convolutional network (GCN) and long short-term memory (LSTM) architectures, the IQ-TGCN takes a node feature matrix as input, derived from the magnitude differences between each node. In comparative assessments against other DL models, our model has consistently exhibited superior performance. To enhance its capabilities further, we integrated deep transfer learning, leading to a remarkable 30% improvement in classification accuracy. Notably, at a SNR of 10 dB, IQ-TGCN reached its pinnacle, attaining an impressive accuracy of 99%, all the while significantly reducing training time by nearly threefold.
... 2019 et al. Da Ke [6] convolution neural network (CNN) LSTM A method was devised for enhancing cognitive radio based detection in low signal-to-noise ratio (SNR) situations. With the use of this technique, frequency and temporal domain information may be extracted from the data using both a CNN and LSTM approach. ...
... An energy transmitter time management scheme for IoT nodes can reduce charging expenses while keeping IoT nodes adequately charged. Deep-learning-method-based passive signal detection [22] can optimize cognitive-radio-based detection across low-signal-to-noise environments. Convolution neural networks and long short-term memory algorithms can be leveraged in signal frequency and time domain feature extraction. ...
The aim of this systematic review was to identify the correlations between spectrum sensing, clustering algorithms, and energy-harvesting technology for cognitive-radio-based internet of things (IoT) networks in terms of deep-learning-based, nonorthogonal, multiple-access techniques. The search results and screening procedures were configured with the use of a web-based Shiny app in the Preferred Reporting Items for Systematic Reviews and Meta-analysis (PRISMA) flow design. AMSTAR, DistillerSR, Eppi-Reviewer, PICO Portal, Rayyan, and ROBIS were the review software systems harnessed for screening and quality assessment, while bibliometric mapping (dimensions) and layout algorithms (VOSviewer) configured data visualization and analysis. Cognitive radio is pivotal in the utilization of an adequate radio spectrum source, with spectrum sensing optimizing cognitive radio network operations, opportunistic spectrum access and sensing able to boost the efficiency of cognitive radio networks, and cooperative spectrum sharing together with simultaneous wireless information and power transfer able increase spectrum and energy efficiency in 6G wireless communication networks and across IoT devices for efficient data exchange.
... Furthermore,they also show the perfor mance of the proposed model on IQ samples and AP samples. • Authors in [15] propose a deep learning method-based passive signal detection.They used a convolution neural network (CNN) and and therefore the long short-term memory (LSTM) approach to extract the frequency and time domain features of the signal. This method can de tect signal when little to none prior information exists. ...
The misuse of frequency bands leads to a spectrum shortage. The cognitive radio appears as a natural solution to this problem. A good
exploitation of the frequency spectrum starts with a good detection through various techniques, each with its advantages and limitations. In this paper we worked on improving the accuracy of spectrum sensing by developing a new cnn model and the transfer learning of data, also we used the automatic modulation recognition technique to insure the previous knowledge of data witch helped in improving the quality of detection and the performance of the cnn model. our method is based on three aspects entitled aspect1, aspect2 and aspect3. In aspect1 we trained the model to preform the modulation recognition with 11 classes. In aspect2 the model was trained with tow classes an performed the spectrum sensing. In aspect 3 we used the pre-trained model from aspect1 to perform the spectrum sensing with data from aspect2.We trained the model with many types of signals from the dataset RadioML2016.10a as well as noise data that we generated. We also use transfer learning strategies to improve the performance of the sensing model. The results show that we were able to achieve maximum accuracy of 97.22% for the sensing and 99 % for the modulation classification as best accuracy which is very competitive and better than many other proposed techniques.
... Ke et. al. [17] experimented with the combination of CNN and LSTM to extract the frequency and time-domain features of the sensing data for PU detection in a low SNR environment. ...
The hidden node problem is one of the most challenging issue in Cooperative Spectrum Sensing (CSS). The system models adopted by the existing Deep Learning-based spectrum sensing methods have not focused on modeling the hidden node scenario in cognitive radio networks. Further, these methods are unable to adapt to the dynamic channel conditions in the wireless environment since they have not considered the effect of fading environment. Motivated from these limitations, we propose GCN-CSS, a novel Graph Convolution Network (GCN) based cooperative spectrum sensing methodology which adapts to the dynamic changes in the Cognitive Radio Network. To the best of the author's knowledge, this is the first work to apply GCN for solving CSS problem. We have considered a practical system model which handles the dynamic channel condition i.e. SUs with multiple antennas experiencing different fading models with different fading severity. We have also catered the scenario of imperfect reporting channel between the SUs and the fusion centre along with the imperfect sensing channel to prove the robustness of the proposed model. With sufficient simulations, the superiority of the proposed methodology is proven in different dynamic scenarios of the wireless environment.