Automatic Digital Modulation Recognition Based on Genetic-Algorithm-Optimized Machine Learning Models

Ansari, Sam; Alnajjar, Khawla A.; Saad, Mohamed; Abdallah, Saeed; El-Moursy, Ali A.

doi:10.1109/access.2022.3171909

Cited by 23 publications

(6 citation statements)

References 44 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…(55) In reality the signal is always not ideal and combined with unwanted signal that is considered as noise which can affect our ability to determine the signal. Their relationship with the transmitter side I[n] and Q[n] is given by [52], [53], [36] 𝑟 𝐼 ).…”

Section: 𝑟[𝑛] = 𝑟 𝐼 [𝑛] + 𝑗𝑟 𝑄 [𝑛]mentioning

confidence: 99%

Radio Modulation Classification Optimization Using Combinatorial Deep Learning Technique

Elkhatib,

Kamalov,

Moussa

et al. 2024

IEEE Access

View full text Add to dashboard Cite

We present an automatic signal modulation classification model using combinatorial deep learning technique. Our proposed deep learning model increase accuracy for low Signal-to-Noise Ratio (SNR) and maintain a high classification accuracy for high SNR signals. Using a hybrid deep learning model combining both ConvLSTM with Transformer-block neural networks, the proposed modulation classifier architecture can learn the signal for both low and high SNR and get better accuracy for signals with high noise. The proposed deep learning modulation classification technique achieves improved classification accuracy of 66% for low SNR signals and 93.5% at high SNR showing that our model is robust under noisy signal modulation. Thus, getting better accuracy in lower SNR signals without sacrifice accuracy for higher SNR signals. An adaptive weighted focal loss function is proposed as an optimized loss function for efficient classification which can be used to control the outliers within a class imbalance and avoid underflow issues. Our deep learning radio modulation classification model works using raw signal without the need of denoising the noisy signal. INDEX TERMS Automatic modulation classification, dynamic spectrum allocation, deep learning techniques, transformer-block convLSTM, feature-based extraction, AI-based wireless communications.

show abstract

Section: 𝑟[𝑛] = 𝑟 𝐼 [𝑛] + 𝑗𝑟 𝑄 [𝑛]mentioning

confidence: 99%

Radio Modulation Classification Optimization Using Combinatorial Deep Learning Technique

Elkhatib,

Kamalov,

Moussa

et al. 2024

IEEE Access

View full text Add to dashboard Cite

show abstract

“…Unfortunately, the reconstructed images suffer from over-smoothness. A genetic algorithm (GA) is a well-known algorithm, inspired by the process of biological evolution [28]. This algorithm mimics Darwinian theory's "survival of the fittest" in nature.…”

Section: Introductionmentioning

confidence: 99%

Imaging of intracerebral hemorrhage with adaptive genetic algorithm in brain electrical impedance tomography

Shi

Liu

et al. 2023

Meas. Sci. Technol.

View full text Add to dashboard Cite

Intracerebral hemorrhage refers to bleeding caused by spontaneous rupture of blood vessels. Accurate diagnosis of hemorrhage is vital in the treatment of a patient. As a new medical imaging technique, electrical impedance tomography (EIT) is able to offer images of conductivity distribution variation caused by pathological change. However, image reconstruction of EIT suffers from the problem of serious ill-posedness. Especially in the brain imaging, irregular and multi-layered head structure together with the low conductivity skull further aggravate the problem. In order to address the problem, a new image reconstruction method is proposed for imaging of hemorrhage in this work. With the solution solved by Tikhonov regularization method as the original conductivity distribution, the proposed method tends to enhance the reconstruction quality by introducing adaptive genetic algorithm. To test the performance of the proposed method, simulation work is conducted. A three-layer head model is established and an inclusion which simulates hemorrhage is placed at six different locations in the brain layer. Images reconstructed by Tikhonov method, Newton-Raphson method and traditional Genetic Algorithm method are used for comparisons. Quantitative evaluation is also performed. The anti-noise performance of the proposed method is estimated by considering noise with signal-to-noise ratio of different levels. Aside from the simulation, phantom experiments are carried out to further verify the performance of the proposed method. The results show that the proposed method performs well in the reconstruction of simulated intracerebral hemorrhage. With the proposed method, the inclusion can be more accurately reconstructed and the background is much clearer than other three traditional methods.

show abstract

“…More specifically, the genetic algorithm has been applied in many machine learning applications, such as in the article by Ansari et al, which deals with the recognition of digital modulation signals. In this article, the genetic algorithm is used to optimize machine learning models by adjusting their features and parameters to achieve better signal recognition accuracy [39]. Additionally, in the study by Ji et al, a methodology is proposed that uses machine learning models to predict amplitude deviation in hot rolling, while genetic algorithms are employed to optimize the machine learning models and select features to improve prediction accuracy [40].…”

Section: Introductionmentioning

confidence: 99%

An Intelligent Technique for Initial Distribution of Genetic Algorithms

Charilogis,

Tsoulos,

Stavrou

2023

Axioms

View full text Add to dashboard Cite

The need to find the global minimum in multivariable functions is a critical problem in many fields of science and technology. Effectively solving this problem requires the creation of initial solution estimates, which are subsequently used by the optimization algorithm to search for the best solution in the solution space. In the context of this article, a novel approach to generating the initial solution distribution is presented, which is applied to a genetic optimization algorithm. Using the k-means clustering algorithm, a distribution based on data similarity is created. This helps in generating initial estimates that may be more tailored to the problem. Additionally, the proposed method employs a rejection sampling algorithm to discard samples that do not yield better solution estimates in the optimization process. This allows the algorithm to focus on potentially optimal solutions, thus improving its performance. Finally, the article presents experimental results from the application of this approach to various optimization problems, providing the scientific community with a new method for addressing this significant problem.

show abstract

Automatic Digital Modulation Recognition Based on Genetic-Algorithm-Optimized Machine Learning Models

Cited by 23 publications

References 44 publications

Radio Modulation Classification Optimization Using Combinatorial Deep Learning Technique

Radio Modulation Classification Optimization Using Combinatorial Deep Learning Technique

Imaging of intracerebral hemorrhage with adaptive genetic algorithm in brain electrical impedance tomography

An Intelligent Technique for Initial Distribution of Genetic Algorithms

Contact Info

Product

Resources

About