Machine Learning Supervised Antenna for Software Defined Cognitive Radios

Abstract Automatic Modulation Classification (AMC) is becoming an essential component in receiver designs for next-generation communication systems, such as Cognitive Radios (CR). AMC enables receivers to classify an intercepted signal’s modulation scheme without any prior information about the signal. This is becoming increasingly vital due to the combination of congested frequency bands and geographically disparate frequency licensing for the railroad industry across North America. Thus, a radio technology is needed that allows train systems to adapt automatically and intelligently to changing locations and corresponding RF environment fluctuations. Three AMC approaches have been proposed in the scientific literature. The performance of these approaches depends especially on the particular environment where the classifiers are employed. In this work, the authors present a performance evaluation of the Feature-based AMC approach, as this is the most promising approach for railroads in real-time AMC operations under various different wireless channel environments. This study is done as the first one for railroads application where it considers different environments models including Non-Gaussian Class A noise, Multipath fast fading, and their combination. The evaluation is conducted for signals using a series of QAM modulation schemes. The authors selected the signal’s Cumulant statistical features for the feature extraction stage in this study, coupled with three different machine learning classifiers: Support Vector Machine (SVM), Deep Neural Network (DNN) and Recurrent Neural Network (RNN) utilizing long-short term memory (LSTM), in order to maintain control over the classifiers’ accuracy and computational complexity, especially for the non-linear cases. Our results indicate that when the signal model noise shows higher non-linear behavior, the RNN classifier on average achieves higher classification accuracy than the other classifiers.

Download Full-text

A RELATIONSHIP BETWEEN COGNITIVE INFORMATION PROCESSING IN LEARNING THEORY AND MACHINE LEARNING TECHNIQUES IN COGNITIVE RADIOS

SOCIETY INTEGRATION EDUCATION Proceedings of the International Scientific Conference ◽

10.17770/sie2018vol1.3191 ◽

2018 ◽

Vol 5 ◽

pp. 465-474 ◽

Cited By ~ 3

Author(s):

Zhaneta Tasheva ◽

Rosen Bogdanov

Keyword(s):

Machine Learning ◽

Information Processing ◽

Learning Theory ◽

Cognitive Radios ◽

Machine Learning Techniques ◽

Learning Theories ◽

Cognitive Approach ◽

Cognitive Information Processing ◽

Learning Techniques ◽

Cognitive Information

The relationship between cognitivism as learning theory in education and machine learning is characterized in this survey paper. The cognitivism describes how learning occurs through internal processing of information and thus leads to understanding and retention. Cognitive information processing plays an active role to understand and process information that learner receives and relates it to already known and stored within learner’s memory. Thus, the cognitive approach defines learning as a change in knowledge which is stored in learner’s memory, and not a change in learner’s behaviour. In regard with importance of various learning problems to designing cognitive communications systems the two main classification categories of learning techniques are explained. Furthermore, the cognitive radio learning algorithms that have been proposed are described. Finally, the similarities and differences among the principles of learning theories and machine learning are discussed.

Download Full-text

Machine Learning for 5G MIMO Modulation Detection

Sensors ◽

10.3390/s21051556 ◽

2021 ◽

Vol 21 (5) ◽

pp. 1556

Author(s):

Haithem Ben Chikha ◽

Ahmad Almadhor ◽

Waqas Khalid

Keyword(s):

Machine Learning ◽

Software Defined Radio ◽

Mimo Systems ◽

False Positive Rate ◽

Cognitive Radios ◽

Machine Learning Techniques ◽

Destination Node ◽

Detection Techniques ◽

Positive Rate ◽

Modulation Detection

Modulation detection techniques have received much attention in recent years due to their importance in the military and commercial applications, such as software-defined radio and cognitive radios. Most of the existing modulation detection algorithms address the detection dedicated to the non-cooperative systems only. In this work, we propose the detection of modulations in the multi-relay cooperative multiple-input multiple-output (MIMO) systems for 5G communications in the presence of spatially correlated channels and imperfect channel state information (CSI). At the destination node, we extract the higher-order statistics of the received signals as the discriminating features. After applying the principal component analysis technique, we carry out a comparative study between the random committee and the AdaBoost machine learning techniques (MLTs) at low signal-to-noise ratio. The efficiency metrics, including the true positive rate, false positive rate, precision, recall, F-Measure, and the time taken to build the model, are used for the performance comparison. The simulation results show that the use of the random committee MLT, compared to the AdaBoost MLT, provides gain in terms of both the modulation detection and complexity.

Download Full-text

Machine Learning Techniques Applied to Multiband Spectrum Sensing in Cognitive Radios

Sensors ◽

10.3390/s19214715 ◽

2019 ◽

Vol 19 (21) ◽

pp. 4715 ◽

Cited By ~ 3

Author(s):

Yanqueleth Molina-Tenorio ◽

Alfonso Prieto-Guerrero ◽

Rafael Aguilar-Gonzalez ◽

Silvia Ruiz-Boqué

Keyword(s):

Machine Learning ◽

Spectrum Sensing ◽

Expectation Maximization ◽

Cognitive Radios ◽

Primary User ◽

Machine Learning Techniques ◽

Learning Techniques ◽

Multiple Primary ◽

Approximation Coefficients ◽

Multiple Primary Users

In this work, three specific machine learning techniques (neural networks, expectation maximization and k-means) are applied to a multiband spectrum sensing technique for cognitive radios. All of them have been used as a classifier using the approximation coefficients from a Multiresolution Analysis in order to detect presence of one or multiple primary users in a wideband spectrum. Methods were tested on simulated and real signals showing a good performance. The results presented of these three methods are effective options for detecting primary user transmission on the multiband spectrum. These methodologies work for 99% of cases under simulated signals of SNR higher than 0 dB and are feasible in the case of real signals.

Download Full-text

Analysis of Machine Learning Algorithms for Spectrum Decision in Cognitive Radios

2018 15th International Symposium on Wireless Communication Systems (ISWCS) ◽

10.1109/iswcs.2018.8491060 ◽

2018 ◽

Cited By ~ 1

Author(s):

Luccas Rafael Martins Pinto ◽

Luiz Henrique Andrade Correia

Keyword(s):

Machine Learning ◽

Learning Algorithms ◽

Cognitive Radios ◽

Machine Learning Algorithms ◽

Spectrum Decision

Download Full-text

A Strong Machine Learning Classifier and Decision Stumps Based Hybrid AdaBoost Classification Algorithm for Cognitive Radios

Sensors ◽

10.3390/s19235077 ◽

2019 ◽

Vol 19 (23) ◽

pp. 5077 ◽

Cited By ~ 4

Author(s):

Siji Chen ◽

Bin Shen ◽

Xin Wang ◽

Sang-Jo Yoo

Keyword(s):

Machine Learning ◽

Spectrum Sensing ◽

Cooperative Spectrum Sensing ◽

Cognitive Radios ◽

Radio Networks ◽

Adaptive Boosting ◽

Adaboost Algorithm ◽

Second Stage ◽

Learning Classifier

Machine learning (ML) based classification methods have been viewed as one kind of alternative solution for cooperative spectrum sensing (CSS) in recent years. In this paper, ML techniques based CSS algorithms are investigated for cognitive radio networks (CRN). Specifically, a strong machine learning classifier (MLC) and decision stumps (DS) based adaptive boosting (AdaBoost) classification mechanism is proposed for pattern classification of the primary user’s behavior in the network. The conventional AdaBoost algorithm only combines multiple sub-classifiers and produces a strong weight based on their weights in classification. Taking into account the fact that the strong MLC and the weak DS serve as different sub-classifiers in classification, we propose employing a strong MLC as the first-stage classifier and DS as the second-stage classifiers, to eventually determine the class that the spectrum energy vector belongs to. We verify in simulations that the proposed hybrid AdaBoost algorithms are capable of achieving a higher detection probability than the conventional ML based spectrum sensing algorithms and the conventional hard fusion based CSS schemes.

Download Full-text