Pub Date : 2022-03-30DOI: 10.1109/dspa53304.2022.9790740
D. Poleshchenko, A. Glushchenko, A. Fomin
This paper is devoted to the efficiency analysis of detectors for the recognition of digits, which are mechanically stamped by a special machine on a steel cast billet. Such detectors are based on pre-trained deep neural networks. In the study, we analyze the performance of four different Faster R-CNN-based detectors. These neural networks have been trained and tested on our own training dataset obtained from the electro-metallurgical plant. According to the experiments, the best results are achieved by the Faster-RCNN Inception-Resnet v2 neural network detector. Its accuracy is about 98% on the test set.
本文研究了用专用机床在铸钢坯上机械冲压数字识别探测器的效率分析。这种检测器基于预训练的深度神经网络。在研究中，我们分析了四种不同的Faster r - cnn检测器的性能。这些神经网络已经在我们自己从电冶炼厂获得的训练数据集上进行了训练和测试。实验结果表明，Faster-RCNN Inception-Resnet v2神经网络检测器的检测效果最好。在测试集上，其准确率约为98%。
{"title":"Application of Pre-Trained Deep Neural Networks to Identify Cast Billet End Stamp before Heating","authors":"D. Poleshchenko, A. Glushchenko, A. Fomin","doi":"10.1109/dspa53304.2022.9790740","DOIUrl":"https://doi.org/10.1109/dspa53304.2022.9790740","url":null,"abstract":"This paper is devoted to the efficiency analysis of detectors for the recognition of digits, which are mechanically stamped by a special machine on a steel cast billet. Such detectors are based on pre-trained deep neural networks. In the study, we analyze the performance of four different Faster R-CNN-based detectors. These neural networks have been trained and tested on our own training dataset obtained from the electro-metallurgical plant. According to the experiments, the best results are achieved by the Faster-RCNN Inception-Resnet v2 neural network detector. Its accuracy is about 98% on the test set.","PeriodicalId":428492,"journal":{"name":"2022 24th International Conference on Digital Signal Processing and its Applications (DSPA)","volume":"27 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-03-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126940485","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-03-30DOI: 10.1109/dspa53304.2022.9790777
A. V. Kleopin, M. A. Zenchenko
A model for oscilloscope measurements has been developed and the broadband pulses waveform recovering and regularization methods have been proposed. In practice due to the influence of noise and measurement uncertainty, input signal recovery requires equations that are ill-posed to be solved, thus regularization was used. We have chosen a stabilizing function and have found a number of regularization parameters to describe a recovery filter. A mathematical simulation and an experimental verification have been conducted to confirm that recovery of a pulse waveform with a bandwidth increased by 1.5 times is performed with an increase in the noise root mean square (RMS) value of a recovered signal by 1.5–3 times.
{"title":"Tikhonov Regularization in Pulse Signal Processing for Oscilloscope Measurements","authors":"A. V. Kleopin, M. A. Zenchenko","doi":"10.1109/dspa53304.2022.9790777","DOIUrl":"https://doi.org/10.1109/dspa53304.2022.9790777","url":null,"abstract":"A model for oscilloscope measurements has been developed and the broadband pulses waveform recovering and regularization methods have been proposed. In practice due to the influence of noise and measurement uncertainty, input signal recovery requires equations that are ill-posed to be solved, thus regularization was used. We have chosen a stabilizing function and have found a number of regularization parameters to describe a recovery filter. A mathematical simulation and an experimental verification have been conducted to confirm that recovery of a pulse waveform with a bandwidth increased by 1.5 times is performed with an increase in the noise root mean square (RMS) value of a recovered signal by 1.5–3 times.","PeriodicalId":428492,"journal":{"name":"2022 24th International Conference on Digital Signal Processing and its Applications (DSPA)","volume":"9 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-03-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130440054","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-03-30DOI: 10.1109/dspa53304.2022.9790741
A. Gvozdarev, Yury A. Bryukhanov
The article presents the study of ADC performance in the presence of additive noise and input signal fading. The ADC device is represented by the quantization unit with antisymmetric characteristics, which resolves the possible overload situation with either the saturation or the reset strategies. The research is mainly focused on the low-resolution quantizers that are interesting from the practical perspectives of implementation in the Massive-MIMO systems with the hundreds of antenna elements. The input signal fading is accounted for by assuming that the instantaneous channel coefficient follows the complex Nakagami-m distribution. Based on the obtained earlier closed-form expressions for the ADC output signal statistics an energy suppression coefficient (as a quantifying measure of the induced distortions) was evaluated and analyzed for various system and channel parameters values. The possible ranges of analyzed parameters values yielding minimum suppression are calculated, and the worst-case scenarios are determined.
{"title":"Multipath Fading Impact on the Quantizer Output Signal Energy Suppression","authors":"A. Gvozdarev, Yury A. Bryukhanov","doi":"10.1109/dspa53304.2022.9790741","DOIUrl":"https://doi.org/10.1109/dspa53304.2022.9790741","url":null,"abstract":"The article presents the study of ADC performance in the presence of additive noise and input signal fading. The ADC device is represented by the quantization unit with antisymmetric characteristics, which resolves the possible overload situation with either the saturation or the reset strategies. The research is mainly focused on the low-resolution quantizers that are interesting from the practical perspectives of implementation in the Massive-MIMO systems with the hundreds of antenna elements. The input signal fading is accounted for by assuming that the instantaneous channel coefficient follows the complex Nakagami-m distribution. Based on the obtained earlier closed-form expressions for the ADC output signal statistics an energy suppression coefficient (as a quantifying measure of the induced distortions) was evaluated and analyzed for various system and channel parameters values. The possible ranges of analyzed parameters values yielding minimum suppression are calculated, and the worst-case scenarios are determined.","PeriodicalId":428492,"journal":{"name":"2022 24th International Conference on Digital Signal Processing and its Applications (DSPA)","volume":"16 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-03-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115390030","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-03-30DOI: 10.1109/dspa53304.2022.9790779
V. Lesnikov, T. Naumovich, A. Chastikov
Modern methods of digital signal processing are based not only on equidistant discretization. The methods of non-uniform discretization, including periodically non-uniform discretization, are widely used. This paper describes the possibility of representing the results of periodically nonuniform sampling of a complex harmonic signal by the results of uniform sampling of a specially formed signal, which is the sum of complex harmonic signals, the number of which, amplitudes, frequencies and phases are determined both by the parameters of the original signal and by the parameters of the sampling scheme.
{"title":"Analysis of Periodically Non-Uniform Sampled Signals","authors":"V. Lesnikov, T. Naumovich, A. Chastikov","doi":"10.1109/dspa53304.2022.9790779","DOIUrl":"https://doi.org/10.1109/dspa53304.2022.9790779","url":null,"abstract":"Modern methods of digital signal processing are based not only on equidistant discretization. The methods of non-uniform discretization, including periodically non-uniform discretization, are widely used. This paper describes the possibility of representing the results of periodically nonuniform sampling of a complex harmonic signal by the results of uniform sampling of a specially formed signal, which is the sum of complex harmonic signals, the number of which, amplitudes, frequencies and phases are determined both by the parameters of the original signal and by the parameters of the sampling scheme.","PeriodicalId":428492,"journal":{"name":"2022 24th International Conference on Digital Signal Processing and its Applications (DSPA)","volume":"25 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-03-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121345193","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-03-30DOI: 10.1109/dspa53304.2022.9790750
A. Tashlinskii, G. Safina, Michael Tsaryov
The stochastic algorithm for multi-time image alignment based on Tsallis mutual information is proposed. The algorithm assumes determining a set of fragments on the reference image, estimating the deformation parameters of these fragments using a local alignment model, filtering unreliable parameter estimates, finding the parameters of the global alignment model by deformations of fragments.
{"title":"Algorithm for Different-Time Image Alignment on the Base of Mutual Information","authors":"A. Tashlinskii, G. Safina, Michael Tsaryov","doi":"10.1109/dspa53304.2022.9790750","DOIUrl":"https://doi.org/10.1109/dspa53304.2022.9790750","url":null,"abstract":"The stochastic algorithm for multi-time image alignment based on Tsallis mutual information is proposed. The algorithm assumes determining a set of fragments on the reference image, estimating the deformation parameters of these fragments using a local alignment model, filtering unreliable parameter estimates, finding the parameters of the global alignment model by deformations of fragments.","PeriodicalId":428492,"journal":{"name":"2022 24th International Conference on Digital Signal Processing and its Applications (DSPA)","volume":"157 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-03-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122458497","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-03-30DOI: 10.1109/dspa53304.2022.9790767
N. Andriyanov, V. Dementiev, A. Tashlinsky, A. Danilov
The paper discusses using Deep Stream technologies in tasks for predicting best locations for deployment the bakeries. Such methods provides the calculation of people going through possible bakery and use convolutional neural networks for detection people and some effective algorithms for counting. The proposed solution allows deploying successful bakeries and keeping money in real production. Furthermore a lot of applied data science models were used for data analysis in these conditions. The paper discusses in detail the regression, factorial, cluster and discriminant analysis on the example of real data on the operation of a chain of bakeries with changes to preserve trade secrets. The analysis made it possible to simplify the decision-making process for managers among many factors. Moreover, the proposed method made it possible to predict profitability when opening a new point and explore various models of its development. Comparison results are provided for 3 models. The choice was made in favor of one of them. This choice resulted in the opening of a profitable bakery with a high profit margin for the retail market. The regression, factor and cluster analysis results show good opportunities to apply the results of the analysis for making management decisions when choosing the location of bakeries.
{"title":"Machine Learning Technologies for Bakery Management Decisions","authors":"N. Andriyanov, V. Dementiev, A. Tashlinsky, A. Danilov","doi":"10.1109/dspa53304.2022.9790767","DOIUrl":"https://doi.org/10.1109/dspa53304.2022.9790767","url":null,"abstract":"The paper discusses using Deep Stream technologies in tasks for predicting best locations for deployment the bakeries. Such methods provides the calculation of people going through possible bakery and use convolutional neural networks for detection people and some effective algorithms for counting. The proposed solution allows deploying successful bakeries and keeping money in real production. Furthermore a lot of applied data science models were used for data analysis in these conditions. The paper discusses in detail the regression, factorial, cluster and discriminant analysis on the example of real data on the operation of a chain of bakeries with changes to preserve trade secrets. The analysis made it possible to simplify the decision-making process for managers among many factors. Moreover, the proposed method made it possible to predict profitability when opening a new point and explore various models of its development. Comparison results are provided for 3 models. The choice was made in favor of one of them. This choice resulted in the opening of a profitable bakery with a high profit margin for the retail market. The regression, factor and cluster analysis results show good opportunities to apply the results of the analysis for making management decisions when choosing the location of bakeries.","PeriodicalId":428492,"journal":{"name":"2022 24th International Conference on Digital Signal Processing and its Applications (DSPA)","volume":"29 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-03-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121403960","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Methods and algorithms for digital one-dimensional and two-dimensional Fourier processing have the widest applications in solving a wide range of practical problems in many areas of scientific research. The paper gives a generalization of standard discrete Fourier transform in the form of parametric discrete Fourier transform, and considers the theory of its application. The article considers new discrete two-dimensional Fourier transform which is discrete two-dimensional Fourier transform with variable parameters, which, being a generalization of standard two-dimensional discrete Fourier transform, has a number of advantages over the standard one. Due to the widespread use of real signals in practice, the paper investigates the properties of the complex-conjugate symmetry of the coefficients of two-dimensional discrete Fourier transform with variable parameters for this class of signals. The concept of cross-complex-conjugate symmetry of the coefficients of two-dimensional discrete Fourier transform with variable parameters for real signals is introduced. The properties of the cross-complex-conjugate symmetry of the coefficients of two-dimensional discrete Fourier transform with variable parameters for real signals are confirmed by the results of mathematical modeling. Methods and algorithms for fast computation of discrete Fourier transform with variable parameters of real signals for various combinations of variable parameters have been developed.
{"title":"Complex-Conjugate Symmetry of Coefficients of Two-Dimensional Discrete Fourier Transform with Variable Parameters of Real Signals","authors":"Ponomareva Olga, Ponomarev Alexey, Smirnova Natalia","doi":"10.1109/dspa53304.2022.9790758","DOIUrl":"https://doi.org/10.1109/dspa53304.2022.9790758","url":null,"abstract":"Methods and algorithms for digital one-dimensional and two-dimensional Fourier processing have the widest applications in solving a wide range of practical problems in many areas of scientific research. The paper gives a generalization of standard discrete Fourier transform in the form of parametric discrete Fourier transform, and considers the theory of its application. The article considers new discrete two-dimensional Fourier transform which is discrete two-dimensional Fourier transform with variable parameters, which, being a generalization of standard two-dimensional discrete Fourier transform, has a number of advantages over the standard one. Due to the widespread use of real signals in practice, the paper investigates the properties of the complex-conjugate symmetry of the coefficients of two-dimensional discrete Fourier transform with variable parameters for this class of signals. The concept of cross-complex-conjugate symmetry of the coefficients of two-dimensional discrete Fourier transform with variable parameters for real signals is introduced. The properties of the cross-complex-conjugate symmetry of the coefficients of two-dimensional discrete Fourier transform with variable parameters for real signals are confirmed by the results of mathematical modeling. Methods and algorithms for fast computation of discrete Fourier transform with variable parameters of real signals for various combinations of variable parameters have been developed.","PeriodicalId":428492,"journal":{"name":"2022 24th International Conference on Digital Signal Processing and its Applications (DSPA)","volume":"12 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-03-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125781074","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-03-30DOI: 10.1109/dspa53304.2022.9790763
A. Novikov, A. Pronkin, S. Vityazev
Methods of periodic noise filtering are usually based on the idea of transforming the coefficients of the direct two-dimensional Fourier transform before performing the inverse Fourier transform in such a way that components which correspond to the periodic noise are excluded. This paper proposes a simple method for detecting local maxima of the amplitude spectrum and their neighborhoods based on the use of an edge detector. The contour image obtained with the edge detector forms a mask. This mask is then used to “cut out” the frequencies of periodic interferences. The offered approach is investigated for various shapes and intensities of the noise. A mixture of a periodic noise and an additive white noise is studied.
{"title":"Edge Detector Application in the Problem of Periodic Interference Filtering","authors":"A. Novikov, A. Pronkin, S. Vityazev","doi":"10.1109/dspa53304.2022.9790763","DOIUrl":"https://doi.org/10.1109/dspa53304.2022.9790763","url":null,"abstract":"Methods of periodic noise filtering are usually based on the idea of transforming the coefficients of the direct two-dimensional Fourier transform before performing the inverse Fourier transform in such a way that components which correspond to the periodic noise are excluded. This paper proposes a simple method for detecting local maxima of the amplitude spectrum and their neighborhoods based on the use of an edge detector. The contour image obtained with the edge detector forms a mask. This mask is then used to “cut out” the frequencies of periodic interferences. The offered approach is investigated for various shapes and intensities of the noise. A mixture of a periodic noise and an additive white noise is studied.","PeriodicalId":428492,"journal":{"name":"2022 24th International Conference on Digital Signal Processing and its Applications (DSPA)","volume":"43 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-03-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129955046","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-03-30DOI: 10.1109/dspa53304.2022.9790760
Peter V. Minin
A new hybrid method for object edge detection is suggested which uses a combination of analog and digital processing. Detection is performed using second derivative approach known as Marr-Hildreth method. Smoothing and physical measurement of gradient is performed at the imaging step in hardware of linear array of differential sensing cells. It is followed by digital computation of divergence of gradient. The result of imaging of sheet-like media is differential image, where one of components of gradient is recorded directly while for the other a single-band modulation of spatial carrier frequency is used. Point spread function (PSF) of the imaging array cell consists of negative and positive lobes of triangular shape oriented in opposite directions. Such lobe arrangement provides direct analog measurement of gradient vector and spatial modulation of one of its components. The divergence of gradient is further calculated by digital filtering which implements demodulation of spatial carrier frequency, calculation of derivatives of gradient components, and their summation. Subpixel interpolation is finally performed to obtain Laplacian image of media. For the experiment an array of flat capacitive differential cells was used to detect edges of objects formed as pieces of plasic film pasted to paper sheet. The size of PSF lobes of the cell was equal to 12 mm. In the Laplacian image acquired, the contour of increased thickness area of the sheet was marked from inside by a bright rim. Objects a few centimeters large were reproduced with correct size and slight distortion of the contour. The very presence of areas from 5 to 10 mm large was detected consistently, but their shape was heavily distorted and the size was bigger than in reality.
{"title":"Hybrid Implementation of Marr-Hildreth Method of Edge Detection","authors":"Peter V. Minin","doi":"10.1109/dspa53304.2022.9790760","DOIUrl":"https://doi.org/10.1109/dspa53304.2022.9790760","url":null,"abstract":"A new hybrid method for object edge detection is suggested which uses a combination of analog and digital processing. Detection is performed using second derivative approach known as Marr-Hildreth method. Smoothing and physical measurement of gradient is performed at the imaging step in hardware of linear array of differential sensing cells. It is followed by digital computation of divergence of gradient. The result of imaging of sheet-like media is differential image, where one of components of gradient is recorded directly while for the other a single-band modulation of spatial carrier frequency is used. Point spread function (PSF) of the imaging array cell consists of negative and positive lobes of triangular shape oriented in opposite directions. Such lobe arrangement provides direct analog measurement of gradient vector and spatial modulation of one of its components. The divergence of gradient is further calculated by digital filtering which implements demodulation of spatial carrier frequency, calculation of derivatives of gradient components, and their summation. Subpixel interpolation is finally performed to obtain Laplacian image of media. For the experiment an array of flat capacitive differential cells was used to detect edges of objects formed as pieces of plasic film pasted to paper sheet. The size of PSF lobes of the cell was equal to 12 mm. In the Laplacian image acquired, the contour of increased thickness area of the sheet was marked from inside by a bright rim. Objects a few centimeters large were reproduced with correct size and slight distortion of the contour. The very presence of areas from 5 to 10 mm large was detected consistently, but their shape was heavily distorted and the size was bigger than in reality.","PeriodicalId":428492,"journal":{"name":"2022 24th International Conference on Digital Signal Processing and its Applications (DSPA)","volume":"4 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-03-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121820832","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-03-30DOI: 10.1109/dspa53304.2022.9790756
A. Sergienko
Noncoherent reception of space-time block code (STBC) is considered. As a reception method, generalized likeli-hood ratio test is used. Two classes of codebooks are presented. The first class of codebooks is based on Alamouti STBC with added pilot symbols that make these codebooks suitable for noncoherent reception. The second class of codebooks is obtained using the autoencoder neural network. Error rate performance of the obtained codebooks is analyzed. The best results are obtained when autoencoder network is initialized with weights describing an Alamouti-based codebook. Power gain of optimized codebook over the Alamouti-based initialization codebook varies from a fraction of decibel to almost 5 dB.
{"title":"Design of Codebooks for Space-Time Block Code with Noncoherent GLRT-Based Reception","authors":"A. Sergienko","doi":"10.1109/dspa53304.2022.9790756","DOIUrl":"https://doi.org/10.1109/dspa53304.2022.9790756","url":null,"abstract":"Noncoherent reception of space-time block code (STBC) is considered. As a reception method, generalized likeli-hood ratio test is used. Two classes of codebooks are presented. The first class of codebooks is based on Alamouti STBC with added pilot symbols that make these codebooks suitable for noncoherent reception. The second class of codebooks is obtained using the autoencoder neural network. Error rate performance of the obtained codebooks is analyzed. The best results are obtained when autoencoder network is initialized with weights describing an Alamouti-based codebook. Power gain of optimized codebook over the Alamouti-based initialization codebook varies from a fraction of decibel to almost 5 dB.","PeriodicalId":428492,"journal":{"name":"2022 24th International Conference on Digital Signal Processing and its Applications (DSPA)","volume":"38 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-03-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115599241","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：