Pub Date : 2022-08-29DOI: 10.23919/eusipco55093.2022.9909967
Manon Mottier, G. Chardon, F. Pascal
Identifying unknown RADAR emitters from re-ceived pulses is an important problem in electronic intelligence. It is a difficult problem, as agile RADAR emitters can have complex characteristics, and measurements are corrupted by various noises (non-Gaussian noise, missing pulses, etc.). In this paper, we introduce a new classification method based on optimal transport distances between collected RADAR pulses and a priori known emitter classes. Compared to previously proposed methods, this method does not require a training step, it can deal with a large number of classes, and it is easily interpretable. The method is tested on data obtained by a realistic RADAR scene simulator.
{"title":"RADAR Emitter Classification with Optimal Transport Distances","authors":"Manon Mottier, G. Chardon, F. Pascal","doi":"10.23919/eusipco55093.2022.9909967","DOIUrl":"https://doi.org/10.23919/eusipco55093.2022.9909967","url":null,"abstract":"Identifying unknown RADAR emitters from re-ceived pulses is an important problem in electronic intelligence. It is a difficult problem, as agile RADAR emitters can have complex characteristics, and measurements are corrupted by various noises (non-Gaussian noise, missing pulses, etc.). In this paper, we introduce a new classification method based on optimal transport distances between collected RADAR pulses and a priori known emitter classes. Compared to previously proposed methods, this method does not require a training step, it can deal with a large number of classes, and it is easily interpretable. The method is tested on data obtained by a realistic RADAR scene simulator.","PeriodicalId":231263,"journal":{"name":"2022 30th European Signal Processing Conference (EUSIPCO)","volume":"32 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-08-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124659367","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 : 2022-08-29DOI: 10.23919/eusipco55093.2022.9909898
M. Vázquez, P. Henarejos, A. Pérez-Neira
This paper proposes an optimization technique for satellite systems with flexible payloads. Unlike current satellites whose per-beam capacity is fixed, forthcoming payloads will have bandwidth and power allocation reconfiguration capabilities allowing the operators to modify the offered capacity. Assuming a generic flexible payload architecture, this paper introduces an op-timization technique that is able to provide an efficient bandwidth and power allocation that fulfil the user terminals rate requests. Furthermore, we introduce a deep learning regression algorithm able to reproduce the mapping of the proposed optimization technique with a very reduced computational complexity. By using the output of the optimization technique as ground truth, we design a deep neural network that behaves very similar to the optimization problem yet with a dramatically reduced computational time. Numerical results show the benefits of the proposed technique and in particular, we observe two order of magnitude computational time decrease when using the deep learning approach compared to the classical optimization technique yet preserving almost the same performance.
{"title":"Learning to Optimize Satellite Flexible Payloads","authors":"M. Vázquez, P. Henarejos, A. Pérez-Neira","doi":"10.23919/eusipco55093.2022.9909898","DOIUrl":"https://doi.org/10.23919/eusipco55093.2022.9909898","url":null,"abstract":"This paper proposes an optimization technique for satellite systems with flexible payloads. Unlike current satellites whose per-beam capacity is fixed, forthcoming payloads will have bandwidth and power allocation reconfiguration capabilities allowing the operators to modify the offered capacity. Assuming a generic flexible payload architecture, this paper introduces an op-timization technique that is able to provide an efficient bandwidth and power allocation that fulfil the user terminals rate requests. Furthermore, we introduce a deep learning regression algorithm able to reproduce the mapping of the proposed optimization technique with a very reduced computational complexity. By using the output of the optimization technique as ground truth, we design a deep neural network that behaves very similar to the optimization problem yet with a dramatically reduced computational time. Numerical results show the benefits of the proposed technique and in particular, we observe two order of magnitude computational time decrease when using the deep learning approach compared to the classical optimization technique yet preserving almost the same performance.","PeriodicalId":231263,"journal":{"name":"2022 30th European Signal Processing Conference (EUSIPCO)","volume":"21 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-08-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124670559","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 : 2022-08-29DOI: 10.23919/eusipco55093.2022.9909648
{"title":"30th European Signal Processing Conference (EUSIPCO 2022)","authors":"","doi":"10.23919/eusipco55093.2022.9909648","DOIUrl":"https://doi.org/10.23919/eusipco55093.2022.9909648","url":null,"abstract":"","PeriodicalId":231263,"journal":{"name":"2022 30th European Signal Processing Conference (EUSIPCO)","volume":"28 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-08-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129451594","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 : 2022-08-29DOI: 10.23919/eusipco55093.2022.9909730
Oded Bialer, A. Jonas, O. Longman
Conventional automotive radar perform range-Doppler coherent integration (stretch processing) under the assumption that the range of each object is constant during the integration interval. This assumption yields an efficient computation algorithm. However, when the object's relative speed is high and/or the coherent integration interval is large, the range migration is significant with respect to the range resolution, and as a result, the detection performance of the conventional range-Doppler coherent integration degrades significantly. The Radon-Fourier Transform (RFT) is the optimal method (in the sense of detection performance) for coherent integration with range migration, however, its complexity is large and may not be practical for implementation. In this paper, we develop a range-Doppler coherent integration algorithm that takes into account the range migration with efficient computation. We utilize the fact that range migration is a function of the Doppler frequency and derive an approximation to the RFT. The proposed algorithm significantly outperforms conventional coherent integration when the object's range migration is significant. Furthermore, it attains the performance of the RFT but with significantly lower complexity.
{"title":"Efficient and Robust Automotive Radar Coherent Integration With Range Migration","authors":"Oded Bialer, A. Jonas, O. Longman","doi":"10.23919/eusipco55093.2022.9909730","DOIUrl":"https://doi.org/10.23919/eusipco55093.2022.9909730","url":null,"abstract":"Conventional automotive radar perform range-Doppler coherent integration (stretch processing) under the assumption that the range of each object is constant during the integration interval. This assumption yields an efficient computation algorithm. However, when the object's relative speed is high and/or the coherent integration interval is large, the range migration is significant with respect to the range resolution, and as a result, the detection performance of the conventional range-Doppler coherent integration degrades significantly. The Radon-Fourier Transform (RFT) is the optimal method (in the sense of detection performance) for coherent integration with range migration, however, its complexity is large and may not be practical for implementation. In this paper, we develop a range-Doppler coherent integration algorithm that takes into account the range migration with efficient computation. We utilize the fact that range migration is a function of the Doppler frequency and derive an approximation to the RFT. The proposed algorithm significantly outperforms conventional coherent integration when the object's range migration is significant. Furthermore, it attains the performance of the RFT but with significantly lower complexity.","PeriodicalId":231263,"journal":{"name":"2022 30th European Signal Processing Conference (EUSIPCO)","volume":"3 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-08-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129660888","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 : 2022-08-29DOI: 10.48550/arXiv.2209.05900
D. Krause, A. Mesaros
Sound event detection (SED) and Acoustic scene classification (ASC) are two widely researched audio tasks that constitute an important part of research on acoustic scene analysis. Considering shared information between sound events and acoustic scenes, performing both tasks jointly is a natural part of a complex machine listening system. In this paper, we investigate the usefulness of several spatial audio features in training a joint deep neural network (DNN) model performing SED and ASC. Experiments are performed for two different datasets containing binaural recordings and synchronous sound event and acoustic scene labels to analyse the differences between performing SED and ASC separately or jointly. The presented results show that the use of specific binaural features, mainly the Generalized Cross Correlation with Phase Transform (GCC-phat) and sines and cosines of phase differences, result in a better performing model in both separate and joint tasks as compared with baseline methods based on logmel energies only.
{"title":"Binaural Signal Representations for Joint Sound Event Detection and Acoustic Scene Classification","authors":"D. Krause, A. Mesaros","doi":"10.48550/arXiv.2209.05900","DOIUrl":"https://doi.org/10.48550/arXiv.2209.05900","url":null,"abstract":"Sound event detection (SED) and Acoustic scene classification (ASC) are two widely researched audio tasks that constitute an important part of research on acoustic scene analysis. Considering shared information between sound events and acoustic scenes, performing both tasks jointly is a natural part of a complex machine listening system. In this paper, we investigate the usefulness of several spatial audio features in training a joint deep neural network (DNN) model performing SED and ASC. Experiments are performed for two different datasets containing binaural recordings and synchronous sound event and acoustic scene labels to analyse the differences between performing SED and ASC separately or jointly. The presented results show that the use of specific binaural features, mainly the Generalized Cross Correlation with Phase Transform (GCC-phat) and sines and cosines of phase differences, result in a better performing model in both separate and joint tasks as compared with baseline methods based on logmel energies only.","PeriodicalId":231263,"journal":{"name":"2022 30th European Signal Processing Conference (EUSIPCO)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-08-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130399335","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 : 2022-08-29DOI: 10.23919/eusipco55093.2022.9909831
Izumi Ito
Convolutional neural networks (CNNs) are widely used in many areas. They feature convolutional layers that focus on spatial local node connections rather than full node connections. This makes networks much more efficient for spatial information. The convolution is a mathematical operation on two functions and can be calculated using the discrete Fourier transform (DFT). Due to the close relation to the DFT, the discrete cosine transforms (DCTs) can be used for the calculation. In this paper, we focus on the convolution using DCTs for improvement of the performance of CNNs. The periodicity and symmetry inherent in the DCTs generate larger output feature maps. The proposed method in simple CNNs is demonstrated and the efficacy of the proposed method is testified using CIFAR-10 dataset.
{"title":"Convolution Using Discrete Cosine Transforms for Improving Performance of Convolutional Neural Networks","authors":"Izumi Ito","doi":"10.23919/eusipco55093.2022.9909831","DOIUrl":"https://doi.org/10.23919/eusipco55093.2022.9909831","url":null,"abstract":"Convolutional neural networks (CNNs) are widely used in many areas. They feature convolutional layers that focus on spatial local node connections rather than full node connections. This makes networks much more efficient for spatial information. The convolution is a mathematical operation on two functions and can be calculated using the discrete Fourier transform (DFT). Due to the close relation to the DFT, the discrete cosine transforms (DCTs) can be used for the calculation. In this paper, we focus on the convolution using DCTs for improvement of the performance of CNNs. The periodicity and symmetry inherent in the DCTs generate larger output feature maps. The proposed method in simple CNNs is demonstrated and the efficacy of the proposed method is testified using CIFAR-10 dataset.","PeriodicalId":231263,"journal":{"name":"2022 30th European Signal Processing Conference (EUSIPCO)","volume":"3 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-08-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129176735","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 : 2022-08-29DOI: 10.23919/eusipco55093.2022.9909567
Micael Bernhardt, Jaakko Marin, T. Riihonen
We assess the operation of a special multifunction full-duplex transceiver that uses its frequency-shifting constant-envelope transmitted signal as the downconversion carrier (in contrast to a tone for conventional direct conversion). While self-interference suppression is greatly simplified by using this architecture, the spectra of other downconverted signals turn up sweeping through the frequency domain as the cost of that. Adequate characterization and compensation of these consequences is the key to guarantee the required performance of emerging multifunction systems. We develop solutions to these effects and evaluate the behavior of the transceiver concept by varying transmission- and reception-related parameters.
{"title":"Characterization of Full-Duplex Constant-Envelope Transceiver for Emerging Multifunction Systems","authors":"Micael Bernhardt, Jaakko Marin, T. Riihonen","doi":"10.23919/eusipco55093.2022.9909567","DOIUrl":"https://doi.org/10.23919/eusipco55093.2022.9909567","url":null,"abstract":"We assess the operation of a special multifunction full-duplex transceiver that uses its frequency-shifting constant-envelope transmitted signal as the downconversion carrier (in contrast to a tone for conventional direct conversion). While self-interference suppression is greatly simplified by using this architecture, the spectra of other downconverted signals turn up sweeping through the frequency domain as the cost of that. Adequate characterization and compensation of these consequences is the key to guarantee the required performance of emerging multifunction systems. We develop solutions to these effects and evaluate the behavior of the transceiver concept by varying transmission- and reception-related parameters.","PeriodicalId":231263,"journal":{"name":"2022 30th European Signal Processing Conference (EUSIPCO)","volume":"2 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-08-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123815196","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 : 2022-08-29DOI: 10.23919/eusipco55093.2022.9909788
G. Enzner, Christoph Urbanietz, R. Martin
The acquisition of head-related impulse responses (HRIRs) has traditionally been a time-consuming acoustic measurement process. Novel continuous-azimuth recording techniques have dramatically accelerated the acquisition, but conversion into continuous Spatial-Fourier representations (SpaFoR) of HRIRs provides a host of cumbersome implementation challenges. The direct closed-form least-squares approach is unfortunately not practical and we will therefore explore the retrieval of SpaFoR model parameters of HRIR by contemporary machine-learning tools. Specifically, we employ the standard stochastic-gradient learning with Tensorflow on a graphics processing unit (GPU) and compare its performance with previous covariance-based least-squares on the general purpose processor. Apart from the sought simplification and acceleration, our paper is dedicated to hyperparameter optimization in order to make sure the final state of the machine learning approach still attains the accuracy of the optimal least-squares solution. The paper finally applies the proposed method to a real acoustic HRIR recording to illustrate the validity of the system identification obtained by learning.
{"title":"Optimized Learning of Spatial-Fourier Representations from Fast HRIR Recordings","authors":"G. Enzner, Christoph Urbanietz, R. Martin","doi":"10.23919/eusipco55093.2022.9909788","DOIUrl":"https://doi.org/10.23919/eusipco55093.2022.9909788","url":null,"abstract":"The acquisition of head-related impulse responses (HRIRs) has traditionally been a time-consuming acoustic measurement process. Novel continuous-azimuth recording techniques have dramatically accelerated the acquisition, but conversion into continuous Spatial-Fourier representations (SpaFoR) of HRIRs provides a host of cumbersome implementation challenges. The direct closed-form least-squares approach is unfortunately not practical and we will therefore explore the retrieval of SpaFoR model parameters of HRIR by contemporary machine-learning tools. Specifically, we employ the standard stochastic-gradient learning with Tensorflow on a graphics processing unit (GPU) and compare its performance with previous covariance-based least-squares on the general purpose processor. Apart from the sought simplification and acceleration, our paper is dedicated to hyperparameter optimization in order to make sure the final state of the machine learning approach still attains the accuracy of the optimal least-squares solution. The paper finally applies the proposed method to a real acoustic HRIR recording to illustrate the validity of the system identification obtained by learning.","PeriodicalId":231263,"journal":{"name":"2022 30th European Signal Processing Conference (EUSIPCO)","volume":"12 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-08-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114068852","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 : 2022-08-29DOI: 10.23919/eusipco55093.2022.9909667
Stéphane Dilungana, Antoine Deleforge, C. Foy, S. Faisan
This paper presents a method to jointly estimate the frequency-dependent absorption coefficients of the walls, ceiling and floor in a room from several impulse response measurements. The principle of the approach is to search among the observations for temporal windows of fixed size in which there is only one manifestation of acoustic reflection, based on the geometry of the setup which is assumed known up to some error. A probablistic procedure inspired by RANSAC that rejects putative outliers is devised for this purpose. Once the windows have been selected, the parameters of interest are estimated from the magnitude spectrograms of room impulse responses by minimizing a constrained cost function. Extensive simulation results on random shoebox rooms reveal that absorption coefficients can be efficiently recovered with the procedure, and that increasing the number of measurements improve the results while enhancing the robustness to noise and to geometrical uncertainty.
{"title":"Geometry-Informed Estimation of Surface Absorption Profiles from Room Impulse Responses","authors":"Stéphane Dilungana, Antoine Deleforge, C. Foy, S. Faisan","doi":"10.23919/eusipco55093.2022.9909667","DOIUrl":"https://doi.org/10.23919/eusipco55093.2022.9909667","url":null,"abstract":"This paper presents a method to jointly estimate the frequency-dependent absorption coefficients of the walls, ceiling and floor in a room from several impulse response measurements. The principle of the approach is to search among the observations for temporal windows of fixed size in which there is only one manifestation of acoustic reflection, based on the geometry of the setup which is assumed known up to some error. A probablistic procedure inspired by RANSAC that rejects putative outliers is devised for this purpose. Once the windows have been selected, the parameters of interest are estimated from the magnitude spectrograms of room impulse responses by minimizing a constrained cost function. Extensive simulation results on random shoebox rooms reveal that absorption coefficients can be efficiently recovered with the procedure, and that increasing the number of measurements improve the results while enhancing the robustness to noise and to geometrical uncertainty.","PeriodicalId":231263,"journal":{"name":"2022 30th European Signal Processing Conference (EUSIPCO)","volume":"66 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-08-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114627913","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 : 2022-08-29DOI: 10.23919/eusipco55093.2022.9909876
Anish Datta, S. Bandyopadhyay, Shruti Sachan, A. Pal
Today's world extensively depends on analytics of high dimensional sensor time-series, and, extracting informative representation. Sensor time-series across various applications such as healthcare and human wellness, machine maintenance etc., are generally unlabelled, and, getting the annotations is costly and time-consuming. Here, we propose an unsupervised feature selection method exploiting representation learning with a choice of best clustering and recommended distance measure. Proposed method reduces the feature space, to a compressed latent representation, known as Auto-encoded Compact Sequence of features, by retaining the most informative parts. It further selects a set of discriminative features, by computing the sim-ilarity / dissimilarity among the features in latent space using the recommended best distance measure. We have experimented using diverse time-series from UCR Time Series Classification archive, and observed, proposed method consistently outperforms state-of-the-art feature selection approaches.
{"title":"Unsupervised Feature Recommendation using Representation Learning","authors":"Anish Datta, S. Bandyopadhyay, Shruti Sachan, A. Pal","doi":"10.23919/eusipco55093.2022.9909876","DOIUrl":"https://doi.org/10.23919/eusipco55093.2022.9909876","url":null,"abstract":"Today's world extensively depends on analytics of high dimensional sensor time-series, and, extracting informative representation. Sensor time-series across various applications such as healthcare and human wellness, machine maintenance etc., are generally unlabelled, and, getting the annotations is costly and time-consuming. Here, we propose an unsupervised feature selection method exploiting representation learning with a choice of best clustering and recommended distance measure. Proposed method reduces the feature space, to a compressed latent representation, known as Auto-encoded Compact Sequence of features, by retaining the most informative parts. It further selects a set of discriminative features, by computing the sim-ilarity / dissimilarity among the features in latent space using the recommended best distance measure. We have experimented using diverse time-series from UCR Time Series Classification archive, and observed, proposed method consistently outperforms state-of-the-art feature selection approaches.","PeriodicalId":231263,"journal":{"name":"2022 30th European Signal Processing Conference (EUSIPCO)","volume":"91 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-08-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124388116","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：