Product Graph U-networks using Temporal pooling for
Spatiotemporal imputation
Aman Atman,Santosh Nannuru
KDD-Undergraduate and Master’s Consortium, KDD-UMC, 2025
@inproceedings{bib_Prod_2025, AUTHOR = {Atman, Aman and Nannuru, Santosh }, TITLE = {Product Graph U-networks using Temporal pooling for
Spatiotemporal imputation}, BOOKTITLE = {KDD-Undergraduate and Master’s Consortium}. YEAR = {2025}}
Missing values in time-series is a frequently occurring problem in
multi-sensor data analysis. The current state-of-the-art imputation
networks don’t generalize well across the benchmarks. Further,
some directly operate on space-time product graphs which may
minimize the information loss but are prohibitively expensive. We
propose Product Graph U-networks using Temporal pooling for Spa-
tiotemporal imputation (P-GUTS). It uses pooled representations
of the space-time sequence in the bottleneck part of the U-network
and is memory efficient. Since datasets can have differing levels of
redundancy, relying on a single view (pooled representation) of the
data can be limiting. Therefore, P-GUTS uses multiple U-Nets with
different pooling factors and aggregates the representations. We
demonstrate the effectiveness of P-GUTS on multiple real-world
imputation datasets and robustness across increasing missing rates.
P-GUTS also gives encouraging results in spatiotemporal forecast-
ing.
Mixture-of-Experts for Time Series Imputation
Aman Atman,Santosh Nannuru
Canadian AI Graduate Student Symposium, Caiac, 2025
@inproceedings{bib_Mixt_2025, AUTHOR = {Atman, Aman and Nannuru, Santosh }, TITLE = {Mixture-of-Experts for Time Series Imputation}, BOOKTITLE = {Canadian AI Graduate Student Symposium}. YEAR = {2025}}
Time-series data from sensor networks often has missing entries due to sensor failures,
poor communication, scheduled downtime, etc. Missing data imputation improves perfor-
mance of downstream tasks such as forecasting and anomaly detection. Therefore, highly
accurate imputation methods are desirable.
Multi-variate time series imputation methods using interpolation [1–3], or matrix factor-
ization [4, 5] have limited accuracy. Auto-regressive methods like ARIMA [6] and RNNs
like GRIN [7] are subject to error propagation and are slow. More recently, attention [8] has
been successfully used to design effective networks for various time series tasks – imputation
(SPIN, SPIN-H and Transformer) [9], forecasting [10, 11] and anomaly detection [12].
However, as evident from Figure 2, these state-of-the-art (SOTA) networks are effective
only on certain datasets. It motivates a more general meta-network which can effectively
aggregate these experts using a gating network. We provide the scope of improvement when
these frozen experts are selected optimally using an oracle. The possible reduction in error
can be as high as 28 − 40 percent by using existing imputation networks.
Sequential DOA Trajectory Estimation using Deep Complex Network and Residual Signals
Shreyas Jaiswal,Peter Gerstoft,Santosh Nannuru
International Conference on Acoustics, Speech, and Signal Processing, ICASSP, 2025
@inproceedings{bib_Sequ_2025, AUTHOR = {Jaiswal, Shreyas and Gerstoft, Peter and Nannuru, Santosh }, TITLE = {Sequential DOA Trajectory Estimation using Deep Complex Network and Residual Signals}, BOOKTITLE = {International Conference on Acoustics, Speech, and Signal Processing}. YEAR = {2025}}
We propose a data-driven method for direction-of-arrival (DOA) trajectory estimation. We use a deep complex architecture which leverages complex-valued representations to capture both magnitude and phase information in the received sensor array data. The network is designed to output the DOA trajectory parameters and amplitudes of the strongest source. Deviating from conventional methods, which attempt to estimate parameters for all sources simultaneously -- leading to assignment ambiguity and the problem of uncertain output dimensions, we adopt a sequential approach. The estimated source signal contribution is subtracted from the input to obtain a residual signal. This residual signal is then fed back into the network to identify the next strongest source and so on, making the proposed network reusable. We evaluate our network on simulated data of varying complexity. Results demonstrate the feasibility of such a reusable network and potential improvements can be explored in future.
Sparse Attention-based Imputation Network for
Time series
Aman Atman,Santosh Nannuru
Pacific Asia Conference on Knowledge Discovery and Data Mining Workshops, PAKDD-W, 2025
@inproceedings{bib_Spar_2025, AUTHOR = {Atman, Aman and Nannuru, Santosh }, TITLE = {Sparse Attention-based Imputation Network for
Time series}, BOOKTITLE = {Pacific Asia Conference on Knowledge Discovery and Data Mining Workshops}. YEAR = {2025}}
Time series data related to traffic and air quality are useful
indicators for urban planning but they frequently have missing values.
Imputation of multi-sensor time series data is thus a vital pre-processing
step for forecasting, anomaly detection and other downstream tasks. We
develop an efficient architecture – Sparse Attention-based Imputation
Network for Time series (SAINT) – which outperforms the state-of-
the-art imputation networks. Efficiency is achieved by separating the
computations on the space-time product graph sequentially into channel
independent temporal attention and sparse space-time transformer. This
channel independent network can reduce overfitting to effectively repre-
sent general temporal patterns. Sparse space-time transformer performs
message passing on the spatial graph conditioned on time. We consider
real-world datasets for evaluation – PEMS-BAY, METR-LA and AQI –
which are gathered from sensor networks in major cities. We demonstrate
the effectiveness and robustness of SAINT across complex missing data
scenarios. Additionally, SAINT generalizes well to short-term forecasting
and is practical for long-term forecasting with limited resources.
Network inference from heat diffusion dynamics
Santosh Nannuru,Ashuthosh Bharadwaj
European Signal Processing Conference, EUSIPCO, 2024
@inproceedings{bib_Netw_2024, AUTHOR = {Nannuru, Santosh and Bharadwaj, Ashuthosh }, TITLE = {Network inference from heat diffusion dynamics}, BOOKTITLE = {European Signal Processing Conference}. YEAR = {2024}}
We address the problem of network connectivity inference by observing interacting time-series signals at the nodes. The evolving nodal signals depend on the connectivity structure of the network thus allowing us to successfully invert for it. Our approach is demonstrated using the example of heat diffusion dynamics. To learn the graph, we combine our knowledge of the dynamical equation and the observed graph signals. Specifically, we perform approximate numerical discretization of the dynamical equation leading to a system of linear equations in the unknown graph topology. Simulations indicate feasibility of this method for accurate network inference.
Impact of diffusion on synchronization pattern of epidemics in non-identical metapopulation networks.
Anika Roy,Ujjwal Shekhar,Aditi Bose,Subrata Ghosh,Santosh Nannuru,Syamal Kumar Dana,Chittaranjan Hens
@inproceedings{bib_Impa_2024, AUTHOR = {Roy, Anika and Shekhar, Ujjwal and Bose, Aditi and Ghosh, Subrata and Nannuru, Santosh and Dana, Syamal Kumar and Hens, Chittaranjan }, TITLE = {Impact of diffusion on synchronization pattern of epidemics in non-identical metapopulation networks.}, BOOKTITLE = {Chaos}. YEAR = {2024}}
In a prior study, a novel deterministic compartmental model known as the SEIHRK model was introduced, shedding light on the pivotal role of test kits as an intervention strategy for mitigating epidemics. Particularly in heterogeneous networks, it was empirically demonstrated that strategically distributing a limited number of test kits among nodes with higher degrees substantially diminishes the outbreak size. The network's dynamics were explored under varying values of infection rate. In this research, we expand upon these findings to investigate the influence of migration on infection dynamics within distinct communities of the network. Notably, we observe that nodes equipped with test kits and those without tend to segregate into two separate clusters when coupling strength is low, but beyond a critical threshold coupling coefficient, they coalesce into a unified cluster. Building on this clustering phenomenon, we develop a reduced equation model and rigorously validate its accuracy through comprehensive simulations. We show that this property is observed in both complete and random graphs.
Graph learning methods to extract empathy supporting regions in a naturalistic stimuli fMRI
Sasanka Grs,Ayushi Kumari Agrawal,Santosh Nannuru,Kavita Vemuri
Technical Report, arXiv, 2024
@inproceedings{bib_Grap_2024, AUTHOR = {Grs, Sasanka and Agrawal, Ayushi Kumari and Nannuru, Santosh and Vemuri, Kavita }, TITLE = {Graph learning methods to extract empathy supporting regions in a naturalistic stimuli fMRI}, BOOKTITLE = {Technical Report}. YEAR = {2024}}
Functional MRI research using naturalistic stimuli (like movies) can help examine brain networks supporting empathy. Applying graph learning methods to whole-brain time-series signals, we propose a novel fMRI signal processing pipeline that includes high-pass filtering, voxel-level clustering, and windowed graph learning with a sparsity-based approach. The fMRI dataset is from passive viewing of an 8-minute movie shown to 14 healthy volunteers. The emotion rating of the movies by age-matched 40 participants is the ground truth. We considered a total of 54 regions extracted from the AAL Atlas for the study. The sparsity-based graph learning method consistently outperforms in capturing variations in the emotion scale, achieving over 88% match of the graph cluster labels averaged across participants. Temporal analysis reveals a high fit to the emotion scale supported by the method’s effectiveness in capturing dynamic connectomes through graph clustering, indicating the role of contextual inference in building empathy. While the edge-weight dynamics analysis underscores the superiority of sparsity-based learning, the connectome-network analysis highlights the crucial role of the Insula, Amygdala, and Thalamus in empathy, with lateral brain connections facilitating synchronized responses. Spectral filtering analysis by band-pass filter isolated the regions linked to emotional and empathy processing in scenes with high emotional context. We also observe strong similarities between the movies in graph cluster labels, connectome-network analysis, and spectral filtering-based analyses, revealing robust neural correlates of empathy. These findings contribute to a better understanding of the neural dynamics associated with empathy and identify regions specific to empathetic responses to naturalistic stimuli.
Denoising and completion of Euclidean distance matrix from multiple observations
Sai Sumanth Natva,Santosh Nannuru
National Conference on Communications, NCC, 2024
@inproceedings{bib_Deno_2024, AUTHOR = {Natva, Sai Sumanth and Nannuru, Santosh }, TITLE = {Denoising and completion of Euclidean distance matrix from multiple observations}, BOOKTITLE = {National Conference on Communications}. YEAR = {2024}}
For a point set, Euclidean distance matrix (EDM) is the matrix consisting of squared distances between every pair of points in the set. It frequently appears in wide ranging applications such as sensor networks, acoustic arrays, crystallography, and self localization among others. Often the measured EDM are inaccurate (due to noise) and incomplete (due to limited availability of measurements) requiring denoising and completion algorithms, frequently both. In literature, methods based on rank alternation and semi-definite relaxation (SDR) have been successful but are restricted to processing a single noisy and/or incomplete input EDM. In this paper we propose extensions of these two algorithms which can process multiple noisy and incomplete EDM simultaneously. Simulations show that the proposed joint algorithms (which process multiple EDM) outperform the corresponding individual algorithms (which process individual EDM and combine the results). We focus on the challenging case of multiplicative noise.
Grid-free algorithms for direction-of-arrival trajectory localization
Ruchi Pandey,Santosh Nannuru
The Journal of the Acoustical Society of America, JASA, 2024
Abs | | bib Tex
@inproceedings{bib_Grid_2024, AUTHOR = {Pandey, Ruchi and Nannuru, Santosh }, TITLE = {Grid-free algorithms for direction-of-arrival trajectory localization}, BOOKTITLE = {The Journal of the Acoustical Society of America}. YEAR = {2024}}
Direction-of-arrival (DOA) estimation algorithms are crucial in localizing acoustic sources. Traditional localization methods rely on block-level processing to extract the directional information from multiple measurements processed together. However, these methods assume that DOA remains constant throughout the block, which may not be true in practical scenarios. Also, the performance of localization methods is limited when the true parameters do not lie on the parameter search grid. In this paper, two trajectory models are proposed, namely the polynomial and harmonic trajectory models, to capture the DOA dynamics. To estimate trajectory parameters, two gridless algorithms are adopted: (i) Sliding Frank–Wolfe (SFW), which solves the Beurling LASSO problem, and (ii) Newtonized orthogonal matching pursuit (NOMP), which is improved over orthogonal matching pursuit (OMP) using cyclic refinement. Furthermore, our analysis is extended to include multi-frequency processing. The proposed models and algorithms are validated using both simulated and real-world data. The results indicate that the proposed trajectory localization algorithms exhibit improved performance compared to grid-based methods in terms of resolution, robustness to noise, and computational efficiency.
Improving audio event localization accuracy via derivative prediction
Ruchi Pandey,Shreyas Jaiswal,Huy Phan,Santosh Nannuru
European Signal Processing Conference, EUSIPCO, 2023
@inproceedings{bib_Impr_2023, AUTHOR = {Pandey, Ruchi and Jaiswal, Shreyas and Phan, Huy and Nannuru, Santosh }, TITLE = {Improving audio event localization accuracy via derivative prediction}, BOOKTITLE = {European Signal Processing Conference}. YEAR = {2023}}
Accurate localization of sound sources is essential in many acoustic sensing and monitoring applications. In the absence of temporal continuity models, many methods produce unrealistic direction of arrival (DOA) estimates involving sudden changes. To address this, we propose an approach that trains a neural network to predict DOA derivatives in Cartesian coordi- nates (x′, y′, z′), which capture the rate of change in DOA (x, y, z) over time. By combining the predicted DOAs with the predicted derivatives, our method can suppress sudden DOA changes and generate smooth motion trajectories. We introduce an update rule that combines the predicted DOAs with the predicted derivatives to obtain the final DOAs. We validate our approach using the TAU-NIGENS Spatial Sound Events (TNSSE) 2021 dataset. Our results demonstrate that incorporating DOA derivatives improves the accuracy of DOA estimation, particularly in low signal-to- noise ratio scenarios
Localization of DOA trajectories--Beyond the grid
Ruchi Pandey,Santosh Nannuru
Technical Report, arXiv, 2023
@inproceedings{bib_Loca_2023, AUTHOR = {Pandey, Ruchi and Nannuru, Santosh }, TITLE = {Localization of DOA trajectories--Beyond the grid}, BOOKTITLE = {Technical Report}. YEAR = {2023}}
The direction of arrival (DOA) estimation algorithms are crucial in localizing acoustic sources. Traditional localization methods rely on block-level processing to extract the directional information from multiple measurements processed together. However, these methods assume that DOA remains constant throughout the block, which may not be true in practical scenarios. Also, the performance of localization methods is limited when the true parameters do not lie on the parameter search grid. In this paper we propose two trajectory models, namely the polynomial and bandlimited trajectory models, to capture the DOA dynamics. To estimate trajectory parameters, we adopt two gridless algorithms: i) Sliding Frank-Wolfe (SFW), which solves the Beurling LASSO problem and ii) Newtonized Orthogonal Matching Pursuit (NOMP), which improves over OMP using cyclic refinement. Furthermore, we extend our analysis to include wideband processing. The simulation results indicate that the proposed trajectory localization algorithms exhibit improved performance compared to grid-based methods in terms of resolution, robustness to noise, and computational efficiency.
Deep Architecture for DOA Trajectory Localization
Shreyas Jaiswal,Ruchi Pandey,Santosh Nannuru
International Conference on Acoustics, Speech, and Signal Processing, ICASSP, 2023
Abs | | bib Tex
@inproceedings{bib_Deep_2023, AUTHOR = {Jaiswal, Shreyas and Pandey, Ruchi and Nannuru, Santosh }, TITLE = {Deep Architecture for DOA Trajectory Localization}, BOOKTITLE = {International Conference on Acoustics, Speech, and Signal Processing}. YEAR = {2023}}
We propose a data-based joint localization and tracking task called trajectory localization with source trajectories identified for a block (multiple measurements) of array data. This is in contrast to localization tasks where directions of arrival (DOA) are estimated per measurement. We employ parametric motion models with focus on linear trajectories. Deep learning based U-Net architecture is proposed to estimate the linear trajectory parameters. The results show that the proposed method gives better and fast trajectory estimates as compared to the trajectory localization (TL) methods of conventional beamforming (TL-CBF) and sparse Bayesian learning (TL-SBL).
PARTICLE FLOW GAUSSIAN SUM PARTICLE FILTER
Comandur Rajasekhar Karthik,Yunpeng Li,Santosh Nannuru
International Conference on Acoustics, Speech, and Signal Processing, ICASSP, 2023
@inproceedings{bib_PART_2023, AUTHOR = {Karthik, Comandur Rajasekhar and Li, Yunpeng and Nannuru, Santosh }, TITLE = {PARTICLE FLOW GAUSSIAN SUM PARTICLE FILTER}, BOOKTITLE = {International Conference on Acoustics, Speech, and Signal Processing}. YEAR = {2023}}
The particle flow Gaussian particle filter (PFGPF) uses an invertible particle flow to generate a proposal density. It approximates the predictive and posterior distributions as Gaussian densities. In this paper, we use a bank of PFGPF filters to construct a Particle flow Gaussian sum particle filter (PFGSPF), which approximates the prediction and posterior as Gaussian mixture model. This approximation is useful in complex estimation problems where a single Gaussian approximation is inadequate. We compare the performance of this proposed filter with the PFGPF and others in challenging numerical simulations. Index Terms— Gaussian mixture, invertible particle flow, Gaussian sum particle filer, Gaussian particle filter, PFGPF.
Identification of Local Neighbourhoods in a Network for Graph-based Signal Reconstruction
Loay Rashid,Santosh Nannuru
World Forum on Internet of Things, WF-IoT, 2022
@inproceedings{bib_Iden_2022, AUTHOR = {Rashid, Loay and Nannuru, Santosh }, TITLE = {Identification of Local Neighbourhoods in a Network for Graph-based Signal Reconstruction}, BOOKTITLE = {World Forum on Internet of Things}. YEAR = {2022}}
The Internet of Things (IoT) relies heavily on exchange of data collected by nodes in a sensor network. This network data can be represented as a signal on a graph allowing tools from graph signal processing (GSP) to analyse it. Here, we utilize graph signal reconstruction algorithms to extrapolate missing sensor data from partially sampled network data. An important class of graph signal reconstruction algorithms operate by partitioning the graph vertices into local sets to achieve lower reconstruction errors and faster convergence. However, methods for identification of these local neighbourhoods have not been clearly studied in the literature. Here, we propose two flexible algorithms to generate local sets on a graph. These algorithms are based on the distance between the sampled and unsampled vertices. Our algorithms are competitive with the methods in literature while offering flexibility in the number of sampled vertices. We carry out simulation-based analysis of sampling strategies and proposed local set generation algorithms using local-set-based reconstruction algorithms.
Particle Flow Gaussian Particle Filter
Comandur Rajasekhar Karthik,Yunpeng Li,Santosh Nannuru
International Conference on Information Fusion, FUSION, 2022
@inproceedings{bib_Part_2022, AUTHOR = {Karthik, Comandur Rajasekhar and Li, Yunpeng and Nannuru, Santosh }, TITLE = {Particle Flow Gaussian Particle Filter}, BOOKTITLE = {International Conference on Information Fusion}. YEAR = {2022}}
State estimation in non-linear models is performed by tracking the posterior distribution recursively. A plethora of algorithms have been proposed for this task. Among them, the Gaussian particle filter uses a weighted set of particles to construct a Gaussian approximation to the posterior. In this paper, we propose to use invertible particle flow methods, derived under the Gaussian boundary conditions for a flow equation, to generate a proposal distribution close to the posterior. The resultant particle flow Gaussian particle filter (PFGPF) algorithm retains the asymptotic properties of Gaussian particle filters, with the potential for improved state estimation performance in highdimensional spaces. We compare the performance of PFGPF with the particle flow filters and particle flow particle filters in two challenging numerical simulation examples. Index Terms—Particle filters, particle flow filters, Gaussian particle filters, particle flow particle filters.
Experimental Validation of Wideband SBL Models for DOA Estimation
Ruchi Pandey,Santosh Nannuru,Peter Gerstoft
European Signal Processing Conference, EUSIPCO, 2022
@inproceedings{bib_Expe_2022, AUTHOR = {Pandey, Ruchi and Nannuru, Santosh and Gerstoft, Peter }, TITLE = {Experimental Validation of Wideband SBL Models for DOA Estimation}, BOOKTITLE = {European Signal Processing Conference}. YEAR = {2022}}
Sparse Bayesian learning (SBL) has been successful in direction of arrival (DOA) estimation due to its robustness and high resolution using a few snapshots. Most wideband SBL algorithms make the simplifying assumption that distinct sources have the same power spectrum across frequency bands. However, this assumption may not be true in practice (for example speech signals). We analyze three wideband signal models and compare variants of wideband SBL (called SBL1, SBL2, and SBL3) with different assumptions on source signal power spectrum. The localization performance of SBL algorithms is compared with wideband processing of conventional beamforming (CBF) and multiple signal classification (MUSIC). The experimental validation is presented using simulated data and experimental LOCATA data. This comparative study shows that SBL3 which simultaneously enforces sparsity and models frequency-dependent signal spectrum shows superior performance in most scenarios. Index Terms—Compressive sensing, Sparse Bayesian learning, DOA estimation, MUSIC, CBF.
Improving trajectory localization accuracy via direction-of-arrival estimation
Ruchi Pandey,Shreyas Jaiswal,Huy Phan,Santosh Nannuru
Technical Report, arXiv, 2022
@inproceedings{bib_Impr_2022, AUTHOR = {Pandey, Ruchi and Jaiswal, Shreyas and Phan, Huy and Nannuru, Santosh }, TITLE = {Improving trajectory localization accuracy via direction-of-arrival estimation}, BOOKTITLE = {Technical Report}. YEAR = {2022}}
Sound source localization is crucial in acoustic sensing and monitoring-related applications. In this paper, we do a comprehensive analysis of improvement in sound source localization by combining the direction of arrivals (DOAs) with their derivatives which quantify the changes in the positions of sources over time. This study uses the SALSA-Lite feature with a convolutional recurrent neural network (CRNN) model for predicting DOAs and their first-order derivatives. An update rule is introduced to combine the predicted DOAs with the estimated derivatives to obtain the final DOAs. The experimental validation is done using TAU-NIGENS Spatial Sound Events (TNSSE) 2021 dataset. We compare the performance of the networks predicting DOAs with derivative vs. the one predicting only the DOAs at low SNR levels. The results show that combining the derivatives with the DOAs improves the localization accuracy of moving sources. Index Terms— Deep learning, Microphone array, SALSALite, Sound event localization and detection (SELD).
Parametric Models for Doa Trajectory Localization
Ruchi Pandey,Santosh Nannuru
International Conference on Acoustics, Speech, and Signal Processing, ICASSP, 2022
@inproceedings{bib_Para_2022, AUTHOR = {Pandey, Ruchi and Nannuru, Santosh }, TITLE = {Parametric Models for Doa Trajectory Localization}, BOOKTITLE = {International Conference on Acoustics, Speech, and Signal Processing}. YEAR = {2022}}
Directions of arrival (DOA) estimation or localization of sources is an important problem in many applications for which numerous algorithms have been proposed. Most lo- calization methods use block-level processing that combines multiple data snapshots to estimate DOA within a block. The DOAs are assumed to be constant within the block duration. However, these assumptions are often violated due to source motion. In this paper, we propose a signal model that cap- tures the linear variations in DOA within a block. We applied conventional beamforming (CBF) algorithm to this model to estimate linear DOA trajectories. Further, we formulate the proposed signal model as a block sparse model and sub- sequently derive sparse Bayesian learning (SBL) algorithm. Our simulation results show that this linear parametric DOA model and corresponding algorithms capture the DOA tra- jectories for moving sources more accurately than traditional signal models and methods.
Sparse planar arrays for azimuth and elevation using experimental data
Santosh Nannuru, Peter Gerstoft,Guoli Ping,Efren Fernandez-Grande
The Journal of the Acoustical Society of America, JASA, 2021
Abs | | bib Tex
@inproceedings{bib_Spar_2021, AUTHOR = {Nannuru, Santosh and Gerstoft, Peter and Ping, Guoli and Fernandez-Grande, Efren }, TITLE = {Sparse planar arrays for azimuth and elevation using experimental data}, BOOKTITLE = {The Journal of the Acoustical Society of America}. YEAR = {2021}}
Sparse arrays are special geometrical arrangements of sensors which overcome some of the drawbacks associated with dense uniform arrays and require fewer sensors. For direction finding applications, sparse arrays with the same number of sensors can resolve more sources while providing higher resolution than a dense uniform array. This has been verified numerically and with real data for one-dimensional microphone arrays. In this study the use of nested and co-prime arrays is examined with sparse Bayesian learning (SBL), which is a compressive sensing algorithm, for estimating sparse vectors and support. SBL is an iterative parameter estimation method and can process multiple snapshots as well as multiple frequency data within its Bayesian framework. A multi-frequency variant of SBL is proposed, which accounts for non-flat frequency spectra of the sources. Experimental validation of azimuth and elevation [two-dimensional (2D)] direction-of-arrival (DOA)estimation are provided using sparse arrays and real data acquired in an anechoic chamber with a rectangular array. Both co-prime and nested arrays are obtained by sampling this rectangular array. The SBL method is compared with conventional beamforming and multiple signal classification for 2D DOA estimation of experimental data.
SPARSE BAYESIAN LEARNING FOR ACOUSTIC SOURCE LOCALIZATION
Ruchi Pandey,Santosh Nannuru,Aditya Siripuram
International Conference on Acoustics, Speech, and Signal Processing, ICASSP, 2021
@inproceedings{bib_SPAR_2021, AUTHOR = {Pandey, Ruchi and Nannuru, Santosh and Siripuram, Aditya }, TITLE = {SPARSE BAYESIAN LEARNING FOR ACOUSTIC SOURCE LOCALIZATION}, BOOKTITLE = {International Conference on Acoustics, Speech, and Signal Processing}. YEAR = {2021}}
The localization of acoustic sources is a parameter estimation problem where the parameters of interest are the direction of arrivals (DOAs). The DOA estimation problem can be formulated as a sparse parameter estimation problem and solved using compressive sensing (CS) methods. In this paper, the CS method of sparse Bayesian learning (SBL) is used to find the DOAs. We specifically use multi-frequency SBL leading to a non-convex optimization problem, which is solved using fixed-point iterations. We evaluate SBL along with traditional DOA estimation methods of conventional beamforming (CBF) and multiple signal classification (MUSIC) on various source localization tasks from the open access LOCATA dataset. The comparative study shows that SBL significantly outperforms CBF and MUSIC on all the considered tasks.
SPARSE BAYESIAN LEARNING FOR ACOUSTIC SOURCE LOCALIZATION
Ruchi Pandey,Santosh Nannuru,Aditya Siripuram
International Conference on Acoustics, Speech, and Signal Processing, ICASSP, 2021
@inproceedings{bib_SPAR_2021, AUTHOR = {Pandey, Ruchi and Nannuru, Santosh and Siripuram, Aditya }, TITLE = {SPARSE BAYESIAN LEARNING FOR ACOUSTIC SOURCE LOCALIZATION}, BOOKTITLE = {International Conference on Acoustics, Speech, and Signal Processing}. YEAR = {2021}}
The localization of acoustic sources is a parameter estimation problem where the parameters of interest are the direction of arrivals (DOAs). The DOA estimation problem can be formulated as a sparse parameter estimation problem and solved using compressive sensing (CS) methods. In this paper, the CS method of sparse Bayesian learning (SBL) is used to find the DOAs. We specifically use multi-frequency SBL leading to a non-convex optimization problem, which is solved using fixed-point iterations. We evaluate SBL along with traditional DOA estimation methods of conventional beamforming (CBF) and multiple signal classification (MUSIC) on various source localization tasks from the open access LOCATA dataset. The comparative study shows that SBL significantly outperforms CBF and MUSIC on all the considered tasks.
ENERGY-AWARE NOISE REDUCTION FOR WIRELESS ACOUSTIC SENSOR NETWORKS
Jie ZHANG,Prof. dr. ir. R. Heusdens,Dr. ir. R. C. Hendrik,Prof. dr. J. Jensen,Ing. T. Gerkmann,ir. A. Bertrand,. ir. G.J.T. Leus, A. Hanjalic,Santosh Nannuru
Frontiers in Neurology, FIN, 2020
@inproceedings{bib_ENER_2020, AUTHOR = {ZHANG, Jie and Heusdens, Prof. Dr. Ir. R. and Hendrik, Dr. Ir. R. C. and Jensen, Prof. Dr. J. and Gerkmann, Ing. T. and Bertrand, ir. A. and Leus, . Ir. G.J.T. and Hanjalic, A. and Nannuru, Santosh }, TITLE = {ENERGY-AWARE NOISE REDUCTION FOR WIRELESS ACOUSTIC SENSOR NETWORKS}, BOOKTITLE = {Frontiers in Neurology}. YEAR = {2020}}
This paper develops efficient channel estimation techniques for millimeter-wave (mmWave) massive multiple-input multiple-output (MIMO) systems under practical hardware limitations, including an arbitrary array geometry and a hybrid hardware structure. Taking on an angle-based approach, this work adopts a generalized array manifold separation approach via Jacobi-Anger approximation, which transforms a non-ideal, non-uniform array manifold into a virtual array domain with a desired uniform geometric structure to facilitate super-resolution angle estimation and channel acquisition. Accordingly, structure-based optimization techniques are developed to effectively estimate both the channel covariance and the instantaneous channel state information (CSI) within a short sensing time. The different time-varying scales of channel path angles versus path gains are capitalized to design a two-step CSI estimation scheme that can quickly sense fading channels. Theoretical results are provided on the fundamental limits of the proposed technique in terms of sample efficiency. For computational efficiency, a fast iterative algorithm is developed via the alternating direction method of multipliers. Other related issues such as spurious-peak cancellation in non-uniform linear arrays and extensions to higher-dimensional cases are also discussed. Simulations testify the effectiveness of the proposed approaches in hybrid mmWave massive MIMO systems with arbitrary arrays.
2D beamforming on sparse arrays with sparse Bayesian learning
Santosh Nannuru, Peter Gerstof
International Conference on Acoustics, Speech, and Signal Processing, ICASSP, 2019
@inproceedings{bib_2D_b_2019, AUTHOR = {Nannuru, Santosh and Gerstof, Peter }, TITLE = {2D beamforming on sparse arrays with sparse Bayesian learning}, BOOKTITLE = {International Conference on Acoustics, Speech, and Signal Processing}. YEAR = {2019}}
Sparse arrays such as co-prime and nested arrays can identify more sources than the number of sensors. This is because their difference co-arrays contain a uniformly spaced virtual array with more elements than the number of sensors in the array. In this paper we demonstrate this using two dimensional co-prime and nested sparse arrays combined with sparse Bayesian learning (SBL) for 2D beamforming in azimuth and elevation. SBL can directly process the sparse array data and significantly outperform conventional beam-forming and MUSIC as seen from simulations.
Sparse Bayesian learning with multiple dictionaries
Santosh Nannuru,Kay L. Gemba,Peter Gerstof,William S. Hodgkiss, Christoph Mecklenbrauker
Signal Processing, SP, 2019
@inproceedings{bib_Spar_2019, AUTHOR = {Nannuru, Santosh and Gemba, Kay L. and Gerstof, Peter and Hodgkiss, William S. and Mecklenbrauker, Christoph }, TITLE = {Sparse Bayesian learning with multiple dictionaries}, BOOKTITLE = {Signal Processing}. YEAR = {2019}}
Sparse Bayesian learning (SBL) has emerged as a fast and competitive method to perform sparse processing. The SBL algorithm, which is developed using a Bayesian framework, iteratively solves a non-convex optimization problem using fixed point updates. It provides comparable performance and is significantly faster than convex optimization techniques used in sparse processing. We propose a multi-dictionary SBL algorithm that simultaneously can process observations generated by different underlying dictionaries sharing the same sparsity profile. Two algorithms are proposed and corresponding fixed point update equations are derived. Noise variances are estimated using stochastic maximum likelihood. The multi-dictionary SBL has many practical applications. We demonstrate this using direction-of-arrival (DOA) estimation. The first example uses the proposed multi-dictionary SBL to process multi …
DOA Estimation in heteroscedastic noise
Peter Gerstoft,Santosh Nannuru,Christoph F. Mecklenbräuker,Geert Leus
Signal Processing, SP, 2019
@inproceedings{bib_DOA__2019, AUTHOR = {Gerstoft, Peter and Nannuru, Santosh and Mecklenbräuker, Christoph F. and , Geert Leus }, TITLE = {DOA Estimation in heteroscedastic noise}, BOOKTITLE = {Signal Processing}. YEAR = {2019}}
The paper considers direction of arrival (DOA) estimation from long-term observations in a very noisy environment. The concern is to derive methods obtaining reasonable DOAs at very low SNR. The noise is assumed zero-mean Gaussian and its variance varies in time and space, causing stationary data models to fit poorly over long observation times. Therefore a heteroscedastic Gaussian noise model is introduced where the variance varies across observations and sensors. The source amplitudes are assumed independent zero-mean complex Gaussian distributed with unknown variances (i.e. the source powers), inspiring stochastic maximum likelihood (ML) DOA estimation. The DOAs of plane waves are estimated from multi-snapshot sensor array data using sparse Bayesian learning (SBL) where the noise is estimated across both sensors and snapshots. This SBL approach is more flexible and performs better …
Multi-frequency sparse Bayesian learning with noise models
Kay L. Gemba,Santosh Nannuru,Peter Gerstoft
The Journal of the Acoustical Society of America, JASA, 2019
@inproceedings{bib_Mult_2019, AUTHOR = {Gemba, Kay L. and Nannuru, Santosh and Gerstoft, Peter }, TITLE = {Multi-frequency sparse Bayesian learning with noise models}, BOOKTITLE = {The Journal of the Acoustical Society of America}. YEAR = {2019}}
Page 1. Multi-frequency sparse Bayesian learning with noise models Kay L. Gemba1, Santosh Nannuru, and Peter Gerstoft ASA SD FALL 2019 Marine Physical Laboratory of the Scripps Institution of Oceanography University of California at San Diego 1gemba@ucsd.edu 1 Page 2. Presentation objectives We investigate SBL performance and present results for the MFP and beamforming applications using 3 noise models: 1. SBL behaves similarly to an adaptive processor and displays robustness to modest model-data mismatch. SBL performance is compared to the white noise gain constraint (WNGC), MUSIC, and Bartlett processors. 2. Compare SBL performance, implemented with 3 noise models: 1. SBL1: stationary noise, 2. SBL2: non-stationary noise, which is useful when the noise variance evolves across snapshots, 3. SBL3: non-stationary noise in a spatially, non-homogeneous field (example: surface noise) …
DOA ESTIMATION IN HETEROSCEDASTIC NOISE WITH SPARSE BAYESIAN LEARNING
Peter Gerstof,Santosh Nannuru,Christoph F. Mecklenbrauker,Geert Leus
International Conference on Acoustics, Speech, and Signal Processing, ICASSP, 2018
@inproceedings{bib_DOA__2018, AUTHOR = {Gerstof, Peter and Nannuru, Santosh and Mecklenbrauker, Christoph F. and Leus, Geert }, TITLE = {DOA ESTIMATION IN HETEROSCEDASTIC NOISE WITH SPARSE BAYESIAN LEARNING}, BOOKTITLE = {International Conference on Acoustics, Speech, and Signal Processing}. YEAR = {2018}}
The paper considers direction of arrival (DOA) estimation from long-term observations in a noisy environment. In such an environment the noise source might evolve, causing the stationary models to fail. Therefore a heteroscedastic Gaussian noise model is introduced where the variance can vary across observations and sensors. The source amplitudes are assumed independent zero-mean complex Gaussian distributed with unknown variances (i.e. the source powers), leading to stochastic maximum likelihood (ML) DOA estimation. The DOAs of plane waves are estimated from multi-snapshot sensor array data using sparse Bayesian learning (SBL) where the noise is estimated across both sensors and snapshots. Simulations demonstrate that taking the heteroscedastic noise into account improves DOA estimation.
Sparse Bayesian learning for DOA estimation using co-prime and nested arrays
Santosh Nannuru,Peter Gerstof,Ali Koochakzadeh,Piya Pal
Sensor Array and Multichannel Signal Processing Workshop, SAM, 2018
@inproceedings{bib_Spar_2018, AUTHOR = {Nannuru, Santosh and Gerstof, Peter and Koochakzadeh, Ali and Pal, Piya }, TITLE = {Sparse Bayesian learning for DOA estimation using co-prime and nested arrays}, BOOKTITLE = {Sensor Array and Multichannel Signal Processing Workshop}. YEAR = {2018}}
Sparse Bayesian learning (SBL) has been used to obtain source direction-of-arrivals (DoAs) from uniform linear array (ULA) data. The maximum number of sources that can be resolved using a ULA is limited by the number of sensors in the array. It is known that sparse linear arrays such as co-prime and nested arrays can resolve more sources than the number of sensors. In this paper we demonstrate this using SBL. We compute the mean squared error in source power estimation as various parameters are varied.
Multi-frequency sparse Bayesian learning for matched field processing in non-stationary noise
Kay L Gemba,Santosh Nannuru,Peter Gerstoft
The Journal of the Acoustical Society of America, JASA, 2018
@inproceedings{bib_Mult_2018, AUTHOR = {Gemba, Kay L and Nannuru, Santosh and Gerstoft, Peter }, TITLE = {Multi-frequency sparse Bayesian learning for matched field processing in non-stationary noise}, BOOKTITLE = {The Journal of the Acoustical Society of America}. YEAR = {2018}}
Using simulations and data, we localize a quiet source in the presence of an interferer. The SWellEx-96 Event S59 consists of a submerged source towed along an isobath over a 65 min duration with an interferer traversing the source track. This range independent, multi-frequency scenario includes mismatch, non-stationary noise, and operational uncertainty. Mismatch is defined as a misalignment between the actual source field observed at the array and the modeled replica vector. The noise process changes likely with time. This is modelled as a heteroscedastic Gaussian process, meaning that the noise variance is non-stationary across snapshots. Sparse Bayesian learning (SBL) has been applied previously to the matched field processing application [Gemba et al, J. Acoust. Soc. Am., 141:3411-3420, 2017]. Results demonstrate that SBL exhibits desirable robustness properties and improved localization …
