AGAnkit Gupta

Ankit Gupta

*This is an algorithmically generated author record based on publicly available information.

Ankit Gupta has a total of 15 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • NeuromorphicRx: From Neural to Spiking Receiver

    by Ankit Gupta, Onur Dizdar, Yun Chen, Fehmi Emre Kadan, Ata Sattarzadeh, Stephen Wang

    In this work, we propose a novel energy-efficient spiking neural network (SNN)-based receiver for 5G-NR OFDM system, called neuromorphic receiver (NeuromorphicRx), replacing the channel estimation, equalization and symbol demapping blocks. We leverage domain knowledge to design the input with spiking encoding and propose a deep convolutional SNN with spike-element-wise residual connections. We integrate an SNN with artificial neural network (ANN) hybrid architecture to obtain soft outputs and employ surrogate gradient descent for training. We focus on generalization across diverse scenarios and robustness through quantized aware training. We focus on interpretability of NeuromorphicRx for 5G-NR signals and perform detailed ablation study for 5G-NR signals. Our extensive numerical simulations show that NeuromorphicRx is capable of achieving significant block error rate performance gain compared to 5G-NR receivers and similar performance compared to its ANN-based counterparts with 7.6× less energy consumption.


  • Frequency Transfer Learning for AI-aided Beam Management in cmWave Spectrum

    by Yun Chen, Onur Dizdar, Ankit Gupta, Stephen Wang

    This paper focuses on the AI-aided beam management (AI-BM) scheme design for FR3 centimeter-Wave (cmWave) spectrum. A spindle-like decoder-encoder neural network is proposed to conduct spatial prediction with reduced measurements. To facilitate the AI-BM deployment for multi-frequency scenarios in 6G, an efficient frequency transfer learning solution is further proposed based on the antenna array theory. Moreover, a fully 3GPP compliant real-time proof-of-concept (PoC) platform is developed to evaluate the proposed scheme. By experimental results, we validate the advanced performance and robustness of the proposed AI-BM scheme.


  • AI-Aided Delay-Doppler Estimation for Bistatic OFDM ISAC Systems

    by Longfei Yin, Julia Equi, Fehmi Emre Kadan, Onur Dizdar, Ankit Gupta, Yun Chen, Gábor Fodor, Stephen Wang

    This letter introduces a novel artificial intelligence (AI)-aided delay-Doppler estimation method for bistatic integrated sensing and communications systems employing orthogonal frequency division multiplexing (OFDM) under dense clutter and inter-cell interference. An attention-based recurrent neural network is designed to extract features from individual OFDM symbols using Long Short-Term Memory layers followed by an attention layer to determine the relative importance, aiming to improve sensing accuracy and robustness. We demonstrate by numerical results that the proposed AI-aided approach can outperform conventional baseline methods and is robust against the degrading effects of heavy clutter and colored noise.


  • AI-Empowered Bi-Static Sensing for Rank-Limited Communications

    by Fehmi Emre Kadan, Ankit Gupta, Onur Dizdar, Ata Sattarzadeh, Yun Chen, Longfei Yin, Stephen Wang

    This letter proposes an AI-empowered localization method for passive objects using rank-limited communication signals in bi-static scenario. Specifically, we consider linearly precoded communication signals that are rank-limited by the total number of independent data streams, resulting in traditional subspace methods failing to find correct angular information. A deep neural network (DNN) based solution is proposed to accurately localize the targets under reflections from clutters and inter-cell interference. The proposed DNN model applies data fusion over multiple user equipments to overcome the limitations of using rank-limited signals. The numerical results show the efficiency of the developed technique over baselines.

  • End-to-End Learning-Based Full-Duplex Amplify-and-Forward Relay Networks

    by Ankit Gupta, Mathini Sellathurai, Tharmalingam Ratnarajah

    Full duplex (FD) relaying can provide double spectral efficiency. Despite advanced self-interference cancellation techniques, residual self-interference (RSI) limits the performance significantly. We present an autoencoder (AE)-based block coded modulation (BCM) and differential BCM (d-BCM) for an FD amplify-and-forward (FD-AF) relay network that can tackle the deteriorating impacts of RSI. Existing works treat AE frameworks as black-box with minimal/no focus on training convergence, limiting AE’s practical deployment. Focussing on training convergence, firstly, we show that training of AE converges above minimum signal-to-noise-ratio (SNR) and below maximum RSI level, and CSI helps in faster convergence. Secondly, we establish a relationship between training hyper-parameters and AE-based BCM/d-BCM design, by showing that, for any given hyper-parameters, training of the AE has converged to its maximum potential of decoding if AEs encoder has designed 2 k codewords, with an emphasis on the minimum required training samples. To open the black-box AE, we reveal five observations in the AE-based designed codewords concerning Euclidean distance, packing density, hamming distance, and kurtosis, that resemble the desired observations of theoretical random coded modulations. By extensive simulations, we show that the proposed AE outperforms the conventional methods considerably for varying SNR, RSI, transmission rate, channel estimation errors, and small/practical block lengths.


  • Ranking Actionable Mental Health Advice through a Personalized Video Recommendation System

    by Ankit Gupta, Nada Basit

    In order to improve the quality of life in today’s digital culture, one of the most important areas of health to consider is mental health. Over the past several years, mental health has become increasingly important as the digital landscape has gotten more complicated, and as more and more people seek ways to relax, connect with others, and find mental balance in their life. While there have been advances in technology, particularly in information retrieval and recommendation systems, very few of these advances have been applied to the mental health industry. In this paper, we bridge this gap by introducing a ranking system that suggests personalized videos for users with the intention of helping them improve their mental health. By updating the relative ranking of videos by marginal improvement for different users, then aligning the distribution of videos depending on the areas that the user needs the most help with, the website helps provide personalized and actionable mental health advice.


  • A Novel Average Autoencoder-Based Amplify-and-Forward Relay Networks With Hardware Impairments

    by Ankit Gupta, Mathini Sellathurai

    In this paper, we propose a novel Average autoencoder (AE)-based amplify-and-forward (AF) relay networks impacted by the I/Q imbalance (IQI) and additional hardware impairments (AHI), where the source and destination nodes are equipped with neural network (NN)-based encoder and decoder, while a conventional AF relay node assists the transmission. The average AE employs multiple small NN-based decoders at the destination node, each decoding a soft probabilistic output that is averaged to obtain the final soft probabilistic output at the destination node. By considering multiple small NN decoders, we reduce the implementation complexity significantly while improving the performance compared to the AE with a single large but NN-based decoder. Within this Average AE framework, we propose a coded modulation design (CMD) with zero-forcing-based IQI compensation that considers the availability of the channel state information (CSI) and IQI knowledge. However, the IQI and CSI need to be estimated separately. Thus, we also propose a CMD with no IQI compensation that requires only the CSI knowledge. Finally, we propose a differential CMD that removes the necessity of both the CSI and IQI knowledge. Under low signal-to-interference-and-noise-ratio regimes, we show that the proposed Average AE outperforms the optimal maximum likelihood detector by considerable margin.


  • Machine Learning-Based Urban Canyon Path Loss Prediction Using 28 GHz Manhattan Measurements

    by Ankit Gupta, Jinfeng Du, Dmitry Chizhik, Reinaldo A. Valenzuela, Mathini Sellathurai

    Large bandwidth at millimeter wave (mm-wave) is crucial for fifth generation (5G) and beyond, but the high path loss (PL) requires highly accurate PL prediction for network planning and optimization. Statistical models with slope-intercept fit fall short in capturing large variations seen in urban canyons, whereas ray tracing, capable of characterizing site-specific features, faces challenges in describing foliage and street clutter and associated reflection/diffraction ray calculation. Machine learning (ML) is promising but faces three key challenges in PL prediction: 1) insufficient measurement data; 2) lack of extrapolation to new streets; 3) overwhelmingly complex features/models. We propose an ML-based urban canyon PL prediction model based on extensive 28 GHz measurements from Manhattan where street clutters are modeled via a light detection and ranging (LiDAR) point cloud dataset and buildings by a mesh-grid building dataset. We extract expert knowledge-driven street clutter features from the point cloud and aggressively compress the 3-D building information using a convolutional autoencoder. Using a new street-by-street training and testing procedure to improve generalizability, the proposed model using both clutter and building features achieves a prediction error [root-mean-square error (RMSE)] of 4.8 ± 1.1 dB compared to 10.6 ± 4.4 and 6.5 ± 2.0 dB for 3GPP line of sight (LOS) and slope-intercept prediction, respectively, where the standard deviation indicates street-by-street variation. By only using four most influential clutter features, the RMSE of 5.5 ± 1.1 dB is achieved.


  • Open Access

    End-to-End Learning-Based Framework for Amplify-and-Forward Relay Networks

    by Ankit Gupta, Mathini Sellathurai

    We study end-to-end learning-based frameworks for amplify-and-forward (AF) relay networks, with and without the channel state information (CSI) knowledge. The designed framework resembles an autoencoder (AE) where all the components of the neural network (NN)-based source and destination nodes are optimized together in an end-to-end manner, and the signal transmission takes place with an AF relay node. Unlike the literature that employs an NN-based relay node with full CSI knowledge, we consider a conventional relay node that only amplifies the received signal using CSI gains. Without the CSI knowledge, we employ power normalization-based amplification that normalizes the transmission power of each block of symbols. We propose and compare symbol-wise and bit-wise AE frameworks by minimizing categorical and binary cross-entropy loss that maximizes the symbol-wise and bit-wise mutual information (MI), respectively. We determine the estimated MI and examine the convergence of both AE frameworks with signal-to-noise ratio (SNR). For both these AE frameworks, we design coded modulation and differential coded modulation, depending upon the availability of CSI at the destination node, that obtains symbols in 2n-dimensions, where n is the block length. To explain the properties of the 2n-dimensional designs, we utilize various metrics like minimum Euclidean distance, normalized second-order and fourth-order moments, and constellation figures of merit. We show that both these AE frameworks obtain similar spherical coded-modulation designs in 2n-dimensions, and bit-wise AE that inherently obtains the optimal bit-labeling outperforms symbol-wise AE (with faster convergence under low SNR) and the conventional AF relay network with a considerable SNR margin.