Satyam Agarwal
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Satyam Agarwal has a total of 28 co-authors. Here are the top co-authors by number of shared publications:
- Brijesh Kumbhani
Department of Electrical Engineering, Indian Institute of Technology Ropar, Rupnagar, Punjab, India
8 shared publications - Nishant Gupta
Department of Communication and Computer Engineering, The LNMIIT Jaipur, Jaipur, India
7 shared publications - Arhum Ahmad
Department of Electrical Engineering, Indian Institute of Technology Ropar, Rupnagar, Punjab, India
5 shared publications - Akanksha Gupta
IIT Ropar
4 shared publications - Deepak Mishra
University of New South Wales (UNSW)
4 shared publications - Sam Darshi
Department of Electrical Engineering, Indian Institute of Technology Ropar, Ropar, India
4 shared publications
Latest Publications
- Open Access
- 12 May 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Open Journal of the Communications Society
- Vol. 7, 5407-5432
A Multi-Stage Detection Framework for Wideband Spectrum Sensing
We present WISDOM, a hardware-aware, blind wideband spectrum-sensing framework designed to bridge the gap between theoretical detection algorithms and the reality of low-cost radio front-ends. Unlike conventional detectors that suffer from the SNR wall, WISDOM operates entirely in the digital domain using a multi-stage statistical pipeline. The framework introduces three core novelties: (i) a Noise-Avoidance Criterion (ηn) derived via Ordered-Statistic Constant False Alarm Rate on a guaranteed noise-only spectral region manufactured by the pre-processing chain; (ii) an inverted “Count-and-Slope” decision rule that identifies signals by detecting structured local spectral deviations rather than absolute power thresholds; and (iii) an analytically scalable Selective Spectral Refinement. Extensive validation on Software Defined Radio testbeds demonstrates sensitivity of -85 dBm across a 50 MHz sensing bandwidth and acceptable performance against hardware impairments.
- 12 May 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Open Journal of the Communications Society
- Vol. 7, 5407-5432
Open AccessA Multi-Stage Detection Framework for Wideband Spectrum Sensing
We present WISDOM, a hardware-aware, blind wideband spectrum-sensing framework designed to bridge the gap between theoretical detection algorithms and the reality of low-cost radio front-ends. Unlike conventional detectors that suffer from the SNR wall, WISDOM operates entirely in the digital domain using a multi-stage statistical pipeline. The framework introduces three core novelties: (i) a Noise-Avoidance Criterion (ηn) derived via Ordered-Statistic Constant False Alarm Rate on a guaranteed noise-only spectral region manufactured by the pre-processing chain; (ii) an inverted “Count-and-Slope” decision rule that identifies signals by detecting structured local spectral deviations rather than absolute power thresholds; and (iii) an analytically scalable Selective Spectral Refinement. Extensive validation on Software Defined Radio testbeds demonstrates sensitivity of -85 dBm across a 50 MHz sensing bandwidth and acceptable performance against hardware impairments.
- 5 May 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal of Radio Frequency Identification
- Vol. 10, 391-400
Transforming RF Reflections Into Respiratory Intelligence Through Radar-ML Synergy
Radar-based contactless monitoring of respiratory rate (RR) plays an important role in a person’s life as it provides a new way of patient observation that does not require physical contact, which medical professionals need for monitoring patients in neonatal intensive care units (NICU) because it safeguards their well-being while maintaining their comfort and stopping the spread of diseases. In this paper, we conduct a comprehensive evaluation of various machine learning (ML) models that use continuous wave (CW) radar data to determine noncontact respiratory rate (RR) measurements. In a previous paper, we showed that the random forest (RF) algorithm can accurately predict RR but is not up to mark for medical purposes. This study systematically investigates different ML algorithms, which include linear regression (LR), support vector regression (SVR), gradient boosting regression (GBR), XGBoost regression (XGBR), and light gradient boosting machine regression (LGBMR) to evaluate their ability to make predictions and their system performance. These algorithms alone do not provide an accurate prediction of RR monitoring, so we introduce a new hybrid algorithm called stacked ensemble regression (SER), which utilizes an SER framework that integrates various base regressors through a meta-learning method to achieve shared benefits, which allow it to capture both vast linear patterns and small nonlinear changes within respiration data. The proposed SER method achieves an accuracy of 92.24%, which surpasses the performance of XGBR, which achieves an accuracy of 89.65%. The proposed framework creates a strong and expandable system that enables monitoring respiratory patterns through both physical and remote methods while developing radar-based respiratory assessment technology for medical use in hospitals and home care environments.
- 5 May 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal of Radio Frequency Identification
- Vol. 10, 391-400
Transforming RF Reflections Into Respiratory Intelligence Through Radar-ML Synergy
Radar-based contactless monitoring of respiratory rate (RR) plays an important role in a person’s life as it provides a new way of patient observation that does not require physical contact, which medical professionals need for monitoring patients in neonatal intensive care units (NICU) because it safeguards their well-being while maintaining their comfort and stopping the spread of diseases. In this paper, we conduct a comprehensive evaluation of various machine learning (ML) models that use continuous wave (CW) radar data to determine noncontact respiratory rate (RR) measurements. In a previous paper, we showed that the random forest (RF) algorithm can accurately predict RR but is not up to mark for medical purposes. This study systematically investigates different ML algorithms, which include linear regression (LR), support vector regression (SVR), gradient boosting regression (GBR), XGBoost regression (XGBR), and light gradient boosting machine regression (LGBMR) to evaluate their ability to make predictions and their system performance. These algorithms alone do not provide an accurate prediction of RR monitoring, so we introduce a new hybrid algorithm called stacked ensemble regression (SER), which utilizes an SER framework that integrates various base regressors through a meta-learning method to achieve shared benefits, which allow it to capture both vast linear patterns and small nonlinear changes within respiration data. The proposed SER method achieves an accuracy of 92.24%, which surpasses the performance of XGBR, which achieves an accuracy of 89.65%. The proposed framework creates a strong and expandable system that enables monitoring respiratory patterns through both physical and remote methods while developing radar-based respiratory assessment technology for medical use in hospitals and home care environments.
- Open Access
- 27 April 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 14, 65064-65089
A Machine Learning Framework for DDoS Attack Detection in SDN-Enabled Mobile Wireless Networks
Attacks on network components and devices pose a significant threat to service continuity, necessitating robust detection mechanisms. This paper presents a Distributed Denial of Service (DDoS) attack detection framework tailored for heterogeneous mobile wireless networks within a Software-Defined Networking architecture. A two-tier model is proposed: localized attack detection at access points (APs) using a Multi-Layer Perceptron (MLP) classifier, and centralized detection under mobility at the controller using a Long Short-Term Memory (LSTM) model. The system incorporates novel traffic features such as flow count, speed of source IP, source and destination IP address entropy, proportion of bidirectional flows, and handover frequency, which together enhance detection in mobile environments. An LSTM model analyzes inter-AP traffic correlation over time to address mobility-driven DDoS attack amplification. The proposed approach is evaluated under diverse traffic types (TCP, UDP, ICMP) and varying attack intensities. The MLP model selected for integration into the framework demonstrates consistently strong detection capability across the evaluated scenarios, achieving accuracy values in the range of 95%–99% and showing improved performance relative to existing state-of-the-art schemes. Furthermore, multi-run statistical validation confirms stable behavior under randomized initialization and mobility-driven conditions, while controller-level correlation analysis enhances robustness against mobility-driven attack propagation.
- 27 April 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 14, 65064-65089
Open AccessA Machine Learning Framework for DDoS Attack Detection in SDN-Enabled Mobile Wireless Networks
Attacks on network components and devices pose a significant threat to service continuity, necessitating robust detection mechanisms. This paper presents a Distributed Denial of Service (DDoS) attack detection framework tailored for heterogeneous mobile wireless networks within a Software-Defined Networking architecture. A two-tier model is proposed: localized attack detection at access points (APs) using a Multi-Layer Perceptron (MLP) classifier, and centralized detection under mobility at the controller using a Long Short-Term Memory (LSTM) model. The system incorporates novel traffic features such as flow count, speed of source IP, source and destination IP address entropy, proportion of bidirectional flows, and handover frequency, which together enhance detection in mobile environments. An LSTM model analyzes inter-AP traffic correlation over time to address mobility-driven DDoS attack amplification. The proposed approach is evaluated under diverse traffic types (TCP, UDP, ICMP) and varying attack intensities. The MLP model selected for integration into the framework demonstrates consistently strong detection capability across the evaluated scenarios, achieving accuracy values in the range of 95%–99% and showing improved performance relative to existing state-of-the-art schemes. Furthermore, multi-run statistical validation confirms stable behavior under randomized initialization and mobility-driven conditions, while controller-level correlation analysis enhances robustness against mobility-driven attack propagation.
- 2 March 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Communications
- Vol. 74 (00906778) , 5867-5878
Start-of-Packet Detection Using Machine Learning
With the rapid proliferation of wireless communication systems, accurate Start-of-Packet (SOP) detection is essential for efficient packet reception and decoding. Conventional correlation-based techniques struggle under low signal-to-noise ratio (SNR) and phase/frequency offsets. This paper proposes a Proximity-based Support Vector Machine (P-SVM), a novel SVM-based approach that incorporates Proximity-based weighting and a Distance adjustment function to enhance robustness under such impairments. Evaluated using both simulations and SDR-based experiments, P-SVM consistently outperforms conventional and state-of-the-art methods in Probability of detection, Mean Absolute Error, and Probability of false alarm. Moreover, P-SVM offers reduced implementation complexity, making it suitable for practical deployment.
- 2 March 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Communications
- Vol. 74 (00906778) , 5867-5878
Start-of-Packet Detection Using Machine Learning
With the rapid proliferation of wireless communication systems, accurate Start-of-Packet (SOP) detection is essential for efficient packet reception and decoding. Conventional correlation-based techniques struggle under low signal-to-noise ratio (SNR) and phase/frequency offsets. This paper proposes a Proximity-based Support Vector Machine (P-SVM), a novel SVM-based approach that incorporates Proximity-based weighting and a Distance adjustment function to enhance robustness under such impairments. Evaluated using both simulations and SDR-based experiments, P-SVM consistently outperforms conventional and state-of-the-art methods in Probability of detection, Mean Absolute Error, and Probability of false alarm. Moreover, P-SVM offers reduced implementation complexity, making it suitable for practical deployment.
- 1 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Communications Letters
- Vol. 30, 2954-2958
Reg-GelNet: Deep Learning Receiver for Asynchronous Distributed Chirp Spread Spectrum
- 1 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Communications Letters
- Vol. 30, 2954-2958
Reg-GelNet: Deep Learning Receiver for Asynchronous Distributed Chirp Spread Spectrum
- 17 November 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Communications Letters
- Vol. 30 (10897798) , 203-207
Practicalities to Communication Receiver for JSC Waveform: Whether to Filter or to Undersample?
In this letter, we consider low-complexity symbol detection schemes for the UAV communication receiver using a phase shift keying-linear frequency modulated (PSK-LFM) joint sensing and communication (JSC) waveform. The JSC waveform typically employs a large bandwidth as it helps in performing radar functionality. However, UAV communication receivers typically operate at a much lower bandwidth as compared to radar. Therefore, this letter presents a communication receiver performance for two schemes: one that filters and samples a portion of the signal, while the other undersamples the entire received signal for bit detection. We analytically derive BER expressions for both schemes under imperfect CSI and assess their performance. The results show that undersampling provides better BER performance than with filtering at the same sampling rate, highlighting the robustness of the proposed receiver design.
- 17 November 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Communications Letters
- Vol. 30 (10897798) , 203-207
Practicalities to Communication Receiver for JSC Waveform: Whether to Filter or to Undersample?
In this letter, we consider low-complexity symbol detection schemes for the UAV communication receiver using a phase shift keying-linear frequency modulated (PSK-LFM) joint sensing and communication (JSC) waveform. The JSC waveform typically employs a large bandwidth as it helps in performing radar functionality. However, UAV communication receivers typically operate at a much lower bandwidth as compared to radar. Therefore, this letter presents a communication receiver performance for two schemes: one that filters and samples a portion of the signal, while the other undersamples the entire received signal for bit detection. We analytically derive BER expressions for both schemes under imperfect CSI and assess their performance. The results show that undersampling provides better BER performance than with filtering at the same sampling rate, highlighting the robustness of the proposed receiver design.
- 30 October 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Wireless Communications Letters
- Vol. 15 (21622337) , 420-424
WiFiML-PD: Machine Learning-Based Packet Detection for IEEE 802.11ax Systems
With rising demand for reliable wireless connectivity in dense environments, accurate packet detection in IEEE 802.11ax (WiFi-6) systems is crucial. Conventional correlation-based methods degrade under low SNR, multipath fading, and RF impairments. This letter proposes WiFiML-PD, a machine learning-based packet detection scheme employing a contiguity-based Support Vector Machine (SVM) trained on Legacy Short Training Field (L-STF) preamble. By addressing class imbalance and leveraging temporal context, the proposed model enhances packet detection performance while requiring lower computational complexity than conventional and state-of-the-art methods. Extensive simulations and real-time hardware validation demonstrate its robustness and practical viability for deployment in next-generation WiFi-6 networks.
- 30 October 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Wireless Communications Letters
- Vol. 15 (21622337) , 420-424
WiFiML-PD: Machine Learning-Based Packet Detection for IEEE 802.11ax Systems
With rising demand for reliable wireless connectivity in dense environments, accurate packet detection in IEEE 802.11ax (WiFi-6) systems is crucial. Conventional correlation-based methods degrade under low SNR, multipath fading, and RF impairments. This letter proposes WiFiML-PD, a machine learning-based packet detection scheme employing a contiguity-based Support Vector Machine (SVM) trained on Legacy Short Training Field (L-STF) preamble. By addressing class imbalance and leveraging temporal context, the proposed model enhances packet detection performance while requiring lower computational complexity than conventional and state-of-the-art methods. Extensive simulations and real-time hardware validation demonstrate its robustness and practical viability for deployment in next-generation WiFi-6 networks.
- 15 October 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Intelligent Vehicles
- Vol. 11 (1) , 236-248
Design and Analysis of Low-Complexity Communication Receiver for PSK–LFM Joint Sensing and Communication Waveform
In this article, we propose novel strategies for preamble detection and synchronization to detect the symbols from the phase shift keying-linear frequency modulated (PSK-LFM) joint sensing and communication waveform at the communication receiver of an uncrewed aerial vehicle (UAV). Since radar waveform requires huge bandwidth (order of 100 MHz–2 GHz) for better range resolution, processing the same waveform at the UAV's communication receiver necessitates high analog-to-digital converter (ADC) sampling rates. To reduce the ADC sampling requirements, in this paper, we propose two signal processing schemes for the communication receiver. In first, a part of the received signal is filtered using a low pass filter (LPF) and sampled for preamble detection, symbol synchronisation, and detecting the symbols. In the other case, the entire signal undergoes undersampling for subsequent processing. Furthermore, we obtain bit error rate (BER) performance for all the cases by considering time, phase, and carrier frequency offsets. We show that processing the undersampled signal yields superior BER performance compared to the filtering approach, even when both operate at an equivalent sampling rate. Furthermore, for all cases, we compare the simulation results with and without offsets, along with the analytical results obtained without offsets.
- 15 October 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Intelligent Vehicles
- Vol. 11 (1) , 236-248
Design and Analysis of Low-Complexity Communication Receiver for PSK–LFM Joint Sensing and Communication Waveform
In this article, we propose novel strategies for preamble detection and synchronization to detect the symbols from the phase shift keying-linear frequency modulated (PSK-LFM) joint sensing and communication waveform at the communication receiver of an uncrewed aerial vehicle (UAV). Since radar waveform requires huge bandwidth (order of 100 MHz–2 GHz) for better range resolution, processing the same waveform at the UAV's communication receiver necessitates high analog-to-digital converter (ADC) sampling rates. To reduce the ADC sampling requirements, in this paper, we propose two signal processing schemes for the communication receiver. In first, a part of the received signal is filtered using a low pass filter (LPF) and sampled for preamble detection, symbol synchronisation, and detecting the symbols. In the other case, the entire signal undergoes undersampling for subsequent processing. Furthermore, we obtain bit error rate (BER) performance for all the cases by considering time, phase, and carrier frequency offsets. We show that processing the undersampled signal yields superior BER performance compared to the filtering approach, even when both operate at an equivalent sampling rate. Furthermore, for all cases, we compare the simulation results with and without offsets, along with the analytical results obtained without offsets.
- 29 September 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Cognitive Communications and Networking
- Vol. 12, 1269-1281
Harnessing Imperfection: Deep Learning-Based Receiver for Non-Coherent Distributed Transmission
This paper investigates the feasibility of distributed multi-transmission in non-coherent systems using deep learning (DL), a concept we term non-coherent distributed transmission (NCDT). In this approach, multiple transmitters non-coherently send the same content to a receiver. We leverage DL to learn and capture the complex patterns arising from signal combinations with frequency and phase offsets. By engineering relevant features and designing a deep neural network with an adaptable auto-tuning mechanism, we enable precise mapping of received signals to transmitted symbols within the NCDT framework. Additionally, we develop a likelihood ratio test (LRT)-based detector as a baseline for comparison. Our simulations and experiments demonstrate that the proposed DL-based receiver achieves lower bit error rates (BER) than the other state-of-the-art DL-based and LRT-based detectors while eliminating the need for synchronization, channel state information (CSI) estimation, or precoding. Experimental validation using software-defined radios corroborates our simulation results, confirming that NCDT can be practically integrated into wireless systems without requiring hardware modifications. Notably, our results show that the proposed method achieves comparable BER performance at nearly half the transmission power, corresponding to an approximate 3 dB gain.
- 29 September 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Cognitive Communications and Networking
- Vol. 12, 1269-1281
Harnessing Imperfection: Deep Learning-Based Receiver for Non-Coherent Distributed Transmission
This paper investigates the feasibility of distributed multi-transmission in non-coherent systems using deep learning (DL), a concept we term non-coherent distributed transmission (NCDT). In this approach, multiple transmitters non-coherently send the same content to a receiver. We leverage DL to learn and capture the complex patterns arising from signal combinations with frequency and phase offsets. By engineering relevant features and designing a deep neural network with an adaptable auto-tuning mechanism, we enable precise mapping of received signals to transmitted symbols within the NCDT framework. Additionally, we develop a likelihood ratio test (LRT)-based detector as a baseline for comparison. Our simulations and experiments demonstrate that the proposed DL-based receiver achieves lower bit error rates (BER) than the other state-of-the-art DL-based and LRT-based detectors while eliminating the need for synchronization, channel state information (CSI) estimation, or precoding. Experimental validation using software-defined radios corroborates our simulation results, confirming that NCDT can be practically integrated into wireless systems without requiring hardware modifications. Notably, our results show that the proposed method achieves comparable BER performance at nearly half the transmission power, corresponding to an approximate 3 dB gain.
- 19 September 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Communications Letters
- Vol. 29 (11) , 2751-2754
Clustering in Asynchronous Distributed Transmission for IoT Applications
Asynchronous distributed transmission (ADT) supports scalable, energy-efficient Internet of Things (IoT) communication but suffers from time, frequency, and phase offsets at the receiver. Bit error rate (BER) performance in ADT depends on the frequency offsets between the transmitting nodes, making it a suitable parameter for system design. Motivated by this, we propose a clustering framework that groups IoT devices into a fixed number of transmitters in ADT with high intra-cluster frequency diversity to enhance BER performance. We model the clustering as a constrained combinatorial optimization problem, which is NP-hard, and solve it by proposing a Frequency-Aware Size-Constrained (FASC) clustering algorithm. FASC achieves near-optimal performance with reduced computational cost and consistently outperforms state-of-the-art schemes in terms of BER and convergence, making it well-suited for real-world ADT-based IoT systems.
- 19 September 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Communications Letters
- Vol. 29 (11) , 2751-2754
Clustering in Asynchronous Distributed Transmission for IoT Applications
Asynchronous distributed transmission (ADT) supports scalable, energy-efficient Internet of Things (IoT) communication but suffers from time, frequency, and phase offsets at the receiver. Bit error rate (BER) performance in ADT depends on the frequency offsets between the transmitting nodes, making it a suitable parameter for system design. Motivated by this, we propose a clustering framework that groups IoT devices into a fixed number of transmitters in ADT with high intra-cluster frequency diversity to enhance BER performance. We model the clustering as a constrained combinatorial optimization problem, which is NP-hard, and solve it by proposing a Frequency-Aware Size-Constrained (FASC) clustering algorithm. FASC achieves near-optimal performance with reduced computational cost and consistently outperforms state-of-the-art schemes in terms of BER and convergence, making it well-suited for real-world ADT-based IoT systems.