Khalil Alipour
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
Khalil Alipour has a total of 69 co-authors. Here are the top co-authors by number of shared publications:
- Bahram Tarvirdizadeh
Advanced Service Robots (ASR) Laboratory, Department of Mechatronics Engineering, School of Intelligent Systems Engineering, College of Interdisciplinary Science and Technology, University of Tehran, Tehran, Iran
25 shared publications - Mohammad Ghamari
Department of Electrical Engineering, California Polytechnic State University, San Luis Obispo, CA, USA
10 shared publications Bahram Tarvirdizadeh
Advanced Service Robots (ASR) Laboratory, Department of Mechatronics Engineering, School of Intelligent Systems Engineering, College of Interdisciplinary Science and Technology, University of Tehran, Tehran, Iran
7 shared publicationsMohammad Ghamari
Department of Electrical Engineering, California Polytechnic State University, San Luis Obispo, CA, USA
6 shared publications- Alireza Hadi
Advanced Service Robots Laboratory, Department of Mechatronics, School of Intelligent Systems Engineering, College of Interdisciplinary Science and Technology, University of Tehran, Tehran, Iran.
5 shared publications - Yasin Hasanpoor
Department of Mechatronics Engineering Faculty of New Sciences and Technologies Advanced Service Robots (ASR) Lab University of Tehran Tehran Tehran Iran
4 shared publications
Latest Publications
- Open Access
- 20 November 2026
- Published by Springer Nature in Scientific Reports
Optimal actor-critic learning-based cooperative control of nonlinear multi-agent systems subject to communication delays
- 20 November 2026
- Published by Springer Nature in Scientific Reports
Open AccessOptimal actor-critic learning-based cooperative control of nonlinear multi-agent systems subject to communication delays
- Open Access
- 25 June 2026
- Published by Springer Nature in Scientific Reports
- Vol. 16  (1)
Event-triggered adaptive formation control for multi-agent systems using ratio-of-distance rigidity
This paper investigates adaptive formation control for multi-agent systems with nonlinear dynamics and uncertainties within a ratio-of-distance (RoD) rigidity framework. Distributed control laws are developed to achieve and maintain rigid formation shapes while tracking prescribed maneuvering velocities in both time-invariant and time-varying cases. The RoD-based formulation enables formation stabilization using only relative distance ratios, providing a scalable description of rigid shapes. To improve practical efficiency, an event-triggered control strategy is incorporated. In the proposed approach, each agent monitors its own states and the current states of its neighbors while updating its control input only when necessary. Adaptive terms within the distributed controller compensate for uncertainties in the agent dynamics. For maneuvering tasks with time-varying desired velocities, nonsmooth control terms are introduced to handle velocity variations. Lyapunov-based analysis is carried out to examine the stability of the closed-loop system. It is shown that, under suitable control gains, the formation errors asymptotically converge to zero while all adaptive signals remain bounded. Sufficient conditions on the controller parameters and event-triggering functions are derived to ensure asymptotic formation stabilization and velocity consensus. Simulations confirm the effectiveness of the proposed method under uncertain dynamics with significantly fewer control updates compared with continuous-time control.
- 25 June 2026
- Published by Springer Nature in Scientific Reports
- Vol. 16  (1)
Open AccessEvent-triggered adaptive formation control for multi-agent systems using ratio-of-distance rigidity
This paper investigates adaptive formation control for multi-agent systems with nonlinear dynamics and uncertainties within a ratio-of-distance (RoD) rigidity framework. Distributed control laws are developed to achieve and maintain rigid formation shapes while tracking prescribed maneuvering velocities in both time-invariant and time-varying cases. The RoD-based formulation enables formation stabilization using only relative distance ratios, providing a scalable description of rigid shapes. To improve practical efficiency, an event-triggered control strategy is incorporated. In the proposed approach, each agent monitors its own states and the current states of its neighbors while updating its control input only when necessary. Adaptive terms within the distributed controller compensate for uncertainties in the agent dynamics. For maneuvering tasks with time-varying desired velocities, nonsmooth control terms are introduced to handle velocity variations. Lyapunov-based analysis is carried out to examine the stability of the closed-loop system. It is shown that, under suitable control gains, the formation errors asymptotically converge to zero while all adaptive signals remain bounded. Sufficient conditions on the controller parameters and event-triggering functions are derived to ensure asymptotic formation stabilization and velocity consensus. Simulations confirm the effectiveness of the proposed method under uncertain dynamics with significantly fewer control updates compared with continuous-time control.
- Open Access
- 3 June 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 14, 86364-86378
Non-Invasive Blood Pressure Estimation Using PPG and a CNN-BiLSTM-Attention Deep Learning Architecture
Blood pressure (BP) serves as a critical indicator of cardiovascular health, quantifying the force exerted by circulating blood against arterial walls during each cardiac cycle. Traditional BP measurement methods, however, are often impractical for continuous monitoring in daily life. To address this limitation, non-invasive estimation techniques leveraging photoplethysmography (PPG) signals have been proposed. These methods, nevertheless, encounter significant challenges, including a reliance on manual feature engineering, high susceptibility to noise and motion artifacts, and an inherent inability to fully replicate the complex temporal dynamics of BP waveforms, thereby constraining estimation accuracy and reliability. This study presents an advanced deep learning framework for precise, non-invasive BP estimation using PPG signals. The proposed method feeds PPG signals into a hybrid architecture comprising six convolutional layers, two bidirectional long short-term memory (BiLSTM) layers, and an attention mechanism to generate the corresponding BP waveforms. From the predicted BP signals, systolic and diastolic points are extracted and compared against those in the reference arterial blood pressure (ABP) signals to evaluate estimation accuracy. The framework was validated on a comprehensive dataset comprising 2,064 patients from the MIMIC-II database, achieving mean absolute errors (MAE) of 2.44 mmHg for systolic blood pressure (SBP), 1.72 mmHg for diastolic blood pressure (DBP), and 1.10 mmHg for mean arterial pressure (MAP) across 5-fold cross-validation. Despite the increased complexity of generating entire BP signals, the proposed approach demonstrates superior performance compared to existing BP estimation methods, highlighting its robustness and effectiveness for scalable, non-invasive cardiovascular health monitoring.
- 3 June 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 14, 86364-86378
Open AccessNon-Invasive Blood Pressure Estimation Using PPG and a CNN-BiLSTM-Attention Deep Learning Architecture
Blood pressure (BP) serves as a critical indicator of cardiovascular health, quantifying the force exerted by circulating blood against arterial walls during each cardiac cycle. Traditional BP measurement methods, however, are often impractical for continuous monitoring in daily life. To address this limitation, non-invasive estimation techniques leveraging photoplethysmography (PPG) signals have been proposed. These methods, nevertheless, encounter significant challenges, including a reliance on manual feature engineering, high susceptibility to noise and motion artifacts, and an inherent inability to fully replicate the complex temporal dynamics of BP waveforms, thereby constraining estimation accuracy and reliability. This study presents an advanced deep learning framework for precise, non-invasive BP estimation using PPG signals. The proposed method feeds PPG signals into a hybrid architecture comprising six convolutional layers, two bidirectional long short-term memory (BiLSTM) layers, and an attention mechanism to generate the corresponding BP waveforms. From the predicted BP signals, systolic and diastolic points are extracted and compared against those in the reference arterial blood pressure (ABP) signals to evaluate estimation accuracy. The framework was validated on a comprehensive dataset comprising 2,064 patients from the MIMIC-II database, achieving mean absolute errors (MAE) of 2.44 mmHg for systolic blood pressure (SBP), 1.72 mmHg for diastolic blood pressure (DBP), and 1.10 mmHg for mean arterial pressure (MAP) across 5-fold cross-validation. Despite the increased complexity of generating entire BP signals, the proposed approach demonstrates superior performance compared to existing BP estimation methods, highlighting its robustness and effectiveness for scalable, non-invasive cardiovascular health monitoring.
- Open Access
- 1 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 14, 159698-159713
Prior-Informed Online Data-Driven Safe Control via LMI and Uncertainty-Set Update
- 1 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 14, 159698-159713
Open AccessPrior-Informed Online Data-Driven Safe Control via LMI and Uncertainty-Set Update
- Open Access
- 1 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 14, 129859-129874
Deep Learning Based Multiple Impedance Control of a Space Free-Flying Robot
- 1 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 14, 129859-129874
Open AccessDeep Learning Based Multiple Impedance Control of a Space Free-Flying Robot
- Open Access
- 1 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 14, 118248-118262
Event-Triggered Data-Based Reinforcement Learning Control for Optimal Consensus of Heterogeneous Multi-Agent Systems in Graphical Games
- 1 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 14, 118248-118262
Open AccessEvent-Triggered Data-Based Reinforcement Learning Control for Optimal Consensus of Heterogeneous Multi-Agent Systems in Graphical Games
- Open Access
Optimal Regulation of Nonlinear Input-Affine Systems via an Integral Reinforcement Learning-Based State-Dependent Riccati Equation Approach
The State-Dependent Riccati Equation (SDRE) technique generalizes the classical algebraic Riccati formulation to nonlinear systems by designing an input to the system that optimally(suboptimally) regulates system states toward the origin while simultaneously optimizing a quadratic performance index. In the SDRE technique, we solve the State-Dependent Riccati Equation to determine the control for regulating a nonlinear input-affine system. Since an analytic solution to SDRE is not straightforward, one method is to linearize the system at every state, solve the corresponding Algebraic Riccati Equation (ARE), and apply optimal control until the next state of the system. Completing this task with high frequency gives a result like the original SDRE technique. Both approaches require a complete model; therefore, here we propose a method that solves ARE in every state of the system using a partially model-free approach that learns optimal control in every state of the system, without explicit knowledge of the drift dynamics, based on Integral Reinforcement Learning (IRL). To show the effectiveness of our proposed approach, we apply it to the second-order nonlinear system in simulation and compare its performance with the classical SDRE method, which relies on the system's model and solves the ARE at each state. Our simulation results demonstrate that, with sufficient iterations, the IRL-based approach achieves approximately the same performance as the conventional SDRE method, demonstrating its capability as a reliable alternative for nonlinear system control that does not require an explicit environmental model. Index Terms-Algebraic Riccati Equation (ARE), Integral Reinforcement Learning (IRL), Nonlinear Input-Affine Systems, Optimal Regulation, State-Dependent Riccati Equation (SDRE)
Open AccessOptimal Regulation of Nonlinear Input-Affine Systems via an Integral Reinforcement Learning-Based State-Dependent Riccati Equation Approach
The State-Dependent Riccati Equation (SDRE) technique generalizes the classical algebraic Riccati formulation to nonlinear systems by designing an input to the system that optimally(suboptimally) regulates system states toward the origin while simultaneously optimizing a quadratic performance index. In the SDRE technique, we solve the State-Dependent Riccati Equation to determine the control for regulating a nonlinear input-affine system. Since an analytic solution to SDRE is not straightforward, one method is to linearize the system at every state, solve the corresponding Algebraic Riccati Equation (ARE), and apply optimal control until the next state of the system. Completing this task with high frequency gives a result like the original SDRE technique. Both approaches require a complete model; therefore, here we propose a method that solves ARE in every state of the system using a partially model-free approach that learns optimal control in every state of the system, without explicit knowledge of the drift dynamics, based on Integral Reinforcement Learning (IRL). To show the effectiveness of our proposed approach, we apply it to the second-order nonlinear system in simulation and compare its performance with the classical SDRE method, which relies on the system's model and solves the ARE at each state. Our simulation results demonstrate that, with sufficient iterations, the IRL-based approach achieves approximately the same performance as the conventional SDRE method, demonstrating its capability as a reliable alternative for nonlinear system control that does not require an explicit environmental model. Index Terms-Algebraic Riccati Equation (ARE), Integral Reinforcement Learning (IRL), Nonlinear Input-Affine Systems, Optimal Regulation, State-Dependent Riccati Equation (SDRE)
- Open Access
LSTM-Based Modeling and Reinforcement Learning Control of a Magnetically Actuated Catheter
Autonomous magnetic catheter systems are emerging as a promising approach for the future of minimally invasive interventions. This study presents a novel approach that begins by modeling the nonlinear and hysteretic dynamics of a magnetically actuated catheter system, consists of a magnetic catheter manipulated by servo-controlled magnetic fields generated by two external permanent magnets, and its complex behavior is captured using a Long Short-Term Memory (LSTM) neural network. This model validated against experimental setup's data with a root mean square error (RMSE) of 0.42 mm and 99.8% coverage within 3 mm, establishing it as a reliable surrogate model. This LSTM enables the training of Reinforcement Learning (RL) agents for controlling the system and avoiding damage to the real setup, with the potential for subsequent fine-tuning on the physical system. We implemented Deep Q-Network (DQN) and actor-critic RL controllers, comparing these two agents first for regulation and subsequently for path following along linear and half-sinusoidal paths for the catheter tip. The actor-critic outperforms DQN, offering greater accuracy and faster performance with less error, along with smoother trajectories at a 10 Hz sampling rate, in both regulation and path following compared to the DQN controller. This performance, due to the continuous action space, suits dynamic navigation tasks like navigating curved vascular structures for practical applications.
Open AccessLSTM-Based Modeling and Reinforcement Learning Control of a Magnetically Actuated Catheter
Autonomous magnetic catheter systems are emerging as a promising approach for the future of minimally invasive interventions. This study presents a novel approach that begins by modeling the nonlinear and hysteretic dynamics of a magnetically actuated catheter system, consists of a magnetic catheter manipulated by servo-controlled magnetic fields generated by two external permanent magnets, and its complex behavior is captured using a Long Short-Term Memory (LSTM) neural network. This model validated against experimental setup's data with a root mean square error (RMSE) of 0.42 mm and 99.8% coverage within 3 mm, establishing it as a reliable surrogate model. This LSTM enables the training of Reinforcement Learning (RL) agents for controlling the system and avoiding damage to the real setup, with the potential for subsequent fine-tuning on the physical system. We implemented Deep Q-Network (DQN) and actor-critic RL controllers, comparing these two agents first for regulation and subsequently for path following along linear and half-sinusoidal paths for the catheter tip. The actor-critic outperforms DQN, offering greater accuracy and faster performance with less error, along with smoother trajectories at a 10 Hz sampling rate, in both regulation and path following compared to the DQN controller. This performance, due to the continuous action space, suits dynamic navigation tasks like navigating curved vascular structures for practical applications.
- Open Access
- 27 October 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 13, 185657-185673
Noninvasive Assessment of Arterial Stiffness Using Photoplethysmography: Feature Analysis and Machine Learning-Based Estimation of Carotid-Femoral Pulse Wave Velocity
Cardiovascular diseases are the leading cause of mortality worldwide, underscoring the need for accessible, noninvasive risk assessment. Arterial stiffness, a hallmark of vascular aging, is an established independent predictor of cardiovascular events. While carotid–femoral pulse wave velocity is the clinical gold standard, its use is constrained by specialized equipment and trained personnel. Photoplethysmography, a simple and cost-effective optical technique, offers a promising alternative for wearable technologies. This study analyzes 60 morphological, temporal, and second-derivative features extracted from digital, radial, and brachial arterial signals in a dataset representing 4,374 healthy-adult profiles across six age groups. Correlation and trajectory analyses identified the radial artery as the site with the strongest age association (peak correlation r = 0.75) and revealed site-dependent age trends. Building on these findings, a gradient-boosting regression model trained on radial-artery features estimated carotid–femoral pulse wave velocity using stratified fivefold cross-validation, achieving a mean absolute error of 0.115 ± 0.005 m/s, a root-mean-square error of 0.180 ± 0.011 m/s, and a coefficient of determination of 0.993 ± 0.001. Bland–Altman analysis showed negligible bias (−0.003 m/s) with 95% limits of agreement from −0.357 to 0.351 m/s. Feature-importance analysis corroborated the relevance of interpretable, physiologically meaningful indices. Overall, this work provides a comprehensive cross-site evaluation of photoplethysmography-derived features for arterial-stiffness assessment and demonstrates accurate carotid–femoral pulse wave velocity estimation from the radial artery, supporting integration into wearable systems for continuous vascular-health monitoring.
- 27 October 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 13, 185657-185673
Open AccessNoninvasive Assessment of Arterial Stiffness Using Photoplethysmography: Feature Analysis and Machine Learning-Based Estimation of Carotid-Femoral Pulse Wave Velocity
Cardiovascular diseases are the leading cause of mortality worldwide, underscoring the need for accessible, noninvasive risk assessment. Arterial stiffness, a hallmark of vascular aging, is an established independent predictor of cardiovascular events. While carotid–femoral pulse wave velocity is the clinical gold standard, its use is constrained by specialized equipment and trained personnel. Photoplethysmography, a simple and cost-effective optical technique, offers a promising alternative for wearable technologies. This study analyzes 60 morphological, temporal, and second-derivative features extracted from digital, radial, and brachial arterial signals in a dataset representing 4,374 healthy-adult profiles across six age groups. Correlation and trajectory analyses identified the radial artery as the site with the strongest age association (peak correlation r = 0.75) and revealed site-dependent age trends. Building on these findings, a gradient-boosting regression model trained on radial-artery features estimated carotid–femoral pulse wave velocity using stratified fivefold cross-validation, achieving a mean absolute error of 0.115 ± 0.005 m/s, a root-mean-square error of 0.180 ± 0.011 m/s, and a coefficient of determination of 0.993 ± 0.001. Bland–Altman analysis showed negligible bias (−0.003 m/s) with 95% limits of agreement from −0.357 to 0.351 m/s. Feature-importance analysis corroborated the relevance of interpretable, physiologically meaningful indices. Overall, this work provides a comprehensive cross-site evaluation of photoplethysmography-derived features for arterial-stiffness assessment and demonstrates accurate carotid–femoral pulse wave velocity estimation from the radial artery, supporting integration into wearable systems for continuous vascular-health monitoring.
- Open Access
- 1 October 2025
- Published by SAGE Publications in Health Informatics Journal
- Vol. 31  (4)
Exploring the potential of five machine learning regression algorithms for noninvasive blood pressure estimation with photoplethysmography
Objective: Blood pressure (BP) is a vital factor for human health and survival, and its elevation and fluctuations can have dangerous consequences on an individual’s well-being. Traditional BP measurement methods—including cuff-based devices and invasive arterial catheters—are unsuitable for continuous monitoring in daily life: cuffs are intermittent and uncomfortable, whereas arterial lines provide continuous data but are invasive and confined to clinical settings (e.g., ICUs/ORs). In response to this requirement, we propose a cuff-less, continuous, and noninvasive system for BP measurement using photoplethysmograph (PPG) signals and machine learning (ML) algorithms. Methods: In this investigation, we analyzed PPG signals acquired from a diverse cohort, with participants ranging in age from 21 to 86 years and including both healthy subjects and those with health conditions. The data underwent rigorous preprocessing and feature extraction procedures. To address computational efficiency and mitigate overfitting, we applied five distinct feature selection methods to refine the feature set. Subsequently, each method’s selected features were independently trained and tested using five ML regression algorithms to estimate systolic blood pressure (SBP) and diastolic blood pressure (DBP). Results: Our findings reveal that the ensemble-based extra trees (ET) algorithm, coupled with the SelectFromModel feature selection approach, surpassed competing algorithms in estimative performance. The ET algorithm achieved notably low root mean squared errors (RMSEs) of 5.21 for SBP and 2.65 for DBP, demonstrating its exceptional capability in the estimation of BP. Conclusion: The proposed approach demonstrates strong potential for accurate, non-invasive BP estimation. These findings have important implications for the development of wearable and mobile health technologies that support continuous, real-time BP monitoring for the prevention and management of hypertension and cardiovascular diseases.
- 1 October 2025
- Published by SAGE Publications in Health Informatics Journal
- Vol. 31  (4)
Open AccessExploring the potential of five machine learning regression algorithms for noninvasive blood pressure estimation with photoplethysmography
Objective: Blood pressure (BP) is a vital factor for human health and survival, and its elevation and fluctuations can have dangerous consequences on an individual’s well-being. Traditional BP measurement methods—including cuff-based devices and invasive arterial catheters—are unsuitable for continuous monitoring in daily life: cuffs are intermittent and uncomfortable, whereas arterial lines provide continuous data but are invasive and confined to clinical settings (e.g., ICUs/ORs). In response to this requirement, we propose a cuff-less, continuous, and noninvasive system for BP measurement using photoplethysmograph (PPG) signals and machine learning (ML) algorithms. Methods: In this investigation, we analyzed PPG signals acquired from a diverse cohort, with participants ranging in age from 21 to 86 years and including both healthy subjects and those with health conditions. The data underwent rigorous preprocessing and feature extraction procedures. To address computational efficiency and mitigate overfitting, we applied five distinct feature selection methods to refine the feature set. Subsequently, each method’s selected features were independently trained and tested using five ML regression algorithms to estimate systolic blood pressure (SBP) and diastolic blood pressure (DBP). Results: Our findings reveal that the ensemble-based extra trees (ET) algorithm, coupled with the SelectFromModel feature selection approach, surpassed competing algorithms in estimative performance. The ET algorithm achieved notably low root mean squared errors (RMSEs) of 5.21 for SBP and 2.65 for DBP, demonstrating its exceptional capability in the estimation of BP. Conclusion: The proposed approach demonstrates strong potential for accurate, non-invasive BP estimation. These findings have important implications for the development of wearable and mobile health technologies that support continuous, real-time BP monitoring for the prevention and management of hypertension and cardiovascular diseases.