Institution profile

Technical University of Iasi

Academic institutioneurope · ro
Official website
Research library4linked papers
Opportunities0open roles
Selected work

Representative Papers

An analysis of higher-order kinematics formalisms for an innovative surgical parallel robot

Mar 27, 2025Mechanism and Machine Theory

Conventional first- and second-order kinematic models for surgical parallel robots suffer from insufficient trajectory accuracy, smoothness, and dynamic response—critical limitations in minimally invasive pancreatic surgery. Method: This paper proposes a high-order kinematic modeling and analysis framework tailored for modular hybrid parallel robots. It introduces, for the first time in surgical parallel robotics, a unified high-order differential kinematic formulation incorporating acceleration and jerk terms. Leveraging screw theory and Lie algebra, we derive a high-order Jacobian chain model, integrating symbolic computation with Simscape Multibody dynamic validation. Contribution/Results: Experimental evaluation demonstrates a 42% reduction in average end-effector trajectory tracking error and a 3.8× increase in transient response bandwidth. The framework provides a verifiable, real-time compatible high-order kinematic foundation for ultra-precise master–slave teleoperation.

1 citationsRead paper

Distinguishing AI-Generated Music from Edited Audio as a Hard-Negative Robustness Task

Aug 14, 2026

This study addresses the challenge of misclassification in AI music detection caused by edited audio serving as hard negative samples. We propose a robust detection framework that explicitly models edited audio as hard negatives and employs a pretrained PaSST architecture to process raw waveforms. To prevent information leakage, an anchor-song-based data partitioning strategy is designed, while Grad-CAM is utilized to reveal detectable spectral fingerprint features. Experimental results demonstrate that the model achieves a video-level balanced accuracy of 0.811 and an F1 score of 0.836 for AI-generated segments. These findings confirm the framework’s effectiveness in distinguishing between AI-generated and edited audio, validating the critical role of spectral cues in achieving robust detection performance.

0 citationsRead paper

Koopman Representations for Early Outbreak Warning and Minimal Counterfactual Intervention in Multi-Agent Epidemic Simulations

May 03, 2026

This work proposes a Koopman operator–based early warning and minimal intervention framework to address the high sensitivity of multi-agent epidemic systems to minor perturbations near critical transmission thresholds. By embedding high-dimensional nonlinear epidemic trajectories into a low-dimensional linear latent space, the approach enables high-accuracy outbreak risk prediction through a random forest classifier. Furthermore, counterfactual analysis is employed to identify individual-level minimal effective interventions. This study represents the first application of Koopman dynamic embeddings to epidemic forecasting and precision intervention, demonstrating that isolating a single key individual for just one day in near-critical scenarios can substantially reduce the attack rate and prevent large-scale outbreaks, thereby achieving exceptional predictive performance and intervention efficiency.

0 citationsRead paper
Recent publications

Latest Papers

Distinguishing AI-Generated Music from Edited Audio as a Hard-Negative Robustness Task

Aug 14, 2026

This study addresses the challenge of misclassification in AI music detection caused by edited audio serving as hard negative samples. We propose a robust detection framework that explicitly models edited audio as hard negatives and employs a pretrained PaSST architecture to process raw waveforms. To prevent information leakage, an anchor-song-based data partitioning strategy is designed, while Grad-CAM is utilized to reveal detectable spectral fingerprint features. Experimental results demonstrate that the model achieves a video-level balanced accuracy of 0.811 and an F1 score of 0.836 for AI-generated segments. These findings confirm the framework’s effectiveness in distinguishing between AI-generated and edited audio, validating the critical role of spectral cues in achieving robust detection performance.

0 citationsRead paper

Koopman Representations for Early Outbreak Warning and Minimal Counterfactual Intervention in Multi-Agent Epidemic Simulations

May 03, 2026

This work proposes a Koopman operator–based early warning and minimal intervention framework to address the high sensitivity of multi-agent epidemic systems to minor perturbations near critical transmission thresholds. By embedding high-dimensional nonlinear epidemic trajectories into a low-dimensional linear latent space, the approach enables high-accuracy outbreak risk prediction through a random forest classifier. Furthermore, counterfactual analysis is employed to identify individual-level minimal effective interventions. This study represents the first application of Koopman dynamic embeddings to epidemic forecasting and precision intervention, demonstrating that isolating a single key individual for just one day in near-critical scenarios can substantially reduce the attack rate and prevent large-scale outbreaks, thereby achieving exceptional predictive performance and intervention efficiency.

0 citationsRead paper

An analysis of higher-order kinematics formalisms for an innovative surgical parallel robot

Mar 27, 2025Mechanism and Machine Theory

Conventional first- and second-order kinematic models for surgical parallel robots suffer from insufficient trajectory accuracy, smoothness, and dynamic response—critical limitations in minimally invasive pancreatic surgery. Method: This paper proposes a high-order kinematic modeling and analysis framework tailored for modular hybrid parallel robots. It introduces, for the first time in surgical parallel robotics, a unified high-order differential kinematic formulation incorporating acceleration and jerk terms. Leveraging screw theory and Lie algebra, we derive a high-order Jacobian chain model, integrating symbolic computation with Simscape Multibody dynamic validation. Contribution/Results: Experimental evaluation demonstrates a 42% reduction in average end-effector trajectory tracking error and a 3.8× increase in transient response bandwidth. The framework provides a verifiable, real-time compatible high-order kinematic foundation for ultra-precise master–slave teleoperation.

1 citationsRead paper