Institution profile

University of Siena

Academic institutioneurope · it
Official website
Research library89linked papers
Opportunities0open roles
Selected work

Representative Papers

Wearable Haptics for a Marionette-inspired Teleoperation of Highly Redundant Robotic Systems

May 13, 2024IEEE International Conference on Robotics and Automation

To address the challenges of motion-manipulation coupling, weak environmental perception, and low intuitiveness in teleoperation of highly redundant embodied robots (e.g., CENTAURO), this paper proposes a wearable haptic human–robot interface inspired by the marionette metaphor. Our approach uniquely integrates multimodal haptic feedback—vibrotactile and force cues—deeply into a closed-loop motion mapping framework, synergizing virtual physical interaction modeling with a real-time motion–sensor co-mapping algorithm. This enables natural, full-platform kinematic mapping from operator limb motions to robot motion while concurrently rendering proprioceptive and environmental contact states. Experimental evaluation demonstrates that novice users achieve a 37% reduction in task completion time, a 52% decrease in collision misclassification rate, and a 41% reduction in NASA-TLX subjective workload—significantly enhancing operational efficiency, safety, and immersion.

2 citationsRead paper

From Arabic Text to Puzzles: LLM-Driven Development of Arabic Educational Crosswords

Jan 19, 2025

Arabic language education has long suffered from a scarcity of high-quality interactive learning tools, constraining both learning efficacy and user engagement. To address this, we propose the first end-to-end Arabic educational crossword puzzle generation system. Our method introduces Arabic-Clue-Instruct—a novel, domain-specific clue instruction dataset comprising 52,000 high-quality examples—and a unified text-to-clue-to-grid generation framework. We employ multi-model collaborative reasoning using GPT-4-Turbo, GPT-3.5-Turbo, and Llama3-8B-Instruct, integrated with rule-based grid layout optimization and category-aware clue generation to ensure linguistic accuracy and pedagogical appropriateness. Empirical evaluation demonstrates significant improvements in vocabulary retention and learner engagement. All models, datasets, and source code are publicly released to foster reproducibility and community advancement.

1 citationsRead paper

Beyond Fixed Topologies: Unregistered Training and Comprehensive Evaluation Metrics for 3D Talking Heads

Oct 14, 2024arXiv.org

Existing speech-driven 3D talking head methods are constrained by fixed mesh topologies, limiting generalization to arbitrary topologies—e.g., real-world scanned faces. To address this, we propose the first topology-agnostic speech-driven animation framework. Our method introduces: (i) a registration-free training paradigm eliminating reliance on point-to-point correspondences; (ii) a heat-diffusion-based feature prediction mechanism ensuring topology-robust geometric modeling across diverse meshes; (iii) an adaptive graph neural network that learns dynamic graph structures per input; and (iv) a multi-granularity lip-sync evaluation metric suite addressing shortcomings of conventional metrics in temporal alignment and semantic consistency. Experiments demonstrate high-fidelity animation on arbitrary-topology 3D faces—including unseen scanned data—outperforming fixed-topology baselines. We establish the first topology-independent benchmark for 3D talking heads.

1 citationsRead paper
Recent publications

Latest Papers

Learning Social Robot Navigation By Sensing Human Legs

Jul 30, 2026

This work addresses the limited social compliance of existing 2D LiDAR–based robot navigation methods, which typically model pedestrians as static geometric obstacles and neglect the dynamic leg motion actually captured by LiDAR. To overcome this, we propose CALF, an end-to-end neural architecture that integrates convolutional layers, attention mechanisms, and multilayer perceptrons to directly interpret human leg dynamics from raw LiDAR scans and generate socially compliant navigation commands. We introduce, for the first time, human legs as the primary perceptual target, developing a dedicated gait model and a lightweight JAX-based ray-traced simulator, LegNav, enabling efficient policy training via deep reinforcement learning. Our approach outperforms both classical and learning-based baselines in navigation performance and social compliance, trains within one hour on a single consumer-grade GPU, and achieves zero-shot real-world deployment on a TurtleBot 4, producing smooth and socially appropriate trajectories.

0 citationsRead paper

Representation and Reference Selection in Training-Free Synthetic Image Attribution

Jul 13, 2026

This study addresses the coupling between representation space and reference sample selection in training-agnostic synthetic image provenance. It systematically investigates the impact of intermediate-layer features from pretrained models (CLIP, DINOv2) and three reference selection strategies—arbitrary, semantic alignment, and resynthesis—on provenance performance. The findings reveal that intermediate-layer representations achieve an optimal trade-off between preserving generator-discriminative cues and maintaining semantic consistency. Semantic constraints substantially mitigate query–reference mismatch: resynthesis proves most effective under low reference budgets, while semantic alignment offers the best cost–performance balance at moderate budgets. Building on these insights, the work proposes budget-adaptive strategies for constructing optimal reference sets.

0 citationsRead paper