Artificial Intelligence and Advanced Technologies for Next-Generation Sensors and Computer Vision

Research objectives：

This Research Project covers the following topics:

1. Applications for AI, Sensors and Computer Vision in various fields such as healthcare, automation and industry.
 
2. Novel AI models, algorithms, and learning paradigms that enhance sensing accuracy, perception robustness, data interpretation, and scene understanding.
 
3. Novel developments in next-generation sensor technologies, including multi-modal sensing, miniaturized and embedded sensing platforms, and emerging hardware innovations.
 
4. Interdisciplinary research on the integration of AI, sensing, and vision systems for real-world applications across domains such as autonomous systems, robotics, healthcare, smart manufacturing, transportation, security, and environmental monitoring.
 
5. Foster contributions that address challenges related to data fusion, uncertainty quantification, explainability, scalability, low-latency inference, energy efficiency, and system reliability.
 
6. Research bridging theory and practice, including benchmarks, real-world deployment, datasets, reproducible pipelines, and industry case studies

Keywords：

• Sensors
• AI
• Computer Vision
• Signal Processing
• Advanced Technology

Expected Outcomes：

Academic: Peer-reviewed publications

Planned Paper Information:

1: Title: AI-Driven Multi-Sensor Fusion in Human Activity Recognition for Healthcare Monitoring
Authors: Minh Long Hoang
Author: University of Parma, Italy; minhlong.hoang@unipr.it
Abstract: Human activity recognition (HAR) has emerged as a key enabling technology for healthcare monitoring, rehabilitation assessment, and ambient assisted living. However, real-world healthcare environments often require robust perception under heterogeneous sensing conditions, variability in patient behavior, and multi-modal data streams. This paper investigates AI-driven multi-sensor fusion approaches for HAR, leveraging data from wearable and ambient sensing platforms to improve recognition accuracy, temporal stability, and interpretability. Deep learning architectures and feature-level fusion strategies are explored to integrate inertial, biomechanical, and contextual information into a unified activity recognition framework. Experimental evaluations on healthcare datasets demonstrate improved performance for complex activities of daily living, including transitions and fine-grained motion patterns. The results indicate that multi-sensor AI fusion not only enhances classification robustness compared to single modality systems but also provides valuable insights for personalized healthcare and remote patient monitoring. This study highlights the potential of AI-driven HAR systems for next-generation smart healthcare solutions, emphasizing challenges and opportunities related to sensor heterogeneity, real-time processing, and clinical deployment.

2: Smart Sensor-Driven Optimization of PV Systems for Smart Grid Applications
Author: Nicola Delmonte
University of Parma, Italy; nicola.delmonte@unipr.it
Abstract: The integration of photovoltaic (PV) systems into smart grids requires advanced monitoring, control, and optimization strategies to ensure energy reliability, stability, and efficiency. Smart sensors play a pivotal role in enabling real-time data acquisition, environmental awareness, and predictive control for PV generation systems. This paper investigates a smart sensor-driven framework for optimizing PV system performance within smart grid environments. The proposed approach utilizes sensor networks to collect high-resolution data on irradiance, temperature, load demand, and grid conditions, enabling enhanced forecasting, maximum power point tracking (MPPT), and adaptive energy management. By incorporating data-driven optimization algorithms and intelligent control mechanisms, the system improves energy yield, reduces operational losses, and enhances grid integration under dynamic environmental and load conditions. Experimental evaluations and simulation-based analyses demonstrate the effectiveness of smart sensor-assisted optimization in improving PV system efficiency, grid stability, and responsiveness to demand-side variations. The research highlights the potential of smart sensors as key enablers for next-generation PV-based smart grids, supporting sustainable energy infrastructures.