Mihai Nan

About

Mihai Nan is a researcher at the forefront of social robotics and human action recognition, with a focus on enabling robots to understand and proactively respond to human behavior. His work bridges computer vision and human-robot interaction, particularly for applications in assisted living and ambient intelligence. Nan’s most influential contribution is his 2021 study comparing recurrent networks and temporal convolutional networks for skeleton-based action recognition (37 citations), which proposed key improvements to video-based human action analysis. He further advanced the field with a fast temporal graph convolutional model (2022) designed for real-time recognition in social robotics contexts. Nan also developed and evaluated the AMIRO social robotics framework on the Pepper robot (2020, 27 citations), demonstrating how affordable platforms can deliver focused assistance in retail and active aging domains. His 2019 work on human action recognition for social robots (12 citations) established proactivity as a core requirement for assistive systems. Collectively, Nan’s research has garnered over 80 citations, shaping how social robots perceive and interact with humans in real-world environments.

Research Focus

Key Achievements

Top Papers

1. 1
   Comparison between Recurrent Networks and Temporal Convolutional Networks Approaches for Skeleton-Based Action Recognition

   37 citations · 2021
2. 2
   The AMIRO Social Robotics Framework: Deployment and Evaluation on the Pepper Robot

   27 citations · 2020
3. 3
   Human Action Recognition for Social Robots

   12 citations · 2019
4. 4
   Fast Temporal Graph Convolutional Model for Skeleton-Based Action Recognition

   8 citations · 2022

Key Collaborators

Contact & Links