About

Noriaki Hirose is a robotics and machine learning researcher whose work spans vision-based robot navigation, traversability estimation, social robotics, and deep learning-driven control systems. He has made significant contributions to the development of intelligent autonomous robots that can navigate complex, dynamic environments without relying on traditional mapping and localization pipelines. Hirose's most influential work includes "Deep Visual MPC-Policy Learning for Navigation" (2019, 88 citations), which demonstrated how robots could follow image-defined trajectories in a manner reminiscent of human landmark-based navigation. His semi-supervised deep learning framework GONet (2019, 71 citations) advanced traversability estimation using fisheye cameras and Generative Adversarial Networks, enabling safer robot deployment in unstructured environments. He also contributed to GNM (2023, 71 citations), a generalizable navigation model trained across diverse robot platforms, pushing the boundaries of data-efficient learning for robotics. Beyond navigation, Hirose has explored social robot behavior through SoPhie and SACSoN, addressing human-compliant path prediction and scalable social navigation. His earlier work on personal assistive robots and friction modeling via LSTM neural networks reflects a strong foundation in physical control systems. Collectively, his research demonstrates a sustained commitment to bridging perception, learning, and real-world robot deployment.

Research Focus

Key Achievements

14
H-Index
34
Papers
620
Total Citations
18
Avg Citations/Paper
🏆 Most Cited Paper
Deep Visual MPC-Policy Learning for Navigation
88 citations · 2019
📈 Most Prolific Year: 2018 (5 Papers)
🤝 Key Collaborators: 33
🏛 Institutions: Stanford University, University of California, Berkeley, Toyota Central Research and Development Laboratories (Japan), Toyota Motor Corporation (United States), Berkeley College, Toyota Motor Corporation (Switzerland)

Top Papers

  1. 1
  2. 2
  3. 3
  4. 4
  5. 5
  6. 6
  7. 7
  8. 8
  9. 9
  10. 10

Key Collaborators

Contact & Links

Available for collaboration
Content generated · 14 days ago