Marvin Chancán

mchancanl at uni.pe

I am a Postdoctoral Associate at Yale, working on bio-inspired sensing and control of robotic systems in the Intelligent Autonomy Lab led by Ian Abraham.

I received a Ph.D. in Electrical Engineering and Robotics from QUT in 2022, an M.S. in Applied Control from PUC-Rio in 2012, and a B.S. in Mechatronics Engineering from UNI in 2009. I have also worked in the automation and IT industries for many years in several countries of The Americas, Europe and Oceania.

profile photo

Google Scholar             GitHub             LinkedIn             YouTube             Twitter

Research

During my PhD, I focused on the intersection between deep learning, computer vision, neuroscience and reinforcement learning for motion-and-vision-based localization and navigation tasks in the context of mobile robots and autonomous vehicles. Specifically, I studied biological neural circuits underlying these complex tasks in insect and rat mammalian brains, and developed new high performance neural architectures that were shown to be orders of magnitude faster than classical systems. Altogether, this research has set new state-of-the-art standards in terms of accuracy, high-throughput and low-latency. As a result, my papers have been published at top-tier machine learning and robotics venues such as NeurIPS, RA-L, ICRA, and RSS.

News

PhD Thesis

The Role of Motion-and-Visual Perception in Robot Place Learning and Navigation
Marvin Chancán
PhD Thesis, 2022
Project | Talk | ePrint

Publications

         
DeepSeqSLAM: A Trainable CNN+RNN for Joint Global Description and Sequence-based Place Recognition
Marvin Chancán, Michael Milford
Machine Learning for Autonomous Driving Workshop at the 34th Conference on Neural Information Processing Systems (NeurIPS 2020)
Project | Video | Code
Robot Perception enables Complex Navigation Behavior via Self-Supervised Learning
Marvin Chancán, Michael Milford
Self-Supervised Robot Learning Workshop at the Robotics: Science and Systems XVI (RSS 2020)
Project | Video | ArXiv
CityLearn: Diverse Real-World Environments for Sample-Efficient Navigation Policy Learning
Marvin Chancán, Michael Milford
IEEE International Conference on Robotics and Automation (ICRA 2020)
Project | Video | Code
A Hybrid Compact Neural Architecture for Visual Place Recognition
Marvin Chancán, Luis Hernandez-Nunez, A. Narendra, Andrew B. Barron, Michael Milford
IEEE Robotics and Automation Letters (RA-L)
Presented at ICRA 2020
Project | Video | Code

Preprints

   
Sequential Place Learning: Heuristic-Free High-Performance Long-Term Place Recognition
Marvin Chancán, Michael Milford
arXiv:2103.02074, March 2021
Project | ArXiv | Codebase
MVP: Unified Motion and Visual Self-Supervised Learning for Large-Scale Robotic Navigation
Marvin Chancán, Michael Milford
arXiv:2003.00667, March 2020
Project | ArXiv

Service

Honors & Awards

  • 2020 High Achiever HDR Award for high quality PhD research outputs from QUT.
  • 2018 Full President of the Republic Scholarship for a PhD degree.
  • 2017 Full Huiracocha PhD Scholarship from PUCP (declined).
  • 2010 Full Master's degree Scholarship from PUC-Rio.
  • 2009 Graduated 1st in my class at UNI.
  • 2008 Academic Honors Diploma for best performance in BS class at UNI.
  • 2005 Admitted in the top 2% (2000+ applicants) at UNI (top national science & engineering university).

Copyright 2022 © Marvin Chancán