Research

The versatility of Computer Vision (CV) and Machine Learning (ML) has allowed me to conduct interdisciplinary research on the algorithmic and computational aspects of these techniques for process automation (industry and health) and robotics (aerial, industrial, space).

I have developed SW, such as the “HMPMR_ICIA” tracking algorithm used for aerial robotics and vision-based aerial refuelling. “PoLIS”, where CV and ML were integrated to automate power line inspection. “MalariApp”, a mobile App that uses CV, Deep Learning (DL), and Classical ML (supervised and unsupervised) for real-time detection of malaria parasites; ML for surgery time estimation for operating room programming; CV and DL for Industrial Robots within the context of Industry 4.0., among others

In 2015, as Professor at Pontificia Universidad Javeriana, Colombia, I started my research on Robotic Manipulation. Humans’ capacity for dexterous manipulation is exceptional. However, despite rapid technological advances, robots have been slower to harness these capabilities. My research then focused on process automation (waste separation task using co-robots) and robotic perception (CV and DL) for collaborative human-robot scenarios.

As a Research Scientist at SpaceR-SnT (2020), I expanded my research to Space applications.

My current research focuses on perception approaches for the autonomous operation of robots in space, especially for multi-purpose manipulation tasks for planetary and orbital robotics. It deals with fundamental and practical aspects of Machine Learning for space applications, including efforts on domain adaptation (to reduce sim2real gap), perception approaches for robot-environment interaction, and benchmarking (high-fidelity testing environment).

Videos from my research can be found in my youtube channel