My Research

Recently, I have been researching motion planning techniques for autonomous navigation among humans. My goal is to develop a motion planner, for both autonomous vehicles or mobile robots, which allow them to navigate safely in crowded environments. Hence, my research is focused on Optimization and Machine Learning (ML)techniques. Optimization methods are widely used for motion planning because they allow generating optimal control commands that satisfy some pre-defined hard-constraints. These constraints are typically used to ensure that the generated commands satisfy the vehicle/robot limits as well as collision avoidance. However, optimization methods in crowded scenarios suffer from “the curse of the dimensionality” making them unapplicable to complex scenarios. In contrast, ML techniques have been successfully used to learn from big and high dimensional data but do not give any guarantees. Thus, my goal is to combine algorithms from both fields solving the weakness of one with the strength of the other. In summary, my research has three main goals:

1. To develop Safe and Social-Aware motion planners
2. To use Deep Learning models to learn the underlying human navigation model
3. To employ Online Learning and Reinforcement Learning techniques to improve motion planning over time

If you are interested to work in any of these topics feel free to contact me!