Miloš Prágr

Postdoctoral researcher

Computational Robotics Laboratory

Department of Computer Science

Czech Technical University in Prague

Active perception and learning for mobile robot traversability.

― ReSearch ABSTRACT
My research aims to improve the autonomy of mobile robots on long-term, unsupervised missions in outdoor environments. I study traversability, which describes the ease or difficulty experienced by the robot when attempting to traverse a particular terrain. For a mobile robot, understanding its traversability is crucial for its ability to select safe and efficient paths through potentially hazardous environments without supervision. My main focus is on assessing traversability through complex environments, such as terrains whose traversability changes with time or vegetation that appears rigid but can be passed through. To address these challenges, I advocate that traversability models should be adapted incrementally during the deployment, both exploiting the robot's newly accrued experience and actively seeking new information about unknown areas. To this end, my contributions are centered on applying incremental learning and active perception in traversability modeling.
― MEDIA

― EDUCATION

Postgradual student

Thesis - Learning Traversability from Mobile Robot Experience

Computational Robotics Laboratory, Department of Computer Science, Czech Technical University in Prague

2018 - 2024

Master Degree in Computer Vision and Image Processing

Thesis - Terrain Classification and Traversability Assessment from Exteroceptive Data

Faculty of Electrical Engineering, Czech Technical University in Prague

2016 - 2018

Bachelor Degree in Computer and Information Science

Thesis - Scalable Agent Navigation in Crowd Simulation

Faculty of Electrical Engineering, Czech Technical University in Prague

2013 - 2016

― ExpeRIENCE and COmmunity SERVICE

Participating

CTU-CRAS team in V4 Innovation Challenge Day

Debrecen, Hungary

2022

Participating

CTU-CRAS-NORLAB team in DARPA Subterranean Challenge

Virtual Track

2021

Research fellow

Computational Robotics Laboratory, Department of Computer Science, Czech Technical University in Prague

2018 - present

Reviewer

International Journal of Advanced Robotic Systems (IJARS)

Expert Systems With Applications (ESWA)

International Conference on Artificial Neural Networks (ICANN)

Project team member

Mobile Robot Pipeline Inspection for GasNet

2022

Project team member

Powerline Damage Detection from Thermal Imagery for ALEEGO

2021 - 2022

Project team member

Multi-Robot Persitent Monitoring of Dynamic Environments

Czech Science Foundation (GA ČR), Project No. 19-20238S, awarded to Jan Faigl

2019 - 2022

Project team member

Robotic Lifelong Learning of Multi-legged Robot Locomotion Control in Autonomous Data Collection Missions

Czech Science Foundation (GA ČR), Project No. 18-18858S, awarded to Jan Faigl

2018 - 2020

Internship

Robot Learning Lab, University of Washington

Seattle, WA, USA

2023

Teaching assistant

Department of Computer Science, Czech Technical University in Prague

AI in Robotics, Functional Programming, Procedural Programming

2018 - present

Project team member

Evaluation of Geometry-based Traversability Analysis for Off-road Navigation for the Ministry of Defence of the Czech Republic

2023

― Selected PUBLICATIONS

Miloš Prágr, Jan Bayer, and Jan Faigl

Autonomous exploration with online learning of traversable yet visually rigid obstacles

Autonomus Robots, 47, 2023

doi: 10.1007/s10514-022-10075-4

Miloš Prágr, Jan Bayer, and Jan Faigl

Autonomous robotic exploration with simultaneous environment and traversability models learning

Frontiers in Robotics and AI, 9, 2022

doi: 10.3389/frobt.2022.910113

Rudolf Szadkowski, Miloš Prágr, and Jan Faigl

Self-learning event mistiming detector based on central pattern generator

Frontiers in Neurorobotics, 155, 2021

doi: 10.3389/fnbot.2021.629652

Miloš Prágr, Petr Čížek, Jan Bayer, and Jan Faigl

Online incremental learning of the terrain traversal cost in autonomous exploration

In Robotics: Science and Systems (RSS), 2019

doi: 10.15607/RSS.2019.XV.040

Miloš Prágr, Petr Čížek, and Jan Faigl

Traversal cost modeling based on motion characterizationfor multi-legged walking robots

In European Conference on Mobile Robotics (ECMR), 2019

doi: 10.1109/ECMR.2019.8870912

Miloš Prágr and Jan Faigl

Benchmarking incremental regressors in traversal cost assessment

In International Conference on Artificial Neural Networks (ICANN), 2019

doi: 10.1007/978-3-030-30487-4_52

Jan Faigl and Miloš Prágr

On Unsupervised learning of traversal cost and terrain types identification using self-organizing maps

In International Conference on Artificial Neural Networks (ICANN), 2019

doi: 10.1007/978-3-030-30487-4_50

Miloš Prágr, Petr Čížek, and Jan Faigl

Cost of transport estimation for legged robot based on terrain features inference from aerial scan

In IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 2018

doi: 10.1109/IROS.2018.8593374

Jan Faigl, Miloš Prágr, and Jiří Kubík

On Autonomous mobile robot exploration projects in robotics course

In International Conference on Robotics in Education (RiE), 2023

doi: 10.1007/978-3-031-38454-7_1

Miloš Prágr, Jan Bayer, and Jan Faigl

On Predicting Terrain Changes Induced by Mobile Robot Traversal

Accepted for International Conference on Intelligent Robots and Systems (IROS), 2024