Eric Aislan Antonelo
Professor / Coordinator
Professor at UFSC (DAS). Research interests include autonomous systems, imitation learning, reservoir computing, and physics-informed neural networks for modeling and control.
About the group
The group advances machine learning and control methods for intelligent autonomous systems, spanning end-to-end driving, Bird’s-Eye View representations, multimodal imitation learning, physics-informed modeling, and control of nonlinear dynamical systems.
GSAI’s research portfolio is organized around two complementary axes: autonomous systems and physics-informed neural networks for control. This combination allows the group to work on both data-driven intelligence and physically consistent decision-making.
Research pillars
Two core directions shape the current agenda of the group.
Autonomous systems
Research on intelligent vehicles and robots, especially policy learning for navigation in complex environments.
- Autonomous driving in simulated and real-world-inspired environments
- Imitation learning, GAIL, EBMs, and diffusion models
- Bird’s-Eye View representations and mid-level perception
- Robust policy generalization under multimodal behaviors
PINC and physics-informed control
Development of models that incorporate physical laws into learning for simulation and control.
- Physics-informed neural networks for ODE/PDE-governed systems
- Control of nonlinear and controllable dynamical systems
- Physics-informed recurrent and reservoir-based models
- Bridging data efficiency, prior knowledge, and control performance
News
Highlighted project currently under development in the group.
Approved CNPq Universal Project
Aprendizado de Políticas para Navegação Autônoma de Veículos
The project investigates advanced imitation learning approaches for autonomous driving, including GAIL, energy-based models, and diffusion models, combined with Bird’s-Eye View (BEV) representations. The main objective is to improve generalization beyond traditional behavior cloning in dynamic, multimodal urban driving scenarios. Experiments are conducted in CARLA and transferred to a miniature vehicle operating on simulated urban tracks and real campus sidewalks.
Members
Current team of staff and students.
Staff
PhD students
Gustavo Claudio Karl Couto
PhD Student
Research on implicit imitation learning and autonomous vehicles in urban environments.
Yessica Maria Valencia Lemos
PhD Student
Researcher in intelligent autonomous systems and machine learning.
Master students
Felipe Carlos dos Santos
Master Student
Works on Bird’s-Eye View generation from autonomous vehicle cameras using transformers.
Leonardo Pezenatto da Silva
Master Student
Works on vision-based inertial signal estimation and road condition monitoring.
Eduardo Vargas Zummach
Master Student
Works on imitation learning based on generative models for autonomous vehicles.
Undergraduate students
MC
Miguel Castilho Silva
Undergraduate Student
Undergraduate researcher in intelligent autonomous systems.
Gabriel Zipperer
Undergraduate Student
Undergraduate researcher in AI and autonomous systems.
Selected publications
A curated list highlighting recent work aligned with the two main themes of the group.
PINC
Physics-informed neural nets for control of dynamical systems
Neurocomputing · 2024
PINC
Physics-Informed Neural Networks for Control of single-phase flow systems governed by partial differential equations
Engineering Applications of Artificial Intelligence · 2025
PINC
Physics-informed Echo State Networks for modeling controllable dynamical systems
Neurocomputing · 2025
Autonomous Driving
Hierarchical Generative Adversarial Imitation Learning With Mid-Level Input Generation for Autonomous Driving on Urban Environments
IEEE Transactions on Intelligent Vehicles · 2024
Autonomous Driving
Investigating Behavior Cloning from Few Demonstrations for Autonomous Driving Based on Bird’s-Eye View in Simulated Cities
Lecture Notes in Computer Science · 2025
Autonomous Driving
Exploring multimodal implicit behavior learning for vehicle navigation in simulated cities
ENIAC · 2025
Autonomous Driving
Investigating Diffusion Models for Offline Behavior Cloning in Autonomous Driving Scenarios
CROS · 2025
Perception
An Initial Study of Bird’s-Eye View Generation for Autonomous Vehicles using Cross-View Transformers
ENIAC · 2025
Media
External media and demonstration content related to the group’s autonomous driving research.
Featured media
Learning to Autonomously Drive a Vehicle by Imitation
Autonomous driving by imitation
This media item highlights the group’s research on autonomous driving based on imitation learning, including work on urban driving and hierarchical generative adversarial imitation learning.
Previous alumni
Selected previous master’s students supervised or co-supervised in related topics.
- Bruno Maciel Machado — Diffusion Models for Offline Behavior Cloning in Autonomous Driving Scenarios (2025)
- Willian Benoski — Implicit Behavioral Cloning for Multimodal Behavior Learning in Autonomous Driving (2025)
- Gustavo Claudio Karl Couto — Generative Adversarial Imitation Learning with Bird’s-Eye View Input Generation for Autonomous Driving on Urban Environments (2023)
- Irving Giovani — Imitation Learning for Autonomous Driving: Disagreement-Regularization and Behavior Cloning with Beta Distribution (2023)
- Eric Mochiutti — Physics-Informed Echo State Networks for Modeling Controllable Dynamical Systems (2023)
- Rafael de Souza Toledo — Semantic Segmentation of Low-Resolution Images from Damaged Roads (2022)
Contact
Academic home of the group.
Eric Aislan Antonelo
Department of Automation and Systems (DAS), Federal University of Santa Catarina (UFSC), Florianópolis, Brazil.
Homepage: ericantonelo.paginas.ufsc.br