Matheus Paiva Angarola

About me

I am a Computer Science undergraduate student and a FAPESP Research Fellow at the University of São Paulo (USP), under the supervision of Prof. Marcelo Becker. Currently, I am based in Champaign, Illinois, working as a Fully Funded Research Intern at the Distributed Autonomous Systems Lab (DASLab) at the University of Illinois Urbana-Champaign (UIUC) advised by Prof. Girish Chowdhary.

My research focuses on developing adaptive quadruped locomotion and mobile manipulation policies using Mixture-of-Experts (MoE) architectures and Reinforcement Learning. I am deeply passionate about embodied autonomy, investigating how robotic systems can dynamically adjust to unstructured terrains and physical constraints through scalable models, visual perception, and temporal memory. Key research areas include:

  • Developing agile whole-body control strategies for quadrupedal loco-manipulation.
  • Training robust deep reinforcement learning policies for sim-to-real robotic transfer.
  • Leveraging Mixture-of-Experts frameworks for policy adaptation in unstructured environments.
  • Integrating computer vision and temporal memory for real-time perceptive locomotion.

Publications

CTS-MoE Publication
CTS-MoE: Implicit Terrain Adaptation via Mixture-of-Experts for Perceptive Locomotion Francisco Affonso*, Matheus P. Angarola, Ana Luiza Mineiro, Aditya Potnis, Marcelo Becker, Girish Chowdhary Under Review, 2026
IROS 2026 Publication
Towards Capability-Aware Traversability Navigation for Unstructured Environments Gianluca Capezzuto, Felipe Tommaselli, Matheus P. Angarola, Ricardo V. Godoy, Marcelo Becker IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 2026
ICAR 2025 Publication
Learning Terrain-Specialized Policies for Adaptive Locomotion in Challenging Environments Matheus P. Angarola*, Francisco Affonso* and Marcelo Becker IEEE International Conference on Advanced Robotics (ICAR), 2025

Projects

Loco-manipulation Project
Loco-manipulation with Mixture of Experts Personal project aimed at integrating Mixture of Experts (MoE) with loco-manipulation, focused on training with Reinforcement Learning (RL) with future plans for real-world robot deployment. Frameworks: IsaacLab, rsl-rl, and Reinforcement Learning.
Project 1
Traversability Data Generation (IROS 2026) Interactive pipeline and codebase for generating the Capability-Aware Traversability (CAT) dataset. Frameworks: DinoV3, SAM2, SegFormer, and ROS.
Project 3
Deployment low-level system for locomotion (CTS-MOE) Deployment pipeline for perceptive locomotion policies on a Unitree Go1, focusing on optimizing latency and inference times to improve sim-to-real transfer. Frameworks: ROS, LCM, and ONNX Runtime.
Project 1
Autonomous Robotic System for CBR@Work and RoboCup@Work Team project for the CBR@Work 2024 competition, focused on autonomous industrial tasks and smart robotics. Frameworks: ROS 1, Gazebo, ESP-IDF, Jetson Nano, and ESP32.

Experience & Education

Robotics Research Intern 2026 — Present

Distributed Autonomous Systems Lab (DASLab), UIUC

Bachelor’s in Computer Science 2023 — Present

University of São Paulo (ICMC - USP)

Technical Degree in Systems Development 2020 — 2021

ETEC Martin Luther King