I am a Mechanical Engineering student at UFTM, working on robotics and studying how physical machines, communication, simulation and intelligent software fit together.
My long-term direction is to design and understand the physical system, model its virtual counterpart, connect both reliably, measure the difference, and then explore intelligent behavior.
LinkedIn · Public repositories
The diagram describes my learning and development direction. It is not an implemented end-to-end system.
Physical system
↕
Sensors / actuators / embedded systems
↕
ROS 2 / DDS · OPC UA / MQTT (planned study)
↕
Digital Twin synchronization (planned)
↕
Gazebo · Isaac Sim / OpenUSD (future evaluation)
↕
AI / perception / planning / autonomous systems (research direction)
- Building on: Python actuator tooling, ROS 2/Gazebo simulation, mechanical CAD and telemetry.
- Studying: communication architectures, DDS QoS, latency, jitter, state consistency and physical/virtual synchronization.
- Future experiments: OPC UA/MQTT gateways, edge computing and Edge AI, sim-to-real, and autonomous behavior.
- Secondary track: CAD/CAE automation and traceable CAD-to-simulation workflows.
Physical AI is a long-term research direction here; no deployed learning-based controller is claimed.
| Project | Current evidence | Next step |
|---|---|---|
| GAIA Humanoid Simulation | ROS 2 Jazzy/Gazebo Harmonic package using the ROBOTIS OP3 reference model, joint control, camera and IMU; automated simulation checks | Identify a GAIA joint and plan Digital Twin integration. The fixed-base OP3 lab is not a complete GAIA Digital Twin. |
| LX-225 / BusLinker Test Bench | Python serial driver, position/voltage/temperature feedback, logging and mocked serial tests; velocity is estimated | A single-actuator physical/virtual experiment, starting with a documented ROS 2 interface |
| GAIA Humanoid Structure | Mechanical CAD assets, mounting guidance and a model-data checklist | Identify masses, joint axes, limits and provenance before CAD → URDF/USD → simulation |
| ROS 2 Environment | Ubuntu 24.04/Jazzy setup and environment validation documentation | Record reproducible machine configurations for the experiments |
I am developing the architecture of physical-ai-digital-twin-lab, an experimental laboratory for robotics, distributed communication and physical/virtual synchronization.
The first target is one virtual actuator, followed by an LX-225/BusLinker connection and a reproducible comparison of physical and virtual state. Architecture and experiment planning come before hardware integration, middleware benchmarks or AI. The repository contains architecture, a research roadmap and initial Issues; runtime implementation is planned.
- GAIA Robotics: team captain.
- Valkiriaz — H2 Challenge: vice-president, contributing to ROS 2, Digital Twin, telemetry and modeling work.
- URASpace: 3D modeling and physical rocket testing.
Team projects involve multiple contributors. These roles do not imply sole authorship or independent validation of every system.
ESP32-S3 learning firmware · GeekMagic PC telemetry · Leve11 · GAIA portal contributions · GAIA website contributions
I aim to preserve raw data, document engineering decisions, and distinguish implemented, experimental, documented and planned work. Software tests, simulation and physical validation are separate kinds of evidence.