Portfolio

Objective: Allow real aerial vehicles to execute autonomous navigation in indoor environments, following simulated paths created by generative models.
Contribution:
- Developed a waypoint based navigation system for indoor drones fusing ORB-SLAM3 and inertial state estimation
- Integrated the navigation system with ROS packages and Airsim trajectory plans.
- Implemented an user interface to control the drone and visualize the navigation system.
Results:
- The drone was able to navigate autonomously in a 3D environment, avoiding obstacles and following a pre-defined paths.
- The system was deployed and featured in a official Microsoft Demo

Objective: Indoor course: Develop an autonomous micro drone that, under the context of an 4.0 industry, is capable of performing several tasks such as:
Autonomously navigate between and through shelves, register the inventory of products, pickup a box and deploy it at a specific spot.
Outdoor course: Develop a drone capable of autonomously deliver packages of different sizes, detect rooftops, create two dimensions reconstructed maps.
Contribution:
- Indoor team Captain
- Developed the navigation system for the indoor drone using Monocular ORB SLAM combined with odometry data.
- Developed the computer vision and AI solutions for onboard and offboard control.

Objective: Develop an autonomous Arduino-controlled robot, with dimensions smaller than 25x25x25 cm, capable of: follow a broken path; make route decisions based on color signals; dodge obstacles; locate and rescue 5mm diameter spheres.
Contribution:
- Managed the research and development process in the areas of computing, electronics and mechanics.
- Led a team of four team mates during the development period of both competitions (team technical training, deadline management, sponsorship and project promotion)
Results:
- In both years the robots met the requirements.
- In 2016, we held the 1st position (among 40 teams) in the state of São Paulo competition stage.
- In 2015 and 2016 we received the award for best private school team in the state of São Paulo.
- The team represented the state of São Paulo in Uberlândia (2015) and Recife (2016)

Partnership between USP and Aalto University (Finland). Budget: €10.000 sponsored by SAAB.
Objective: Create an underwater localization and communication system to support technical divers in activities requiring self-localization.
Contribution:
- Technical Leadership
- Responsible for the electronic design
- Development of the communication and localization embedded software.
Results:
- Successfully tested proving the concept of data transmission and acoustic triangulation using frequency-modulated acoustic waves and multi-frequency interferometry.
- The project was the cover of the Finnish magazine “Metallitekniikka”

Objective: Develop a 3D printed robotic hand capable of mimic the amount of fingers of a captured image, displayed in front of it.
Contribution:
- Developed the mechanical design.
- Coordinated the elaboration of the software
Results:
- The algorithm used (Convolutional Neural Network) correctly classified the number of fingers in 95% of the test dataset.
- Correctly copied the number of fingers presented in front of camera in a time less than 200ms.

The project is being developed in a partnership with project Tamar.
Objective: Apply concepts of distributed networks and swarm intelligence in low-cost embedded systems for sea turtle life monitoring and organic sensing.
Contribution:
- Responsable for creating the concept of the solution
- Implemneted 2 simulators to validate the concept
- Developed of embedded eletronic, sofware system and web tracking interface
Results:
- The system was validated in simulation.
- The communication system was validated in real experiments
- The system was deployed in real turtle in a controlled environment

In partnership with Fleury Laboratory
Objective: Build a semi-industrial machine that automatizes the process of separating and classifying petri dishes with bacterial cultures, according to the bacterial growing state.
Contribution:
- Petri dishes barcode reader
- Classification algorithm (data enhancement + preprocessing + CNN)
Results: The classification algorithm was able to perform well even under limited available training data (dataset with 79 images and 4 classes). The maximum accuracy was of 97.7% on the trainset and 94.7% on testset.

Objective: Design and build an Arduino based line-following cart capable of, when an obstacle in the way is detected, make a sudden and precise braking and stop less than 5 cm from the object.
Contribution:
- Led a team of six in project execution
- Subdivided the team in three development areas: mechanics, control and sensing.
- Led the process of capturing the team's technical knowledge.
- Integrated results produced by the areas
Results:
- Met all requirements
- The project held the 1st place in the competition (out of 11 teams)

Objective: Design and Build a Radio Controlled Robot for cucumbers and lettuces picking in Vertical Farms using “raw materials“ and “basic manufacture techniques”. Team of 3 interdisciplinary students from Brazil, Argentina and Peru
Results: First Place at the competition

Partnership with brazilian federal police
Objective: Design and prototype a machine to destroy large volumes of smuggled cigarettes seized by the brazilian federal police.
Contribution: - mechanical and electronic design

Objective: Model, simulate and produce a mechanical automaton design. Designed in CAD / CAM software NX10.0 and manufactured with 2 laser cutted 3mm and 6mm MDF boards. The mechanism should be able to be continuously moved by only one torque source.
Contribution:
- Led and oriented a team of 4 people in the development of the project.
- Designed the mechanisms and visual of the project.
- Integrated the mechanical elements.
Results:
- The project met the requirements
- We won the best project award in 2017 (out of 200 teams)

Built a mini USB plotter with spare DVD parts for quick PCB marking and etching

Built a 3D printer with 0.1mm precision and 26 x 26 x 28 cm of usable area, with a budget restricted to $80,00 (USD).

Smaller projects developed between the ages of 14 and 16 years old