Igor Maurell Email & Phone Number

State Of Rio Grande Do Sul, Brazil

Rio Grande Do Sul, Brazil

Rio Grande, Rio Grande Do Sul, Brazil

• Member of the Computer Vision team in a 𝐑&𝐃 𝐩𝐫𝐨𝐣𝐞𝐜𝐭 of 𝐏𝐞𝐭𝐫𝐨𝐛𝐫𝐚𝐬 company.• For the safety in the working environment of 𝐏𝐞𝐭𝐫𝐨𝐛𝐫𝐚𝐬 Floating Production Storage and Offloading (𝐅𝐏𝐒𝐎), the project aims to perform the 𝟑𝐃 𝐒𝐜𝐞𝐧𝐞 𝐔𝐧𝐝𝐞𝐫𝐬𝐭𝐚𝐧𝐝𝐢𝐧𝐠 of the laydown area to avoid collisions between the crane payloads and other structures and people.• Working with 𝟑𝐃 𝐬𝐭𝐞𝐫𝐞𝐨 𝐢𝐧𝐝𝐮𝐬𝐭𝐫𝐢𝐚𝐥 𝐜𝐚𝐦𝐞𝐫𝐚𝐬 and 𝟑𝐃 𝐋𝐢𝐃𝐀𝐑𝐬.• Designed a 𝐑𝐎𝐒𝟐 𝐩𝐞𝐫𝐜𝐞𝐩𝐭𝐢𝐨𝐧 𝐚𝐫𝐜𝐡𝐢𝐭𝐞𝐜𝐭𝐮𝐫𝐞 capable of 𝐬𝐞𝐦𝐚𝐧𝐭𝐢𝐜 𝐮𝐧𝐝𝐞𝐫𝐬𝐭𝐚𝐧𝐝𝐢𝐧𝐠 of the laydown scene through the 𝟑𝐃 𝐫𝐞𝐜𝐨𝐧𝐬𝐭𝐫𝐮𝐜𝐭𝐢𝐨𝐧 of 𝐢𝐦𝐚𝐠𝐞 𝐢𝐧𝐬𝐭𝐚𝐧𝐜𝐞 𝐬𝐞𝐠𝐦𝐞𝐧𝐭𝐚𝐭𝐢𝐨𝐧 using stereo cameras and 𝟑𝐃 𝐋𝐢𝐃𝐀𝐑 𝐏𝐨𝐢𝐧𝐭 𝐂𝐥𝐨𝐮𝐝 𝐬𝐞𝐠𝐦𝐞𝐧𝐭𝐚𝐭𝐢𝐨𝐧.

Rio Grande, Rio Grande Do Sul, Brazil

• 𝐑&𝐃 𝐩𝐫𝐨𝐣𝐞𝐜𝐭 aiming the development of a 𝐝𝐫𝐨𝐧𝐞-𝐛𝐚𝐬𝐞𝐝 solution to 𝐢𝐧𝐬𝐩𝐞𝐜𝐭 the metal tanks of 𝐏𝐞𝐭𝐫𝐨𝐛𝐫𝐚𝐬 Floating Production Storage and Offloading (𝐅𝐏𝐒𝐎) based on 𝐜𝐨𝐫𝐫𝐨𝐬𝐢𝐨𝐧 𝐝𝐞𝐭𝐞𝐜𝐭𝐢𝐨𝐧 and 𝐭𝐡𝐢𝐜𝐤𝐧𝐞𝐬𝐬 𝐦𝐞𝐚𝐬𝐮𝐫𝐞𝐦𝐞𝐧𝐭.• Solution assembly using a 𝐃𝐉𝐈 𝐌𝟑𝟎𝟎 drone with 𝐑𝐎𝐒 𝐃𝐉𝐈 𝐂++ 𝐒𝐃𝐊. (𝐑𝐎𝐒𝟏 and 𝐑𝐎𝐒𝟐)• Connection of an 𝐄𝐌𝐀𝐓 thickness sensor to the 𝐑𝐎𝐒 architecture to show the measures on 𝐑𝐕𝐢𝐳. • Improve 𝐮𝐬𝐚𝐛𝐢𝐥𝐢𝐭𝐲 by using DJI Remote Control to command the EMAT sensor measurement.• Added and 𝐟𝐚𝐮𝐥𝐭 𝐭𝐨𝐥𝐞𝐫𝐚𝐧𝐜𝐞 layer in the solution to make the 𝐄𝐌𝐀𝐓 docking system to automatic retreat, preventing the drone from getting stuck on the wall in case of communication failures.• Make the solution 𝐩𝐥𝐮𝐠-𝐚𝐧𝐝-𝐩𝐥𝐚𝐲 by running the 𝐑𝐎𝐒 architecture as multiple system services.• The solution is being prepared to patent and to be tested in its desired working environment by the company.

Rio Grande, Rio Grande Do Sul, Brazil

After 7 years as developer and team leader of the Fbot Robotics Team with experience in many areas that are necessary for development of robots, I continue in the team as an advisor, keeping contact with the members and helping to validate important project decisions to help then to develop better and more robust solutions.

Rio Grande, Rio Grande Do Sul, Brazil

Fbot@Home is a subteam from Fbot Robotics Team to compete in the Robocup@Home category. This league from Robocup Federation aims to establish a benchmark for comparison of General Purpose Service Robots in domestic environment from various teams around the world.After 3 years of experience in robotics development and competitions, I was chosen by Fbot@Home team to lead then in the next competitions to come. Being the team leader, while I continued to actively develop the robot's capabilities, I also was in charge of the strategic planning, team recruitment, monitoring members' tasks and other general management duties.During the 4 years being leader of the team we have developed the robot DoRIS, achieving the following results in this time:Awards• 🥉 3rd place in the Latin America Robotics Competition (LARC 2019) • 🥉 3rd place in the Latin America Robotics Competition (LARC 2020)• 🥈 2nd place in the Latin America Robotics Competition (LARC 2021)• 🥉 3rd place in the World Robotics Competition (Robocup Thailand 2022)• 🥇 1st place in the Latin America Robotics Competition (LARC 2022)

Rio Grande, Rio Grande Do Sul, Brazil

Fbot@Home is a subteam from Fbot Robotics Team to compete in the Robocup@Home category. This league from Robocup Federation aims to establish a benchmark for comparison of General Purpose Service Robots in domestic environment from various teams around the world.As an undergraduate student, I started working in the Fbot@home team as a computer vision developer, aiming to produce the ROS packages with the necessary compute vision requirements to be used during the tasks of the competition. The main robot vision challenges of the competition were:• Object detection• People detection• People recognition• People tracking• Obstacle detection• Detection from 2D to 3D

Rio Grande, Rio Grande Do Sul, Brazil

As my first contact with Robotics, start my journey being part of the 6 times latin-america champion team in the SSL Robocup League was important to cut corners for a fast development of my capabilities. In this project, the challenge was to build a system to command a team of small size robots to play soccer. This system was big AI software in full development using C++ and Qt, that began in data acquisition and management, passing by team strategical planning, robot tactics and skills execution, path planning, kinematics models, radio communication and some other important concepts about mobile robotics.In this project, I worked rebuilding the data management part of the system, adding an OpenGL visualizer to better understand the decisions of the system and the behaviors of each robot and a robot position filter layer to prevent that small error in the vision system to highly impact the decisions made by the AI. Also, I developed a solution to made the robots moving and PID control tests more easy, using a joystick to provide the velocities for the robots to move and the commands to kick the ball or dribbling using it. Furthermore, I made some adjustments on the AI strategies to adequate it to new competition rules and I worked in the code of the embedded micro controller of the robots, contributing in the PID control code efficiency using intern interruptions and with the use of pipes to improve the RF robots and AI communication.

Rio Grande Do Sul, Brazil

• Leader of the 3D data processing team (P4D) in a 𝐑&𝐃 𝐩𝐫𝐨𝐣𝐞𝐜𝐭 of 𝐏𝐞𝐭𝐫𝐨𝐛𝐫𝐚𝐬 company.• Aiming the automatic 𝐀𝐬 𝐁𝐮𝐢𝐥𝐭 of off-shore structures, the project developed a drone-based solution to collect point clouds of industrial sites.• Developing deep learning based solutions to find the underlying 𝐠𝐞𝐨𝐦𝐞𝐭𝐫𝐢𝐜 𝐩𝐫𝐢𝐦𝐢𝐭𝐢𝐯𝐞𝐬 on these large 𝐩𝐨𝐢𝐧𝐭 𝐜𝐥𝐨𝐮𝐝𝐬 to generate As Built 𝐂𝐀𝐃𝐬, as a reversing engineering process for large-scale structures.• Creation of a tool to generate synthetic industrial 𝐩𝐨𝐢𝐧𝐭 𝐜𝐥𝐨𝐮𝐝𝐬 with labeled geometries from 𝐁𝐑𝐞𝐩 𝐂𝐀𝐃 models using 𝐎𝐩𝐞𝐧 𝐂𝐚𝐬𝐜𝐚𝐝𝐞 𝐓𝐞𝐜𝐡𝐧𝐨𝐥𝐨𝐠𝐲 (𝐎𝐂𝐂𝐓), 𝐆𝐦𝐬𝐡, 𝐎𝐩𝐞𝐧𝟑𝐃, 𝐍𝐮𝐦𝐩𝐲 and 𝐏𝐂𝐋, working with some 3D data representations, such as: 𝐏𝐨𝐢𝐧𝐭 𝐂𝐥𝐨𝐮𝐝𝐬, 𝐌𝐞𝐬𝐡𝐞𝐬 and 𝐂𝐀𝐃𝐬.• Contribution to 𝐏𝐲𝐭𝐡𝐨𝐧𝐎𝐂𝐂, a Python version of OCCT, improving the time efficiency of the 𝐭𝐨𝐩𝐨𝐥𝐨𝐠𝐲 𝐞𝐱𝐩𝐥𝐨𝐫𝐚𝐭𝐢𝐨𝐧 𝐢𝐧 𝐁𝐑𝐞𝐩 𝐦𝐨𝐝𝐞𝐥𝐬 𝐛𝐲 𝟏𝟎𝟎𝟎 𝐭𝐢𝐦𝐞𝐬.• Adapted the deep learning based state-of-the-art methods to large-scale point clouds to train them in the generated dataset using 𝐏𝐲𝐓𝐨𝐫𝐜𝐡 and 𝐓𝐞𝐧𝐬𝐨𝐫𝐅𝐥𝐨𝐰. • Worked with 𝐩𝐨𝐢𝐧𝐭𝐧𝐞𝐭𝐬, 𝐠𝐫𝐚𝐩𝐡 𝐧𝐞𝐮𝐫𝐚𝐥 𝐧𝐞𝐭𝐰𝐨𝐫𝐤𝐬 (𝐆𝐍𝐍𝐬) and 𝐬𝐩𝐚𝐫𝐬𝐞 𝐜𝐨𝐧𝐯𝐨𝐥𝐮𝐭𝐢𝐨𝐧 𝐧𝐞𝐭𝐰𝐨𝐫𝐤𝐬.• Applied concepts like 𝐆𝐞𝐨𝐦𝐞𝐭𝐫𝐢𝐜 𝐃𝐞𝐞𝐩 𝐋𝐞𝐚𝐫𝐧𝐢𝐧𝐠, understanding the differences between euclidean and non-euclidean data domains and the specific symmetry relations of each representation.• Both the generated dataset and the adapted methods contributions are 𝐮𝐧𝐝𝐞𝐫 𝐫𝐞𝐯𝐢𝐞𝐰 for publication in 𝐡𝐢𝐠𝐡 𝐢𝐦𝐩𝐚𝐜𝐭 𝐟𝐚𝐜𝐭𝐨𝐫 𝐜𝐨𝐧𝐟𝐞𝐫𝐞𝐧𝐜𝐞𝐬.• The developed solutions were incorporated into the project web application and are being beta tested by the clients of the company.

Rio Grande, Rio Grande Do Sul, Brazil

• Part of the computer vision team in an 𝐑&𝐃 𝐩𝐫𝐨𝐣𝐞𝐜𝐭 of 𝐏𝐞𝐭𝐫𝐨𝐛𝐫𝐚𝐬 company to transform its scanned documents into text PDF’s.• Worked actively in fine-tuning of 𝐎𝐩𝐭𝐢𝐜𝐚𝐥 𝐂𝐡𝐚𝐫𝐚𝐜𝐭𝐞𝐫 𝐑𝐞𝐜𝐨𝐠𝐧𝐢𝐭𝐢𝐨𝐧 (𝐎𝐂𝐑) techniques, 𝐭𝐚𝐛𝐥𝐞 𝐝𝐞𝐭𝐞𝐜𝐭𝐢𝐨𝐧 and 𝐏𝐃𝐅 𝐠𝐞𝐧𝐞𝐫𝐚𝐭𝐢𝐨𝐧 • Implemented the computer vision functionalities as 𝐦𝐢𝐜𝐫𝐨𝐬𝐞𝐫𝐯𝐢𝐜𝐞𝐬 to be integrated in the project 𝐝𝐢𝐬𝐭𝐫𝐢𝐛𝐮𝐭𝐞𝐝 𝐬𝐲𝐬𝐭𝐞𝐦.

Rio Grande, Rio Grande Do Sul, Brazil

• Adapted a dense monocular 𝐓𝐫𝐚𝐜𝐤𝐢𝐧𝐠 𝐚𝐧𝐝 𝐌𝐚𝐩𝐩𝐢𝐧𝐠 (𝐓𝐀𝐌) method to be used in 𝐮𝐧𝐝𝐞𝐫𝐰𝐚𝐭𝐞𝐫 𝐞𝐧𝐯𝐢𝐫𝐨𝐧𝐦𝐞𝐧𝐭𝐬 through the modification of the 𝐩𝐡𝐨𝐭𝐨𝐦𝐞𝐭𝐫𝐢𝐜 𝐞𝐫𝐫𝐨𝐫 𝐞𝐪𝐮𝐚𝐭𝐢𝐨𝐧.• Developed concepts such as: 𝐑𝐎𝐒, 𝐎𝐩𝐞𝐧𝐂𝐕, 𝐜𝐚𝐦𝐞𝐫𝐚 𝐜𝐚𝐥𝐢𝐛𝐫𝐚𝐭𝐢𝐨𝐧, 𝐞𝐩𝐢𝐩𝐨𝐥𝐚𝐫 𝐠𝐞𝐨𝐦𝐞𝐭𝐫𝐲, 𝐯𝐢𝐬𝐮𝐚𝐥 𝐨𝐝𝐨𝐦𝐞𝐭𝐫𝐲, 𝐟𝐞𝐚𝐭𝐮𝐫𝐞 𝐝𝐞𝐭𝐞𝐜𝐭𝐢𝐨𝐧, 𝐟𝐞𝐚𝐭𝐮𝐫𝐞 𝐦𝐚𝐭𝐜𝐡𝐢𝐧𝐠 and 𝟑𝐃 𝐑𝐞𝐜𝐨𝐧𝐬𝐭𝐫𝐮𝐜𝐭𝐢𝐨𝐧.

Rio Grande, Rio Grande Do Sul, Brazil

• Part of the computer vision researchers team in a 𝐑&𝐃 𝐩𝐫𝐨𝐣𝐞𝐜𝐭 of 𝐏𝐞𝐭𝐫𝐨𝐛𝐫𝐚𝐬 company to automate a linear welding robot.• Developed an algorithm to analyze the welding process using 𝐌𝐚𝐭𝐥𝐚𝐛 through images of 𝐡𝐢𝐠𝐡-𝐬𝐩𝐞𝐞𝐝 𝐏𝐡𝐚𝐧𝐭𝐨𝐦 𝐜𝐚𝐦𝐞𝐫𝐚𝐬.

Rio Grande, Rio Grande Do Sul, Brazil

• Contributed to the development of the back-end of a system to describe scenes and objects for blind people using 𝐂++, 𝐐𝐭, 𝐎𝐩𝐞𝐧𝐂𝐕, 𝐂𝐚𝐟𝐟𝐞 and 𝐈𝐧𝐜𝐞𝐩𝐭𝐢𝐨𝐧.• Developed the front-end native application using 𝐀𝐧𝐝𝐫𝐨𝐢𝐝 𝐒𝐭𝐮𝐝𝐢𝐨 and 𝐉𝐚𝐯𝐚 integrated with 𝐆𝐨𝐨𝐠𝐥𝐞 𝐓𝐚𝐥𝐤𝐁𝐚𝐜𝐤.• Communicate back-end and front-end using a 𝑝𝑢𝑏𝑙𝑖𝑠ℎ/𝑠𝑢𝑏𝑠𝑐𝑟𝑖𝑏𝑒 architecture through 𝐌𝐐𝐓𝐓, which was a project requirement.

Master'S Degree, Mechatronics, Robotics, And Automation Engineering

Master Exchange, Artificial Intelligence

Integrated Master Exchange, Computer Engineering

Bachelor'S Degree, Computer Engineering

Técnico, Mechatronics, Robotics, And Automation Engineering