Hello I'm

Paulo Rezeck

I'm a robotics researcher, maker, and a PhD student in computer science!

More About Me

About

I'm currently working as a robotics researcher at FUNDEP/UFMG, acting as a technical leader to develop a camera-based system for analyzing human activities in confined spaces in the oil and gas industry, Petrobras. Moreover, I have worked in diverse research and development project, including cooperative magnetic-mapping using small UAVs and advanced teleoperation of excavators, both for in the mining industry, ITV/VALE.

Concurrently, I'm also pursuing a Ph.D. in the Computer Vision and Robotics Laboratory (VeRLab) at the Department of Computer Science of the Universidade Federal de Minas Gerais (UFMG) in Brazil. My thesis is on the emergence of intelligent behaviors in robotic swarms using dynamic Gibbs Random Fields.

My master's thesis was on the development of a small and affordable robot for robot swarm research and also for educational purposes.

Education

  • Ph.D., Computer Science, Federal University of Minas Gerais, Brazil, 2019-2022*
  • M.Sc., Computer Science, Federal University of Minas Gerais, Brazil, 2016-2019
  • M.Sc./Internship, Computer Science, University of Tuebingen, Germany, 2014-2015
  • B.Sc, Computer Science, Federal University of Minas Gerais, Brazil, 2011-2016

Interests

  • Multi-Robot Systems
  • Telerobotics for Mining Industry
  • Modeling and Construction of Robotic Systems
  • Robot Vision
  • SLAM, Robot Control and Navigation
  • DRL for Dynamic Control

Skills

  • Programming languages:
  • Development Tools:
  • Maker:

Research

Research

I integrate the group XQuad, the Latin American Finalist of the AlphaPilot Challenge

We performed at the top 9 from over 400 other teams from around the world. These nine teams will compete to design an AI framework capable of piloting racing drones through high-speed aerial courses without any GPS, data relay or human intervention. AlphaPilot teams will battle it out during AIRR's inaugural, four-event season later this year for a chance at winning a $1 million cash prize, sponsored by Lockheed Martin.

Source: DRL and Lockheed Martin 07/08/2019

Research

Advanced Teleoperation of Mining Equipment: Bulldozer

This project aims to investigate the teleoperation of mining equipment, focusing primarily on the bulldozer. In this context, teleoperation can be defined as equipment operation at a distance. It is very desirable that users can be able to interact with the remote environment naturally. In other words, the interaction should provide the user the sensation of being present at the site by offering information from different sensors, that is known as telepresence.

Source: VeRLab 01/08/2017

Research

HeRo: An Open Platform for Robotics Research and Education

In this project, we present HeRo, a novel small and capable platform for robotic research and education. Specially designed for swarm experimentation, which requires a large set of simpler robots, HeRo is a very low-cost alternative to other types of robotic platforms, without removing most capabilities and sensor abilities of commonly used or commercial platforms. HeRo is deeply integrated with the most used robotic framework available today, ROS.

Source: VeRLab 01/02/2017

Publications

Journal of Intelligent Service Robotics , 2021

An Integrated Solution for an Autonomous Drone Racing in Indoor Environments

Rezende, A. M. C., Miranda, V. R. F, Rezeck, P. A. F., Azpurua, H., Santos, E. R. S., Goncalves, V. M., Macharet, D. G., Freitas, G. M.
The use of drones is becoming more present in modern daily life. One of the most challenging tasks associated with these vehicles is the development of perception and autonomous navigation systems. Competitions such as Artificial Intelligence Robotic Racing (AIRR) and Autonomous Drone Racing were launched to drive the advances in such strategies, requiring the development of integrated systems for autonomous navigation in a racing environment. In this context, this paper presents an improved integrated solution for autonomous drone racing, which focuses on simple, robust, and computationally efficient techniques to enable the application in embedded hardware. The strategy is divided into four modules: (i) A trajectory planner computes a path that passes through the sequence of desired gates; (ii) a perception system that obtains the global pose of the vehicle by using an onboard camera; (iii) a localization system which merges several sensed information to estimate the drone’s states; and, (iv) an artificial vector field-based controller in charge of following the plan by using the estimated states. To evaluate the performance of the entire integrated system, we extended the FlightGoggles simulator to use gates similar to those used in the AIRR competition. A computational cost analysis demonstrates the high performance of the proposed perception method running on limited hardware commonly embedded in aerial robots. In addition, we evaluate the localization system by comparing it with ground truth and when there is a disturbance in the location of the gates. Results in a representative race circuit based on the AIRR competition showed the proposed strategy’s competing performance, presenting itself as a feasible solution for drone racing systems.

Journal of Intelligent and Robotic Systems, 2021

Cooperative Localization and Mapping with Robotic Swarms

Pires, A. G., Rezeck, P., Chaves, R. A., Machret, D. G., Chaimowicz, L.
The Cooperative Localization (CL) problem considers the case where groups of robots aim to improve their overall localization by sharing position estimates within the team instead of using landmarks in the environment. Despite being a well-studied problem, very few works deal with the increased complexity when a very large number of robots is used, as is the case in robotic swarms. In this paper, we propose a methodology to perform cooperative localization in robotic swarms while they navigate through the environment. A collective motion strategy maintains the group’s cohesion, which allows each robot to perform the localization using information from its immediate neighbors. The method is based on the Covariance Intersection (CI) algorithm, which is employed to perform the localization in a decentralized way. Experiments in both simulation and real-world scenarios show the feasibility of the proposed approach. Furthermore, it overcomes and has a reduced space usage and time complexity compared to a traditional centralized EKF-based method. We also investigate how the methodology can be used by a robotic swarm to build occupancy grid maps, a task that generally requires more sophisticated robots and is heavily dependent on good localization estimates.

2021 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)

Cooperative Object Transportation using Gibbs Random Fields

Rezeck, P., Assuncao, R. M., Chaimowicz, L.
This paper presents a novel methodology that allows a swarm of robots to perform a cooperative transportation task. Our approach consists of modeling the swarm as a Gibbs Random Field (GRF), taking advantage of this framework's locality properties. By setting appropriate potential functions, robots can dynamically navigate, form groups, and perform co- operative transportation in a completely decentralized fashion. Moreover, these behaviors emerge from the local interactions without the need for explicit communication or coordination. To evaluate our methodology, we perform a series of simulations and proof-of-concept experiments in different scenarios. Our results show that the method is scalable, adaptable, and robust to failures and changes in the environment.

2021 IEEE International Conference on Robotics and Automation (ICRA)

Flocking-Segregative Swarming Behaviors using Gibbs Random Fields

Rezeck, P., Assuncao, R. M., Chaimowicz, L.
This paper presents a novel approach that allows a swarm of heterogeneous robots to produce simultaneously segregative and flocking behaviors using only local sensing. These behaviors have been widely studied in swarm robotics and their combination allows the execution of several complex tasks. Our approach consists of modeling the swarm as a Gibbs Random Field (GRF) and using appropriate potential functions to reach segregation, cohesion and consensus on the velocity of the swarm. Simulations and proof-of-concept experiments using real robots are presented to evaluate the performance of our methodology in comparison to some of the state-of-the-art works that tackle segregative behaviors.

Journal Swarm Intelligence, 2020

Spatial Segregative Behaviors in Robotic Swarms Using Differential Potentials

Santos, V. G., Pires, A. G., Alitappeh, R. J., Rezeck, P., Pimenta, L. C. A., Macharet, D. G., Chaimowicz, L.
Segregative behaviors, in which individuals with common characteristics are placed together and set apart from other groups, are commonly found in nature. In swarm robotics, these behaviors can be important in different tasks that require a heterogeneous group of robots to be divided in homogeneous sets according to their physical (sensors, actuators) or logical (algorithms) capabilities. In this paper, we propose a controller that can spatially segregate a swarm of robots in two specific ways: clusters and concentric rings. By segregation, we mean that the swarm is partitioned in groups, with similar robots belonging to a same group, and these groups are spatially separated from each other. We achieve this by adapting and extending the differential potential concept, an abstraction of the mechanism by which cells achieve segregation. We present stability analysis and perform simulated experiments in 2D and 3D spaces in order to show the robustness of the proposed controller. Experiments with a limited number of real robots are also presented as a proof of concept. Results show that our approach allows a swarm of heterogeneous robots to segregate in a stable, compact, and collision-free fashion.

Journal of Intelligent and Robotic Systems, 2020

Localization Using Ultra Wideband and IEEE 802.15.4 Radios with Nonlinear Bayesian Filters: a Comparative Study

Santos, E. R. S., Azpurua, H., Rezeck, P., Correa, M. F. S., Vieira, M. A. M., Freitas, G. M., Macharet, D. G.
In this work, we study the localization problem considering two Wireless Sensor Network (WSN) metrics commonly used for distance estimation. In this sense, we use Bayesian filters to combine odometry and distance estimations provided by the WSN devices. Our strategies aim at a general application and can be used for both indoor and outdoor environments depending only on the type of metric and radio technology employed. In this work, we investigate two metrics: Ultra Wideband (UWB) and the Received Signal Strength (RSS). The first one is a recent technology and presents better overall accuracy than other metrics, and the second is one of the most well-explored metrics in WSN-based localization approaches. We evaluated the performance of our two approaches and compare them with the Decawave® built-in application. The experiments were performed in simulated and real environments with different scenarios (indoors and outdoors) and sensor configurations. The results show the proposed strategies feasibility by improving the localization accuracy for both types of environments. In indoor environments, the proposed system has a mean position error bellow 0.09 meters and mean orientation error of 0.08 rads. Furthermore, the proposed system is in average 0.03 meters more accurate than the built-in application.

2020 International Conference on Automation Science and Engineering (CASE)

On the Evaluation of Force Feedback Augmented Teleoperation of Excavator-like Mobile Manipulators

Carvalho, L., Rezeck, P., Lima, M., Pinto, L., Freitas, G., Nascimento, E., Macharet, D., Chaimowicz, L., Pessin, G., Campos, M.
This paper reports the results of the investigation on the use of force feedback to assist a human operator in the teleoperation of open-pit mining excavators. The proposed force feedback strategy allows object interactions. Also, it enables the operator to sense attractive forces toward the excavation/deposit site as a guidance system, especially useful in precision mining. The methodology was evaluated with multiple experiments to compare the use of haptic feedback and the guidance system. Moreover, a group of volunteers performed the NASA Task Load Index to measure their efforts to teleoperate an integrated off-the-shelf excavator-like mobile robot using different scenarios. Results showed that the force feedback decreases by 8% the power consumption of the excavator, which can represent a significant cost reduction in real applications. Furthermore, the proposed system proves to be a useful tool to the operator allowing task assistance and safety in interaction with obstacles, which can mitigate up to 70% of the stress on the manipulator when there is direct contact with surfaces, e.g., manipulation of rocks and boulders.

2019 IEEE International Conference on Advanced Robotics (ICAR)

SwarMap: Occupancy Grid Mapping with a Robotic Swarm

Chaves, R., Rezeck, P., Chaimowicz, L.
In recent years, robotic swarms have been studied as an alternative to replace complex and expensive robots in the execution of different types of tasks. In this paper, we propose an approach to cooperatively build occupancy grid maps using large groups of simpler robots. Robots randomly explore the environment and update a shared occupancy grid stored in the cloud using Bayesian Filters. They localize themselves based on odometry and use pair-to-pair neighborhood communication to exchange pose information and reduce the uncertainty through a Kalman Filter. Simulated and real experiments are performed to show the effectiveness and scalability of the proposed approach.

Journal of Field Robotics, 2019

Cooperative Digital Magnetic‐elevation Maps by small Autonomous Aerial Robots

Azpúrua, H., Potje, G. A., Rezeck, P. A., Freitas, G. M., Uzeda Garcia, L. G., Nascimento, E. R., ... & Campos, M. F.
One of the steps to provide fundamental data for planning a mining effort is the magnetic surveying of a target area, which is typically carried out by conventional aircraft campaigns. However, besides the high cost, fixed-wing aerial vehicles present shortcomings especially for drape flights on mountainous regions, where steep slopes are often present. Traditional human-crewed flights have to perform tedious and dangerous trajectories, under strict velocity and attitude constraints. In this paper, we deal with the problem of accomplishing digital magnetic-elevation maps using autonomous and cooperative aerial robots. The proposed approach for autonomous mapping utilizes a custom-built fluxgate sensor and off the shelf cameras adapted for small airborne platforms. We also propose an innovative approach for generating a digital magnetic-elevation model from the gathered data. Our method was evaluated and validated in field tests in an industrial scenario to detect scrap metals in ore piles. Results show that the proposed method could reliably detect magnetic anomalies while generating accurate three-dimensional magnetic maps.

14º Simpósio Brasileiro de Automação Inteligente, 2019

Realimentaçao de Força para Teleoperaçao de Escavadeiras

Lima, M., Cid, A., Cadar, F., Pinto, L., Rezeck, P., Grabe, T., Freitas, G., Macharet, D., Pessin, G., Chaimowicz, L., Campos, M.
É crescente o interesse em ambientes com melhores condições de trabalho, evitando locais insalubres e perigosos. Nesse contexto, a teleoperação aparece com objetivo de facilitar operações remotas e melhorar as capacidades humanas para a manipulação de equipamentos. A indústria de mineração é uma área em que a teleoperação pode ser aplicada, por exemplo, para teleoperar equipamentos como escavadeiras. Este artigo tem como objetivo o desenvolvimento de uma arcabouço para a teleoperação do braço de uma escavadeira com realimentação de força para o usuário. A metodologia foi avaliada através da simulação de tarefas de carregamento. Os resultados mostram que o arcabouço é intuitivo e que a realimentação de força pode melhora a realização da tarefa ao promover uma realimentação imersiva ao operador, incluindo sensações de carga na caçamba e de forças de interação com obstáculos - esta informação pode levar a uma significante redução do estresse ao qual o manipulador submetido.

2018 IEEE Latin American Robotics Symposium (LARS)

Global Localization of Mobile Robots Using Local Position Estimation in a Geo Tagged Wireless Node Sensor Network

Santos, E., Azpurua, H., Rezeck, P., Corrêa, M., Freitas, G., & Macharet, D.
Precise localization of mobile robots in uncertain environments is a fundamental and crucial issue in robotics. In this paper, we propose an outdoor robot localization strategy using an extended Kalman filter (EKF). The method relies on the received signal strength (RSS) produced by wireless nodes, wheel odometry, and GPS to estimate a precise global position of the robot. The proposed method was evaluated using simulated experiments with different scenarios and sensors configurations. The results have shown the feasibility of the methodology.

2018 IEEE Latin American Robotics Symposium (LARS)

Framework for Haptic Teleoperation of a Remote Robotic Arm Device

Rezeck, P., Frade, B., Soares, J., Pinto, L., Cadar, F., Azpurua, H., ... & Campos, M. F.
Considering the increasing use of sophisticated machinery in remote and hazardous environments, the field of teleoperation aim to improve the human capabilities for the remote manipulation of equipment by providing a human operator with the necessary information to endow the interaction with the remote location as transparent as possible. In this context, a haptic interface allows an operator to perceive the remote environment via tactile feedback or other bodily sensations. Furthermore, it is bi-directional, been capable of both reading and writing input from and to the human user. In this work, we propose a framework for the remote operation of a servo motor robotic arm using haptic feedback. The objective is to provide the sense or presence through a tactile interface using torque feedback from the device motors and a 3D visualization of the arm in real time. As a result, we observed how the load carried by the manipulator during the path impacted accuracy and precision. Also, we perceived the influence of the torque variation of the motors causing users fatigue and mitigated this problem by applying a complementary filter in the force feedback.

2018 IEEE Latin American Robotics Symposium (LARS)

An Immersion Enhancing Robotic Head-Like Device for Teleoperation

Rezeck, P., Cadar, F., Soares, J., Frade, B., Pinto, L., Azpurua, H., ... & Campos, M. F. M.
Teleoperation is a research topic which focuses on the extension of human capabilities to manipulate objects and control equipment remotely. It has been used in different contexts and applications such as nuclear or chemical industry, disarming of explosives, search and rescue missions and space robotics. In this context, the human operator may perceive the environment through different sensory information, such as vision, hearing, and touch. However, vision remains as one of the most important, since it allows a deep immersion feeling. In this paper, we proposed an immersive system for stereoscopic visualization coupled with a robotic head-like device that serves as a proof of concept framework for future applications in the mining environment. For instance, exploration and open-cut mining tasks can be highly benefited from teleoperation, considering that this kind of environment is too hazardous for humans to work. Our proposal provides depth information to the user in real time and minimizes discomfort in the visualization. Experiments with real users indicate improved performance in manipulation, depth perception, comfort and response time.

2017 IEEE Latin American Robotics Symposium (LARS)

HeRo: An Open Platform for Robotics Research and Education

Rezeck, P. A., Azpurua, H., & Chaimowicz, L.
In this paper, we present a novel platform for swarm robotics that is low cost, easy to assemble using off the shelf components and is deeply integrated with the most used robotic framework available today: ROS (Robot Operating System). The robotic platform is completely open, composed by a 3D printed body and open source software. We describe its architecture, present its main features and evaluate its functionalities executing real experiments using a couple of robots. We concluded that the proposed swarm platform is a feasible and usable system for education and research, given its reduced cost, easy of use and small size.

Journal Sensors, 2016

Autonomous Aeromagnetic Surveys Using a Fluxgate Magnetometer

Macharet, D., Perez-Imaz, H., Rezeck, P., Potje, G., Benyosef, L., Wiermann, A., ... & Campos, M.
Recent advances in the research of autonomous vehicles have showed a vast range of applications, such as exploration, surveillance and environmental monitoring. Considering the mining industry, it is possible to use such vehicles in the prospection of minerals of commercial interest beneath the ground. However, tasks such as geophysical surveys are highly dependent on specific sensors, which mostly are not designed to be used in these new range of autonomous vehicles. In this work, we propose a novel magnetic survey pipeline that aims to increase versatility, speed and robustness by using autonomous rotary-wing Unmanned Aerial Vehicles (UAVs). We also discuss the development of a state-of-the-art three-axis fluxgate, where our goal in this work was to refine and adjust the sensor topology and coupled electronics specifically for this type of vehicle and application. The sensor was built with two ring-cores using a specially developed stress-annealed CoFeSiB amorphous ribbon, in order to get sufficient resolution to detect concentrations of small ferrous minerals. Finally, we report on the results of experiments performed with a real UAV in an outdoor environment, showing the efficacy of the methodology in detecting an artificial ferrous anomaly.

ASME 2016 Dynamic Systems and Control Conference (DSCC)

Segregation of Heterogeneous Robotics Swarms via Convex Optimization

Edwards, V., Rezeck, P., Chaimowicz, L., & Hsieh, M. A.
The division of labor amongst a heterogeneous swarm of robots increases the range and sophistication of the tasks the swarm can accomplish. To efficiently execute a task the swarm of robots must have some starting organization. Over the past decade segregation of robotic swarms has grown as a field of research drawing inspiration from natural phenomena such as cellular segregation. A variety of different approaches have been undertaken to devise control methods to organize a heterogeneous swarm of robots. In this work, we present a convex optimization approach to segregate a heterogeneous swarm into a set of homogeneous collectives. We present theoretical results that show our approach is guaranteed to achieve complete segregation and validate our strategy in simulation and experiments.

2016 IEEE International Conference on Automation Science and Engineering (CASE)

Multi-robot 3D Coverage Path Planning for First Responders teams

Perez-Imaz, H. I., Rezeck, P. A., Macharet, D. G., & Campos, M. F.
It is known that the first hours after a disaster are critical to maximize the rescue and recovery of victims. However, in disaster scenarios, with the communication infrastructure damaged and cluttered environment, first responders can be in disadvantage without proper communication and global information about the mission. In this sense, the use of aerial robots have a clear advantage over other kind of robots, since they are able to navigate over large areas faster and with a privileged view from above. This paper proposes a methodology for three-dimensional area coverage using multiple small UAVs based upon a customized cell decomposition algorithm using regular hexagons. The approach has been thoroughly evaluated in different simulated scenarios. We also report on the results of experiments performed with real UAVs in an outdoor environment.

2013 IEEE Latin American Robotics Symposium (LARS)

On the Development of a Robotic System for Telepresence

Rezeck, P. A., Vieira, M. A., Chaimowicz, L., & Campos, M. F.
In a large number of tasks, such as exploration, inspection and surgery, robots are being used to replace or represent humans in remote environments. To investigate the challenges of this emerging research field, we developed a robotic system for telepresence. It is composed of two main modules: a robotic avatar, which is able to represent a human in a remote environment and a user interface that allows a remote operator to task and control the avatar. Through the use of cameras, microphones and other sensors on the avatar, the operator can assess the environment and interact with other humans in the remote area. The robot can execute commands given by the operator and also act autonomously in certain tasks such as navigation and obstacle avoidance. In this paper, we describe the main components of the system in terms of hardware and software and present some proof-of-concept experiments to demonstrate its capabilities.