We are Gentlebots,
a team of researchers in Robotics
from Rey Juan Carlos University
and University of León
experienced in participating in several
Robotics Competitions.

Social
Robotics


Proxemics
People Mood Perception
Social Navigation

Perception


3D Objects and People Detection
Probabilistic Mapping
Visual Attention

Human-Robot
Interaction


Multi-modal communication
Dialogue
Screen and QR communication

Cognitive
Architecture


Behaviour Generation
BICA
PDDL Planning

Introduction

Our history

Gentlebots team is composed of two universities and is focused on software development that allows robots to exhibit intelligent behaviors. For this reason, the majority of the participations in competitions have been in leagues whose robot is standard for all the participants of the league, where the focus is in the programming of the algorithms. Our team has a great tradition of participating in robotics competitions:

The scientific production related to these participations is summarized in 5 doctoral theses, 20 articles in high impact journals, and more than 40 articles in national and international congresses. It is an excellent performance for a team that has never exceeded 10 members at a time.

We are involved in organizing @Home competitions. Francisco J. Rodrı́guez is a member of the Organization Committee (OC) of RoboCup@Home. Vicente Matellán is one of the organizers of the ERL, making available to this competition an officially approved domestic environment of the league within his laboratory in Leon. He has already organized two editions of this competition.

Focusing on the @Home period, since 2012, our team has always developed open source software using ROS. Our developments have always been released as ROS packages and made available to the RoboCup community.

Members

Many people. One team.

Permanents


Francisco J. Rodríguez Lera


Doctor Engineer in Robotics
University of León

fjrodl@unileon.es

Francisco Martín Rico


Doctor Engineer in Robotics
Rey Juan Carlos University

francisco.rico@urjc.es

Vicente Matellán Olivera


Doctor Engineer in Robotics
University of León

vmato@unileon.es



Current non-permanent


Claudia Álvarez Aparicio


Doctor Engineer in Robotics
University of León

calvaa@unileon.es

José Miguel Guerrero


Doctor Engineer in Robotics
Rey Juan Carlos University

josemiguel.guerrero@urjc.es

Rodrigo Pérez Rodríguez


Doctor Engineer in Telematics
Rey Juan Carlos University

rodrigo.perez@urjc.es

Miguel Ángel Gonzalez Santamarta


MSc in Computer Science &
PhD Candidate
University of León

mgons@unileon.es

Juan Diego Peña Narváez


MSc in Robotics &
PhD Candidate
Rey Juan Carlos University

juan.pena@urjc.es

Juan Carlos Manzanares Serrano


Degree in Robotics Software
(Student)
Rey Juan Carlos University

jc.manzanares.2018@alumnos.urjc.es

Irene González Fernández


MSc in Robotics
University of León

igonf@unileon.es

David Sobrín Hidalgo


MSc in Cybersecurity &
PhD Candidate
University of León

dsobh@unileon.es



Gone but not forgotten


Jonatan Ginés Clavero
Rey Juan Carlos University
Ángel Manuel Guerrero Higueras
University of León
Pablo Castellanos López
Rey Juan Carlos University
Fernando González Ramos
Rey Juan Carlos University
Ivan Lopez Broceño
Rey Juan Carlos University
Manuel Fernandez Carmona
Rey Juan Carlos University
David Vargas Frutos
Rey Juan Carlos University
Lorena Bajo Rebollo
Rey Juan Carlos University
Daniel Arévalo Rodrigo
Rey Juan Carlos University
María Belén Tualombo Chimbo
Rey Juan Carlos University
Pablo Moreno
Rey Juan Carlos University
Miguel Alamillo Reguero
Rey Juan Carlos University
Victor Martín
Rey Juan Carlos University
Carlos E. Agüero
Rey Juan Carlos University
Fernando Casado
University of León
Jesús Balsa
University of León
Miguel Cazorla Quevedo
University of Alicante
Boyán Bonev
University of Alicante
David Herrero
University of Murcia
Humberto Barbera
University of Murcia
Domeneç Puig
Rovira i Virgili University
Tomás González
Rovira i Virgili University
Álvaro Moreno
Rey Juan Carlos University
Javier Gutierrez-Maturana
Rey Juan Carlos University

Development

Cognitive Architecture


In the implementation of our cognitive architecture, we use two main elements: ROSPlan and BICA. ROSPlan is an IA planning framework, use popf among other planners. BICA is a toolbox to create control solutions for robots. Virtually all the elements of our design are BICA components that perform different functions and is mapped to a ROS node. These components are executed concurrently in a hierarchical way in order to generate complex behaviors.

Perception


3D Detection
We have developed a perception system based on Deep Learning. We use YOLO2 to detect objects and people in images. This software is a C/C++ implementation of darknet, which is a convolutional neural network. This software is fully integrated in ROS through the darknet_ros package. The output is a list of bounding boxes labeled with the class and probability of the object detected.

We have developed Darknet3D. This component receives the 2D bounding boxes and a point cloud, and then generates 3D bounding boxes. For this purpose, we use dedicated hardware (Jetson TX2) in order to avoid the overload of the robot computer.

Visual Attention
When our robot wants to detect objects, or track an element, it uses our visual attention system. This system is controlled using and updating the knowledge graph. For each type of target, there is a specialized care component. Each specialized care component provides 3D points that the robot must attend. The attention system visits in round robin every 3D point that is provided.

Social Navigation


Starting from the standard ROS navigation package, we have modified the map and location system to have a system that, based on a map of walls and doors, adds the permanent objects that the robot perceives during its operation. This is called long-term navigation, and is based on the fusion of several maps that reflect the presence of obstacles over time. We use perception of both the laser and the 3D camera of the robot to perceive the obstacles.

Our approach builds a static map starting from the construction plans of a building. A long-term map is started from the static map, and updated when adding and removing furniture, or when doors are opened or closed. A short-term map represents dynamic obstacles such as people. This approach is appropriate for fast deployment and long-term operations in office or domestic environments, able to adapt to changes in the environment. We combine our long-term mapping approach with a social navigation algorithm for robots that is acceptable to people. Robots will detect both the personal areas of humans and their areas of activity, to carry out their tasks, generating navigation routes that have less impact on human activities. The main novelty of this approach is that the robot will perceive moods of people, to adjust the size of proxemic areas.





Human-Robot Interaction


Our robot performs a multi-modal communication with humans. Use three simultaneous forms of communication to avoid communication errors that may block the robot at some point.

Dialogue generation
The main form of communication is through a natural voice dialogue. Our system uses DialogFlow and Google Speech. DialogFlow has a web interface where, as in a chatbot, instances, contexts, and the elements necessary for a sophisticated dialogue are encoded. It is the only component of our software that requires access to a service outside the robot. This is why we have added the other communication systems, to avoid relying on wireless connectivity.


Alternative Ways for Communication
The second form of communication is the touch screen. In addition to the state of the robot, the text that the robot is saying is shown. The third form of communication is through phrases and words coded in a deck of QR symbols.

Context Awareness
Context-awareness is a fundamental component in human robot interaction. Understanding the user activity context the robot improves the overall decision-making process. This is because, given a context, the robot can reduce the set of feasible actions. We consider that using an on-board microphone and gathering environmental sounds we can improve the context recognition. For this purpose, we have designed, developed and tested a computational lightweight Environment Recognition Component (ERC). In this iteration our ERC is able to recognize a set of four home bells. This component provides information to a Context-Awareness Component (CAC) that implements a hierarchical Bayesian network to tag user’s activities based on the American Occupational Therapy Association (AOTA) classification.

ROS2, Nav2 and ROS Security


We are Nav2 active contributors, the ROS navigation stack evolution for ROS2, and we were part of one of the IROS Best RoboCup Paper Award finalists, The Marathon2.


We are planning to develop some of the modules in ROS2 in order to an efficient and convenient distribution of the components. Since our goal is to use our systems in real environments, we will implemented our software with security mechanisms that prevent intrusions, attacks and not allowed manipulations.

Publications



In @Home competitions



Defining Adaptive Proxemic Zones for Activity-Aware Navigation

Clavero, J. G., Rico, Francisco M., Rodríguez-Lera, Francisco J., Hernández, J. M. G., & Olivera, V. M. (2020, November). In Workshop of Physical Agents (pp. 3-17). Springer, Cham.

The Marathon 2: A Navigation System

Steve Macenski, Francisco Martín, Ruffin White, Jonatan Ginés. 2020 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS 2020), October 2020.

Show 26 more




In SPL competitions



Effective visual attention for behavior-based robotic applications

Francisco Martín, Luis Rubio, Carlos E. Agüero y José M. Cañas Workshop of Physical Agents 2012. Madrid (Spain), Sept 2013.

Effective real-time visual object detection

Francisco Martín and Manuela Veloso. Progress in Artificial Intelligence. Springer. Volume 1, Issue 4. pp 259-265. 2012.

Show 15 more




In 4-legged competitions



Portable autonomous walk calibration for 4-legged robots

Boyan Bonev, Miguel Cazorla, Francisco Martín,Vicente Matellán. Applied Intelligence, Volume 36, Number 1, pp. 136-147. DOI 10.1007/s10489-010-0249-9 ISI Impact Factor: 1.936 (2012) Q2

Visual Based Localization of a Legged Robot with a topological representation

Francisco Martín Rico, José María Cañas Plaza, Vicente Matellán. In ''Robot Localization and Map Building'',Hanafiah Yussof (Ed.), INTECH,. March 2010. DOI:10.5772/9261

Show 28 more