ROBOT2022: Fifth Iberian Robotics Conference : Advances in Robotics, Volume 2 / edited by Danilo Tardioli, Vicente Matellán, Guillermo Heredia, Manuel F. Silva, Lino Marques.

Format:

Book

Contributor:

Tardioli, Danilo, editor.

Series:

Lecture Notes in Networks and Systems, 2367-3389 ; 590

Language:

English

Subjects (All):

Robotics.

Automation.

Computational intelligence.

Computational Intelligence.

Local Subjects:

Robotics.

Automation.

Computational Intelligence.

Physical Description:

1 online resource (634 pages)

Edition:

1st ed. 2023.

Place of Publication:

Cham : Springer International Publishing : Imprint: Springer, 2023.

Summary:

This book contains a selection of papers accepted for presentation and discussion at ROBOT 2022—Fifth Iberian Robotics Conference, held in Zaragoza, Spain, on November 23-25, 2022. ROBOT 2022 is part of a series of conferences that are a joint organization of SEIDROB—Sociedad Española para la Investigación y Desarrollo en Robótica/Spanish Society for Research and Development in Robotics, and SPR—Sociedade Portuguesa de Robótica/Portuguese Society for Robotic. ROBOT 2022 builds upon several previous successful events, including three biennial workshops and the four previous editions of the Iberian Robotics Conference, and is focused on presenting the research and development of new applications, on the field of Robotics, in the Iberian Peninsula, although open to research and delegates from other countries. ROBOT 2022 featured four plenary talks on state-of-the-art subjects on robotics and 15 special sessions, plus a main/general robotics track. In total, after a careful review process, 98 high-quality papers were selected for publication, with a total of 219 unique authors, from 22 countries. .

Contents:

Intro

Preface

Organization

Local Organizing Committee

Program Committee Manager

Publicity Chairs

Program Committee

Additional Reviewers

Contents

Intelligent Robotics

Deep Learning for Victims Detection from Virtual and Real Search and Rescue Environments

1 Introduction

2 Materials and Methods

2.1 Materials

2.2 Reconstructed Environments

2.3 Convolutional Neural Network

3 Experiments and Results

3.1 Trained CNN Metrics

3.2 Cross Evaluation of Models for Victim Detection

4 Conclusions

References

Full-stack S-DOVS: Autonomous Navigation in Complete Real-World Dynamic Scenarios

2 Background

2.1 Related Work

2.2 Dynamic Object Velocity Space (DOVS)

2.3 Contribution

3 Approach

3.1 Decentralization and ROS

3.2 Localization

3.3 Global Planner

3.4 Obstacle Tracker

3.5 Modifications in S-DOVS

4 Experimental Results

4.1 Simulation Experiments

4.2 Real-world Experiments

5 Conclusion

Artificial Stupidity in Robotics: Something Unwanted or Somehow Useful?

2 Artificial Stupidity in Robotics

2.1 AS by Accident

2.2 AS by Design

3 Definition of the Problem

4 Analysis of Methods

5 Conclusions

Machine Learning in Robotics

Hand Gestures Recognition for an Intelligent Wheelchair Steering Command

2 State of the Art

2.1 Hand Gestures

2.2 Leap Motion

3 Methodology

3.1 Data Set Description

3.2 Data Division

3.3 Feature Extraction

3.4 Feature Selection and Dimensionality Reduction

3.5 Classification

3.6 Application

4 Results and Discussion

4.1 Data Set

4.2 Real Time Application

Object Segmentation for Bin Picking Using Deep Learning

2 Related Work.

3 Deep Learning Training

3.1 Dataset Generation

4 Proposed Bin Picking System

4.1 2D Object Detection

4.2 3D Point Cloud Segmentation

4.3 Pose Estimation and Grasping

5 System Tests and Evaluation

5.1 Object Detection and Segmentation Efficiency and Performance

6 Discussion and Future Work

Robot Navigation Anticipative Strategies in Deep Reinforcement Motion Planning

2 Related Work

2.1 AutoRL for Robot Navigation

2.2 Hybrid Policy with SFM

2.3 Anticipative Planning

3 Anticipative Models

3.1 Problem Formulation

3.2 Anticipative Turn (AT)

3.3 Anticipative Robot and Pedestrian's Propagation (ARP)

3.4 Anticipative Pedestrian's Propagation (APP)

4 Simulations and Experiments

4.1 Metrics

4.2 Simulation Results

4.3 Real-Life Experiments

4.4 Implementation Details

An Inductive Logic Programming Approach for Entangled Tube Modeling in Bin Picking

2 Literature Review

2.1 Pose and Shape Estimation in Bin Picking

2.2 Inductive Logic Programming with 3D Data

3 Dataset Generation

4 Learning Theories Using Inductive Logic Programming

5 Results

6 Conclusions

Mobile Robot Olfaction

Aerial Robot with Electronic Nose for Air Quality Monitoring

2 Material and Methods

2.1 UAV and E-Nose Integration

2.2 Data Processing

2.3 Experimental Setup

3 Results and Discussion

3.1 Calibration

3.2 Field Test

Simulating a Gas Source Localization Algorithm with Gas Dispersion Produced by Recorded Outdoor Wind

2 Gas Dispersion Simulator Using Outdoor Wind

3 Simulated GSL Algorithm

4 Tested Modification and Simulation Setup

5 Simulation Result and Discussion

6 Conclusion

References.

Potential Field Based Swarm Exploration of Processes Governed by Partial Differential Equations

2 Fundamentals

2.1 Process Model

2.2 Probabilistic Framework

2.3 Artificial Potential Fields for Multi-robot Control

3 Process Mapping via Potential Field Navigation

4 Simulation Trials

5 Experimental Evaluation

Consensus-Based Distributed Source Term Estimation with Particle Filter and Gaussian Mixture Model

2 Problem Formulation

3 Main Results

3.1 Local Particle Filtering

3.2 Kullback-Leibler Average

3.3 GMM-Based Approximation of PDF

4 Case Study

Evolving Swarm Formations for Odour Source Localisation

2 Evolution of Swarm Formations

3 Experimental Setup

3.1 Baseline Formations

3.2 Formation Leader Controller

3.3 Genetic Algorithm

3.4 Evaluation Environment

5 Conclusions and Future Work

Evaluating Cognitive Odour Source Localisation Strategies in Natural Water Streams

2 Information-Theoretic Framework for Odour Source Localisation

2.1 Dispersion Model

2.2 Inference

2.3 Decision Making

3 Testing Environment

3.1 Evaluation Metrics

GadenTools: A Toolkit for Testing and Simulating Robotic Olfaction Tasks with Jupyter Notebook Support

2 On the Importance of Gas Dispersion Simulators and Datasets

3 Technologies Employed

3.1 Python

3.2 Jupyter Notebooks

4 GadenTools

4.1 The Python Module

4.2 The Notebook Tutorial

Learning from the Past: Sequential Deep Learning for Gas Distribution Mapping

3.1 Data Generation.

3.2 Preparation of Data Sets

3.3 Neural Network Architectures

3.4 Model Implementation and Training

4 Results

5 Discussion

6 Conclusion and Outlook

A Novel Odor Source Localization Method via a Deep Neural Network-Based Odor Compass

2 Related Works

3 System Design

3.1 Overall Structure

3.2 Hardware Platform

4 DNN-Based Odor Compass Model

4.1 Model Design

4.2 Data Collection

4.3 Training and Validation Test

5 OSL Strategy and Experiment

5.1 OSL Strategy

5.2 OSL Experiment

Multi-robot for Industrial Applications

Modeling and Simulation of a Crossdocking System with an Integrated AS/RS

2 Policies and Concepts

2.1 Crossdocking

2.2 Automated Storage and Retrieval Systems

3 Crossdocking System and Model Architecture

3.1 Crossdocking System Architecture

3.2 Automated Warehouse (AS/RS) Architecture

4 Modelling Software and Model Development

4.1 Implementation of the CS System

4.2 Implementation of the AS/RS

5 Activity Scenarios and Results Obtained

5.1 Simulated Scenarios

5.2 Analysis of the Results

6 Conclusions and Future Developments

A Novel Algorithm for Maximization of the Secondary-Task Function for Redundant Serial Manipulators

2 Problem Overview

2.1 Proposed Algorithm

3 Experimental Scenario

4 Results and Analysis

Multi-robot Coordination for a Heterogeneous Fleet of Robots

3 Path Planning

3.1 Time Enhanced A* (TEA*)

3.2 Robot Priority Swapping

3.3 Moving Idle Robots

4 System Implementation

4.1 Graph

4.2 Graph Decomposition

4.3 Supervision

5 Results and Discussion

6 Conclusions and Future Work

Rehabilitation and Assistive Robotics

A Neuromechanical Model of Knee Flexion-Extension Based on Linear Regressors for Neuro-Controlled Exoskeletons

2 Methods

2.1 Experimental Setup

2.2 EMG and Kinematic Recordings

2.3 Neuromechanical Model

3 Results and Validation

4 Discussion

5 Conclusion and Future Work

A Preliminary Approach to a Wheelchair with Embedded Exoskeleton Capabilities

3 Design Trade-Off

4 Proposed Architecture

4.1 Identification of Modules and Determination of the Degrees of Freedom (DoF) to Implement in the Wheelchair

4.2 Definition of the Kinematic Pairs and Actuators for the Modular Chain

5 Simulation Features

5.1 Working Environment

5.2 Model Characterization

Machine Learning and Inertial Sensors to Estimate Vertical Ground Reaction Force During Gait

2.1 Lower Limb Joint Measurement

2.2 Experimental Setup and Acquired Data

2.3 Experimental Protocol

2.4 Machine Learning Models Training

2.5 Evaluation of Model Performance

3 Results

Intelligent Wheelchairs Rolling in Pairs Using Reinforcement Learning

3 Methods and Materials

3.1 Simulated Environment

3.2 Training the Intelligent Wheelchairs

3.3 Space and Actions

3.4 Reward System

4.1 Metrics Using only the Leader Chair

4.2 Metrics Using Two Chairs

Influence of Session Duration in Robot-Assisted Neurorehabilitation Therapies: A Pilot Study with Patients

2.1 Subjects

2.2 Experimental Setup.

2.3 Arm Rehabilitation Robot.

Notes:

Includes bibliographical references and index.

Other Format:

Print version: Tardioli, Danilo ROBOT2022: Fifth Iberian Robotics Conference

ISBN:

3-031-21062-X