Advances in Physical Agents II

Name: Advances in Physical Agents II | Proceedings of the 21st International Workshop of Physical Agents (WAF 2020), November 19-20, 2020, Alcalá de Henares, Madrid, Spain
Brand: Springer
Price: 149.79 EUR
Availability: OnlineOnly

Proceedings of the 21st International Workshop of Physical Agents (WAF 2020), November 19-20, 2020, Alcalá de Henares, Madrid, Spain

Luis M. Bergasa Manuel Ocaña Rafael Barea Elena López-Guillén Pedro Revenga(Herausgeber*in)

Springer (Verlag)

Erschienen am 2. November 2020

XI, 362 Seiten

E-Book

PDF mit Wasserzeichen-DRM

Systemvoraussetzungen

978-3-030-62579-5 (ISBN)

149,79 €inkl. 7% MwSt.

Systemvoraussetzungen

für PDF mit Wasserzeichen-DRM

E-Book Einzellizenz

Als Download verfügbar

Beschreibung

Weitere Details

Weitere Ausgaben

Inhalt

Intro
Preface
Organization
Organising Committee
Scientific Committee
Contents
Autonomous Navigation, Localization and Mapping
Defining Adaptive Proxemic Zones for Activity-Aware Navigation
1 Introduction
1.1 Contribution
1.2 Paper Organization
2 Framework Description
2.1 Asymmetric Gaussian Function as Human Activity Representation
2.2 Adaptation of the Proxemic Areas, the Cooperation Zone
3 Building the Social Layer
4 Experimental Results
4.1 Approaching People
4.2 Collaborative Actions
5 Conclusions and Future Works
References
Lifelong Object Localization in Robotic Applications
1 Introduction
2 Related Work
3 Segmentation Methods
3.1 Lexical Labels in RGB-D Images
3.2 Semantic Segmentation
4 Lifelong Object Localization
4.1 Short-Term Validation
4.2 Long-Term Predictions
4.3 Passive and Active Search
5 Use Case
6 Conclusions
References
Estimation of Customer Activity Patterns in Open Malls by Means of Combining Localization and Process Mining Techniques
1 Introduction
2 BAI4SOW Global Vision
3 Localization System
4 Test-Bed and Results
4.1 Smartphone Application Development
4.2 Test-Bed Description
4.3 User Behavioural Analytics Results
5 Conclusions and Future Works
References
Train Here, Drive There: Simulating Real-World Use Cases with Fully-Autonomous Driving Architecture in CARLA Simulator
1 Introduction
2 Related Works
3 Autonomous Navigation Architecture
4 Simulation Stage
4.1 Environment
4.2 Vehicle
4.3 Sensors
5 Experimental Results
6 Conclusions and Future Works
References
DQN-Based Deep Reinforcement Learning for Autonomous Driving
1 Introduction
2 Related Work
3 Framework Overview
4 Method
4.1 MDP Formulation
4.2 Deep Q-Learning Algorithm for Reinforcement Learning
5 Deep Q-Network Agents
5.1 DQN Architectures
6 Results
7 Conclusions
References
Mobile and Social Robots
Concept Drift Detection and Adaptation for Robotics and Mobile Devices in Federated and Continual Settings
1 Introduction
2 Related Work
3 Concept Drift in Federated Settings: Problem Definition
4 Concept Drift Detection and Adaptation for FedAvg
4.1 Drift Detection
4.2 Drift Adaptation
4.3 Benefits of Explicit Drift Detection: Saving Resources on Devices
5 Experimental Results
6 Conclusions
References
QoS Metrics-in-the-Loop for Better Robot Navigation
1 Introduction
2 Adaptive Robot Navigation at Design-Time
2.1 Nominal Behaviour Tree
2.2 Modelling QoS Metrics
2.3 Modelling Robot Reconfigurations
3 Composition and Runtime Self-adaptation
4 Experimental Evaluation
4.1 Experimental Setting
4.2 Runtime Monitoring and Actuation
5 Conclusions and Future Work
References
DSRd: A Proposal for a Low-Latency, Distributed Working Memory for CORTEX
1 Introduction
2 The Deep State Representation
3 Eventual Consistency and CRDTs
4 Design and Implementation
5 Examples and Experiments
6 Conclusions and Future Work
References
Multi-agent System Model of Taxi Fleets
Abstract
1 Motivation
2 Cooperative Taxis Case
2.1 Environment
2.2 The Lua Framework
3 Belief-Desire-Intention Agents
3.1 State Machine Stack Structure
3.2 Extended Finite-State Machines
3.3 Extended Finite-State Stack Machines
4 Experimentation
5 Conclusion
References
Human-Robot Interaction
Can a Social Robot Learn to Gesticulate Just by Observing Humans?
1 Introduction
2 Motion Capturing Systems
3 Generative Models and Their Applications
4 OpenPose Based Motion Mapping
4.1 Arms Motion
4.2 Head Motion
4.3 Hands Motion
5 Experimental Setup
5.1 Robot and Computer Platforms
5.2 Generative Model
6 Results and Motion Analysis
7 Conclusions and Further Work
References
Evaluation of a Multi-speaker System for Socially Assistive HRI in Real Scenarios
1 Introduction
2 The Use of ASR for Socially Assistive HRI
3 Implementation of an Acoustic Signal Enhancement Module for the Source of Interest in Real Time
3.1 Beamforming Techniques
3.2 Masking Techniques
3.3 Simulated Evaluations of Intelligibility
4 Overview of the General System and Implementation
5 Evaluation
5.1 Off-line Analysis of Intelligibility
5.2 Experimental Procedure and Variables
5.3 Setup for the Real Environment Conditions
6 Results
7 Conclusions and Future Research
References
Graph Neural Networks for Human-Aware Social Navigation
1 Introduction
2 Graph Neural Networks
3 Formalisation of the Problem
4 Experimental Results
5 Conclusions
References
A Toolkit to Generate Social Navigation Datasets
1 Introduction
2 Navigation Datasets
3 Proposal
4 Use Case
4.1 Use Case Description
4.2 Use Case Design and Implementation
4.3 Experimental Results
5 Conclusions
References
JBCA: Designing an Adaptative Continuous Authentication Architecture
1 Introduction
2 Background
2.1 Continuous Authentication
2.2 Data Analysis
2.3 Multi-feature Approaches
2.4 Access Control
3 Architecture Overview
4 Clients
4.1 CA Clients
4.2 Trust Clients
4.3 Partner Clients
5 Authorization Server
5.1 Authorization Token
5.2 User Identity
5.3 Client Roles
5.4 Scope
6 AI Models
6.1 Quality
6.2 Requirements
6.3 Privacy Concerns
6.4 Ethical Concerns
7 Evaluation
7.1 Ensemble Learning
7.2 Adaptative Model
8 Use Cases
8.1 ``Offline'' Use Case
8.2 EDR Use Case
8.3 SIEM Use Case
8.4 Joint Use Case
9 Conclusions and Future Work
References
Perception Systems
Optimization of a Robotics Gaze Control System
1 Introduction
2 Overview of the Related Work
3 Gaze Control with a Competitive Network
4 Analysis of the System
5 Experiments and Results Discussion
6 Conclusions
References
Open-Loop Sidescan Sonar Mosaic and ANN Velocity Estimation
1 Introduction
2 Positioning
3 Mosaicing
3.1 Transformations
3.2 Corrections
3.3 Updating the Mosaic
4 Results
5 Conclusions and Future Work
References
360 Real-Time 3D Multi-object Detection and Tracking for Autonomous Vehicle Navigation
1 Introduction
2 Related Works
2.1 Vision-Only 3D Object Detection
2.2 3D LiDAR Object Detection
2.3 2D Multi-object Tracking
2.4 3D Multi-object Tracking
3 Architecture
3.1 3D Object Detector
3.2 Multi-object Tracking
4 Evaluation
4.1 3D DAMOT Evaluation Metrics
4.2 Object Detector Evaluation in KITTI
4.3 CARLA: A Realistic Urban Environment Simulator
4.4 DAMOT Evaluation in CARLA
5 Conclusions
References
Towards Fine-Tuning of VQA Models in Public Datasets
1 Introduction
2 VQA Implementation
2.1 Pythia's Overview
2.2 Implementation Review
2.3 Contributions
3 Fine-Tuning a VQA Model
3.1 Training Strategies
3.2 Hyperparameters
3.3 Training Loss
4 Experiments and Results
4.1 Public VQA Dataset
4.2 Evaluation Metrics
4.3 Experimentation
4.4 Quantitative Results
4.5 Qualitative Results
5 Conclusions and Future Works
References
Integrating OpenFace 2.0 Toolkit for Driver Attention Estimation in Challenging Accidental Scenarios
1 Introduction
2 System Architecture
2.1 Video Acquisition
2.2 Gaze Estimation
2.3 Visual Attention Model
2.4 Heat Map Visualisation
3 Experimental Results
3.1 Visual Attention Model Evaluation
3.2 DADA2000 Evaluation
4 Conclusions and Future Works
References
Deep-Learning and Robotics
Embedded Deep Learning Solutionbreak for Person Identification and Followingbreak with a Robot
1 Introduction
2 Related Works
3 Design
3.1 Perception Module
3.2 Actuation Module
4 Experiments
4.1 Person Detection Experiments
4.2 Face Detection Experiments
4.3 Face Recognition Experiments
4.4 Deep Learning Optimization with TensorRT
4.5 Typical Execution
5 Conclusions
References
Survival Loss: A Neuron Death Regularizer
1 Introduction
2 Related Work
3 Adam and L2: Dying ReLU's
4 The Role of Dead Filters in Continual Learning
4.1 Survival Loss Regularizer
4.2 Benchmark Results
4.3 Hyper-parameters Discussion
5 Conclusions
References
Reinforcement Learning Experiments and Benchmark for Solving Robotic Reaching Tasks
1 Introduction
2 Environments and Methods
2.1 Manipulator Description
2.2 Reaching Task and Environment Definition
2.3 Experiments and Benchmark
3 Results and Discussion
3.1 Training Convergence
3.2 Evaluation Metrics
4 Conclusion and Future Work
References
Age and Gender Recognition from Speech Using Deep Neural Networks
1 Introduction
2 Problem Statement
3 Framework Description
4 Experimental Work
4.1 Training Corpus
4.2 Experimental Settings
4.3 Results and Discussion
5 Conclusions
References
Monocular 3D Hand Pose Estimation for Teleoperating Low-Cost Actuators
1 Introduction
2 Related Works
3 Tridimensional Hand Pose Estimation Pipeline
3.1 Hand Area of Interest Localization
3.2 Tridimensional Hand Joints Estimation
3.3 Low-Cost Robotic Hand Actuator
4 Dataset Description and Data Balance Technique
5 Experimentation
5.1 Baseline Experiment
5.2 Experiments with a Balanced Version of the Dataset Using a Clustering Procedure
5.3 Experiments with Different Backbones
5.4 Experiments Involving Data Augmentation
5.5 Experiments on the Full Pipeline
6 Conclusion and Future Work
References
Author Index

Systemvoraussetzungen

Als PDF speichern Als Link merken