Perception and Computational Efficiency for Autonomous Vehicles
University of Nebraska-Lincoln - USA
Perception is a critical computational task in autonomous vehicles. Autonomous vehicles place stringent and somewhat conflicting demands on perception systems: high accuracy, low latency, and performance on limited computational resources. The conflict between these requirements is particularly acute in the case of unmanned aerial vehicles (UAVs) but is also true of ground vehicles. This talk will describe two related efforts to improve and manage efficiency in perception for autonomous vehicles. Work with Krishna Muuva and Justin Bradley of UNL looks at UAV-UAV tracking. We show that tracking performance saturates above a given level of perceptual accuracy. Work with Deep Samal and Saibal Mukhopodhyay of Georgia Tech compares the results of multiple evaluators to improve the accuracy of LIDAR tracking.
Marilyn Wolf is Elmer E. Koch Professor of Engineering and Director of the School of Computing at the University of Nebraska – Lincoln. She received her BS, MS, and PhD in electrical engineering from Stanford University in 1980, 1981, and 1984, respectively. She was with AT&T Bell Laboratories from 1984 to 1989. She was on the faculty of Princeton University from 1989 to 2007 and was Farmer Distinguished Chair at Georgia Tech from 2007 to 2019. Her research interests include cyber-physical systems, embedded computing, embedded video and computer vision, and VLSI systems. She has received the IEEE Kirchmayer Graduate Education Award, the IEEE Computer Society Goode Memorial Award, the ASEE Terman Award, and the IEEE Circuits and Systems Society Education Award. She is a Fellow of the IEEE and ACM and an IEEE Computer Society Golden Core member.
Building Intelligent and Autonomous Vision Systems
University of the West of Scotland - Scotland
Nowadays, the rise of smart applications (e.g. smart agriculture, smart manufacturing, smart cities, e-education, e-health, etc.) boosts the development of innovative IT technologies based on Artificial Intelligence (AI), leading to intelligent and autonomous systems, which themselves use new algorithms, complex software, or advanced embedded systems. In particular, intelligent vision systems (IVS), which are systems able to automatically process visual inputs such as raw still pictures or live video feeds, whatever they are equipped with camera(s) or directly access image databases, have become ubiquitous in our Society, from smart lifts to collaborative robots. Therefore, intelligent vision systems must be both efficient and ethical. Indeed, due to their expanding number and range of applications as well as their growing autonomy, there is an increased expectation for these intelligent technologies to involve explainable algorithms, dependable software, transparent agents, trustworthy systems, etc. Hence, this keynote will present both scientific research and societal challenges as well as technical solutions and emerging standards to build trustworthy intelligent vision systems to be deployed in real-time and in real-world, constrained and unconstrained environments, in the context of Industry 4.0 and Society 5.0.
Joanna Olszewska BSc(Hons) MSc(EPFL) PhD(UCL) is a British computer scientist. She is an Assistant Professor in the School of Computing and Engineering at the University of the West of Scotland, UK, where she leads research on algorithms and software for trustworthy intelligent vision systems. Senior Member of IEEE, she stands on the IEEE Artificial Intelligence Standard Committee, she is part of the IEEE Global Initiative for Ethical Considerations in Artificial Intelligence and Autonomous Systems, and she is Co-Chair of the IEEE RAS Technical Committee on Verification of Autonomous Systems. She is a Chartered Engineer, a Chartered Scientist, and a Fellow of the BCS. She has been TPC member of over 100 international conferences such as IJCAI, and has chaired over 70 conference/workshop sessions, e.g., at IROS. She has been a Guest Editor for the Knowledge Engineering Review Journal, Cambridge University Press, and she has been appointed as an Associate Editor for the Machine Learning with Applications Journal, Elsevier. She has contributed to 11 ISO/IEC/IEEE standards in various roles, e.g. Vice-Chair of ISO/IEC/IEEE P41062, and she is an author of 100+ peer-reviewed publications and one book. She holds several awards, e.g., ESWA Outstanding Reviewer Award, and distinctions, e.g., ACM Distinguished Speaker.
The Power of Multi-agent Systems to Implement Industrial Cyber-Physical Systems
Polytechnic Institute of Bragança - Portugal
Global markets are imposing strong changing conditions for industrial companies running their businesses, facing strong pressures related to the customization of products in high flexible production systems. The fourth industrial revolution aims to promotes the digitization of the traditional industries, aiming intelligent factories characterized by adaptability, efficiency, functionality, reliability, safety, and usability. This digital transformation is also noticed in other domains like administration, electrical grids, transportation, healthcare and agriculture. In this context, smart products, processes and systems emerge of applying cyber-physical systems, complemented with emerging ICT and Artificial Intelligence technologies, such as Internet of Things, Big data, cloud computing, machine learning, virtual/augmented reality and multi-agent systems.
This presentation discusses the use of multi-agent systems as a key enabling technology to realize intelligent, dynamic, self-organized and large-scale cyber-physical systems. Since data and artificial intelligence are pillars in these emergent systems, the presentation also highlights the use of multi-agent systems to distribute intelligence among cloud, fog and edge computing layers taking advantage of the huge amount of data that can be collected in real-time using Internet of Things technologies. Some illustrative examples of using multi-agent systems to deploy industrial cyber-physical systems exhibiting self-organization and intelligence features will be provided to emphasize their innovative aspect.
Paulo Leitão received the Ph.D. degree in Electrical and Computer Engineering from the University of Porto, Portugal, in 2004. From 1993 to 1999 he developed research activities at the CIM Centre of Porto, from 1999 to 2000 at IDIT - Institute for Development and Innovation in Technology, from 2009 to 2017 at LIACC - Artificial Intelligence and Computer Science Laboratory, and since 2018 at CeDRI – Research Centre in Digitalization and Intelligent Robotics, where he is its scientific coordinator. He joined the Polytechnic Institute of Bragança, Portugal, in 1995, where he is Full Professor at the Department of Electrical Engineering. He served as Head of the Department of Electrical Engineering from 2009 to 2015, Vice-President of Directive Board of School of Technology and Management from 2004 to 2009, President of the Pedagogical Council of School of Technology and Management during 2000 and Vice-President of Scientific Council of School of Technology and Management from 2001 to 2004.
His research interests are in the field of intelligent and reconfigurable systems, cyber-physical systems, multi-agent systems, digital twin, Internet of Things, factory automation and holonic systems. He participate / has participated in several national and international research projects, e.g. under the EU FP7, Horizon 2020 and Horizon Europe frameworks, and Networks of Excellence. He is member of the International Program Committee (IPC) of several scientific events, and served as general co-chair of several international conferences, namely IFAC IMS’10, HoloMAS’11, IEEE ICARSC’16, SOHOMA'16 and IEEE INDIN’18. He has published 8 books and more than 300 papers in high-ranked international scientific journals and conference proceedings (per-review). He is co-author of three patents and received six paper awards at international scientific conferences.
He is Senior member of the IEEE Industrial Electronics Society (IES) and Systems, Man and Cybernetics Society (SMCS), Chair of the IEEE IES Technical Committee on Industrial Agents, associate editor of IEEE Transactions on Industrial Informatics, member at-large of the IEEE IES Administrative Committee (AdCom), and chair of the established IEEE 2660.1 standard.
Machine Learning e Data Science - Aplicações práticas no setor energético
Tiago Kaoru Matsuo
AQTech Power Prognostics
Há uma crescente demanda por energia limpa no mundo, e isso tem impulsionado a melhoria contínua dos processos de operação e manutenção no setor elétrico. Nesse contexto, as exigências relacionadas a disponibilidade dos ativos, reduções de custos de manutenção e extensão da vida útil dos ativos têm se tornado cada vez mais elevadas. A difusão da indústria 4.0 tem impulsionado nesse mercado a adoção de tecnologias de IIoT, computação em nuvem, big data, dentre outros, o que está permitindo o uso de técnicas de machine learning e data science para otimização do O&M das concessionárias de energia. Nesta palestra será apresentada uma visão geral do monitoramento em hidrogeradores e aerogeradores, que tem estabelecido uma base de dados que nos permite trabalhar o uso de técnicas de ML para resolução de problema. Será abordado um caso de Detecção de Anomalias no Anel de Regulação de Turbinas Bulbo utilizando Deep Learning e apresentada uma abordagem de machine learning para detecção e ranqueamento automático de anomalias em um parque de aerogeradores.
Tiago Kaoru Matsuo é Diretor Técnico e Vice-Presidente da AQTech Power Prognostics, empresa de base tecnológica que cresceu desenvolvendo soluções de monitoramento e diagnóstico para o mercado de energia elétrica. É formado técnico em Eletrônica pelo Centro Federal de Educação Tecnológica de Santa Catarina em 2005, graduado em Engenharia Elétrica pela Universidade Federal de Santa Catarina em 2010 e com Mestrado profissional em Mecatrônica no Instituto Federal de Santa Catarina em 2017. Entrou AQTech em 2006, e desde 2015 atua na liderança técnica da empresa. Possui experiência no desenvolvimento de soluções para monitoramento e diagnóstico de máquinas, envolvendo desde a especificação, desenvolvimento, implantação e comissionamento de sistemas baseados em análise de vibrações mecânicas.