Claudio Semini, IIT, Genova, IT
Claudio Semini is the head of the Dynamic Legged Systems (DLS) lab at Istituto Italiano di Tecnologia (IIT), since 2012. He received an M.Sc. degree from ETH Zurich in electrical engineering and information technology, in 2005. He spent 2 years in Tokyo for his research: MSc thesis at the Hirose Lab at Tokyo Tech, and staff engineer at the Toshiba R&D center in Kawasaki. During his Ph.D. (2010) and subsequent PostDoc at IIT, he developed the quadruped robot HyQ. He has published more than 100 publications in international journals and conferences. He is a co-founder of the Technical Committee on Mechanisms and Design of the IEEE-RAS Society. He is/was the coordinator/partner of several EU-, National and Industrial projects (HyQ-REAL, INAIL Teleop, Moog@IIT joint lab, ESA-ANT, etc). His research interests include the construction and control of highly dynamic and versatile legged robots for application in real-world operations, locomotion, agricultural robotics, and others. web: http://dls.iit.it
Title and abstract:
Towards Automation of Selective Vineyard Operations (Project VINUM)
Grapevine is considered to be one of the major fruit crops worldwide based on hectares cultivated and economic value. While the automation of non-selective vineyard operations has allowed to increase production and reduce costs, there are several selective operations that are difficult to automate. The VINUM project (a collaboration between IIT Genova and the Università Cattolica del Sacro Cuore) tackles the problem of automating winter pruning of grapevines.
This talk presents the current progress of the project. I will introduce the overall system consisting of a mobile robot, a manipulator arm with electric shears and perception system, as well as onboard computation and AI. I will illustrate how we use machine learning to perform segmentation and plant organ classification to create a representative model of the grapevine. Pruning rules developed by expert pruners are then used to detect the best pruning points that the manipulator arm uses to perform the cuts. Both wheeled and legged robots are currently evaluated as the mobile platform. Legged robots are promising due to their higher mobility on rough terrains.
Doina Bucur, University of Twente, NL
Doina Bucur (read: /’dɔɪ nɑ ‘bu kur/) is Assistant Professor of Network Data Science at University of Twente, The Netherlands. Her research revolves around complex networked systems. She designs computational methods to predict how a networked system will evolve in time, predictive methods to learn the relationship between network structure and network function, and optimization algorithms to improve the functioning of such systems.
Most of these algorithms build on machine learning and artificial intelligence. Evolutionary algorithms have proven particularly useful for solving combinatorial network problems, for which often no other method is computationally feasible. She works with small and big data from many domains: social, information, geographical, and ecological networks, and also with highly dynamic wireless communication networks.
Dr. Bucur holds a PhD from the University of Aarhus, Denmark, and did post-doctoral research at the University of Oxford, UK.
Title and abstract:
Evolutionary algorithms on complex networks
This talk gives an overview of evolutionary algorithms for solving problems in networked systems, such as social or communication networks.
Optimisation problems for networked systems are particularly hard, for a number of reasons. Social and communication networks are complex (the outcome of information-diffusion processes over networks is difficult to simulate or predict, leading to computationally heavy fitness functions). They are hard to model and simulate realistically, to mimic a real-world network of social users, their user profiles, and information preferences. There is also a great diversity of network structures, sizes, and user demographics, so one solution may not generalise across case studies. We survey challenges and solutions to single- and multi-objective optimisation problems related to information or influence maximisation: algorithmic design, computational challenges, explainability and diversity of the solutions, and helpful insights coming from the fields of network science and machine learning.
Travis Waller, TU Dresden, DE
As of early 2022, Steven “Travis” Waller is the Lighthouse Professor and Chair of Transport Modeling and Simulation at the Technical University of Dresden, Germany, as well as a Professor at the Australian National University (ANU). Until recently he was Head of School at UNSW Sydney where he led a new School vision of “Ethical Civil Infrastructure and Sustainable Environments”. He began his tenure-track and tenured career at the University of Illinois at Urbana-Champaign and, subsequently, at the University of Texas at Austin (where he was promoted to full Professor in 2011).
Travis is a global research leader in the domain of transportation network modelling and simulation particularly integrated planning models (including emerging technology as well as ethics/equity metrics), dynamic traffic assignment, and adaptive network equilibrium. He has published more than 300 peer reviewed papers, supervised 38 completed PhD students and conducted over 60 funded research projects for 40 global sponsors.
In 2003, Prof. Waller was named one of the top 100 innovators in science and engineering in the world under 35 years of age by MIT’s Technology Review magazine for his work on dynamic traffic analysis. In 2004, he received the U.S. National Science Foundation CAREER award for his proposed research and teaching plan on adaptive network equilibrium. In 2006, he was the recipient of the Annual New Faculty Award sponsored by the Council of University Transportation Centers and the American Road and Transportation Builders Association. In 2007, he was named a Fellow of the Clyde E. Lee Endowed Professorship in Transportation Engineering. In 2008, he was named to the Phil M. Ferguson Teaching Fellowship in Civil Engineering. In addition, he received the Fred Burggraf Award in 2009, the Hojjat Adeli Award for Innovations in Computing in 2012, the TRB Pyke Johnson Award in 2019 and named a Fellow of the Institution of Engineers Australia in 2021.
Title and abstract:
Evolutionary Algorithms in the Context of Emerging Pervasive Data for Transport Network Modelling
This talk will discuss research into evolutionary algorithms applied to applications in transport network modeling particularly where pervasive data can be employed. Traditionally, transport planning is an intensive process involving lengthy surveys and finely calibrated strategic models for entire regions. Further, transport planning and optimization often requires the incorporation of specific behavioral models to meet the needs of domain stakeholders, funding mandates and political processes. As a result, the primary research problems of interest involve traveler equilibrium (or other such paradigms). Therefore, the presented research utilizes such models as a sub-component which often necessitates mixed optimization/evolutionary algorithm solution methods. In addition, this talk will focus on a range of problems which leverage the rise of broadly available, pervasive, data which enable a new range of preliminary planning applications that can be examined on a much broader scale than traditional transport planning approaches generally address.