The First Robotics and Digital Twins Workshop

Time: Thursday, April 27th, 2023

Venue: Riddel Hall, Queen's University Belfast, 185 Stranmillis Rd, Belfast BT9 5EE 

ABOUT

The First Robotics and Digital Twins Workshop is funded by the EPSRC and UKRAS Network with the aim is to explore the challenges and opportunities of Digital Twins for Robotics, especially when working in challenging environments such as underwater or nuclear environment. The key enabling technologies of communications, AI/machine learning, control, simulation/modelling, etc for the developments of digital twins and robotics will be discussed. A white paper on 'Digital Twins for Robotics' will be published after the workshop.

Organisers:

Chair: Dr. Mien Van, Queen's University Belfast, UK

Co-Chair: Dr. Long Tran-Thanh, University of Warwick, UK

Co-Chair: Prof. Christopher Edwards, University of Exeter, UK

KEYNOTE SPEAKERS

Keynote Speaker 1: Professor Mehrdad Dianati 

Head of the Intelligent Vehicles Research Directorate at Warwick Manufacturing Group (WMG) and Co-Director of the University of Warwick’s Centre for Doctoral Training in Future Mobility Technologies 

Talk Title: On Resilience of Connected and Cooperative Autonomous Systems 

Abstract: Connected and Cooperative Autonomous Systems promise enormous opportunities in various applications to address some of the grand challenges of the 21st century. The term refers to groups of autonomous systems working together and with humans coherently and cohesively to extend humans’ physical and intellectual capabilities. Such systems can unleash the power of big data and AI and the processing and storage of the cloud (Edge/Core) and connectivity capabilities of 5G/6G communication systems to transform various aspects of human civilisation. However, these systems must demonstrate higher levels of autonomy to allow us to realise their full potential. At the same time, they shall be designed, verified/validated to be trustworthy to be deployed safely and effectively in mass in real-world applications. Trust is a multi-facet and cross-disciplinary challenge. It is a function of the competence and integrity of the systems and the transparency of system developers and operators.  An essential element of competence is resilience, defined as the capability of a system to avoid, withstand, adapt to, and recover from system failures. This talk elaborates on the needs and challenges of achieving resilience in AI-driven connected and cooperative autonomous systems. To this end, some promising approaches to achieving resilience in such systems will be discussed. Also, some highlights of the speaker’s recent and ongoing work on this topic will be presented. 


Speaker Bio   

Professor Mehrdad Dianati is currently the Head of the Intelligent Vehicles Research Directorate at Warwick Manufacturing Group (WMG) and Co-Director of the University of Warwick’s Centre for Doctoral Training in Future Mobility Technologies. Before joining Warwick, he held a professorial post at the 5G/6G Innovation Centre of the University of Surrey. At Surrey, he led a research theme in V2X systems for Connected and Autonomous Vehicles and was the Work Area Leader for the UK’s first flagship 5G programme. In addition to over 20 years of academic experience, he has worked in the industry as a software/hardware developer and Head of R&D for 10 years. His current research focuses on Connected and Cooperative Autonomous Systems involving 1) the development of AI-Native Network-Compute Fabrics in the context of 5G/6G Future Networks for those systems and 2) the exploitation of Cooperative Intelligence in those systems to ensure their efficiency and resilience. The latter aspect of his research involves using Cooperative Perception (Multi-Agent Belief Formation, Propagation) and Distributed Multi-Agent Decision Making and Control for Connected and Automated Vehicles. He has led over 20 research projects as the Principal Investigator (PI), including the recently delivered TASCC-CARMA project, sponsored by EPSRC and JLR with contributions from several leading UK companies, with a total budget of about £5.2 million, investigating the use of Mobile Edge Cloud Computing Systems for the control of Connected and Autonomous Vehicles (CAVs). Also, he has been the local PI in several flagship EU and Innovate UK CAV projects, such as Drive-C2X, L3Piot, Hi-Drive, AutopleX and MACAM, since 2008. Professor Dianati is a Fellow of the Alan Turning Institute. He is the Field Chief Editor of Frontier in Future Transportation. In the past, he was an associate editor for IEEE Transactions on Vehicular Technology and IET Communications and several other journals. He also served as the Guest Editor of several IEEE journals, such as JSAC and Transactions on Intelligent Transportation Systems.

Keynote Speaker 2: Professor Charalampos Tsimenidis

Head of Sustainable Digital Communications and Energy Systems (DIGCOM) Group,

Department of Engineering, Nottingham Trent University

Talk Title: Underwater Acoustic Communications

Abstract: In recent years, there has been an immense interest in developing underwater acoustic communication systems, most of which are related to commercial and industrial applications, and military and defence telemetry applications. Other applications include oceanographic research and environmental monitoring, such as ocean-bottom survey and collection of scientific data acquired by sub-sea sensors without the need for retrieving the equipment. However, for all these applications the principal function is to achieve reliable communication both in point-to-point links, and in network scenarios. In practice, the only feasible method to achieve subsea communications is by means of acoustic signals. Such acoustic links are exposed to adverse physical phenomena governing acoustic wave propagation in the sea. These include ambient noise, frequency-dependent attenuation, temperature and pressure variations, reverberation, and extended multi-path. Any successful acoustic modem design must consider all these effects to select an appropriate configuration for system-related parameters. The transmit power level and operating frequency must be considered in conjunction with the ambient noise and transmission range, the utilized modulation scheme, data rate, and the level of diversity must properly match the expected channel conditions related to time and frequency dispersion, and choice of multiple-access strategy. The focus of this talk will be to provide a review of the state-of-the-art art in point-to-point underwater acoustic communication modems, and an insight into their design and analysis. Emphasis will be placed on transceiver structures that employ well-established communication techniques, such as spatial and time filtering, single- and multi-carrier approaches, and spread-spectrum and error control coding methods that strike a balance between power and bandwidth efficient transmission, multiple-access capability and reliability.


Speaker Bio   

Professor Tsimenidis joined NTU in October 2022 after working for 25 years as an academic at Newcastle University, where he has obtained his PhD in 2002 and was promoted to Prof. of Digital Communications in 2022. His research interests are in the area of digital communications with focus in wireless RF and underwater acoustic communications. He has published over 250 articles and graduated successfully 50 PhD students, and secured more than 15 research projects from industry and government to a value of over £5M. He is a senior member of IEEE and a member of IET. 

Keynote Speaker 3: Dr. Son Tong

ADAS Team Manager

Siemens Digital Industries Software


Talk title: Digital Twin in Autonomous Driving Development 

The presentation discusses some Siemens engineering technologies for autonomous driving development that combines algorithms, virtual and physical testing in a digital twin fashion. The ADAS and autonomous vehicle industries recently rely on virtual testing as one of the safe and efficient technologies for validation, especially to deal with safety critical scenarios. Starting from measurement and real traffic sensor data, we demonstrate the process to create digital twin high-fidelity models of sensor, vehicle dynamics, and traffic scenarios. These models are then deployed for scenario generation and algorithms development (i.e. perception, planning, and control), targeting to safety and other driving metrics such as driving comfort, efficiency. Combine with hardware and physical car testing, the digital twin framework helps to accelerate the development stages, detect issues early on, and capture corner cases. Some applications will be presented: scenario detection and generation, imitation learning, and sim2real reinforcement learning control. 


Speaker Bio  

Son Tong is an ADAS R&D Manager at Siemens Digital Industries Software with focus on control systems, automated driving and AI engineering topics. He has been facilitating technology developments dealing with vehicle control, safety, comfort, and validation processes. The solutions often involve state of art algorithms, simulation and physical testing. Son Tong is active in different Siemens and European Union (EU), Belgian research programs as technical coordinator and PhD co-supervisor. He was awarded Siemens PL Invention of the Year Award, and selected in the AutoSens2019 most influential research award finalist.


Keynote speaker 4: Dr. Matthew Bonney

Lecturer in Space Engineering, Swansea University 

Talk title: Considerations of Multidisciplinary Trust in Digital Twins 

With the ideal of a truly virtual representation of a system, the understanding of all components in the system should be under consideration. Two major components that are commonly under-studied are the incorporation of multidisciplinary aspects of digital twins (cybersecurity, Multiphysics simulations, business predictions, etc.) and the consideration of human factors. For understanding the human factor of digital twins, the concept of “trust” is a fascinating topic of research. This talk will discuss a wide variety of digital twin applications and the various aspects/concerns that these industries have around using digital twins, including the different components of “trust”.

Speaker Bio  

Dr. Bonney is a Lecturer in Space Engineering at Swansea University and a Research Associate at the University of Sheffield. His research interests are Multiphysics simulations, multidisciplinary digital twins, and high-consequence systems. While working on aerospace and nuclear systems, Dr. Bonney worked to develop various engineering tools such as digital twins and ABAQUS plug-ins. He has worked as multi-university project lead for the EPSRC funded Digital Twins for Improved Dynamic Design, technical lead for the Alan Turing Institute funded Digital Twins for High-Value Engineering Applications, and the Demonstrator lead for the Turing Research and Innovation Cluster in Digital Twins’ Hawk demonstrator.

Presenters:

Dr. Ziwei Wang (Lecturer, University of Lancaster)

Abstract: To facilitate working in extreme environments, a human worker could benefit greatly from the assistance of a robotic arm controlled through teleoperation. However, it is currently unclear how the robot should be programmed to effectively support the worker's movements. To address this issue, a haptic teleoperation system that enables complex manipulation and docking tasks will be presented. This system allows for human demonstrations of motor skills, which can then be developed into autonomous operation capabilities. The haptic teleoperation system includes various control modes for object capture and assembly tasks, such as manual, semi-autonomous, and fully autonomous modes. Additionally, the system utilises bio-monitoring to estimate the operator's mental load, holographic user interfaces to enhance usability, and haptic feedback to improve system transparency for the operator. Successful interaction tasks, including collecting rock samples on the moon and peg transfer based on the Da Vinci Surgical Robot System, have validated that the developed techniques can be transferable to other types of human-machine interfaces, including command through the feet.


Dr. Zhang Yao (Lecturer, University of Southampton)

Abstract -- As optimal control problems, the performance of marine renewable energy system such as wave energy converter control relies on the accuracy of the future incoming sea state prediction. However, the inevitable prediction errors can degrade control performance dramatically especially when a long prediction horizon is needed by a non-causal optimal control strategy. This research is to investigate methods that effectively mitigate the control performance degradation caused by prediction errors and also reduce cost by replacing high-cost sensors by advanced robust online estimators. The proposed control scheme does not cause heavy computational load enabling its real-time implementation on standard computational hardware, which is especially critical for the control of marine systems with complicated dynamics. The proposed framework is generic and can be applied to a wide range of marine renewable energy systems.


Dr. Erfu Yang (Senior Lecturer, University of Strathclyde )

Abstract --The use of UAVs/drones is increasing across many industries, such as oil & gas and mining sector to address the challenging application issues in complex environments (confined, dusty, dark and GPS-denied etc). In this talk, a novel aerial robotic scheme for visual inspection in low-illuminated environments is first introduced, where the adaptive gamma correction algorithm with weighting distribution is incorporated into the ORB-SLAM3. Then, an accurate and efficient vision-based corrosion detection algorithm based on deep learning techniques is presented for the visual inspection of metallic structures. Thirdly, a novel efficient VSLAM algorithm named AFE-ORB-SLAM is proposed for addressing the challenges when deploying UAVs in the challenging lighting environment. An image enhancement method is proposed and adopted to the ORB-SLAM3. An adaptive FAST feature point extraction method is also investigated. Finally, a novel deep learning feature-based monocular VSLAM system is developed to address the UAV localisation problem in challenging lighting and textureless environments. In the system, an efficient and effective feature extraction model with a self-training scheme is proposed to handle the scenario that both the traditional and learned feature based VSLAM approaches fail. 


Dr. Pantelis Sopasakis (Lecturer, Queen's University Belfast )

Abstract -- Uncertain dynamical systems give rise to either stochastic or worst-case (minimax) optimal control formulations. In the stochastic case, the distributions of the disturbances are assumed to be known. In worst-case formulations, no probabilistic information is taken into account. On the other hand, multistage risk-averse formulations, allow to exploit inexact probabilistic information that is usually available. This leads to risk-averse optimal control problems, which have gained a lot of attention in the last decade, mostly due to their attractive mathematical properties and practical importance. They can be seen as an interpolation between stochastic and robust optimal control approaches, allowing the designer to trade-off performance for robustness and vice-versa. Due to their stochastic nature, risk-averse problems are of a very large scale, involving up to millions of decision variables. In this work, we propose a splitting for general multistage risk-averse problems and show how to efficiently compute iterates on a GPU-enabled hardware. Moreover, we propose SPOCK - a new algorithm that utilizes the proposed splitting and takes advantage of the SuperMann scheme combined with fast directions from Anderson's acceleration method for enhanced convergence speed. We implement Spock in Julia as an open-source solver, which is amenable to warm-starting and massive parallelization.


Prof. Yan Jin (Professor, Queen's University Belfast )

Abstract: This talk will introduce the state of the art of using parallel robots for high value manufacturing, such as aircraft assembly. The latest research on automated assembly and double-sided machining using parallel kinematic machines will be presented. Research issues and future directions will be discussed. 


Dr. Anh Nguyen (Lecturer, University of Liverpool)

Abstract: Endovascular intervention is a minimally invasive medical procedure used to diagnose and treat various conditions within blood vessels. In recent years, the development of digital twin technology, especially simulation platforms, has shown great potential in improving the accuracy and effectiveness of endovascular interventions. By creating a digital twin of a patient's vasculature, clinicians can perform training and optimize treatment plans in simulation before carrying out the actual procedure. It can lead to more precise and personalized interventions, reducing the risk of commodities and improving patient outcomes. In this talk, the benefits, challenges, and future prospects of digital twins for endovascular intervention will be discussed.

Schedule - Thursday, April 27th, 2023


8:30am - 9:00am: Coffee/Tea

9:00am - 9:05 am: Welcome (Dr. Mien Van,  Dr. Long Tran-Thanh, Prof. Christopher Edwards)

9:05am - 9:30am: Opening talk on Digital Twin Research- Prof. Roger Woods, (Dean of Research, EPS Faculty, Queen's University Belfast)

9:30am-10:10am: Keynote talk: Considerations of Multidisciplinary Trust in Digital Twins, Dr. Matthew Bonney

10:10am-10:30am: Parallel robots for high value manufacturing, Prof. Yan Jin 

10:30am-10:50am: Optimal control with enhanced robustness for marine renewable energy systems, Dr. Zhang Yao

10:50am-11:30am: Poster presentation & Break

 

11:30am-12:10pm: Keynote talk: Digital Twin in Autonomous Driving Development, Dr. Son Tong

12:10am-12:30am: SPOCK: parallelisable solution of large-scale risk-averse optimal control problems, Dr. Pantelis Sopasakis

12:30pm-12:50pm: Robust Sensing, Detection and Localisation for UAV-Based Smart Visual Inspection in Complex Environments, Dr. Erfu Yang

12:50pm-13:15pm: Discussions on the applications of Digital Twins for Robotics and its challenges (Group work, chair: Dr. Mien Van)

Lunch 

13:15: 2:30pm: Lunch



Afternoon Session


2:30pm-3:10pm: Keynote talk: On Resilience of Connected and Cooperative Autonomous Systems, Professor Mehrdad Dianati 


3:10pm-3:50pm: Keynote talk: Underwater Acoustic Communications, Professor Charalampos Tsimenidis

3:50pm:4:10pm: Digital Twins for Endovascular Intervention, Dr. Anh Nguyen

4:10pm-4:30pm: Bimanual Coordinated Teleoperation and Assessment of Operator's Cognitive Workload, Dr. Ziwei Wang

4:30pm- 5:00pm: Discussions on key enabling technologies for digital twins  (Group work, chair: Dr. Long Tran-Thanh)

5:00pm- 5:45pm: Showcase and Demonstration

5:45:00pm- 6:00pm: Closing and Networking



Register

Register to attend for free here by 16th April 2023.