earth

Dependable Intelligent Systems

yipapadopoulos
Professor Yiannis Papadopoulos
Group lead

The Challenge

System dependability is crucially important for a new generation of emerging intelligent, co-operative, autonomous, cyberphysical systems. These include smart vehicles, robotics, telehealth devices, smart energy grids, and the internet of things, their related data repositories and communication channels. 

If such systems fail, they may harm people and lead to the collapse of important infrastructures. Ensuring their dependability is an unsolved problem and the key to enabling their industrial and societal uptake.

The Approach

The group is pioneering novel syntheses of software engineering and bio-inspired technologies for the design, analysis, and operation of dependable intelligent systems. Data analytics and AI are exploited to assess and guarantee dependability properties, including safety and security of systems, their communication channels and data repositories. Applications span from the engineering of critical transport systems to analysis of documents and social media to the synthesis of generative art. A key outcome of our work is the HiP-HOPS method and tools. The method offers a significant innovation covering the system engineering lifecycle, from intelligent allocation of safety and security requirements through automated dependability analysis to evolutionary optimisation of system architectures, the automated production of certification artefacts, and intelligent safety monitoring using agents. Over 200 papers on theory and industrial application have been published on this method alone.

Our work has been transferred to the automotive and other industries through the HiP-HOPS and Safety Designer software tools. Inspired by the hunting habits of penguins, our algorithms and their application to automotive safety have attracted publicity on the BBC and Automotive IQ.

Laptop

OUR AIMS

  • Enable effective, automated analysis of complex intelligent systems, including of properties such as safety, reliability, availability, data-integrity, security, and maintainability.
  • Enable optimisation of such systems using sound mathematical, computational techniques, data-driven algorithms, and machine learning.
  • Enable run-time certification and assurance of safety and security of intelligent systems, and Systems-of-Systems in the context of unpredictability, uncertainties, and new security threats.
  • Apply AI techniques to a wide range of systems from engineering systems though document analysis to social media, education, and art.

Projects

Land Rover grille

Safety of Autonomous Systems with Jaguar Land Rover (SAS-JLR)

The project examines potential safety flaws in advanced driver assistance systems and software used in autonomous cars. The aim is to derive analyses and methods that could help improve safety. (2018 - 2020)

Find out more
Laptop

HiP-HOPS

HiP-HOPS is a project dating back to 1999 that has been partly funded by industry including Daimler, Volvo, Germanischer Lloyd and others. In HiP-HOPS, algorithms examine the design of a system and locate potential flaws. This is done via model-based construction and analysis of fault trees and FMEAs that can record not only combinations but sequences of faults.

Find out more
wind-turbines

Data-driven Reliability-centred Evolutionary Asset Manager (DREAM)

DREAM is a project funded by EDF where HiP-HOPS is extended with machine learning to develop data-driven and reliable maintenance optimisation techniques to help better understand the useful life and maintenance plans of components in the offshore wind industry.

Find out more
Lights on a motorway

Dependability Engineering Innovation for Cyber Physical Systems (DEIS)

DEIS  is an H2020 project investigating intelligent loosely-connected systems where we developed Digital Dependability Identities (DDI), i.e., modular, composable and executable specifications of dependability for use in complex cyber-physical systems and open Systems-of-Systems.

Find out more
computing-technology

MAENAD & EAST-ADL

MAENAD & EAST-ADL form part of a series of European projects.  EAST-ADL is an emerging architecture description language authored as an industry standard for the design of vehicles. Several metamodels of HiP-HOPS were transferred into the Safety Annex of this language which is gaining substantial influence in the automotive sector.

Find out more
Data centre

Spencer-Data Mining KTP

Spencer-Data Mining KTP is a Knowledge Transfer Partnership (KTP) project on text mining for corporate documents with Spencer Ltd, supervised by Nina Dethlefs. We apply research in data analytics, natural language processing and artificial intelligence to analyse a large repository of disparate types of company documents - relating to projects - to identify patterns and extract, organise and summarise information.

Find out more
Offshore Wind Turbines iStock_26960521_XLARGE

Aura CDT on Offshore Wind and the Environment

The Aura CDT is a doctoral training centre with Sheffield, Durham and Newcastle Universities, EPSRC/NERC-funded. The CDT seeks to enable innovation for the offshore wind sector in a unique collaborative approach of cross-disciplinary engagement between engineering and environmental scientists, industry and policy makers. The DEIS group employs several PhDs in the areas of intelligent fault diagnosis and maintenance. 

Find out more
book

Natural Language Generation from Knowledge Graphs

Natural Language Generation from Knowledge Graphs (2017-2019) is a recent project led by Nina Dethlefs and funded by Stanford AI start-up Diffbot. The project developed a novel hierarchical decomposition approach for text generation that can produce coherent and grammatical text for a specific domain - biography texts were chosen - from a knowledge graph of relevant facts.

Find out more
School

Flexible Pedagogies

Flexible Pedagogies is a project led by Neil Gordon developing approaches that utilise intelligent systems to enable effective education. The National Health Service have been using it in their guidance on implementing technology based training, and more recently was influential in shaping the successful Chinese government response to the virus in terms of their education system.

Find out more
View all projects

Philosophy and digital art projects

Exploring digital art and various aspects of philosophy and its intersections with science: TIMAEUS, VIRTUAL STOA,​ GeNeRaTiVe aRt project and ODYSSEY.

Safety of Autonomous Systems with Jaguar Land Rover

SAS-JLR. 2018 - 2020.

Data-driven Reliability-centred Evolutionary Asset Manager

DREAM. 2018 - 2021.

Dependability Engineering Innovation for Cyber Physical Systems

DEIS. 2017 - 2020. Find out more. 

Model-based Analysis & Engineering of Novel Architectures for Dependable Electric Vehicles

MAENAD. 2011 - 2015. Find out more.

Dependable Telehealth

DELTA. 2012 - 2014.

Advancing Traffic Efficiency and Safety through Software Technology phase 2

ATESST2 - 2009 - 2011. Find out more.

SAFEDOR

2005 - 2008. Find out more.

Automated Safety Analysis

ASA. 2007.

Optimal Allocation

OPAL. 2003 - 2005.

Safety of Autonomous Systems with Jaguar Land Rover (SAS-JLR)
The project examines potential safety flaws in advanced driver assistance systems and software used in autonomous cars. The aim is to derive analyses and methods that could help improve safety. (2018 - 2020)
HiP-HOPS

HiP-HOPS (1999-present) is a long-term project that has been partly funded by industry including Daimler, Volvo, Germanischer Lloyd and others. Since 1999 the project has developed the HiP-HOPS method and tools. In HiP-HOPS, algorithms examine the design of a system and locate potential flaws. This is done via model-based construction and analysis of fault trees and FMEAs that can record not only combinations but sequences of faults.

Analysis of sequences is facilitated by PANDORA, a temporal logic invented in Hull. HiP-HOPS incorporates bio- and nature-inspired computational techniques that optimise dependability versus cost, for example via optimal selection among design alternatives for components and subsystems or via optimal allocation and replication of software tasks on controllers. HiP-HOPS enables a dependability-driven mode of design using metaheuristics that cost-optimally allocate system safety requirements as integrity requirements to subsystems and system components. New work on nature-inspired algorithms which imitate the social intelligence of penguins was developed and applied to this problem.

HiP-HOPS creates model-connected safety cases, i.e. electronic certification documents. Techniques have been developed to enable its application on product lines. It is possible, for example, to auto-allocate integrity requirements across a product line in a way that components can meet cost-optimally the dependability requirements of several products. Recent extensions with fuzzy logic, and Bayesian Nets address the dynamic nature of modern systems and uncertainties that arise from limited observability, imperfect data, and unpredictability. Machine learning has been used to repair predictive safety models like fault trees in real time when these deviate from their predictive analyses. These new features address important problems in the dependability of autonomous, cyber-physical and open systems of systems.

Data-driven Reliability-centred Evolutionary Asset Manager (DREAM)
DREAM (2018-2021) is a project funded by EDF where HiP-HOPS is extended with machine learning to develop data-driven, reliability-centred, bio-inspired maintenance optimisation techniques that exploit data-driven prognoses of remaining useful life of components for production of dynamic maintenance plans exploiting HiP-HOPS algorithms. This is being taken up by EDF who operate wind farms in the UK, and EDF France, with exciting potential for further impact in the wind energy industry.
Dependability Engineering Innovation for Cyber Physical Systems (DEIS)
DEIS (2017-2020) is an H2020 project investigating intelligent loosely-connected systems where we developed Digital Dependability Identities (DDI), i.e., modular, composable and executable specifications of dependability for use in complex cyber-physical systems and open Systems-of-Systems. DDIs have drawn heavily from earlier work on HiP-HOPS and EAST-ADL and are a public specification with metamodels and tools in the public domain. SESAME (2021-2024) is a new large H2020 project on security and safety of multirobot systems, with 12 industrial partners. The project will transfer the concept of DDIs via five applications that involve Robots and Autonomous Systems in four domains of application.
MAENAD & EAST-ADL
MAENAD & EAST-ADL (2011-present) belongs in a series of European projects where we co-developed EAST-ADL, an emerging architecture description language authored as an industry standard for the design of vehicles. Several metamodels of HiP-HOPS were transferred into the Safety Annex of this language which is gaining substantial influence in the automotive sector. Hull is sitting on the EAST-ADL association board tasked with the maintenance and evolution of the language. The effort also led to a collaboration with Metacase, a Finnish company which commercialises EAST-ADL products. Within the period, Metacase has developed a HiP-HOPS extension for EAST-ADL within its tool Metaedit+
Spencer-Data Mining KTP
Spencer-Data Mining KTP (2020-2022) is a Knowledge Transfer Partnership (KTP) project on text mining for corporate documents with Spencer Ltd that is supervised by Nina Dethlefs. We apply research in data analytics, natural language processing and artificial intelligence to analyse a large repository of disparate types of company documents - relating to projects - to identify patterns and extract, organise and summarise information. The overall goal is to embed AI into company workflows to increase productivity and support effective and informed decision making.
Aura CDT on Offshore Wind and the Environment

Aura CDT on Offshore Wind and the Environment (2019- present) is a doctoral training centre with Sheffield, Durham and Newcastle, EPSRC/NERC-funded. The Aura Centre for Doctoral Training seeks to enable innovation for the offshore wind sector in a unique collaborative approach of cross-disciplinary engagement between engineering and environmental scientists, industry and policy makers. The centre provides the opportunity to develop and integrate practical examples in taught courses across the first year with industry-led and challenge-led projects, followed by three years of focused doctoral research. The DEIS group employs several PhDs in the areas of intelligent fault diagnosis and maintenance.

Natural Language Generation from Knowledge Graphs

Natural Language Generation from Knowledge Graphs (2017-2019) is a recent project led by Nina Dethlefs and funded by Stanford AI start-up Diffbot. The project developed a novel hierarchical decomposition approach for text generation that can produce coherent and grammatical text for a specific domain - biography texts were chosen - from a knowledge graph of relevant facts. The project was designed to fill a gap in current research that focuses mostly on machine reading and information extraction using knowledge graphs with much less work being done in language generation.

Flexible pedagogies
Flexible Pedagogies is a project led by Neil Gordon developing approaches that utilise intelligent systems to enable effective education. The keystone to this work was a 2014 Research Report for the Higher Education Academy (now AdvanceHE) on Technology Enhanced Learning and Flexible Pedagogies followed by a report on Issues in retention and attainment in Computer Science. This work was used by education institutions across the world as they developed their own approaches to computer supported teaching. This impact went further, the National Health Service using it in their guidance on implementing technology based training, and more recently was influential in shaping the successful Chinese government response to the virus in terms of their education system.

Members

Outputs and publications
  • Chatterjee, J., Dethlefs, N. (2020) A Hybrid Deep Learning Architecture for Anomaly Prediction in Wind Turbines: Towards Transparency and Transductive Transfer Learning. Wind Energy: 23(8), DOI:10.1002/WE.2510.
  • Kabir, S., Papadopoulos, Y. (2020) A Hybrid Modular Computational Intelligence for Safety Assurance of Cooperative Systems of Systems, IEEE Computer 53 (12): 24-34, IEEE
  • Awais M., Raza M., Singh N., Bashir K., Manzoor U., Islam S., Rodrigues J. (2020) “LSTM based Emotion Detection using Physiological Signals: IoT framework for Healthcare and Distance Learning in COVID-19” accepted for publication – IEEE Internet of Things.
  • Gheraibia, Y., Kabir, S., Aslansefat, K., Sorokos, I., & Papadopoulos, Y. (2019) Safety+ AI: A Novel Approach to Update Safety Models Using Artificial Intelligence. IEEE Access, DOI: 10.1109/ACCESS.2019.2941566, IEEE.
  • Kabir, S., Aslansefat, K., Sorokos, I., & Papadopoulos, Y. Konur S. (2020) A Hybrid Modular Approach for Dynamic Fault Tree Analysis, IEEE Access, DOI: 10.1109/ACCESS.2020.2996643.
  • Oliveira A., Braga R., Masiero P., Parker D., Papadopoulos Y., Habli I., Kelly T. (2019) Variability management in safety-critical systems design and dependability analysis, Journal of Software: Evolution and Process, DOI: 10.1002/smr.2202.
  • Gheraibia Y., Moussaoui A., Yin P., Papadopoulos Y. Maazouzi S. (2019) PeSOA: Penguins Search Optimisation Algorithm for Global Optimisation Problems, The Int’l Arab Journal of Information Technology (IAJIT), 16:1-9, May 2019, arXiv:1809.09895.
  • Brayshaw, M., Gordon, N. A., & Grey, S. (2019). Smart, social, flexible and fun: Escaping the flatlands of virtual learning environments. Advances in Intelligent Systems and Computing, 998:1047-1060, Springer.
  • Kabir S., Walker M., Papadopoulos Y. (2019) Applications of Bayesian networks and Petri nets in safety, reliability, and risk assessments: A review, Safety Science 115:154-175, DOI: 10.1016/j.ssci.2019.02.009, Elsevier.
  • Aizpurua J, Papadopoulos Y., Merle G. (2018) Explicit Modelling and Treatment of Repair in Prediction of Dependability, IEEE Transactions on Dependable and Secure Computing, DOI: 10.1109/TDSC.2018.2857810, IEEE.
  • Kabir S., Papadopoulos Y. (2018) A review of applications of fuzzy sets to safety and reliability engineering, International Journal of Approximate Reasoning, 100:29-55, DOI: 10.1016/j.ijar.2018.05.005, Elsevier.
  • Ding X., Luo Y, …, Wang B. (2018) Prior knowledge-based deep learning method for indoor object recognition and application, “Systems Science & Control Engineering”, Journal Systems Science & Control Engineering, DOI:10.1080/21642583.2018.1482477.
  • Xue D., Wang X., Davis D., Wang D. (2018) An Adaptive Ensemble Approach to Ambient Intelligence Assisted People Search, Applied System Innovation, DOI: 10.3390/asi1030033.
  • Kabir S., Yazdi M., Aizpurua J., Papadopoulos Y. (2018) Uncertainty-aware dynamic reliability analysis framework for complex systems, IEEE Access, DOI: 10.1109/ACCESS.2018.2843166, IEEE.
  • Kabir S., Walker M., Papadopoulos Y. (2018) Dynamic system safety analysis in HiP-HOPS with Petri Nets and Bayesian Networks, Safety Science, 105:55-70, DOI: 10.1016/j.ssci.2018.02.001, Elsevier.
  • Lampe J., Rüde E., Papadopoulos Y., Kabir S. (2018) Model-Based Assessment of Energy-Efficiency, Dependability, and Cost-Effectiveness of Waste Heat Recovery Systems Onboard Ship, Ocean Engineering, 157:234-250, DOI: 10.1016/j.oceaneng.2018.03.062, Elsevier.
  • Dethlefs, N. (2017) Domain Transfer for Deep Natural Language Generation from Abstract Meaning Representations. IEEE Computational Intelligence Magazine: Special Issue on Natural Language Generation with Computational Intelligence.
  • Dethlefs, N., Milders, M., Cuayahuitl. H., Al Salkini, T. and L. Douglas (2017) Natural Language in Assistive Technology for the Presentation of Cognitive Stimulation to People with Dementia. Informatics for Health and Social Care. Taylor & Francis.
  • Aizpurua, J. I., Catterson V. M., Papadopoulos Y., Chiacchio F, D. D’Urso (2017), Supporting group maintenance through prognostics-enhanced dynamic dependability prediction, Reliability Engineering & System Safety, DOI: 10.1016/j.ress.2017.04.005, Elsevier, ISSN 0951-8320.
  • Aizpurua, J. I., Papadopoulos, Y., Muxika, E., Chiacchio, F., Manno, G. (2017) On Cost-effective Reuse of Components in the Design of Complex Reconfigurable Systems. Quality Reliability Engineering International, doi:10.1002/qre.2112, Wiley International.
  • Aizpurua, J. I., Catterson V. M., Papadopoulos Y., Chiacchio F, Manno G. (2017) Improved Dynamic Dependability Assessment Through Integration with Prognostics. IEEE Transactions on Reliability, 99:1-21, ISSN 0018-9529.
  • Grey S., Grey D, Gordon N., Purdy J. (2017) Using Formal Game Design Methods to Embed Learning Outcomes into Game Mechanics and Avoid Emergent Behaviour. International Journal of Game-Based Learning, 7(3):63-73.
  • Chalupa D., Balaghan P., Hawick K., Gordon N. (2017) Computational Methods for Finding Long Simple Cycles in Complex Networks. Knowledge-Based Systems 125:96-107.
  • Dethlefs, N., Hastie, H., Cuayahuitl, H., Yu, Y., Rieser, V. and O. Lemon (2016) Information Density and Overlaps in Spoken Dialogue, Computer Speech and Language, 37:82–97.
  • Papadopoulos Y., Walker M., Parker D., Sharvia S., Bottaci L., Kabir S., Azevedo L., Sorokos I. (2016) A Synthesis of Logic and Bio-inspired techniques in the Design of Dependable Systems, Annual Reviews in Control, 41: 170-182, IFAC & Elsevier, ISSN: 1367-5788 (extension of plenary paper at IFAC- DCDS'15).
  • Kabir S., Walker M., Papadopoulos Y. Ruede E., Securius P. (2016), Fuzzy Temporal Fault Tree Analysis of Dynamic Systems, Int'l Journal of Approximate Reasoning, 77:20-37, Elsevier, ISSN 0888-613X 3.
  • Blom H., Chen D.J., Kaiser H., Lönn H., Papadopoulos Y., Reiser M.O., Kolagari T., Tucci S. (2016) EAST-ADL: An Architecture Description Language for Automotive Software-intensive Systems in the Light of Recent use and Research, Int'l Journal of System Dynamics Applications, 5(3):1-19, IGI Publishing, ISSN: 2160-9772.
  • Oliveira A. L., Rosana T. V. Braga, Paulo Cesar Masiero, Papadopoulos Y., Ibrahim Habli, Tim Kelly (in print) Model-Based Safety Analysis of Software Product Lines, Special issue on Critical and Real-time Cyber-physical Systems, Int'l Journal of Critical Embedded Systems, 8(5-6):412-426 (extension of best paper at IEEE SBES'14, ISSN 1741-1068).
  • Zhibao M., Bottaci L., Papadopoulos Y., Mahmud N. (2016) Model Transformation for analysing Dependability of AADL model by using HiP-HOPS, Journal of Systems and Software, DOI: 10.1016/j.jss.2019.02.019, Elsevier.
  • Aizpurua J.I., Muxika E., Papadopoulos Y., Chiacchio F., Manno G. (2016) Application of the D3H2 Methodology for Cost-effective Design of Dependable Systems, Safety, 2(2):9 - doi:10.3390/safety2020009, MDPI Publishers, ISSN 2313-576X.
  • Dethlefs, N. and Cuayahuitl, H. (2015) Hierarchical Reinforcement Learning for Situated Language Generation. Natural Language Engineering 21:391–435. Cambridge University Press.
  • Nefti-Meziani S., Manzoor U., Davis S., Pupala S.K (2015), 3D perception from binocular vision for a low-cost humanoid robot NAO, Robotics and Autonomous Systems 68:129-139.
  • Manzoor U., Zafar B. (2015) Multi-Agent Modelling Toolkit – MAMT, Simulation Modelling Practice and Theory, 49:215-227.
  • Sharvia S., Papadopoulos Y. (2015) Integrating model checking with HiP-HOPS in model-based safety analysis, Reliability Engineering and System Safety, 135:64–80, Elsevier, ISSN 0951-8320.

SN 1741-1068 (in print)

Research Students

Athanasion Retouniotis

Model-Connected Safety Cases

Papadopoulos, Parker

Koorosh Aslansefat

DREAM: Data-driven Reliability-centred, Evolutionary Asset Manager - Optimisation of Maintenance in Wind Farms

Dethlefs, Turner

Luis Torrao

TIMAEUS: Three-dimensional Illuminated Media-Augmented Ethereal-Unreal Sculptures Art Therapy Studio

Papadopoulos, McKie

Annika Schoene

Deep Learning with Knowledge Graphs for Fine-Grained Emotion Classification in Text

Dethlefs, Turner, Cheng

Declan Whiting

Safety and Security of Autonomous Systems

Papadopoulos, Manzoor

Joyjit Chatterjee

Explainable AI in Operations & Maintenance of Wind Turbines

Dethlefs, Turner

Onatkut Dagtekin

Modelling Environmental Data Systems with Deep Learning

Dethlefs, Parsons, Forster

Francis Kambili-Mzembe

Applying Immersive Virtual Reality for Teaching Stem Subjects To Supplement The Public Secondary School Curriculum In Malawi

Gordon, Brayshaw

Lydia Bryan-Smith

Using IoT and Big Data to dynamically map flood risk

Dethlefs, Parsons, McLelland, Kambhampati

Inspired?

Our research impacts the world. Come and join us.

Be part of a vibrant research community at the University of Hull.

Find out more

ink-cloud