System Identification and Data Modelling of Complex Nonlinear and Nonstationary Processes
Lead Research Organisation:
University of Sheffield
Department Name: Automatic Control and Systems Eng
Abstract
Many real-world systems in engineering and science are inherently nonlinear and nonstationary in nature. However, the analysis and modelling techniques commonly used to study many types of typically complex nonlinear and nonstationary processes often assume linearity and stationarity. Conventional signal processing techniques include non-parametric approaches (e.g. statistics and power spectral estimation) and linear parametric modelling approaches (e.g. linear time-invariant and simply time-varying models). It has been noticed that complex signals may contain hidden information which cannot be sufficiently revealed and characterised by using traditional analysis methods. This project is primarily aimed to investigate, adapt and develop system identification and data modelling methods and algorithms for the analysis of nonlinear and nonstationary complex dynamical processes in the time, frequency, time-frequency and spatio-temporal domains.
Planned Impact
There are many potential private and public beneficiaries and a broad range of potential impact arising from this project. 1. Who Will Benefit from the Research and How? 1) Academic Communities: Researchers from a broad range of areas (engineering, neuroscience, life science etc) can benefit from New Methods and Algorithms for example researchers can take up the basic ideas and methods proposed in this project to adapt and extend into new application areas; 2) Medical and Healthcare Communities: Clinical neurophysiology is still suffering from a lack of effective and efficient tools for predicting and detecting impending abnormalities (e.g. epileptic seizures) with sufficient specificity and reliability. One goal neurophysiologists want to achieve is to develop new tools that can automatically detect and predict epileptic seizures; this would enable a possibility to design implantable seizure warning devices that can provide valuable information for specific patients and associated drug delivery. The medical and healthcare community will benefit from the outcomes of the project by means of: a) New knowledge and expertise - The developed methods and algorithms will be applied to neuroimaging data analysis, from which the understanding about the inherent mechanisms of neurological dysfunctions and the associated dynamic interactions between different regions of the human brain can be significantly improved. b) New Medical Device Design. Biomedical systems engineers can design devices that: i) Can be used to predict, detect and diagnose epileptic seizures and other neurological dysfunctional disorders; ii) Can provide clinically important valuable information for specific patients and the associated drug delivery; iii) Can be used to investigate the dynamic interactions of different activated brain areas and thus significantly facilitate studies in cognitive science. 3) Team Members of this Project: They will directly, immediately benefit from the proposed research via Research and Academic Training - The project will be carried out in an internationally leading university, department and research group. The proposed project will provide a research and academic training platform for members involved in the proposed research. It can be envisaged that with a period of research and academic training within an excellent research and academic environment it will significantly facilitate and promote a future career development for research members here. 2. Plans to Maximise Potential Impacts 1) Enhancement of Interdisciplinary Co-operations: Interdisciplinary co-operations (e.g. with hospital or industry) will facilitate and promote knowledge delivery and transfer. 2) Peer-Reviewed Journal Publications: One of the most effective routes for communication with academics is by means of international peer-review journals. 3) Conference Presentations: General conferences, meetings, symposiums etc. are a good place to exchange information and ideas with researchers in interdisciplinary fields. 4) Web Presence: It will also utilise a web resource to provide a document repository e.g. pre-prints, multimedia presentations and illustrations. It will sufficiently make use of departmental and university webpage to communicate the project outcomes to a broad audience including our undergraduate and postgraduate students. 5) Other Routes: Other routes for communicating our work and findings include: internal and external seminars, workshops, within and inter-research-group meetings and discussions, inter-group day off visit and study.
Publications
Akinola T
(2019)
Non-linear system identification of solvent-based post-combustion CO2 capture process
in Fuel
Ayala Solares J
(2015)
Nonlinear model structure detection and parameter estimation using a novel bagging method based on distance correlation metric
in Nonlinear Dynamics
Ayala Solares J
(2016)
Modeling and prediction of global magnetic disturbance in near-Earth space: A case study for K p index using NARX models MODELING AND PREDICTION OF K p INDEX
in Space Weather
Balikhin M
(2011)
Using the NARMAX approach to model the evolution of energetic electrons fluxes at geostationary orbit NARMAX MODELLING OF RADIATION BELT ELECTRON FLUXES
in Geophysical Research Letters
Bashir F
(2015)
Using nonlinear models to enhance prediction performance with incomplete data
in ICPRAM 2015 - 4th International Conference on Pattern Recognition Applications and Methods, Proceedings
Bigg GR
(2014)
A century of variation in the dependence of Greenland iceberg calving on ice sheet surface mass balance and regional climate change.
in Proceedings. Mathematical, physical, and engineering sciences
Billings CG
(2013)
The prediction of in-flight hypoxaemia using non-linear equations.
in Respiratory medicine
Billings S
(2015)
Identification of nonlinear systems with non-persistent excitation using an iterative forward orthogonal least squares regression Algorithm
in International Journal of Modelling, Identification and Control
Boynton R
(2011)
Using the NARMAX OLS-ERR algorithm to obtain the most influential coupling functions that affect the evolution of the magnetosphere DATA-DEDUCED COUPLING FUNCTIONS
in Journal of Geophysical Research: Space Physics
Fei He
(2013)
Spectral Analysis for Nonstationary and Nonlinear Systems: A Discrete-Time-Model-Based Approach
in IEEE Transactions on Biomedical Engineering
Gu Y
(2019)
System Identification and Data-Driven Forecasting of AE Index and Prediction Uncertainty Analysis Using a New Cloud-NARX Model
in Journal of Geophysical Research: Space Physics
Gu Y
(2023)
Modelling Short-Term Appliance Energy Use with Interpretable Machine Learning: A System Identification Approach
in Arabian Journal for Science and Engineering
Gu Y
(2018)
Nonlinear predictive model selection and model averaging using information criteria
in Systems Science & Control Engineering
Gu Y
(2021)
Nonlinear Modeling of Cortical Responses to Mechanical Wrist Perturbations Using the NARMAX Method.
in IEEE transactions on bio-medical engineering
Gu Y
(2018)
A robust model structure selection method for small sample size and multiple datasets problems
in Information Sciences
Guo Y
(2016)
Ultra-Orthogonal Forward Regression Algorithms for the Identification of Non-Linear Dynamic Systems
in Neurocomputing
Guo Y
(2015)
Identification of continuous-time models for nonlinear dynamic systems from discrete data
in International Journal of Systems Science
Guo Y
(2014)
An iterative orthogonal forward regression algorithm
in International Journal of Systems Science
He F
(2014)
A nonlinear generalization of spectral Granger causality.
in IEEE transactions on bio-medical engineering
He F
(2016)
Nonlinear interactions in the thalamocortical loop in essential tremor: A model-based frequency domain analysis.
in Neuroscience
He F
(2014)
A nonlinear causality measure in the frequency domain: nonlinear partial directed coherence with applications to EEG.
in Journal of neuroscience methods
He F
(2013)
Identification and frequency domain analysis of non-stationary and nonlinear systems using time-varying NARMAX models
in International Journal of Systems Science
Jiang R
(2016)
Privacy-Protected Facial Biometric Verification Using Fuzzy Forest Learning
in IEEE Transactions on Fuzzy Systems
Li P
(2013)
Nonlinear Model Identification From Multiple Data Sets Using an Orthogonal Forward Search Algorithm
in Journal of Computational and Nonlinear Dynamics
Li Y
(2012)
Time-varying linear and nonlinear parametric model for Granger causality analysis.
in Physical review. E, Statistical, nonlinear, and soft matter physics
Li Y
(2019)
A Parametric Time-Frequency Conditional Granger Causality Method Using Ultra-Regularized Orthogonal Least Squares and Multiwavelets for Dynamic Connectivity Analysis in EEGs.
in IEEE transactions on bio-medical engineering
Li Y
(2018)
Time-Varying System Identification Using an Ultra-Orthogonal Forward Regression and Multiwavelet Basis Functions With Applications to EEG.
in IEEE transactions on neural networks and learning systems
Li Y
(2015)
Identification of nonlinear time-varying systems using an online sliding-window and common model structure selection (CMSS) approach with applications to EEG
in International Journal of Systems Science
Li Y
(2011)
Time-varying model identification for time-frequency feature extraction from EEG data.
in Journal of neuroscience methods
Li Y
(2019)
Multimodal hyper-connectivity of functional networks using functionally-weighted LASSO for MCI classification.
in Medical image analysis
Li Y
(2011)
Identification of Time-Varying Systems Using Multi-Wavelet Basis Functions
in IEEE Transactions on Control Systems Technology
Lv S
(2015)
An improved image distortion correction algorithm
in Multimedia, Communication and Computing Application - Proceedings of the International Conference on Multimedia, Communication and Computing Application, MCCA 2014
Marshall AM
(2016)
Quantifying heterogeneous responses of fish community size structure using novel combined statistical techniques.
in Global change biology
Nikentari N
(2022)
Multi-Task Learning for Time Series Forecasting Using NARMAX-LSTM
Nikentari N
(2022)
Tide Level Prediction Using NARX-based Recurrent Neural Networks
Rastätter L
(2013)
Geospace environment modeling 2008-2009 challenge: D st index
in Space Weather
Sarrigiannis PG
(2015)
Direct Functional Connectivity between the Thalamus (Vim) and the Contralateral Motor Cortex: Just a Single Case Observation or a Common Pathway in the Human Brain?
in Brain stimulation
Sarrigiannis PG
(2014)
Quantitative EEG analysis using error reduction ratio-causality test; validation on simulated and real EEG data.
in Clinical neurophysiology : official journal of the International Federation of Clinical Neurophysiology
Senawi A
(2017)
A new maximum relevance-minimum multicollinearity (MRmMC) method for feature selection and ranking
in Pattern Recognition
Description | 1. Inspired new research collaborations 2. Found new multidisciplinary applications |
Exploitation Route | The output of the research has been applied to neurophysiological, environmental, societal and ecological systems modelling and analysis, as well as application in engineering. |
Sectors | Aerospace, Defence and Marine,Agriculture, Food and Drink,Digital/Communication/Information Technologies (including Software),Education,Environment,Healthcare,Other |
Description | In the university open days, we explained to the prospectus students how control and systems engineering approach can help understand and improve life science and healthcare. |
First Year Of Impact | 2013 |
Sector | Aerospace, Defence and Marine,Agriculture, Food and Drink,Digital/Communication/Information Technologies (including Software),Education,Environment,Healthcare,Pharmaceuticals and Medical Biotechnology,Other |
Impact Types | Cultural,Societal,Economic |
Description | Ecosystem modelling - Norwegian Research Council, CoDINA: DIet and food web dyNAmics |
Organisation | Norwegian Institute of Marine Research |
Country | Norway |
Sector | Academic/University |
PI Contribution | The signal and data modelling methods and algorithms developed by my research team was/is applied to model the dominant relationships between cod population, environment and prey species, and to consider changes in these over time in the Barents Sea. |
Collaborator Contribution | Providing ecosystem data and background knowledge |
Impact | 1. paper publication Marshall, A. M., Bigg, G. R., van Leeuwen, S. M., Pinnegar, J. K., Wei, H.-L., Webb, T. J. and Blanchard, J. L. (2016), 'Quantifying heterogeneous responses of fish community size structure using novel combined statistical techniques', Global Change Biology,22(5), 1755-1768, DOI:10.1111/gcb.13190. 2. Success in PhD Scholarship Application NERC Doctoral Training Partnership: ACCE: Adapting to the Challenges of a Changing Environment 09/2017 - 03/2021 |
Start Year | 2015 |
Description | Towards a better understanding of epilepsy through EEG study |
Organisation | Royal Hallamshire Hospital |
Country | United Kingdom |
Sector | Hospitals |
PI Contribution | I conducted leading research in this field using control and systems engineering approach, and produced a number of world-leading outputs. In collaboration with NHS neurologists in the Royal Hallamshire Hospital, the team has developed the first reliable method to detect the onset of epileptic seizure several seconds before the seizure occurs. An electroencephalogram (EEG) is routinely used in the evaluation of brain disorders including the diagnosis and treatment of epilepsy. The group has introduced fundamentally new algorithms to model time varying processes, to track rapid parameter variations, and to map these to frequency domain behaviours. The algorithms have now been incorporated into existing EEG software packages and following full ethical approval has been used in clinical practice at the Sheffield Teaching Hospital Trust. |
Collaborator Contribution | The neurophysiologists in the Royal Hallamshire Hospital helped conduct experiments and recording of EEG data that used for the research, and test and verify how the proposed control and systems engineering methods work for most challenging problems in their areas. |
Impact | Sarrigiannis, P.G., Zhao, Y., He, F., Wei, H.L., Billings, S.A., Lawrence, S., Rowe, J., Romanowski, C., Hoggard, N., Rao, D.G., Grünewald, R., Khan, A., Hadjivassilliou, M., Yianni, J. (2015) 'Direct functional connectivity between the thalamus (Vim) and the contralateral motor cortex: Just a single case observation or a common pathway in the human brain?', Brain Stimulation, 8(6), 1230-1233. Li, Y., Wei, H.-L., Billings, S.A., Sarrigiannis, P.G. (2015) 'Identification of nonlinear time-varying systems using an online sliding-window and common model structure selection (CMSS) approach with applications to EEG', International Journal of Systems Science (in press) Sarrigiannis, P.G., Zhao, Y., Wei, H.L., Billings, S. A., Fotheringham, J., and Hadjivassiliou, M. (2014) 'Quantitative EEG analysis using error reduction ratio-causality test: validation on simulated and real EEG data', Clinical Neurophysiology, 125(1), 32-46. He, F., Billings, S. A., Wei, H.L., and Sarrigiannis, P.G. (2014) 'A nonlinear generalization of spectral Granger causality', IEEE Transactions on Biomedical Engineering, 61(6), 1693-1701. He, F., Wei, H.L., Billings, S. A., Sarrigiannis, P.G. (2014) 'A nonlinear causality measure in the frequency domain: Nonlinear partial directed coherence with applications to EEG', Journal of Neuroscience Methods, 225, 71-80. Zhao, Y., Billings, S. A., Wei, H.L., and Sarrigiannis, P.G. (2013) 'A parametric method to measure time-varying linear and nonlinear causality with applications to EEG data', IEEE Transactions on Biomedical Engineering, 60(11), 3141-3148. He, F., Billings, S. A., Wei, H.L., Sarrigiannis, P.G., and Zhao, Y. (2013) 'Spectral analysis for nonstationary and nonlinear systems: a discrete-time model based approach', IEEE Transactions on Biomedical Engineering, 60(8), 2233-2241. Zhao, Y., Billings, S.A., Wei, H.L., He, F., and Sarrigiannis, P.G. (2013) 'A new NARX-based Granger linear and nonlinear casual influence detection method with applications to EEG data', J. Neuroscience Methods, 212(1), 79-86. Zhao, Y., Billings, S.A., Wei, H.L., and Sarrigiannis, P.G. (2012) 'Tracking time-varying causality and directionality of information flow using an error reduction ratio test with applications to electroencephalography data', Physical Review E, 86(5), art. no. 051919. Li, Y., Wei, H.L., Billings, S.A., and Sarrigiannis, P.G. (2011) 'Time-varying model identification for time-frequency feature extraction from EEG data', Journal of Neuroscience Methods,196(1),151-158. |
Start Year | 2011 |
Description | 2015 International Conference on Computer Science and Education |
Form Of Engagement Activity | A talk or presentation |
Part Of Official Scheme? | No |
Geographic Reach | International |
Primary Audience | Professional Practitioners |
Results and Impact | The audience were interested in the methods presented. |
Year(s) Of Engagement Activity | 2015 |
Description | Informal workshop - towards a further understanding of epilepsy and dementia through EEG analyses |
Form Of Engagement Activity | Participation in an activity, workshop or similar |
Part Of Official Scheme? | No |
Geographic Reach | Local |
Primary Audience | Professional Practitioners |
Results and Impact | All participants unanimously agreed that some enhanced and relatively larger scale collaborations are needed to further and well address challenges from neurophysiology studies using control and systems engineering approaches. |
Year(s) Of Engagement Activity | 2016 |