System Identification and Information Processing for Complex Systems

Lead Research Organisation: University of Sheffield
Department Name: Automatic Control and Systems Eng

Abstract

The main aim of this Platform Grant proposal is to build upon the expertise of the Sheffield team in signal processing and system identification for nonlinear and complex systems to develop new research themes and application areas. The applicants developed the narmax methodology, a theory for nonlinear systems in the frequency domain, and a class of algorithms for the identification and analysis of spatio-temporal systems. Recent developments at Sheffield mean we are now in a unique position to develop new multi-disciplinary research themes to exploit and build upon this expertise. The new multi-disciplinary research directions that we plan to develop as part of this research proposal include studies of drosophila brain, stem cell dynamics, diffuse optical tomography of brain, space weather systems, and crystal growth phenomena. Each of these offers unique challenges in signal processing and system identification and each requires scoping or feasibility studies to establish the potential to develop each new research area into a core research proposal. The core theme of this Platform Grant proposal is therefore to provide funding to conduct initial scoping research studies into each of these new multi-disciplinary projects with the objective that this will allow the development of each theme to the point where a responsive mode application can be submitted to take the research forward.Career development for members of our group including RA's is a core part of our strategy. The Platform Grant will provide stability for the collaboration by our research team with other disciplines, and should provide a critical mass of researchers and a longer term stable environment around which the career development of outstanding young researchers can be nurtured. The Platform Grant will enable the group to maintain the momentum that has been generated by the formation of two new research centres, and to build upon and exploit the multi-disciplinary partnerships that have been developed to date. The opportunity to take risks, try out completely new approaches, and completely new application fields produces an excitement in the research, enhances moral and longer term strategic research planning and development, allows flexibility, creates opportunities, and provides support for everyone in our group.

Publications

10 25 50

publication icon
Li Y (2012) Time-varying linear and nonlinear parametric model for Granger causality analysis. in Physical review. E, Statistical, nonlinear, and soft matter physics

publication icon
Li Y (2011) Identification of Time-Varying Systems Using Multi-Wavelet Basis Functions in IEEE Transactions on Control Systems Technology

 
Description New ways to model and analyse complex nonlinear dynamic systems.
Exploitation Route They are generic system theory results and so can be used by others and applied in many different fields.
Sectors Aerospace, Defence and Marine,Agriculture, Food and Drink,Environment,Healthcare,Pharmaceuticals and Medical Biotechnology

 
Description The results from this research have been used in a wide range of multidisciplinary applications. Space weather: Models and algorithms developed as part of this research have been used to develop the most accurate models for forecasting geomagnetic indices, solar wind parameters and the radiation environment in the geospace. The forecasts are continuously updated based on real-time data and made available online. NASA is using the models to forecast radiation bursts that could damage or render 'blind' satellites. Systems biology Modelling methodologies developed as part of this grant were instrumental in the development of advanced mathematical models of the fruit fly photoreceptors. These models are now publicly available as part of the fruit fly brain observatory, a major resource for modelling, simulation and visualisation of the Drosophila brain. Modelling and analysis methodologies developed as part of this grant were also instrumental in furthering our understanding of stem cell heterogeneity, adaptation and differentiation. Systems medicine: Modelling and analysis methodologies arising from this grant have been used to understand disease mechanisms using time-course genome-wide gene expression data sets - polycystic kidney disease, acute coronary syndrome- as well as (currently) the impact of ageing on macrophage activity. Signal and information processing algorithms arising from this research formed the basis for novel algorithms for predicting the onset of epilepsy and diagnosing Alzheimer's disease based on EEG data Econometrics: New methods and models for volatility forecasting are being used to better estimate value-at-risk Conflict modelling: Research carried out as part of the project led to new dynamic spatiotemporal modelling tools for the identification of complex underlying processes in conflict, such as diffusion, relocation, heterogeneous escalation, and volatility. The models were used to provide a strikingly statistically accurate forward prediction of armed opposition group activity in Afghanistan in 2010, based solely on data from previous years. The models have been made available to NATO following publication. Climate change: Nonlinear system identification and coupled ocean-iceberg modelling, was used to demonstrate that variability in the number of Greenland icebergs recorded over more than 100 years is predominantly caused by fluctuation in the Greenland ice sheet calving discharge rather than open ocean iceberg melting.
First Year Of Impact 2014
Sector Aerospace, Defence and Marine,Agriculture, Food and Drink,Chemicals,Communities and Social Services/Policy,Digital/Communication/Information Technologies (including Software),Environment,Financial Services, and Management Consultancy,Healthcare,Pharmaceuticals and Medical Biotechnology,Security and Diplomacy
Impact Types Societal

 
Description Artificial Intelligence Technology for Cell Therapy Manufacturing
Amount £80,000 (GBP)
Funding ID 2106198 
Organisation Medical Research Council (MRC) 
Sector Public
Country United Kingdom
Start 09/2018 
End 08/2022
 
Description CMMI-EPSRC - Right First Time Manufacture of Pharmaceuticals (RiFTMaP)
Amount £1,543,632 (GBP)
Funding ID EP/V034723/1 
Organisation Engineering and Physical Sciences Research Council (EPSRC) 
Sector Public
Country United Kingdom
Start 07/2021 
End 08/2024
 
Description DAFNI-ROSE
Amount £1,083,755 (GBP)
Funding ID EP/V054082/1 
Organisation Engineering and Physical Sciences Research Council (EPSRC) 
Sector Public
Country United Kingdom
Start 05/2021 
End 06/2023
 
Description Open Science Prize
Amount $80,000 (USD)
Organisation Wellcome Trust 
Sector Charity/Non Profit
Country United Kingdom
Start 03/2016 
End 12/2016
 
Description Platform Grant
Amount £1,210,660 (GBP)
Funding ID EP/H00453X/1 
Organisation Engineering and Physical Sciences Research Council (EPSRC) 
Sector Public
Country United Kingdom
Start 01/2010 
End 01/2015
 
Description Program Grant
Amount $900,000 (USD)
Funding ID RGP0001/2012 
Organisation Human Frontier Science Program (HFSP) 
Sector Charity/Non Profit
Country France
Start 05/2012 
End 06/2015
 
Description Responsive Mode
Amount £480,313 (GBP)
Funding ID EP/G042209/1 
Organisation Engineering and Physical Sciences Research Council (EPSRC) 
Sector Public
Country United Kingdom
Start 01/2010 
End 01/2014
 
Description Responsive Mode
Amount £615,105 (GBP)
Funding ID BB/K010123/1 
Organisation Biotechnology and Biological Sciences Research Council (BBSRC) 
Sector Public
Country United Kingdom
Start 03/2013 
End 04/2016
 
Description SURE: Sheffield Undergraduate Research Experience
Amount £1,080 (GBP)
Organisation University of Sheffield 
Sector Academic/University
Country United Kingdom
Start 05/2018 
End 08/2018
 
Description Sustainable microwave manufacturing of functional inorganic materials (SuMMa)
Amount £1,683,171 (GBP)
Funding ID EP/W018950/1 
Organisation Engineering and Physical Sciences Research Council (EPSRC) 
Sector Public
Country United Kingdom
Start 03/2022 
End 02/2025
 
Description System Identification and Data Modelling of Complex Nonlinear and Nonstationary Processes
Amount £101,004 (GBP)
Funding ID EP/I011056/1 
Organisation Engineering and Physical Sciences Research Council (EPSRC) 
Sector Public
Country United Kingdom
Start 05/2011 
End 08/2012
 
Description The Active Building Centre Research Programme (ABC RP)
Amount £9,324,025 (GBP)
Funding ID EP/V012053/1 
Organisation Engineering and Physical Sciences Research Council (EPSRC) 
Sector Public
Country United Kingdom
Start 03/2020 
End 09/2022
 
Description UK Collaboratorium for Research in Infrastructure & Cities: Urban Observatories (Strand B)
Amount £8,000,000 (GBP)
Funding ID EP/P016782/1 
Organisation Engineering and Physical Sciences Research Council (EPSRC) 
Sector Public
Country United Kingdom
Start 03/2017 
End 03/2022
 
Description UKCRIC National Infrastructure Database, Modelling, Simulation and Visualisation Facilities
Amount £8,000,000 (GBP)
Funding ID EP/R012202/1 
Organisation Engineering and Physical Sciences Research Council (EPSRC) 
Sector Public
Country United Kingdom
Start 03/2017 
End 09/2021
 
Title Bioinformatics tools develoment 
Description Development of bioinformatics pipeline for inference of dynamic models of Gene Regulatory Networks 
Type Of Material Improvements to research infrastructure 
Year Produced 2017 
Provided To Others? No  
Impact Analysis of genome wide gene expression array data from polycystic kidney disease studies 
 
Title Early Olfaction Parkinson's Disease Model 
Description This app investigates the effect of the Parkinson's Disease in the early olfactory system. The antennal lobe is the first neuropil in the olfaction sensory pathway, which can be decomposed to ~50 subregions, called glomerulus. Each glomerulus consists of some 50 olfactory sensory neurons (input), and 3-5 projection neurons (output). Glomeruli are interconnected by over 100 local neurons. In total, the antennal lobe has about 2500 olfactory sensory neurons, 150 projection neurons, and 100 local neurons. A healthy model of the antennal lobe developed by the research team is simulated for a given input odorant profile. The ensemble response of projection neurons of each glomerulus is viusalised in both biological view and as an interactive plot. A separate disease model based on the hypothesis that the Parkinson's Disease causes the abnormal inhibitory neurotransmitter (GABA) release was also implemented. The diseased model is simulated with the same input odorant as the healthy model, and its response is provide below alongside the result of the healthy model. 
Type Of Material Model of mechanisms or symptoms - human 
Year Produced 2018 
Provided To Others? Yes  
Impact NA at this time 
URL https://neuroapps.fruitflybrain.org/parkinsons/olfaction/
 
Title FPGA System for Protein Identification and Quantitation 
Description Developed high-performance FPGA-based bioinformatics solution for high-throughput proteomics. Solution consists of a PC server equipped with FPGA board, FPGA implementation of data processing and database search algorithms, server side user interface and data processing software. 
Type Of Material Improvements to research infrastructure 
Year Produced 2010 
Provided To Others? Yes  
Impact The hardware-implemented algorithms for de-noising, baseline correction, peak identification and deisotoping, running on a Xilinx Virtex 2 FPGA at 180 MHz, generate a mass fingerprint over 100 times faster than an equivalent algorithm written in C, running on a Dual 3 GHz Xeon workstation. 
URL https://www.liverpool.ac.uk/pfg/Pubs/files/00fe1160f83e43a71c582afd3e5ef1dc-28.html
 
Title Fruit Fly Brain Observatory 
Description Mental and neurological disorders pose major medical and socioeconomic challenges for society. Understanding human brain function and disease is arguably the biggest challenge in neuroscience. To address this challenge, smaller but sufficiently complex brains like that of the fruit fly have been increasingly used for investigating the mechanisms of human neurological and psychiatric disorders, such as Epilepsy or Parkinson's disease, at molecular, cellular and circuit level. The Fruit Fly Brain Observatory (FFBO) is an open source software platform that stores and processes data related to the neural circuits of the fly brain including location, morphology, connectivity and biophysical properties of every neuron; seamlessly integrates the structural and genetic data from multiple sources that can be queried, visualized and interpreted; automatically generates models of the fly brain that can be simulated efficiently using multiple Graphics Processing Units (GPUs) to help elucidate the mechanisms of human neurological disorders and identify drug targets. 
Type Of Material Improvements to research infrastructure 
Year Produced 2016 
Provided To Others? Yes  
Impact Using computational disease models, researchers make targeted modifications that are difficult to perform in vivo with current genetic techniques. Models of neural circuits affected by disease will enable parallel recording of the responses of multiple components of a model circuit that are currently difficult - if not impossible - to perform in vivo. These capabilities significantly accelerate the development of powerful new ways to predict the effects of pharmaceuticals upon neural circuit functions. The software platform has been used to develop and simulate models of Parkinson's disease, retinal degeneration and epilepsy. 
URL http://fruitflybrain.org/
 
Title Method for modelling adaptation in sensory systems 
Description Adaptation is a fundamental characteristic of sensory processing. It enables sensory neurones to map efficiently the extensive range of environmental signals onto their limited dynamic range in order to prevent saturation and to maximize the amount of information collected. We developed a novel two-step approach for identifying both the nonlinear dynamical model and the time evolution of the gain of a self-adaptive sensory system based on experimental data. 
Type Of Material Improvements to research infrastructure 
Year Produced 2012 
Provided To Others? Yes  
Impact We have demonstrated that the processing of visual stimuli at photoreceptor level involves nonlinear transformations which extract and encode efficiently the biologically relevant higher-order statistical properties of natural stimuli. In particular, we have shown that photoreceptors detect and encode local phase correlations, which occur at the location of an edge or line, as well as long-range phase correlations, which characterize symmetry and texture properties of natural images. 
URL http://www.dflybrain.org/GainControlModel.html
 
Title Modelling cell heterogeneity tool 
Description Modelling tool for characterising dynamical mechanisms which underpin heterogeneity of cell populations. It allows inferring a dynamical model based on sequences of distribution functions generated using Fluorescence Activated Flow Cytometry. 
Type Of Material Model of mechanisms or symptoms - human 
Year Produced 2013 
Provided To Others? Yes  
Impact Identification of substates within the stem cell compartment. 
 
Title Motility Analysis 
Description We developed a framework for characterizing quantitatively different cell lines in terms of their motility and colony-forming properties using individual cell trajectories extracted from time-lapse microscopy images. 
Type Of Material Data analysis technique 
Provided To Others? No  
Impact No major impact yet. We have started to apply this framework to characterize motility of polycystic kidney cells under the the influence of different drugs. 
 
Title NeuroApp - retina degeneration 
Description This app that is part of the Fruit Fly Brain Observatory implements a simulation model of retinal degeneration and a rescue scheme by optogenetics means. The compound eye of the fruit fly consists of 700-800 facets, called ommatidia. We model the retina here to have 721 ommatidia, positioned on a hemispherical surface. Each ommatidium hosts 8 photoreceptors. For simplicity, we consider only one photoreceptor in each ommatidium. Each photoreceptor has a microvillar structure called rhabdomere, which is the functional equivalent of the rod and cone outer segment in vertebrate retina. The rhabdomere contains ~30000 microvilli where rhodopsin (light receptor) are hosted. Retinal degeneration often results in reduced size of rhabdomere and partially or completely lacking of rhodopsin. Consequently, the diseased eye has low or no light sensitivity. This simulate a diseased eye by reducing the number of microvilli in each photoreceptor to 5% of the health photoreceptors. The simulated responses of the diseased photoreceptor array can be visualised in real time. 
Type Of Material Model of mechanisms or symptoms - human 
Year Produced 2018 
Provided To Others? Yes  
Impact NA 
URL https://neuroapps.fruitflybrain.org/retinal_degeneration/
 
Title NeuroApp: Epilepsy model 
Description This application implements a whole-brain simulation model of epileptic activity using the fruit-fly brain connectome. In the Healthy model, when stimulus is onset at 2000ms, average rate rise to 2Hz and decrease to below 0.5Hz. Epilepsy is induced by diseased sodium channel that leads to an average firing rate change rapidly from 5Hz to 25Hz. In a rescued model. high activity only emerges in 2000~2100ms. After 2100ms brain reaches a low activity level. 
Type Of Material Model of mechanisms or symptoms - human 
Year Produced 2018 
Provided To Others? Yes  
Impact NA at this time 
URL https://neuroapps.fruitflybrain.org/epilepsy/
 
Title NeuroApp: Parkinsons Disease Model 
Description Parkinson's is a progressive neurological condition. It is estimated that 6.3 million people have Parkinson's worldwide. Some people with Parkinson's disease notice that as the disease progresses their vision loses sharpness or becomes blurred. Difficulties related to the eyes and vision often progress alongside other Parkinson's symptoms. This model-based study investigates the role of Parkinson's disease in the Visual System, using a mathematical model of the early visual system of Drosophila. It has been shown experimentally that both human, and fly models of the disease show a loss of visual acuity, caused by underlying changes in the photoreceptors ability to deal with changing light contrast levels. The disease state is modelled as a change in the capacity of photoreceptors to adapt to changes in contrast in the visual stimuli. 
Type Of Material Model of mechanisms or symptoms - human 
Year Produced 2018 
Provided To Others? Yes  
Impact NA at this time 
URL https://neuroapps.fruitflybrain.org/parkinsons/vision/
 
Title NeuroGFX 
Description NeuroGFX provides an environment to easily explore circuit structure and function ultimately leading to biological validation. On the whole brain level NeuroGFX lays out the guidelines for the development of whole brain emulation. On the neuropil level, NeuroGFX allows users to study the I/O of each LPU. The canonical circuits (circuit motifs) are also identified on this level and NeuroGFX can be used to study the effect of different compositions mediated by local neurons. NeuroGFX features a set of highly intuitive tools for exploring the function of neural circuit models, which can be accessed through. It provides a graphical user interface (GUI), allowing the user that can be used to (i) associate circuit diagrams with biological data, (ii) graphically construct an in silico experiment and execute manipulated circuits on GPUs, (iii) visualize the execution results in the context of biological brain structure. These capabilities are supported by a seamless integration of the NeuroArch database and the Neurokernel engine in the FFBO architecture. 
Type Of Material Improvements to research infrastructure 
Year Produced 2018 
Provided To Others? Yes  
Impact This tool is actively used by a number of research groups in US and UK 
URL https://neurogfx.fruitflybrain.org/
 
Title NeuroNLP 
Description NeuroNLP provides a modern web-based portal for navigating biological data relating to fruit fly brain circuits. It is equipped with a user-friendly, graphical interface to aggregate cell-type, connectome, synaptome and physiology data in the NeuroArch database, with the ability to simultaneously query against and retrieve information from disparate datasets. NeuroNLP features a novel natural language interface that constructs complex queries against the underlying database from plain English instructions such as "show GABAergic neurons that have dendrites in left antennal lobe and axons in both left lateral horn and right dorsolateral protocerebrum" (or simply "show GABAergic neurons that have dendrites in al and axons in both lh and DLP"). This provides highly intuitive access to the integrated fruit fly brain circuit data, without the presumption of knowledge of a query language, syntax or cumbersome user interfaces. The results of the queries are presented using powerful 3D visualization and can be shared using a tag for publication and collaboration. In addition, any neuron in the scene can be explored in greater detail using the information panel, which provides a one stop access to all data associated with a particular neuron. 
Type Of Material Improvements to research infrastructure 
Year Produced 2018 
Provided To Others? Yes  
Impact The tool is being used both by researchers as well as educators 
URL https://neuronlp.fruitflybrain.org/
 
Title New algorithms for image segementation and tracking 
Description New algorithms for image segmentation and tracking based on geometric active contours and information theoretic criteria were developed and used to process images obtained from time-lapse experiments and generate data automatically. Previously, because the existing commercial/open source software available did not work well with the phase contrast images generated by the time-lapse experiments carried out in the lab, the imaging data used to be sent to a group in Germany for analysis. 
Type Of Material Improvements to research infrastructure 
Year Produced 2010 
Provided To Others? Yes  
Impact The new algorithms were used to analyse time-lapse images of normal and adapted hESC colonies providing information about the movement type (random difussion, directed diffusion etc) and interactions (number and duration of contacts with other cells) of individual stem cells in vitro. A journal paper which will incorpoare rhe results of these studies is currently in preparation. 
 
Title New identification approach for neural circuits 
Description We developed two new algorithmic tool for identifying sensory circuit models consisting of linear and nonlinear filters in series with spiking neuron models, based on experimental data. 
Type Of Material Improvements to research infrastructure 
Year Produced 2018 
Provided To Others? Yes  
Impact The new tool was used successfully to analyse data from the Allen Brain Atlas. Too early to quantify the full impact. 
URL https://www.mitpressjournals.org/doi/abs/10.1162/neco_a_01051
 
Description Ann-Shyn Chiang Laboratrory Taiwan 
Organisation National Tsing Hua University (Taiwan)
Country Taiwan, Province of China 
Sector Academic/University 
PI Contribution Developed analysis and visualisation tools for the fruit fly connectome. See http://fruitflybrain.org/.
Collaborator Contribution Constructed a brain-wide wiring diagram at single-cell resolution
Impact http://fruitflybrain.org/
Start Year 2016
 
Description Bionet 
Organisation Columbia University
Department Electrical Engineering
Country United States 
Sector Academic/University 
PI Contribution We developed, and refined models of the retina for simulation on the Neurokernel platform.
Collaborator Contribution Developed the initial open source simulation platform Neurokernel. Provided tools and resources for our group to use.
Impact Open Science Prize Application This is a multi-disciplinary collaboration to develop models and a software platform for simulating the entire brain of the fruit fly: biology, computer science, control and systems engineering
Start Year 2015
 
Description NVIDIA 
Organisation NVIDIA
Country Global 
Sector Private 
PI Contribution Evaluation of new GPU architectures
Collaborator Contribution Training and technical support. In-kind contribution of £30K
Impact None yet
Start Year 2015
 
Description Oxford 
Organisation University of Oxford
Department Department of Physiology, Anatomy and Genetics
Country United Kingdom 
Sector Academic/University 
PI Contribution We are developing modelling and simulation tools to make experimentally plausible and accurate models of the fruit fly brain.
Collaborator Contribution They support the drosophila brain project with data, evaluation & testing of the software, and to help validate models.
Impact No outputs yet. Multi-disciplinary: biology, systems engineering, control engineering
Start Year 2015
 
Description Stem Cell Australia 
Organisation University of Melbourne
Department Centre for Neuroscience Research
Country Australia 
Sector Academic/University 
PI Contribution Modelling and analysis of experimental data
Collaborator Contribution Single cell QPCR data Microdiscetion and subfractionation experiments
Impact Multidisciplinary collaboration funded by a HFSP program grant started in June 2012 We presented new research results at the 2013 Awardees Meeting in Strasbourg
Start Year 2012