CONTROL 2014: UKACC INTERNATIONAL CONFERENCE ON CONTROL (CONTROL 2014)

PROGRAM FOR THURSDAY, JULY 10TH

Days:
previous day
next day
all days
08:30-09:30 Session 5: Plenary - Richard Parry Jones
 08:30 Research Challenges for Control Systems Engineering in Road and RailSPEAKER: Richard Parry-Jones
10:00-12:00 Session 6A: IMechE Mini Symposium - Unmanned Aircraft Systems
 10:00 Advanced UAS technologies for aerial manipulation and landing on mobile platformsSPEAKER: Antidio Viguria Jiménez 10:24 Bio-Inspired Flight Control Systems for UASSPEAKER: Ian Cowling 10:48 VTOL flying wing control systems developmentSPEAKER: Mark Shaw With Arthur Henry 11:12 Civil UAV applications Using Crowd-sourced Imagery AnalysisSPEAKER: Darren AnsellABSTRACT. This presentation describes the cross-disciplinary ‘Aerosee’ project, involving UCLan’s School of Computing, Engineering and Physical Sciences and Media Innovation Studio, along with commercial firm eMigs. Aerosee is unique in uniting Unmanned Aerial Vehicle technology with the ‘power of the crowd’. It capitalises on some of the key attributes of the burgeoning UAV technology field, and examine how the devices can be used for socially-motivated purposes, such as by the emergency services. The presentation will discuss the first trial which explored how the use of UAV (or ‘Drone’) technology, together with ‘crowd-sourced’ imagery analysis can reduce the time taken to locate and rescue a person in distress. 11:36 Research and Test Environment for Unmanned Aircraft SystemsSPEAKER: Matthew Coombes
10:00-12:00 Session 6B: IET Mini Symposium - Monitoring and control of poorly-defined and uncertain systems
 10:00 The smart grid – essential contribution to efficiency or Achilles heel of a future electricity system?SPEAKER: Roger KempABSTRACT. The principles behind Britain’s electricity system have been largely unchanged for a century. Steam turbines were demonstrated by Sir Charles Parsons in 1884 and, in Edwardian times, Charles Merz established an ac grid in Newcastle on Tyne distributing power at 5.5 kV, three-phase. The present grid, although hundreds of times larger, uses basically the same technology. The Climate Change Act 2008 and the Industrial Emissions Directive 2010/75/EU have triggered a paradigm shift in this system. So far, we have seen only small changes but plans for phasing-out coal-fired power stations and the widespread adoption of renewable energy will radically change the way electricity is generated and used. Instead of the flow of energy being from a few dozen large centralised power stations via high voltage grids to millions of consumers, we will have generation that includes thousands of wind turbines and millions of solar panels, feeding into the 11 kV and 400 V distribution networks. In addition, we can expect to see millions of electric vehicles, each having a high-capacity battery charger, and a wholesale switch from gas central heating to electrically-powered heat pumps which could double the peak load on the grid. Since the 1950s, electricity generation policy has been “predict and provide”. Consumers do not have to think about when they switch on the kettle; there will always be enough electricity. However to continue this policy when much of the generating capacity uses renewable energy, that is only available when the sun shines or the wind blows, and when loads such as electric vehicle chargers could equally well be supplied at 3 am, rather than when the driver happens to return from work, would be horrendously expensive. Luckily the smart grid can come to the rescue! No-one has really defined what a smart grid is – or to be more accurate, thousands of people have defined it to themselves and made the assumption that their understanding is the same as for everyone else. The common assumption is that it is (or will be) some form of intelligent network that recognises when electricity is in short supply and reduces the demand from loads that can readily be cut off for a period. Other visions of a smart grid include the capability to modulate the output of solar panels or trigger reverse power flows in EV battery chargers (vehicle-to-grid or V2G) to compensate for a sudden loss of supply. Companies are already marketing home energy management computers that allow consumers to control loads remotely via an App on their phones. However there are downsides: if the smart grid sends a message to a million 5 kW battery chargers, sold with a default setting of 10 p/kWh, that electricity prices have dropped from 11 to 9 p/kWh, the surge in demand could destabilise the system – any feedback system with unquantifiable delays risks instability. An erroneous message asking V2G battery chargers over a wide area to provide power could defeat protection systems. Attacks on a smart grid are also a possibility. With household energy bills more than £1000 p.a. there is an incentive to buy black market software that hacks into a smart meter – and less risky than a pair of crocodile clips in the meter terminal box. During the 1939-45 war, both sides sent waves of bombers to damage power stations. In a future conflict the task could be achieved more easily and surreptitiously by the equivalent of the Stutnix virus, used against Iranian centrifuges. Without smart grids we are unlikely to be able to meet our climate change emissions targets. With them we may be introducing instability or opportunities for malicious intervention that causes disruption to power supplies. How do we weigh the risks of a technology for which we do not even have a shared definition? 10:40 Challenges for monitoring and controlling biological systemsSPEAKER: Jean-Marie AertsABSTRACT. Thanks to the (r)evolution in sensors and computing power, the possibilities for measuring and processing biological signals in real-time are ever growing. As a consequence, it becomes more and more realistic to apply schemes of model-based monitoring and control to biological systems in practice. However, unlike many mechanical systems, biological systems are very complex, individually different, time-varying and dynamic, making the management of these systems often very challenging. A first challenge is to model these biological systems using compact, accurate and biologically meaningful models that can be applied in monitoring and controlling schemes. It is argued that dynamic data-based mechanistic modelling approaches are potential candidates for model-based management of biological systems. A second challenge is to extract model features that can be used as biomarkers for monitoring the status of individual biological systems. A third challenge is to define proper control targets that take into account the physiological/physical boundary conditions of the considered biological system. Examples of animal and human applications, ranging from cellular up to ecosystem level, will be discussed. 11:20 Poorly-Defined Environmental Systems: Data-Based Mechanistic Modelling, Forecasting and ControlSPEAKER: Peter YoungABSTRACT. Engineering is basically about the way humans apply their knowledge to design, build and maintain objects of various kinds. Although engineering research can take various forms, depending on what branch of engineering we are considering, it normally conforms to the scientific method', following what Popper (1959) has termed a ‘hypothetico-deductive’ procedure. There is, however, an important difference between such standard engineering research and research based on the application of engineering principles and methods to naturally occurring environmental systems (recently termed Geoegineering). Natural systems are not made by humans: they may be shaped and modified by us but we did not create them and so, in contrast to normal engineering constructs, their internal behavioural mechanisms are quite often poorly defined and understood. The climate scientist, for instance, can hypothesize about the causes and mechanisms of global warming' but it is clear from the controversy on this topic, that such hypotheses are not universally accepted. Moreover, while Popper stresses the importance of planned experimentation in hypothetico-deductive reasoning, it is rarely possible to conduct well planned experiments on natural systems in order to remove ambiguity and so clarify our understanding. An alternative, systems-based approach is to observe the normal operational behaviour and try to infer the most important underlying mechanisms from such observational data and our, hopefullly inspired, perceptions of what these dominant mechanisms may be. Such data-based analysis is the domain of inductive' or hypothetico-inductive' reasoning (see e.g. Young, 2013), in which the modelling is based on model structures that are initially identified from the naturally occurring data, thus avoiding undue reliance on prior hypotheses and ensuring that the resulting models are fully identifiable from the available data. This seminar will outline a particular hypothetico-inductive approach to the modelling, forecasting and control of environmental systems and show how this Data-Based Mechanistic (DBM) procedure (see e.g. Young 1998, 2011 and the prior references therein) has been applied in a study of globally averaged climate data. It will show how the behaviour of the large simulation models produced by climate scientists is much simpler than their size and complexity would suggest. In particular, the globally averaged response characteristics of even the largest climate models, the Global Circulation Models (GCMs), are defined almost completely by a relatively small number of dominant modes'; and it is these dominant modes of behaviour that are most important in forecasting and control system design.
10:00-12:00 Session 6C: Special Session - Automotive Control
10:00-12:00 Session 6D: Invited Session - Integrated Operation in Large Scale Industrial Sites, Results from the Energy-Smartops Consortium
Chairs:
12:45-13:45 Session 7: Plenary - Geoff Downton
 12:45 Drilling and ControlSPEAKER: Geoff DowntonABSTRACT. Within a tick of the imperceptibly slow moving geological clock, a borehole is plunged deep into the Earth. How we steer the ensuing transient response of the rocks and the fluids towards a safe and beneficial outcome is at the very heart of the drilling process. Wells are being drilled deeper and faster and in more difficult environments, and so the complexity of the interactions between different domains of physics is increasing. Fortunately, this is a challenge well suited to the interests and tools of the control engineering community. The “factory floor” of a well is clearly very long and very thin—effectively a line along which the mechanical drilling string and the hydraulic drilling fluids must transmit  power, information and control to propagate the borehole to its target resting place. The Earth needs to be persuaded to reveal its structural secrets, and so long arrays of sensors and actuators follow the drill bit into the ground to illuminate, monitor and control the path to be drilled. In comparison to the drilling industry, the downstream processing industries are highly automated. A chemical plant designer has great freedom to create the ideal production environment. Such is not the case for drilling. The impetus for control engineers to take an interest in drilling has been further strengthened by a drive to increase the level of automation for the standard reasons of efficiency, consistency, performance and management of complexity. This drive for controlled drilling presents both challenges and opportunities.
14:00-16:00 Session 8A: InstMC Mini Symposium - Actively Controlled Automotive Safety Systems
14:00-16:00 Session 8B: IChemE Mini Symposium - Industrial Process Control: Latest Challenges and Applications
 14:00 Innovation Still Required to Solve Upstream Oil & Gas Control IssuesSPEAKER: Paul Oram 14:24 Practical problems in implementing small scale MBPC in the chemical industrySPEAKER: Philip Masding 14:48 TBASPEAKER: Dave Lovett 15:12 Making predictive control easier and cheaper to implementSPEAKER: Anthony Rossiter
14:00-16:00 Session 8C: Invited Session - Sustainable Control of Off-shore Wind Turbines
Chairs:
14:00-16:00 Session 8D: Invited Session - Real-time Adaptive Networked Control of Rescue Robots 1
16:30-18:00 Session 9A: Control Education and Future Prospects
Chairs:
 16:30 Exploring future directions of ControlSPEAKER: Argyrios ZolotasABSTRACT. Control system research is entering a new era, which may be considered a milestone in the history of Systems and Control. The traditional framework of a control system structure – i.e. plant, sensor, actuator and a controller – presents a number of limitations to some current research, i.e. (a) System to be controlled comprise network of subsystems; (b) Control task is achieved by a network of distributed local controllers (agents); (c) New research topics on control of networked behaviour, including consensus, formation and synchronisation, are clearly beyond the scope of the traditional “one controller” framework. It is worth noting that with these new trends some new concepts are emerging, e.g. consensusability, formationability, computability. The intended contributions of this paper are: (1) A new framework of control research represented by a new block diagram and its mathematical treatment. It is pointed out that, many current different studies can be considered as some special cases under this general framework; (2) a brief overview of various research fitting the proposed framework and some open questions. This paper is largely motivated by past, current and future applications of power system control, and is based on a recent invited presentation, of the first author, at a special workshop on “Bridging the Gap – Control Theory and Control Engineering Practice”. 16:50 Fuzzy Logic-based Gain Scheduling of Exact FeedForward Linearization Controller for Magnetic Ball Levitation SystemSPEAKER: Abdullah ElgammalABSTRACT. This paper presents an Exact FeedForward Linearization controller combined with fuzzy-based gain scheduling for single DOF magnetic ball levitation system. The proposed method improves the previously reported techniques in terms of transient response behavior and tracking error. A Luenberger State estimator for the nonlinear/uncertain plant is implemented. Finally, the control law and the estimator are combined and applied experimentally to the magnetic ball levitation system. The results show the effectiveness and the robustness of this scheme regarding the increase of the dynamic range of the tracked signals and the presence of payload variation. 17:10 A Survey of Techniques and Opportunities in Power System Automatic Generation ControlSPEAKER: Eyefujirin EjegiABSTRACT. This paper gives a concise survey of the control methods which have been applied to automatic generation control of power systems. Different control approaches have been categorized and a summary has been given of the context of application and strengths and weaknesses. Such a summary is an effective starting point for determining where future control research is likely to bring the most benefit to the emerging large scale and interconnected structures in the power supply market. 17:30 Using clickers in lectures to help identify and teach the control topics students find difficultSPEAKER: Anthony RossiterABSTRACT. It is important that academic staff have a good awareness of the preparedness of their classes for new learning. This papers shows how clicker technology has been used both to help staff learn about student preparedness but also to encourage better engagement with learning of control related topics. Some of the data produced will be of generic interest to the community as it gives hard evidence of what is often suspected anecdotally.
16:30-18:00 Session 9B: Rail Applications of Control
Chairs:
 16:30 A Model of a Repoint Track Switch for ControlSPEAKER: Nick WrightABSTRACT. Track switching provides necessary flexibility to a rail network, allowing vehicles to change routes when necessary. Track switches, however, have historically been prone to failure. To increase asset reliability, a concept for a novel design of switch has been developed which allows multi-channel actuation through a novel actuation and locking mechanism, under a project titled ‘Repoint’. This paper describes a mathematical model of the operation a novel Repoint track switch. The model was derived from a first principles physical analysis of the Repoint concept design. The structure of the model mimics the physical structure of the design. Each physical component has an individual sub-model. The model has been used to estimate the actuator drive requirements for a case study mainline switch installation. It has been found that a Repoint track switch could be run from an existing UK signalling power supply. It is anticipated that this model will be used as the basis for a control system design activity for a technology demonstrator installation currently under construction. 16:50 Applications of the inerter in railway vehicle suspensionsSPEAKER: Roger M. GoodallABSTRACT. This paper is a review of currently identified applications of the inerter in railway vehicle suspensions. We survey the current state of research in the field, including main results, discussions and conclusions regarding overall applicability. 17:10 Detection and isolation of actuator failure for actively controlled railway wheelsetsSPEAKER: Mohammad MirzapourABSTRACT. This paper studies a model-based approach for the condition monitoring of an actively controlled railway system, with a focus on actuator failures to detect and isolate failure modes in such a system. It seeks to establish the necessary basis for fault detection to ensure system reliability in the event of an abnormal change in one of the two actuators. Computer simulation is used to demonstrate the effectiveness of the method. 17:30 Towards Self-powered Lateral Active Suspension for Railway VehiclesSPEAKER: Peng WangABSTRACT. This paper investigates the energy consumption of the lateral secondary active suspension for railway vehicles and studies the conditions for self-powered active suspension control. The influence of the performance index weighting factors to both the ride quality and the energy consumption is discussed based on the linear quadratic regulator (LQR) algorithm. Then the energy balance condition is analyzed. With low energy harvesting restriction, the self-powered condition can be achieved. At last, simulation results are provided to validate the feasibility of the self-powered suspension.
16:30-18:00 Session 9C: Medical Applications of Control
Chairs:
 16:30 Modelling the Self-Tolerance Mechanisms of T Cells: An Adaptive Sliding Mode Control ApproachSPEAKER: Anet J. N. AneloneABSTRACT. Failure of the mechanisms of tolerance to self antigen by the immune system causes an autoimmune response. The immune system changes its immunological steady-state according to a switching logic based on a function encompassing the interactions between antigen and T cells. The resultant immune response function is regarded as the controller for the immune system. Synergies with variable structure control are immediately apparent. A sliding mode is investigated on a suitable surface which is a function of the concentration of self antigen and T cells. Both immunological and control engineering constraints have been considered in the design of the controller. The magnitude of the discontinuity is tuned to reduce chattering as well as to represent immunological reality. Robust stability and robust performance is achieved. The healthy state of the immune system is maintained using adaptive sliding mode control to model the self-tolerance mechanisms. 16:50 Model of the Glucose-Insulin System of Type-1 Diabetics and Optimization-based Bolus CalculationSPEAKER: Tilman UtzABSTRACT. In this contribution, a mathematical model of the glucose-insulin system in the case of type-1 diabetes is presented. This model and is subsequently used to determine an optimal so-called bolus dosage adapted to the diabetic patient's future behaviour. This model on the one hand is chosen with a level of detail comprising the major part of daily situations. On the other hand, it is still relatively easy to parametrize in order to be usable within a so-called bolus calculator. The validity of the model is shown using real patient measurement data and the bolus optimization is carried out in a simulation environment. 17:10 Repetitive Control Based Tremor Suppression using Electrical StimulationSPEAKER: Engin Hasan CopurABSTRACT. Tremor is a common, debilitating movement disorder commonly occurring in neurological disorders, with invasive and pharmacological treatment methods often ineffective. Functional electrical stimulation (FES) holds potential for tremor suppression, but previous control methods have proven of limited success. This paper establishes the feasibility of using repetitive control, a framework which is able to eliminate the effect of periodic disturbances by including an internal model of the oscillation within the control structure. A model of the wrist is developed and two different repetitive control algorithms are applied to suppress tremor via stimulation of wrist flexors and extensors. Experimental results are compared with filter-based methods to establish efficacy of the proposed approach. 17:40 Phase-lead ILC based electrode array stimulation using reduced input subspaceSPEAKER: Christopher FreemanABSTRACT. An iterative learning control (ILC) approach is developed for electrode array based functional electrical stimulation. Building on the phase-lead ILC' form that has proved clinically effective with single electrodes, the first principled extension to multivariable systems is undertaken, resulting in a set of update laws which require minimal plant information that can be identified in a short initial steady-state test procedure. To further reduce the amount of model information required, a reduced input subspace is embedded in the ILC update. Experimental results using a 24 element electrode array confirm efficacy of the proposed approach for clinical practice.
16:30-18:00 Session 9D: Applied Control
Chair: Paul King
 16:30 Sliding-Mode Control of an Underactuated Oilwell Drill-String with Parametric UncertaintySPEAKER: Yang LiuABSTRACT. The stick-slip oscillation of a multibody drill-string is studied in this paper from the point of view of control of underactuated system using a lumped-parameter model. The model has one control input acting on the rotary table from the land surface and multi-degree-of-freedom downhole parts comprising a series of hollow drill pipes, a number of much stiffer drill collars, and a drill bit suffering highly nonlinear friction to be controlled. Two sliding-mode controllers are studied to suppress the stick-slip oscillation while track a desired rotary speed for the drill-string with estimated physical parameters. The stabilities of the proposed control methods are proved by using the Lyapunov direct method. Extensive simulation results are shown to demonstrate their validity and robustness to parameter variations, disturbances, and measurement noises. 16:50 Real-time measurement of ring-rolling geometry using low-cost hardwareSPEAKER: Stephen DuncanABSTRACT. Ring-rolling is an industrial forming process to produce high-strength metal rings up to 6m diameter. Thick-walled cylindrical workpieces of material, typically metallic alloys, are compressed between two or more internal and external rollers and rotated until a target geometry, often outer diameter, is achieved. The process is inherently unstable and the process is often constrained and/or controlled to improve its stability. This paper presents an image processing algorithm for the measurement of ring geometry through photogrammetry in real-time. These measurements will form part of the feedback control system for a ring-rolling process. An off-the-shelf USB webcam is used to capture images of the ring during forming and the scene is controlled to maximise contrast and minimise occlusions of the ring. The image processing tasks include object identification, edge detection, outlier rejection and distortion rectification. The process has been implemented on a desktop-scale forming machine and has been shown to work at rates suitable for control the ring-rolling process using feedback. 17:10 Exact Linearization by Feedback of State Dependent Parameter Models Applied to a Mechatronics DemonstratorSPEAKER: Essam M. ShabanABSTRACT. The paper develops an exact linearization by feedback approach for State Dependent Parameter (SDP), Proportional-Integral-Plus (PIP) control. The method is demonstrated using a simple automated belt driven by a DC motor equipped with a single board Reconfigurable Input-Output (sbRIO-9631) card, within a Field Programmable Gate Array (FPGA), and with a real time processor for control. The demonstrator is first modeled using a discrete-time SDP model structure, in which the parameters are functionally dependent on measured system states. An exact linearization step returns a linear model with unity coefficients, which is subsequently used to design a PIP control algorithm based on linear system design strategies, including pole assignment and optimal linear quadratic design. Preliminary experimental results demonstrate that the new approach yields an acceptable control performance for the nonlinear system. 17:30 Kinetic Modelling of an Industrial Semi-batch Alkoxylation Reactor for Control StudiesSPEAKER: Marcus NohrABSTRACT. Abstract—Semi-batch processes are gaining further importance in the chemical industry, due to the trend to produce more speciality products with higher margins. Besides a constantly high product quality, a maximum space-time-yield is a principal goal of process control for semi-batch processes. This requires to drive the process along its limitations, which often is impossible with standard PID control schemes. More sophisticated control methodologies potentially allow for a maximisation of the spacetime- yield while keeping the product quality high. However the scientific development and test of such methodologies is often conducted using very simplified reactor models, which may allow for a demonstration of the basic feasibility but neglect additional effects, which hamper the transfer to industrial processes. Therefore a more detailed model of a semi-batch process, suitable for the development and test of control schemes is presented in this contribution. Using this model, a simulation study elucidates the limitations of classical temperature control concepts when aiming for a maximisation of the space-time-yield.