This thesis addresses the problem of tracking rigid objects in video sequences. A novel approach to reducing the template size of shapes is presented. The reduced shape template can be used to enhance the performance of tracking, detection and recognition algorithms. The main idea consists of pre-calculating all possible positions and orientations that a shape can undergo for a given state space. From these states, it is possible to extract a set of points that uniquely and robustly characterises the shape for the considered state space. An algorithm, based on the Hough transform, has been developed to achieve this for discrete shapes, i.e. sets of points, projected in an image when the state space is bounded. An extended discussion on particle filters, that serves as an introduction to the topic, is presented, as well as some generic improvements. The introduction of these improvements allow the data to be better sampled by incorporating additional measurements and knowledge about the velocity of the tracked object. A partial re-initialisation scheme is also presented that enables faster recovery of the system when the object is temporarily occluded. A stencil estimator is introduced to identify the position of an object in an image. Some of its properties are discussed and demonstrated. The estimator can be efficiently evaluated using the bounded Hough transform algorithm. The performance of the stencilled Hough transform can be further enhanced with a methodology that decimates the stencils while maintaining the robustness of the tracker. Performance evaluations have demonstrated the relevance of the approach. Although the methods presented in this thesis could be adapted to full 3-D object motion, motions that maintain the same view of the object in front of a camera are more specifically studied.

Many researchers approach the problem of building autonomous systems by looking to biology for inspiration. This has given rise to a wide-range of artificial systems mimicking their biological counterparts—artificial neural networks, artificial endocrine systems, and artificial musculoskeletal systems are prime examples. While these systems are succinct and work well in isolation, they can become cumbersome and complicated when combined to perform more complex tasks. Autonomous behaviour is one such complex task. This thesis considers autonomy as the complex behaviour it is, and proposes a bottom-up approach to developing autonomous behaviour from cognition. This consists of investigating how cognition can provide new approaches to the current limitations of swarm systems, and using this as the basis for one type of autonomous behaviour: artificial homeostasis. Distributed cognition, a form of emergent cognition, is most often described in terms of the immune system and social insects. By taking inspiration from distributed cognition, this thesis details the development of novel algorithms for cognitive decision-making and emergent identity in leaderless, homogenous swarms. Artificial homeostasis is provided to a robot through an architecture that combines the cognitive decision-making algorithm with a simple associative memory. This architecture is used to demonstrate how a simple architecture can endow a robot with the capacity to adapt to an unseen environment, and use that information to proactively seek out what it needs from the environment in order to maintain its internal state.

The focus of the research is to investigate and develop enhanced version of swarm intelligence firefly algorithm and ecology-based invasive weed algorithm to solve global optimisation problems and apply to practical engineering problems. The work presents two adaptive variants of firefly algorithm by introducing spread factor mechanism that exploits the fitness intensity during the search process. The spread factor mechanism is proposed to enhance the adaptive parameter terms of the firefly algorithm. The adaptive algorithms are formulated to avoid premature convergence and better optimum solution value. Two new adaptive variants of invasive weed algorithm are also developed seed spread factor mechanism introduced in the dispersal process of the algorithm. The working principles and structure of the adaptive firefly and invasive weed algorithms are described and discussed. Hybrid invasive weed-firefly algorithm and hybrid invasive weed-firefly algorithm with spread factor mechanism are also proposed. The new hybridization algorithms are developed by retaining their individual advantages to help overcome the shortcomings of the original algorithms. The performances of the proposed algorithms are investigated and assessed in single-objective, constrained and multi-objective optimisation problems. Well known benchmark functions as well as current CEC 2006 and CEC 2014 test functions are used in this research. A selection of performance measurement tools is also used to evaluate performances of the algorithms. The algorithms are further tested with practical engineering design problems and in modelling and control of dynamic systems. The systems considered comprise a twin rotor system, a single-link flexible manipulator system and assistive exoskeletons for upper and lower extremities. The performance results are evaluated in comparison to the original firefly and invasive weed algorithms. It is demonstrated that the proposed approaches are superior over the individual algorithms in terms of efficiency, convergence speed and quality of the optimal solution achieved.

Two-dimensional colloidal crystals, as photonic band gap materials, have a wide variety of interesting and valuable industrial applications as photonic materials as well as simple models to study the basic processes of the atomic model such as phase transition, stability, crystallisation, ordering, and nucleation and growth. Recently, colloidal photonic crystals have got a great interest as templates for the fabrication of two-dimensional (2D) arrays for lithography applications. The dependence of these applications efficiency upon the quality of colloidal ordering during self-assembly process was the motivation for many researchers to perform several investigations in this field. These efforts have been oriented to develop a detailed explanation for mechanisms that took place during the complex self-assembly process of colloids, which may lead to developing a more controllable method to fabricate these structures with better quality. However, we are not yet able to fully understand what exactly happens during the colloidal self-assembly process. Hence, we are still unable to optimise processing conditions and so exploiting the fascinating colloidal optical characteristics is still limited till now. Several techniques have been used to fabricate highly ordered two-dimensional monolayer photonic crystals such as dip coating, electrophoretic deposition, selfassembly at the gas/liquid interface and electric-field induced assembly. However, these techniques have many drawbacks such as the incompatibility to scale-up from laboratory-scale tests to industrial scale mass fabrication. Also, inability to control the thickness of the final film limits the use of these techniques as fabrication methods for uniform colloidal crystals. On the other hand, spin coating was found to be more feasible due to its advantages over other techniques. Spin coating offers a cheap, simple and straightforward technique for the fabrication of two and three-dimensional colloidal crystals. Spin coating provides easy control of the uniformity, domain size and thickness of the fabricated thin films through tuning the operating parameters such as spinning speed, acceleration rate, solids content and solvent volatility. However, the short duration of the process (5-30 s) and rapidly rotating sample (1000-10000 rpm) makes in situ studies challenging and as such we do not yet fully understand colloidal self-assembly, so are unable to optimise processing conditions effectively. I have developed a laser scattering setup, which facilitates collecting laser scattering patterns diffracted by silica colloids in real time during the spin coating process. Tracking the development of these scattering patterns in real time may help to discover in details the stepwise evolution of the geometrical arrangements of monolayer colloidal crystals (MCCs) during the self-assembly process. Monitoring the colloidal self-assembly mechanisms may help to produce better quality colloidal crystals with a minimum defects density and also may help pave the way to fabricate complete three-dimensional photonic band gaps colloidal crystals with valuable photonic industrial applications. This work aims to study the critical factors affecting the degree of ordering of colloids as they self-assemble through the development of the in situ laser scattering experimental techniques. In addition, samples are investigated with scanning electron microscopy (SEM) to characterise the impact of each factor on the colloidal thin films morphology produced. Further understanding of colloidal self-assembly will allow processing conditions to be optimised so that highly uniform, long range and defectfree colloidal thin films may be easily fabricated.

This thesis presents investigations into the development of a versatile scansorial mobile robot and real-time realisation of a control system for different configurations of the robot namely climbing mode, walking mode and steering mode. The mobile robot comprises of a hybrid leg and wheel mechanism with innovative design that enables it to interchange its configuration to perform the specific tasks of pole climbing in climbing mode, walking and step climbing in walking mode, and skid steering and inclined slope climbing in steering mode. The motivation of this research is due to the surrounding environment which is not always structured for exploration or navigation missions, and thus poses significant difficulty for the robot to manoeuvre and accomplish the intended task. Hence, the development of versatile scansorial robot with a flexible and interchangeable configuration can provide a broad range of applications and locomotion system and to achieve the mission objective successfully. The robot design consists of four arms/legs with wheel attached at each end-effector and has two link manipulation capability. In climbing mode, the arms are configured as grippers to grip the pole and wheels accelerate to ascend or descend. The climbing angle is monitored to retain the level of the robot while climbing. However, in walking mode, the arms are configured as legs and the wheels are disabled. By implementing a periodic walking gait, the robot is capable of performing stable walking and step climbing. In steering mode, the arms are configured as suspension and the wheels are used for manoeuvring. In this mode, the skid steering system is used to enable the robot perform the turn. The versatile scansorial robot’s configurations and locomotion capabilities are assessed experimentally in real time implementation using the physical prototype. The experiments provided demonstrate the versatility of the robot and successfully fulfill the aims and objectives of the research.

Fuzzy-model-based (FMB) control scheme is an efficient approach to conduct stability analysis for nonlinear systems. Both Takagi-Sugeno (T-S) FMB and polynomial fuzzy-model-based (PFMB) control systems have been widely investigated. In this thesis, the stability analysis of FMB control systems is conducted via Lyapunov stability theory. The main contribution of the thesis is improving the applicability of T-S FMB and PFMB control strategies by relaxing stability conditions and designing fuzzy observer-controller, which is presented in the following three parts: 1) The stability conditions of FMB control systems are relaxed such that the FMB control strategy can be applied to a wider range of nonlinear systems. For T-S FMB control systems, higher order derivatives of Lyapunov function (HODLF) are employed, which generalizes the commonly used first order derivative. For PFMB control systems, Taylor series membership functions (TSMF) are brought into stability conditions such that the relation between membership grades and system states is expressed. 2) Two types of T-S fuzzy observer-controller are designed such that the T-S FMB control strategy can be applied to systems with unmeasurable states. For the first type, the T-S fuzzy observer with unmeasurable premise variables is designed to estimate the system states and then the estimated states are employed for state-feedback control of nonlinear systems. Convex stability conditions are obtained through matrix decoupling technique. For the second type, the T-S fuzzy functional observer is designed to directly estimate the control input instead of the system states, which can reduce the order of the observer. A new form of fuzzy functional observer is proposed to facilitate the stability analysis such that the observer gains can be numerically obtained and the stability can be guaranteed simultaneously. 3) The polynomial fuzzy observer-controller with unmeasurable premise variables is designed for systems with unmeasurable states. Although the consideration of the polynomial fuzzy model and unmeasurable premise variables enhances the applicability of the FMB control strategy, it leads to non-convex stability conditions. Therefore, two methods are applied to derive convex stability conditions: refined completing square approach and matrix decoupling technique. Additionally, the designed polynomial fuzzy observer-controller is extended for systems where only sampled-output measurements are available. Furthermore, the membership functions of the designed polynomial observer-controller are optimized by the improved gradient descent method. Simulation examples are provided to demonstrate and verify the theoretical analysis.

No description available.

The original contribution of this thesis is to provide insight into research of non-traditional control techniques for automotive power train applications, culminating in experimental evidence of much improved performance and reduced commissioning costs. This includes much work on the technique of Observer Based Robust Control (OBRC) which, before the research documented in this thesis commenced, was only in its infancy with some promise being shown through the simulation of electric drive applications (Dodds, 2007). The thesis, therefore, contributes to the process of bringing this new control technique nearer maturity. OBRC is based on an observer designed to provide information enabling effective control of an automotive power train application and its performance assessment. Comparison with traditional and other robust control techniques is included. The observer in OBRC is is designed to estimate the equivalent disturbance input, referred to the control input to a plant. This represents plant modelling errors as well as external disturbances. The equivalent disturbance estimate is applied to the real plant input to cancel its effect, thereby reducing the control problem to that of controlling the known real-time model of the plant employed in the observer. One of the disadvantages of conventional robust control methods, such as those based on sliding mode control, is that relatively high gain control loops are closed around the uncertain plant. This increases the risk of instability due to the dynamic elements, such as sensor lags, that are not included in the assumed plant model. The initial reason for investigating OBRC is that the high gain loops are applied to the known plant model in the observer and that the stability of these loops, taken in isolation, can therefore be guaranteed. It was found that the observer gains are limited only by the finite sampling frequency of the digital processor. In theory, infinite observer gains would yield ideal robustness. However, in practice only finite gains are possible. The aforementioned application of the equivalent disturbance estimate to the real plant input effectively transfers the high gain loops from the plant model in the observer to the real plant. This meant that closed loop stability could not be guaranteed under all circumstances. In view of this, it was decided that sliding mode control should not be excluded from the set of controllers for comparison. Since the control chatter associated with basic sliding mode control has to be eliminated for the vehicle application, polynomial control (a continuous version of the discrete RST controller) with robust pole assignment is included. This polynomial control can be regarded as equivalent to the sliding mode control with a boundary layer, but without the uncertainty associated with the choice of the boundary layer width. Two more controllers, based on the Internal Model Control (IMC) and H-infinity, are included for comparison on the basis that their design methodologies do not demand high gains. The various control techniques are demonstrated and compared via their application to Diesel Drivelines for commercial road vehicles. One of the operational problems with conventional PI engine speed controllers is the need for time consuming initial controller tuning. This requires different sets of gains for each gear selection, including idle (i.e., neutral) and later retuning to compensate for changes in the driveline characteristics with component aging. A major advantage of OBRC in this application is the elimination of the tuning procedure. Of particular interest is the fact that the order of the system is increased by two when a gear is engaged due to a vibration mode created by the finite torsional compliance of the propeller shaft and other driveline components. Since this driven mechanical load can be represented by its inverse dynamic model in a feedback path whose output acts at the same point as the control variable, the OBRC compensates for this automatically without the need for any parametric changes. The simulations and experimental work were carried out on the DAF 12 litre diesel engine. The comparative study was carried out, not only with respect to the main application of Diesel Drivelines, but also using academic examples that are even more demanding of the controllers’ capabilities. A key parameter in an engine configuration is the saturation limit on the injected fuel rate, which is highly dependent on the engine capacity from 10 litres to 16 litres. One example was introduced with a saturation block and the abilities of the various controllers under this constraint were assessed. The H-infinity controller could not handle such a saturation constraint, which is common practice in automotive applications and therefore this had to deemed unsuitable. The remaining controllers were able to operate with fuel rate saturation. The overall conclusion is that the controllers based on OBRC, polynomial control and IMC are capable of a similar performance with appropriate controller parameter settings. However all are subject to the trade-off between the conflicting requirements of short response times and robustness.

This thesis presents research into the design of brushless permanent magnet (BPM) interior rotor synchronous servomotor for both performance and manufacturability. The investigation has been supported by experimental evidence gathered with the aid of two prototype servomotors, having the same frame sizes and stator stack lengths, designed by the author and manufactured by the sponsoring company, Control Techniques Dynamics (CTD) Ltd. One servomotor, upon which the research is focussed, has the relatively new segmented stator structure containing concentrated windings. The other servomotor has a conventional structure with a solid stator containing distributed windings and this has provided the means of assessing the degree of improvement attainable with the new structure. It has been established that the new structure enables more cost-effective manufacture than attainable with the conventional structure. A significant contribution of the research programme is the special notch for retention of the surface mounted permanent magnets in the rotor. This innovation enables a more uniform and smaller air gap, which greatly improves the dynamic and thermal performance for a given frame size and stator stack length, thereby advancing the state of the art. This, together with the greatly reduced stator winding overhang in the new structure, enables a physically smaller motor for a given application, thereby reducing active material usage. Also the design methodology has focused on reducing energy and eliminating waste in the manufacturing process. Regarding the thermal aspects, only natural cooling has been considered. Conduction, convection and radiation heat transfer in the servomotor has been investigated theoretically, by simulation and by experiments to identify where design improvements can be made. The most critical area identified is the paper wall insulation and comprehensive experiments have been carried out to identify the commercially available material with the highest thermal conductivity to maximise the removal of heat from the stator windings. An array of strategically located thermocouples was used to obtain temperature distributions. The performance indicators for comparison of the new and conventional servomotors are cogging torque, iron loss and dynamic torque. The cogging torque proved to iii increase with the permanent magnet retaining feature for the new servomotor and becomes unpredictable but this disadvantage was offset by material saving and its maximum torque being approximately 40% greater than that of the conventional servomotor. The iron losses have been measured at speeds ranging from 500 r/m to 6000 r/m and at surface housing reference temperatures approximately 810C. At maximum reference winding temperature ranging from 1000C and 1250C, the dynamic torque rating performances of the new and conventional servomotor prototypes have been compared, the torque/speed characteristics being generated from a near stall speed of 50 r/m to the rated 3000 r/m. The contributions of this research programme may be summarised as follows: 1. Above all this thesis provides valuable design steps for permanent magnet synchronous servomotors using three world leading software packages (OPERA_2D, SPEED and Motor-CAD) for machine design and will therefore be a useful reference for electrical machine designers, especially beginners. 2. Material saving and simplified BPM synchronous servomotor manufacturing process for segmented stator, concentrated winding configuration. 3. Novel permanent magnet retaining method permitting more uniform and smaller air gap giving performance enhancement. 4. Systematic finite element analysis (FEA) based optimisation is applied to implement a new design principle for BPM synchronous servomotors regarding the stator parameters. 5. A simplified frequency dependent equation for open-circuit iron loss calculation. 6. Experimental and theoretical investigation of thermal impact of stator wall paper insulation entailing comparison of several different materials. 7. Direct benefit to industry through adoption of the new servomotor design by the company who supported the research.

A Human machine interface (HMI) acts like a bridge between motor function and the brain. Bypassing these natural pathways allows disabled individuals to perform actions that might otherwise be too difficult or impossible. The loss or impairment of voluntary muscle control can have a detrimental effect on an individual’s quality of life. By utilising natural physiological motion with sensor fusion techniques, grasp intention can be predicted, leading to an assistive HMI concept that can reduce the cognitive load that traditional HMIs impose on the user. This thesis investigates a novel design concept for a heterogeneous sensor suite, by fusing mechanomyogram (MMG) sensors for muscle activation, computer vision for object recognition, and inertial measurement sensors for predicting grasp intention. The developed architecture focuses on the prediction of intentional grasp activity of 1 amputee and 10 healthy subjects, using the natural physiological motion of the arm when reaching to grasp 3 objects with up to 3 different grasp patterns. 84 motion features are extracted and used as a classification tool for predicting intention, yielding an average grasp classification accuracy of 100%, 82.5% and 88.9% for bottle, lid and box objects across all subjects. The novel heterogeneous sensor suite is applied to automate the grasp control of a myoelectric hand prosthesis. Real-time task-based experiments evaluated the performance of the proposed system, comparing it against conventional control using MMG sensors, yielding an 8.5% average faster completion time, as well as a reduction in overall cognitive and physical burden. The results of this research provide excellent potential for the use of natural motion to replace discrete muscle input as a selection and intention prediction tool, where the emphasis is on reducing the cognitive load imposed by assistive technologies.