Since the early pioneering work of Rosenfeld and Sklansky, machine vision has been used successfully in a number of practical applications. In this work the areas of application have been reviewed to determine the extent to which computer vision has been utilised in industry. Its application so far has been restricted principally to the mass production field of car manufacture, in the areas of body panel inspection and to a lesser degree to unit assembly. The developnent of robots for picking unsorted components fron containers and for inspection depends upon the reliability and the speed of the vision control system employed. This project investigates a preliminary approach for fast shape recognition of randomly positioned and oriented components. The results obtained indicate that this goal can be achieved by the use of the techniques of contour tracing, Fourier and surface texture analysis and shape fron shading methods.

No description available.

No description available.

No description available.

The two most advanced applications of contactless magnetic levitation are high-speed magnetic bearings and magnetically levitated vehicles (Maglev) for ground transportation using superconducting magnets and controlled d.c. electromagnets. The repulsion force from superconducting magnets provide stable levitation with low damping, while the suspension force generated by electromagnets is inherently unstable. This instability, due to the in verse force-distance relationship, requires the addition of feedback controllers to sustain stable suspension. The problem of controlling magnetically levitated systems using d.c. electromagnets under different operating conditions has been studied in this thesis with a design process primarily driven by experimental results from a representative single-magnet test rig and a multi-magnet vehicle. The controller-design stages are presented in detail and close relationships have been constructed between selection of performance criteria for the derivation process and desired suspension characteristics. Both linear and nonlinear stabilising compensators have been developed. Simulation and experimental results have been studied in parallel to assess operational stability and the main emphasis has been given to assessing performance under different operational conditions. For the experimental work, a new digital signal processor-based hardware platform has been designed, built with interface to Matlab/Simulink. The controller design methods and algorithmic work presented in this thesis can be divided into: non-adaptive, adaptive, optimal linear and nonlinear. Adaptive algorithms based on model reference control have been developed to improve the performance of the suspension system in the presence of considerable variations in external payload and force disturbances. New design methods for Maglev suspension have been developed using robust control theory (%oo and fi synthesis). Single- and multi-magnet control problems have been treated using the same framework. A solution to the Hoo controller-optimisation problem has been derived and applied to Maglev control. The sensitivity to robustness has been discussed and tools for assessing the robustness of the closed-loop system in terms of sustaining stability and performance in the presence of uncertainties in the suspension model have been presented. Multivariable controllers based on %00 and /i synthesis have been developed for a laboratory scale experimental vehicle weighing 88 kg with four suspension magnets, and experimental results have been derived to show superiority of the proposed design methods in terms of ability to deal with external disturbances. The concept of Hoo control has been extended to the nonlinear setting using the concepts of energy and dissipativity, and nonlinear state-feedback and out put-feed back controllers for Maglev have been developed and reported. Simulation and experimental results have been presented to show the improved performance of these controllers to attenuate guideway-induced disturbances while maintaining acceptable suspension qualities and larger operational bandwidth.

The Bees Algorithm is a new population-based optimisation procedure which employs a combination of global exploratory and local exploitatory search. This thesis introduces an enhanced version of the Bees Algorithm which implements a fuzzy logic system for greedy selection of local search sites. The proposed fuzzy greedy selection system reduces the number of parameters needed to run the Bees Algorithm. The proposed algorithm has been applied to a number of benchmark function optimisation problems to demonstrate its robustness and self-organising ability. The Bees Algorithm in both its basic and enhanced forms has been used to optimise the parameters of a fuzzy logic controller. The purpose of the controller is to stabilise and balance an under-actuated two-link acrobatic robot (ACROBOT) in the upright position. Kalman filtering, as a fast convergence gradient-based optimisation method, is introduced as an alternative to random neighbourhood search to guide worker bees speedily towards the optima of local search sites. The proposed method has been used to tune membership functions for a fuzzy logic system. Finally, the fuzzy greedy selection system is enhanced by using multiple independent criteria to select local search sites. The enhanced fuzzy selection system has again been used with Kalman filtering to speed up the Bees Algorithm. The resulting algorithm has been applied to train a Radial Basis Function (RBF) neural network for wood defect identification. The results obtained show that the changes made to the Bees Algorithm in this research have significantly improved its performance. This is because these enhancements maintain the robust global search attribute of the Bees Algorithm and improve its local search procedure.

The research presented in this thesis considers the feasibility of utilising dsPICs (digital signal controllers) in the development of effective monitoring systems which have the capability to adapt to changes in operating conditions and can be quickly calibrated to suit a range of applications, thus helping to reduce the development time constraint. The capability of these monitoring solutions to detect and isolate faults occurring in pneumatic processes is investigated and their effectiveness verified. Three applications are considered gas pipe leakage, linear actuator operations and gripper action. In each case, solutions are developed based upon the dsPIC. The solutions utilise the analysis of pressure transients to overcome the limitation in the dsPIC memory. The deployment of minimal sensors and electronics was essential to optimise the cost of the system. Leak detection techniques are developed with application to gas fitting pipes. The speed at which correct decisions are determined was the essence of this work. The solutions are tested, compared and their capability validated using pipes which had been rejected according to industrial standards. In this application a dsPIC digital signal controller and a pressure sensor were deployed, thus ensuring a low cost monitoring solution. Linear actuator "end of stroke" monitoring has, previously, largely been possible using limit switches. A more challenging method based upon the deployment of a pressure sensor is outlined. Monitoring model surfaces were obtained and their capability to determine the health of the process was proved, at various supply pressures. With regard to the gripper monitoring, a performance surface by which the gripper action can be monitored is generated and embedded within the dsPIC. Various faults are simulated and their effect on the gripper performance investigated. Leakage and blockage are also investigated at various places in the pneumatic circuit to allow for an algorithm to be devised. Faults may be detected and isolated, and their locations identified to allow for timely recovery treatment, thus supporting an enhanced process monitoring strategy.

This thesis documents a study on the sources of the differences found between results from machine and walking testing of shock absorbing systems. A complex programme of experiments was conducted at the Institute of Biomechanics of Valencia to explore the four most outstanding statements proposed with this respect: 1. - No accurate simulation of impacts by machine test. This was investigated by comparing results from testing materials simulating impact forces with results from walking tests. 2. - In use materials degrade and their properties change and existing machine testing methods could not replicate material properties during walking. A new testing method was developed to measure the recovery ability of materials by simulating plantar pressures and results compared with walking tests. 3. - Shoe effect on walking kinematics and heel pad confinement has greater influence on shock absorption than material properties. An instrumented pendulum was developed to study the heel pad. Insole materials were evaluated in walking tests, in pendulum tests and in different machine testing including the new method developed simulating plantar pressures and the results compared. 4. - Accommodation to impact conditions occurs according to a controlled proprioceptive feedback model. Accommodation, impact perception, comfort, walking and passive biomechanical variables and material properties were studied in relation to system's input, output and goal. Accurate simulation of impacts improved the ability of machine test to predict the walking performance of materials, but not upper body shock transmission. Properties of materials such as recovery ability, stiffness and hardness play an important role in concepts and passive interaction but mainly by influencing accommodation. Accommodation was identified as the source of differences of results between machine and walking tests of shock absorbing materials. The human body was described as comprising two independent mechanical systems: One system, governed by the elasticity and hardness of materials, it is defined by impact forces and accelerations that are inversely related to upper body transmission and control the perceived impact through foot position and knee bend. The other system is defined by heel pad stiffness, insole properties at initial loading and passive interaction that regulate upper body shock transmission by ankle inversion for comfort control. Passive interaction is defined in this thesis as the mechanical coupling between insole and heel pad that determines the properties of the system either through heel pad confinement or compression. Machine tests appear to predict results with respect to the first system but not the second, which required passive human testing. For insole use, high-energy absorption materials are preferred. These are capable of increasing elastic deformation to reduce impact forces and accelerations without increasing initial-maximal stiffness by passive interaction thus avoiding any increase of head transmission due to accommodation. Heel pad properties were described by three mechanical components accounting for 93.08% of total variance: These are an elastic component, a viscoelastic component and a component related to elastic deformation at low stiffness. Differences were found between shod and barefoot test results. With barefoot there was an initial low stiffness (18-50 kNm*1) response that was not evident in the shod tests which showed elastic deformation related to final stiffness. With barefoot, the elastic component accounted for impact forces variance (> 70%) and initial deformation component for peak force time (> 60%), while shod impact forces were related mainly to the elastic deformation component (> 60%) whereas rate of loading and acceleration were related to the initial-maximal stiffness component (>20%). Differences in heel pad mechanics due to age, gender and obesity were observed. Although the heel pad properties degraded with age, losses appeared to be compensated by obesity.

No description available.

This thesis describes the development of comprehensive algorithms for the analysis and prediction of the transient behaviour of marine electromechanical systems with particular reference to dynamically positioned vessels. The type of oscillatory behaviour studied may be divided into two categories. The first is concerned with predicting response to normal operational conditions commensurate with maintaining on-station duties, involving various load profiles on both thrust and main propulsion systems. The second concerns the prediction of transient behaviour in response to abnormal operating conditions such as that brought about by various system faults. The method of solution is illustrated by reference to simulated results for a practical system for which a comprehensive data set was made available. Digital computer programs are derived for both linear and non-linear time-domain formulations which are designed to be suitable for the analysis of general marine power installations. These programs are based on the developed algorithms which incorporate a full representation of the component parts of the system including rotational and static electrical plant and all relevant drive mechanisms and system controlling functions.