Electrical and Electronic Engineering - Theses
Permanent URI for this collection
Search Results
1 - 4 of 4
-
ItemModelling of C. elegans & cover time of T cells in the liver( 2017)Mathematical modelling is an essential approach for biologists, enabling them to investigate complex biological systems. Today, computational models are widely used to simulate biological processes, suggest hypothesis, and set up new experiments. This work is an attempt to discover the value of mathematical tools for understanding complex biological systems, such as electrical activity in neurons, and how they enable behaviours like locomotion, and such as the function of the immune system, and how immune cells patrol throughout the body. This thesis consists of two main parts. The overall objective of the first part is to use mathematical descriptions and computer simulations for understanding how biological neural networks work to generate different complex behaviours. The first part deals with different aspects of modelling of C. elegans. After a quick introduction to the biology and especially the neuronal network of the worm, several computational models of the nervous system and behaviours of C. elegans from the literature will be discussed. We will simulate the transmission of signals in the neural circuit of C. elegans from sensory neurons to motor neurons. Then, motivated by the hypothesis that the same neuronal structure can be used to generate different locomotive reactions to various sensory signals, we will show how a generic model of the neuronal network of the worm is able to produce two different behaviours based on two distinct sensory inputs from the environment. Finally, we will use Algorithmic Information Theory to analyse the complexity of the worms’ behaviours when they are subjected to changes in the temperature or food concentration in their environment. The broad objective of part two again is to apply mathematical tools to describe the function of complex biological systems like immune cells. The focus of the second part is on the movement of T cells. We will analyse data sets of recorded movement of T cells on the epidermis, dermis and lymph nodes to investigate their movement pattern. Then, in order to produce a coarse estimate of the cover time of T cells in the mouse liver, we will construct a discretised model of the mouse liver with nodes visited by random walkers representing T cells. Using this first-generation mathematical model, we will show how to compute the time needed to cover a given percentage of the liver in the given time by certain number of immune cells.
-
ItemSignalling and crosstalk in cytokine pathways: mathematical modelling and quantitative analysis( 2016)Cancer is a leading cause of death all over the world. Focusing at the intracellular level, there are several cytokine signalling networks involved in inflammation and tumorigenesis. These pathways regulate the cell biological responses to the environment, both directly and via crosstalk. Recently, there have been several reports which have confirmed the strong relationship between inflammatory diseases and tumor development. Two cytokines, in particular, have significant roles in wound healing, inflammation and cancer Transforming growth factor β (TGF-β) and Interleukin-6 (IL-6). Cytokine signalling pathways are an interconnected complex system of biochemical reactions which can be represented by kinetic equations. These signalling pathways can share proteins and genes which make the intracellular signalling networks extremely complex. Systems biology and mathematical modelling are new approaches for the study of complex systems such as intracellular signalling networks. The focus of this research is to model mathematically new descriptions of the TGF-β and IL-6 pathways based on new logistics and then integrate them into a single, robust, self-regulated model which can be used to investigate tumor development in the stomach and colon. At each level, experimental data sets were used iteratively in order to both parameterize and examine the models (e.g. the model and experimental data of Zi et al. (2011) and further model simulations have been used in Chapter 4 to parameterize our TGF-β model.). Different approaches in Systems biology and their applications in cell signalling research are studied. TGF-β and IL-6 signalling pathways and their components are reviewed next. The previous mathematical models of TGF-β and IL-6 signalling are briefly discussed. Additionally, the role of individual signalling in cancer progression and inflammation is studied. We developed a mathematical model which captures the details of TGF-β signalling. The detailed model consists of over 40 differential equations and highlights the necessity for the reduction and simplification methods. The TGF-β signalling model is simplified and reduced via analytical reduction methods to 6 differential equations and is further validated with experiment. For the first time an explicit negative feedback loop has been included in the model. Another contribution of the TGF-β model is that the inherent time-delays in signalling networks are incorporated in detail. In the final chapter different input patterns are studied for TGF-β signalling. Our model of TGF-β signalling indicates that the positive feedback loop is one mechanism by which stability could be achieved. The thesis reports for the first time the coupling of the positive and negative feedback loops for TGF-β signal transduction. Furthermore, our TGF-β signalling model proposes predictions for the responses of cancer cells to TGF-β stimulation, which suggest new experimental protocols for future work. We also developed a mathematical model that describes the IL-6 signalling system thoroughly. The large number of equations involved in this model highlights the need for simplification. Similar to TGF-β signalling model, IL-6 signalling model is simplified and reduced using mathematical methods. In order to develop a realistic model specific kinetics are used for the different reactions. Time-delays are incorporated in the IL-6 transduction mathematical model for the first time. After being validated with different experimental data, the reduced IL-6 signalling model predicts the behaviour of cancer cells in response to IL-6 stimulation. Different pulsatile ligand inputs are studied using IL-6 model and new hypotheses for the TGF-β and IL-6 signalling crosstalk are raised. Our initial hypothesis that IL-6 signalling regulates TGF-β signalling via SMAD7, is examined using our integrated IL-6:TGF-β model. The simulations produced by the integrated model confirm the importance of the negative feedback loop of TGF-β signalling (SMAD7) via IL-6 downstream signalling, previously suggested by Jenkins et al. (2005). Various kinetic models are examined for the link between the two signalling pathways and several predictions are proposed for the pulsatile inputs and different stimulation patterns. The results of the integrated model are compared with the individual TGF-β and IL-6 models. IL-6-induced activation of SMAD7 leads to suppression of TGF-β signalling and causes double peak responses of PSMAD2 in the short-term, however, the long-term responses of the cells to TGF-β stimulation remain unchanged by IL-6 signalling. The integrated model is also validated experimentally. Conclusively, we found that the regulation of TGF-β signal transduction by IL-6 signalling occurs within the first 300 minutes after stimulation, i.e. within the transient phase of the response. This thesis includes both theoretical and experimental work, performed by the applicant. Theoretical part of the thesis consists of designing and developing models, analytical analysis of the models and conducting numerical simulations with the numerical simulation of the models. In the experimental part, various experimental protocols were developed and examined in order to parameterize, test and validate the proposed models.
-
ItemManaged DC power reticulation systems( 1999-11)Electric power engineering, as it applies to low-voltage power reticulation in buildings and industrial sites, is ripe for a ‘paradigm shift’ to bring it properly into the Electronic Age. The conventional alternating-current approach, now over a hundred years old, is increasingly unsatisfactory from the point of view of plant and appliance requirements. Alternative approaches can deliver substantial cost savings, higher efficiencies, power quality improvements, and greater safety. Power reticulation systems in the future can be expected to differ from present systems in two key respects. The first is a greatly increased role for direct current; the second is the augmentation of the power system with a wide range of ‘management’ technologies. Combining these two trends, which can already be observed today, leads to consideration of ‘managed DC’ power reticulation systems, operating from AC bulk supply mains via AC-DC converters.
-
ItemTopics in robust video analytics( 2011)Robustness is one of the main challenges in building good algorithms for video analytics. The ability to maintain accuracy and precision of the algorithm to a wide range of environments is very beneficial. Robust algorithms should be able to perform properly with limited assumptions either indoors or outdoors. In this thesis, we improve algorithm robustness in three main areas: foreground segmentation, observation detection for single object tracking, and a hierarchical multiple hypothesis tracker for multiple-object tracking. We enhance foreground detection by using fusion methods of masked grey world, probabilistic gradient information and extended conditional random field. The aim is to make the algorithm robust to 1) gradual and sudden illumination changes, 2) colour similarity between foreground object and background, 3) shadow and afterimage noise and 4) moving background object. A 2-stage mask grey world is employed to overcome sudden and gradual illumination changes while edge information is used to solve the colour similarity issue. An extended conditional random field is applied to remove shadow and afterimage components. A patch-based approach is proposed to enhance observation detection for single object tracking. The algorithm is a fusion between a feature-based descriptor and template approach for detecting the same object in consecutive frames. It is built to be robust to illumination change, blur, moderate size change, non-rigid object, low ambient illumination and homogenous textured object. We develop two methods of patch-based observation detection, one deterministic and the other probabilistic. The algorithm works by detecting points of interest before patches are built at matched points. Position and size smoothing are performed so that the detected observation encapsulates the tracked object completely and continuously over the track interval. A third component of the thesis deals with multiple-object tracking by implementing a hierarchical multiple hypothesis tracker. The main novelties of our approach are anchor-based track initialization, prediction assistance for unconfirmed track and two virtual measurements for confirmed track. The system is built mainly to deal with the problems of merge, split, fragments and occlusion. Histogram intersection testing is performed to limit tracked bounding box expansion. Simulation results show that all our algorithms perform well in the surroundings mentioned above. The main weakness of our algorithms is the heavy processing requirement. The algorithm is suitable for incorporation in higher level applications such as traffic monitoring, people counting, robotic vision and many more.