Search

Subtle changes in microbial populations that inhabit different areas of the human body - known as microbiomes or microbiota - can contribute to disease development, and restoring these imbalances may provide a cure. Localised and systemic diseases such as Inflammatory Bowel Disease (IBD) and depression have been linked with alterations to microbiota across the human body. Our understanding of how both diseases develop contains significant gaps, and the microbiome - described by some as our "second genome" - offers a compelling new area for knowledge discovery. This thesis aimed to advance the field of microbiome research and is an account of the work conducted whilst investigating the human gut and oral microbiome for links with IBD and depression. In this thesis, a hybrid model and aggregating ensemble feature selection (EFS) approach are applied to microbiome census data gathered from subjects with IBD. Microbial ecology techniques are applied to identify alterations to the oral microbiome in depressed subjects, and a multimodal Computational Intelligence (CI) classification paradigm known as a Super Self-Organising Map (sSOM) is applied to predict depression from a saliva sample. Finally, a rough set characterisation approach was developed and applied to gut and oral microbiome census data in depressed subjects to avoid destructive data normalisation and to enable knowledge discovery. The outcomes from the development of the hybrid model and aggregating EFS approach include the accurate non-invasive prediction of IBD, and the identification of novel and robust alterations to the gut microbiome in an adult cohort of IBD patients. The result provides a potential alternative to invasive colonoscopy, improve the time to diagnosis and treatment of IBD, and delivers new insights into the aetiology of IBD. The investigation of the oral microbiome identified novel alterations in depressed subjects for the first time. The changes to the structure and composition of the oral microbiome were significant enough to enable the accurate prediction of depression from a saliva sample. The results contribute to the microbiome-gutbrain axis theory by associating alterations to the oral microbiome with depression for the first time, and offer an alternative to subjective criteria for diagnosing depression, which currently relies on patient self-report and clinical judgement. The rough set microbiome characterisation approach replicated existing results and identified previously undescribed alterations to the gut microbiome in depressed subjects. The results provide an alternative approach to destructive normalisation techniques that are often applied to microbiome census data (identifying an optimal approach is an open research question), and contribute to our understanding of the microbiome-gut-brain axis, which could lead to psychobiotic treatments of depression in the future.

Humans perceive the world through information gathered by their five senses. Attempting to replicate some of these senses in intelligent systems has been a focus of research for many years. Due to high quality vision sensors such as cameras and laser scanners being readily available at a relatively low cost for some time now, vision sensing has been heavily researched for many decades enabling systems to distinguish a lot of information such as the size, shape and colour of objects or materials. However, there are attributes of objects, materials and the environment that cannot be determined by vision sensing alone such as compressibility, thermal properties or sub-surface vibration. This thesis presents methods which demonstrate that tactile sensing can be used to assess a human’s current state of health by measuring their vital signs using a biomimetic fingertip, namely BioTAC. It involves three main contributions. The first contribution is a method for classifying materials from tactile sensing alone. Using machine learning approaches, the high sensitivity of the BioTAC tactile sensors is demonstrated via the ability to classify different (and similar) material with high accuracy based on surface texture and thermal properties. The second contribution focuses on the use of the BioTAC fingertip to accurately measure the vital signs of a human by mimicking medical professionals. Algorithms have been developed and evaluated for determining a human’s Beats Per Minute (BPM), Pulse to Pulse Interval (PPI), Respiratory Rate (RR), Breath to Breath Interval (BBI) via tactile sensing. Furthermore, algorithms have been developed to measure Capillary Refill Time (CRT) by using a combination of tactile sensing for the control of a robot fingertip and vision sensing to analyse changes in the subjects skin colour. The final contribution is a fuzzy classification algorithm capable of classifying the human’s health status based on their BPM, RR and CRT. Significant contributions in the field of tactile sensing presented in this thesis demonstrate that a robotic system can determine a human’s current status of health. This could play a vital role in helping to rescue victims of a disaster or emergency by performing medical triage and determining an order of treatment. In turn, emergency personnel will be able to make a more informed decision on how they should allocate their valuable resources thus preventing unnecessary risk and reducing the further loss of human life.

Collaborative manipulation of objects is usually a trivial activity for humans but is still very challenging for robots. Such tasks involve many complex aspects, such as human and object safety, social and handling context, grasping stability, slip detection, and ergonomics. Although huge research efforts have been devoted over decades to endow robots with the skills required for grasping, manipulation, sharing of objects, and collaboration with humans, there is still a need for reliable systems capable of reacting to unexpected events. As for humans, the sense of touch is essential for robots to perform many tasks as it provides information that can not be obtained through contactless sensing modalities. Thus, recent trends in robotics research explore the use of tactile sensing in human-robot object manipulation. An important aspect that is often overlooked in the existing literature is that tactile sensing is inherently sequential and therefore should be approached as a continuous process. The aim of this thesis is to explore continuous tactile sensing to enhance robot collaboration capabilities for object manipulation. The contribution of this work is threefold: firstly, an innovative multimodal technique that identifies the surface materials of objects using continuous tactile sensing is developed. Secondly, continuous tactile sensing is used to provide contact information to a control system that grasps objects of unknown geometry. Finally, an approach to hand over objects between xiii a robot and a human, relying on continuous tactile sensing, is developed to ensure the safety of the robot and the object during the transfer. In this thesis, the proposed approaches are evaluated on real physical robotic platforms. A comparison with the state-of-the art techniques in material recognition shows that the proposed multimodal approach enhances identification speed and accuracy. The experimental results also show excellent performance of the proposed approach for grasping objects even when information about their geometry is not available. Finally, the proposed object handover algorithm is proven to adapt to unexpected force perturbations on the object and release it in a timely manner without dropping. This work entails significant progress towards the development of autonomous robots that collaborate with humans in everyday tasks.