No description available.

Aviation security personnel encounter problems when interpreting x-ray images of hand luggage. This research seeks to determine whether the performance of the human operator can be improved, in terms of both reliability and accuracy, through the employment of a novel multiple-view x-ray imaging technique. Thus, a series of experiments were undertaken with the aim of providing evidence for the feasibility of using KDEX displays to aid in the recognition of threatening objects in airport carry-on luggage; and furthermore demonstrate the real-world value of this technique. This thesis describes experiments comparing how introducing depth information affects the performance of aviation security personnel attempting to detect various weapons in x-ray images of hand luggage. Specifically, multiple 2-dimensional (2D) x-ray luggage scans were acquired and processed to create the perception of 3-dimensionality (3D) in kinetic displays. These results were compared with weapon detection in standard static 2D scans of the same luggage. Threatening objects hidden in this luggage were more readily detected in kinetic 3-dimensional images than in the standard images. Initial results were obtained using greyscale images and limited to various types of knives. Subsequent experiments evolved to primarily use pseudo-colour x-ray images. These were encoded using the industry standard colours; orange for organic materials, green for mixture, and blue for metals. Additionally, as well as knives, the threat categories were expanded to include handguns, scissors and grenades. In view of the remarkable improvement in performance afforded by the kinetic images future studies are warranted. For example, various features of the displays (e.g. differences among hues and contrasts) should be manipulated as parameters to determine if further improvements may be achieved. Additionally, different pseudo-colour schemes could be experimented with in order to potentially maximize the gains afforded by the kinetic display. Overall, when deployed iin the field this technique may well revolutionize x-ray visual security screening.

No description available.

Recent developments in technology have meant that operators of complex systems, such as those found in the modern aircraft cockpit, now have access to an unprecedented volume of information. Significant attention within the Human Factors and Ergonomics community has therefore been focussed upon developing methods by which externally represented information can be made as accessible as possible within the interface. However, commensurate attention has not been paid to recent experimental work demonstrating that even millisecond changes to the accessibility of information provided within the interface can have surprisingly large consequences for the deployment of human memory during low-level routine interactive behaviour (Gray, Simms, Fu, & Schoelles, 2006). Therefore, a series of nine studies explored the impact of information accessibility upon more complex human behaviour, with particular emphasis placed upon learning, memory and planning. Three experiments contained within Chapter 2 point to caveats associated with the use of information fusion as a means of increasing the availability of onscreen information during a series of simulated flight navigation missions. Paradoxically, increasing information availability was found to lead to greater problems for retention of visual-spatial information. The cost of accessing problem solving information was manipulated across three further experiments reported in Chapter 3. Least-effort tradeoffs concerning the use of memory, previously observed during routine copying (Gray et al., 2006), were extended to problem solving and were found to also influence planning behaviour, as reflected by eye-tracker and verbal protocol data. The final three experiments constituting Chapter 4 demonstrated expedited learning during repeated problem solving when task-relevant information was harder to access. Performance deficits were observed when interactive behaviour was characterised by excessive reliance upon highly accessible externally represented information. The results are contrasted to similar work conducted in the training literature, and future directions for exploiting information access costs to facilitate learning, memory and planning are discussed.

Distant talking scenarios, such as hands-free calling or teleconference meetings, are essential for natural and comfortable human-machine interaction and they are being increasingly used in multiple contexts. The acquired speech signal in such scenarios is reverberant and affected by additive noise. This signal distortion degrades the performance of speech recognition and diarization systems creating troublesome human-machine interactions. This thesis proposes a method to non-intrusively estimate room acoustic parameters, paying special attention to a room acoustic parameter highly correlated with speech recognition degradation: clarity index. In addition, a method to provide information regarding the estimation accuracy is proposed. An analysis of the phoneme recognition performance for multiple reverberant environments is presented, from which a confusability metric for each phoneme is derived. This confusability metric is then employed to improve reverberant speech recognition performance. Additionally, room acoustic parameters can as well be used in speech recognition to provide robustness against reverberation. A method to exploit clarity index estimates in order to perform reverberant speech recognition is introduced. Finally, room acoustic parameters can also be used to diarize reverberant speech. A room acoustic parameter is proposed to be used as an additional source of information for single-channel diarization purposes in reverberant environments. In multi-channel environments, the time delay of arrival is a feature commonly used to diarize the input speech, however the computation of this feature is affected by reverberation. A method is presented to model the time delay of arrival in a robust manner so that speaker diarization is more accurately performed.

The main objective of this research was to advance understanding of factors influencing the transmission of vibration through glove materials to the hand and to the fingers. The transmissibility of a glove to the hand depends on two primary factors: the biodynamic response of the hand and the dynamic characteristics of the glove material. Although some of the factors affecting the biodynamic response of the hand and the dynamic characteristics of glove materials have been investigated previously, there is insufficient understanding to be able to predict glove transmissibilities. The apparent mass of the hand, the dynamic stiffness of three glove materials (Foam A, Foam B, and Gel A), and their transmissibilities to the hand were measured with five variables: (a) material thickness (6.4, 12.8, and 19.2 mm), (b) push force (10, 15, and 20 N), (c) contact area (with diameters 12.5. 25.0, and 37.5 mm), (d) vibration magnitude (1, 2, and 4 m/s2 r.m.s.), (e) with and without arm support, and (f) different frequency ranges (10 to 300 Hz and 2 to 50 Hz). With the hand pushing down on a flat surface and vertical vibration, measurements were obtained at the palm of the hand, the thenar eminence, and the index finger. It is concluded that the apparent mass of the hand and the dynamic stiffnesses of the glove materials at high frequencies are predominantly affected by contact area and contact force. A change in contact area or contact force can therefore increase or decrease glove transmissibility. At frequencies greater than 20 Hz, the apparent mass at the palm and the transmissibilities of the glove materials to the palm were similar with and without arm support. At frequencies between 20 and 100 Hz, as the dynamic stiffness of the material decreased, the transmissibility of the material to the palm decreased whereas the transmissibility to the index finger increased. Changes in the vibration magnitude will not have a large effect on glove transmissibility. Using the measured apparent mass of the hand and the measured dynamic stiffnesses of the glove materials, the transmissibility of the materials to the hand were predicted using an impedance model. The predicted transmissibilities were similar to, and showed similar trends to, the measured transmissibilities. The predicted transmissibilities seemed to reflect individual changes in the dynamic response of the hand, suggesting a method of predicting inter-subject variability in glove transmissibility. Simple lumped parameter models of the hand were developed (with two and four degrees-of-freedom) and combined with the dynamic characteristics of glove material (represented by a standard linear viscoelastic model) to predict glove transmissibility to the palm of the hand. The two degree-of-freedom model (representing the material and the hand) consistently provided a prediction of glove transmissibility to the palm that was inferior to the predictions of the four degree-of-freedom model (representing the material, the hand, and the arm), suggesting the exclusion of the dynamic response of the arm affects the ability of a model to predict glove transmissibility to the palm. The research shows that the transmission of vibration through a glove to the palm of the hand or to the finger can be predicted from the dynamic stiffness of the glove material and the apparent mass of the hand. Such predictions will assist the optimisation of glove dynamics, reduce the time to assess the performance of a glove, and reduce the need for testing with human subjects.

Prolonged exposure of the hand to tool-induced vibrations is associated with the occurrence of debilitating conditions such as vibration white finger. The primary aim of this work is to gain a better understanding of the effects of different aspects of exposure to finger transmitted vibration (FTV) related to operators using hand-held vibrating tools. To achieve this, firstly, a new method for measuring finger transmitted vibration was developed and assessed, including a tool vibration test rig and measurement protocol. The effect on FTV measurement of using a small accelerometer attached to the back of the finger was investigated using 2D finite element modelling. Comparisons were also made using a laser vibrometer. Analysis showed that the new test rig is capable of measuring FTV at frequencies ranging from 10 to 400 Hz, under different grip force levels, and that adding a small accelerometer mass (0.3 grams) did not significantly affect measurements. A human participant study then carried out using the new rig. Various characteristic measurements were collected in tandem, including anthropometry, skin characterisation and behaviour under loading to investigate the effect of different factors on FTV. The results showed that FTV varied among individuals and the key finding was that exposure to vibration has a significant effect on finger temperature even for a short period of testing. Anti-vibration (AV) glove materials were investigated using dynamic mechanical analysis (DMA) and tested using human participants. The results showed that the mechanical properties of AV materials change under real world industrial conditions such as excitation frequencies and temperature. Finally, a new artificial test-bed was developed to replicate the transmitted vibration of the index finger. Studies were conducted on a range of 5 test-beds, to allow comparison with the human measurements, including indentation, vibration transmissibility and FE modelling. FE modelling showed that the distribution of dynamic strain was found to be highest in the vasculature region of the finger, indicating that this could be one of the contributing factors of VWF. One of the finger test-bed was selected as best replicating the mechanical properties of the real finger. The artificial test-bed provided better consistency than human participants, for testing parameters, such as grip force, and can be used in future for testing AV gloves with no need for human subjects.ii Further investigations are suggested to be made to enhance the limitations of this project, including material analysis, testing protocol and finite element modelling.

Increasing numbers of older people are living independently for longer. The ability to use domestic information-processing appliances, such as washing machines and microwave ovens, to carry out activities of daily living is an important aspect of independent living. The focus of this research was on the design of operational interfaces on domestic information-processing appliances for older adults. Inclusive and user centred design are used to create operational interfaces that address users’ ergonomic needs. An early study carried out as part of this research identified unclear relationships between operational interfaces and instructional materials, such as cooking instructions on food packaging or washing instructions on clothes, as a major concern. Two impediments to the flow of a task were identified: interaction breakdown (where the task is stalled) and focus shift (where a user is distracted from the task). Given the importance of these to the use and therefore design of the operational interface, a coding scheme was developed to enable systematic analysis of participants’ interactions with operational interfaces and associated instructional information. The coding scheme covers participants’ interactions with operational interfaces and the task being carried out. The research concluded that inability to use operational interfaces was because of limited connections between visual instructions on the interfaces and instructional materials. Also, habitual behaviours demonstrated in the routine activities could be used to design improved visual instructions and information in sequential series on operational interfaces.

This thesis examines the concept of safety in two groups of professionals working in the fields of aviation and medicine. These professional activities are uniquely structured but have been compared in the literature as having certain similar characteristics. However, recent attempts at strengthening patient safety by transferring methods from aviation have had only a small effect, with evidence suggesting that professional engagement has been poor. One possibility is that there are large disparities in conceptual understanding about safety across these groups. This was derived from socio-cultural theories of knowledge showing how collective experiences are formed into meaningful categories through situated learning and structured internally through semiotic mediation. A cross cultural comparison of the safety concept was carried out using linguistic data to capture the concept. Forty-one interviews were conducted with participants across two groups comprising senior airline pilots and hospital consultants. Grounded theory analysis was used to code and analyse the da-ta. Taxonomic structures of the two safety concepts, comprising their main semantic sub-categories, are presented, along with models explaining their internal relation-ships. In aviation, a core category of control was identified involving principles of stability, invariance, and causal attribution. This is dominated by the sub-categories of institutional control and personal autonomy, which are mediated by the availability of information and predictability of events. In the medical sample a core category of clinical success was revealed. This encompassed overall quality of outcomes based on Bayesian thinking about different risks along possible treatment paths. The thesis shows how conceptual knowledge is formed through mediational means within the context of specific purposive activities. A tentative theory of conventionalisation is proposed to explain why top-down interventions for transferring practices between cultures fail when differences between key concepts are large. Interventions, such as change laboratories to scaffold learning towards re-conceptualisation are recommended.

Video-based single human tracking has found wide application but multiple human tracking is more challenging and enhanced processing techniques are required to estimate the positions and number of targets in each frame. In this thesis, the particle probability hypothesis density (PHD) lter is therefore the focus due to its ability to estimate both localization and cardinality information related to multiple human targets. To improve the tracking performance of the particle PHD lter, a number of enhancements are proposed. The Student's-t distribution is employed within the state and measurement models of the PHD lter to replace the Gaussian distribution because of its heavier tails, and thereby better predict particles with larger amplitudes. Moreover, the variational Bayesian approach is utilized to estimate the relationship between the measurement noise covariance matrix and the state model, and a joint multi-dimensioned Student's-t distribution is exploited. In order to obtain more observable measurements, a backward retrodiction step is employed to increase the measurement set, building upon the concept of a smoothing algorithm. To make further improvement, an adaptive step is used to combine the forward ltering and backward retrodiction ltering operations through the similarities of measurements achieved over discrete time. As such, the errors in the delayed measurements generated by false alarms and environment noise are avoided. In the nal work, information describing human behaviour is employed iv Abstract v to aid particle sampling in the prediction step of the particle PHD lter, which is captured in a social force model. A novel social force model is proposed based on the exponential function. Furthermore, a Markov Chain Monte Carlo (MCMC) step is utilized to resample the predicted particles, and the acceptance ratio is calculated by the results from the social force model to achieve more robust prediction. Then, a one class support vector machine (OCSVM) is applied in the measurement model of the PHD lter, trained on human features, to mitigate noise from the environment and to achieve better tracking performance. The proposed improvements of the particle PHD lters are evaluated with benchmark datasets such as the CAVIAR, PETS2009 and TUD datasets and assessed with quantitative and global evaluation measures, and are compared with state-of-the-art techniques to con rm the improvement of multiple human tracking performance.