The impact of computer use is evident in every day life (Harrington, Carter, Birrell and Gompertz, 2000:264). Lind (2002:18) explains that global trends continue to show that the most severe work-related health problems that exist amongst computer users are musculoskeletal disorders. As technology has lead to increases in automation, so it has lead to increases in work-related illnesses. Although studies have explored the effects of ergonomics (Thibodeau, 1995:322) in static working positions there has been little evidence supporting a solution in overcoming poor occupational postures. Many health practitioners argue that occupationally caused, or aggravated, musculoskeletal disorders are steadily increasing. Thus, even though computers have improved productivity and made work easier for the population in general, they have adverse effects as well. Designing the proper tools or a setup of the work place is of prime importance for the elimination of chronic diseases attributed to sedentary lifestyles. Regular physical activity had long been regarded as an important component of a healthy lifestyle. This notion has recently been reinforced by scientific evidence linking regular physical activity with a wide array of physical and mental health benefits, synonymous with an improvement in wellness (Pratt, Macera, and Wang, 2000:63). According to Pratt et al. (2000:63) higher direct medical costs associates with physical inactivity. Further cross-sectional epidemiologic studies and controlled experimental investigations conducted by Okura, Nakata and Tanaka (2003:1131) had demonstrated that physically active adults, in contrast to their sedentary counterparts, tend to develop and maintain higher levels of physical fitness. These studies had not only demonstrated the positive results of physical activity, such as an improvement in blood lipid profile, body composition, glucose tolerance and insulin sensitivity, but had also shown that participation in such activity decreased the risk of developing several chronic hypokinetic diseases, including coronary heart disease (CHD), hypertension, non-insulin dependant diabetes mellitus (type II), osteoporosis, colon cancer, anxiety and depression. In addition, low levels of habitual physical activity and the subsequent low levels of physical fitness were associated with a marked increase in all-cause mortality rates. Okura et al. (2003:1131) confirm that effects of exercise intensity on physical fitness and risk factors for coronary Herat disease.

Graduation date: 1998

Contextual influence on movement was examined for a selection of everyday activities. Non-medicated and medicated Parkinson's disease (PD) patients and control subjects reached for a drinking glass target from both seated and standing postures, and stepped over a surface-level obstacle while walking on a constrained path. Contextual challenge was increased in the seated reach by filling the glass with water, in the standing reach by increasing the depth of the gap between the target and stationary foot position, and in the obstacle negotiation trials by raising the gait path surface above the floor level. In all cases, behaviour among PD patients was uniquely disrupted by contextual challenge. In addition, benefits of conventional medication therapy for PD patients were limited in challenging contexts. The results suggest an adapted movement control mechanism at work in PD patients, with the neural resources used in this adapted response prone for interference during contextual challenges. / xiv, 186 leaves ; 29 cm.

The purpose of this study was to determine the extent to which a novel assistive device improved two dimensional base of support (BOS) as compared to unassisted walking. Telescoping walking poles (T poles) are designed to allow a gait pattern more closely resembling unassisted gait and provide enhanced stability. Nine healthy male and female subjects who had no recent lower extremity pathology or surgery and were between the ages of 59 and 78 were recruited from the community to participate. Subjects were instructed to use the T poles in a consistent manner and were monitored through weekly supervised training sessions. At the conclusion of the study, motion analysis was performed using video collected from six cameras. Subjects were adorned with reflective markers on the toe and heel of their shoes as well as the tips of the T poles. These markers were used to define the BOS. Two dimensional analyses revealed a significant increase in base of support, an average of 316% over the entire gait cycle as well as 531% at heel strike and 126 % at toe off with the T poles. These results support the T poles as an effective intervention in enhancing stability in an elderly population of recreational walkers. / School of Physical Education

Thesis (Ph.D.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, 1980. / MICROFICHE COPY AVAILABLE IN ARCHIVES AND ENGINEERING. / Bibliography: leaves 263-276. / by Roger William Wicke. / Ph.D.

Predicting and understanding the dynamic of human motion has many applications such as motion synthesis, augmented reality, security, education, reinforcement learning, autonomous vehicles, and many others. In this thesis, we create a novel end-to-end pipeline that can predict multiple future poses from the same input, and, in addition, can classify the entire sequence. Our focus is on the following two aspects of human motion understanding: Probabilistic human action prediction: Given a sequence of human poses as input, we sample multiple possible future poses from the same input sequence using a new GAN-based network. Human motion understanding: Given a sequence of human poses as input, we classify the actual action performed in the sequence and improve the classification performance using the presentation learned from the prediction network. We also demonstrate how to improve model training from noisy labels, using facial expression recognition as an example. More specifically, we have 10 taggers to label each input image, and compare four different approaches: majority voting, multi-label learning, probabilistic label drawing, and cross-entropy loss. We show that the traditional majority voting scheme does not perform as well as the last two approaches that fully leverage the label distribution. We shared the enhanced FER+ data set with multiple labels for each face image with the research community (https://github.com/Microsoft/FERPlus). For predicting and understanding of human motion, we propose a novel sequence-to-sequence model trained with an improved version of generative adversarial networks (GAN). Our model, which we call HP-GAN2, learns a probability density function of future human poses conditioned on previous poses. It predicts multiple sequences of possible future human poses, each from the same input sequence but seeded with a different vector z drawn from a random distribution. Moreover, to quantify the quality of the non-deterministic predictions, we simultaneously train a motion-quality-assessment model that learns the probability that a given skeleton pose sequence is a real or fake human motion. In order to classify the action performed in a video clip, we took two approaches. In the first approach, we train on a sequence of skeleton poses from scratch using random parameters initialization with the same network architecture used in the discriminator of the HP-GAN2 model. For the second approach, we use the discriminator of the HP-GAN2 network, extend it with an action classification branch, and fine tune the end-to-end model on the classification tasks, since the discriminator in HP-GAN2 learned to differentiate between fake and real human motion. So, our hypothesis is that if the discriminator network can differentiate between synthetic and real skeleton poses, then it also has learned some of the dynamics of a real human motion, and that those dynamics are useful in classification as well. We will show through multiple experiments that that is indeed the case. Therefore, our model learns to predict multiple future sequences of human poses from the same input sequence. We also show that the discriminator learns a general representation of human motion by using the learned features in an action recognition task. And we train a motion-quality-assessment network that measure the probability of a given sequence of poses are valid human poses or not. We test our model on two of the largest human pose datasets: NTURGB-D, and Human3.6M. We train on both single and multiple action types. The predictive power of our model for motion estimation is demonstrated by generating multiple plausible futures from the same input and showing the effect of each of the several loss functions in the ablation study. We also show the advantage of switching to GAN from WGAN-GP, which we used in our previous work. Furthermore, we show that it takes less than half the number of epochs to train an activity recognition network by using the features learned from the discriminator.

Changes in postural stability following sensory manipulation were investigated among Parkinson's disease patients and healthy older adults. Sixteen Parkinson's disease patients (PD; mean age 68.2 + 2.7 years) and sixteen older adults (control; mean age 67.6 + 2.6 years) performed quiet standing trials that progressed through baseline, sensory manipulation, and reintegration. Postural control following visual deprivation was assessed following alternate removal and reinsertion of visual information. Postural recovery following sensory incongruence was assessed following the termination of visual, somatosensory, and visuosomatosensory incongruence. PD patients' balance was disrupted following visual deprivation, and was initially disrupted when visual information was returned. PD patients' pstural recovery was comparable to control subjects when sensory incongruence ended. These findings indicate that situations of visual deprivation in particular are initially disruptive for PD patients, and imply initial difficulty for sensory reorganization in these patients. Our results provide insight into environmental situations imposing greater fall risk among the parkinsonian population. / xii, 81 leaves ; 29 cm.

The effects of postural threat on gait initiation and steady state gait among Parkinson’s disease (PD) patients and age-matched adults were examined. Ten healthy adults (CTRL; mean age= 68.8 ± 8.4, range 56-80 years) and ten PD patients (PDOFF / ON ; mean age= 69.7 ± 10.3, range 54-81 years) initiated gait and continued with steady state walking along a walkway of two different height conditions. PD patients were first tested in a non-medicated state followed by testing in a medicated state. The results showed that gait initiation and steady state gait deficits inherent to PD are exacerbated in a postural threatening environment. As well, medication efficacy for overcoming parkinsonian deficits may be context dependent. These findings confirm the dynamic nature of movement deficits characteristic of parkinsonian patients and provide empirical evidence for specific environments that can create movement difficulties for people with PD. / x, 59 leaves : ill. ; 29 cm. --

By understanding and mimicking characteristics of postural control used by animals, scientist and engineers may develop standing autonomous robots that work safely within home environments, and treatment strategies that help people overcome postural impairments. To increase our understanding of postural control we developed physical and computational models of standing posture to explain the interrelation of stance width and feedback gain in controlling the stability and dynamics of the postural response. These models facilitated precise analysis of mechanical dynamics and their effects on compliant feedback control, and provided a physical implementation to verify predictions developed from simulation. We show that a scaling of active feedback gain is required to maintain postural stability. These results are consistent with previous studies that have shown that a correlation exists between increased stance width and decreased postural responses. However, these studies have not quantified the relation between stance and the active control of standing posture. This scaling of gains that we show is dependent on the changing kinematic relations of the mechanical structure as it undergoes stance width adjustments. Specifically, we show that increasing stance width increases the leverage of the mechanical system. Feedback gains must be reduced by the reciprocal of the increase in mechanical leverage in order to maintain a consistent postural response; otherwise, the system may become unstable with increasing oscillations. We also showed that increasing magnitudes of intrinsic stiffness increases postural stability by facilitating stable responses over larger ranges of active feedback gain and increasing the stability of responses by decreasing settling time, oscillations, and displacement magnitude. The conclusions of this study were that the variation of mechanical leverage is responsible for changing the dynamics of the response during stance width variation, and that scaling of feedback gains with the changing mechanical leverage of stance width variations is required to maintain consistent response dynamics across stance widths.

Thèse doctorat : Génie Mécanique. Biomécanique : Villeurbanne, INSA : 2004. / Thèse rédigée en anglais sauf le chapitre VIII. Titre provenant de l'écran-titre. Bibliogr. p. 247-254.