No description available.

The acquisition of capable robot manipulation skills is a critical prerequisite for the widespread deployment of robots in real-world environments, from household tasks to industrial applications. However, current robot manipulation systems remain limited in their ability to handle the diversity of objects, materials, and manipulation actions required in the real world. Data-driven methods have shown impressive results toward generalizing across a variety of problems, but existing approaches often require costly data collection using real robot platforms, hindering the scalability of skill acquisition. In this dissertation, we aim to push the limits of robotic manipulation task diversity by providing mechanisms to acquire new skills in a scalable manner. Achieving the "right" data with large quantity and high quality is of vital importance. We approach this problem by leveraging physics simulators. Different from commonly used rigid body simulators, we have customized simulators to support deformable objects with diverse materials and dynamics. Aerodynamics and fracture effects are also included to enable a wider range of manipulation actions such as blowing and cutting. With sophisticated system design, including proper representation selection and customized hardware design, the policies trained in simulation can be seamlessly applied to real robots. More specifically, this dissertation presents a series of works to address the challenges of diversity and scalability in robot manipulation skill acquisition. First, we introduce UMPNet, a universal policy network that can infer closed-loop action sequences for manipulating a wide range of articulated objects using only visual input. Second, we present DextAIRity, a system that leverages active airflow to enable safe and effective deformable object manipulation, expanding the repertoire of skills beyond traditional contact-based methods. Third, we describe RoboNinja, a cutting system for multi-material objects. With an interactive state estimator and an adaptive cutting policy, RoboNinja successfully removes the soft part of an object while preserving the rigid core.

From a parent guiding their toddler when learning to brush their teeth to a physical therapist assisting a client with their range of motion, physically interacting with other people is ubiquitous in our daily life. While some researchers have shown that haptic human-human interaction benefits performance during training as well as later individual performance (Takagi et al. 2017), others have failed to replicate these benefits (Beckers et al. 2018). Participants in these interaction groups were not aware they were haptically linked to a partner and each participant had independent control over their own virtual cursor when tracking the target. Yet, we are typically aware when we are interacting with others and often do so with tasks where we have shared control over the same control point (e.g., a toothbrush). Here, we tested the effectiveness of training alone versus training with a virtual partner when individuals were made aware of their interaction in a redundant reaching task. Participants (N = 100) completed 50 baseline trials followed by 200 trials with a clockwise cursor rotation in one of four randomly assigned groups. Two of the groups performed the adaptation trials with a virtual partner that represented either the fast (Human + Fast Agent Group) or slow (Human + Slow Agent Group) state of the two-state model (Smith et al. 2006) with 30-deg rotation. The two remaining groups performed the task alone with either the 30-deg rotation (Solo full rotation) or a 15-deg rotation (Solo half rotation). Results showed that participants in the fast agent group contributed less to correcting the rotational error early in the adaptation block, but were responsible for most of the correction later in this block, with performance most similar to the solo full rotation group. Conversely, participants in the slow agent group corrected for a greater proportion of the initial errors, but their contribution began to drift during adaptation, with performance resembling that of the solo half rotation group. This pattern of results were consistent with our theory-driven simulations. / Thesis / Master of Science (MSc) / Working with a physiotherapist is the gold-standard in rehabilitating many injuries, but this can be very time consuming and repetitive in nature. This makes it worthwhile to explore other alternatives to supplement standard rehabilitation, such as working with a virtual partner. In our experiment, we tested two partners based on human models. Participants were paired with one of the virtual partners and had to reach a target using a handle, adjusting their reach to a rotation. The partners differ in how fast they help the participant adjust for the rotation. It was found that those who completed the task with a fast-learning partner corrected less error initially and more later on, while those with the slow-learning partner corrected more error initially and less later on. These results suggest that we can influence participant behaviour with different virtual partners.

Thesis (MScSportSc)--University of Stellenbosch, 2004. / ENGLISH ABSTRACT: The purpose of this investigation was to determine whether a sensory motor development programme could have an effect on some of the underlying physical and perceptual abilities that support school readiness. The control group consisted of 23 children and the two intervention groups of 79 children in total. All of the children were enrolled in a pre-primary school programme in a local community. They were all six years old by the end of the intervention. The Movement Assessment Battery for Children (Henderson & Sugden, 1992) was used to assess the children on their manual dexterity, eye-hand coordination, static balance and dynamic balance. The intervention programme consisted of two phases of 10 weeks each, during which the sensory motor development activities were presented to the two intervention groups. Results of the investigation revealed there were significant improvements for some of the children on selected variables that underlie school readiness. It can be concluded that participation in a sensory motor development programme can make a significant contribution to school readiness for many children. / AFRIKAANSE OPSOMMING: Die doel van hierdie ondersoek was om te bepaal of 'n sensories-motoriese ontwikkelingsprogram enige uitwerking op bepaalde onderliggende fisiese en perseptuele vermoëns wat tot skoolgereedheid by jong kinders bydra, sou kon hê. Die kontrolegroep het uit drie-en-twintig kinders bestaan. Die twee tussentredende groepe het altesaam uit nege-en-sewentig kinders bestaan. Die kinders was almal pre-primêre skoolprogramleerders vanuit 'n plaaslike gemeenskap. Teen die einde van die intrede het al die betrokke kinders sesjarige ouderdom bereik. Movement Assessment Battery for Children (Hendersen & Sugden, 1992) is as riglyn gebruik om die kinders se handvaardigheid, oog-hand-koërdinasie, statiese en dinamiese balans te evalueer. Die tussentredeprogram het bestaan uit twee fases van tien weke elk. Die sensoriese-motoriese ontwikkelingsaktiwiteite is terselfdertyd vir die twee tussentredegroepe aangebied. Die resultate van die ondersoek het beduidende vordering in sommige kinders getoon, spesifiek ten opsigte van bepaalde veranderlikes onderliggend aan skoolgereedheid. Hierdie bevinding dui daarop dat sodanige deelname in 'n sensories-motoriese ontwikkelingsprogram wel 'n betekenisvolle bydrae tot die ontwikkeling van skoolgereedheid in baie kinders kan maak.

The purpose of the evaluation was to test effectiveness of a visual, visual-motor, and auditory perceptual skills training program devised by Dr. Jerome Rosner, and to confirm or deny application of this training to improved classroom skills. Subjects were 38 kindergarten children, 20 in the Contrast Group, and 18 in the Experimental Group. Both groups received the same training in the basic curriculum of readiness skills. The Experimental Group also received training in the Visual Analysis and Auditory Analysis Skills programs. Pretests and posttests were administered, tabulated, and analyzed. Differences in raw score means were sufficient to indicate more than a chance factor and all tests demonstrated a plus factor for children in the Experimental Group.

Thesis (Ph.D.)--McMaster University, 2005. / Supervisor: Timothy D. Lee. Includes bibliographical references. Also available online.

The use of mental practice to facilitate the acquisition of a motor skill was investigated. A transfer design was used to determine if the facilitation of mirror tracing performance could be attributed to learning. Following two trials on the mirror trace subjects performed either mental practice, a reading task, an attentional focus task, or massed practice. Subjects then performed trial 3. Then 52 subjects transferred to the nonpreferred hand for trials 4 and 5. The remaining subjects continued to trace with the preferred hand for trials 4 and 5. Subjects in the mental practice group traced faster than subjects in the reading task and massed practice groups, although they did not trace significantly faster than the attentional focus group. Mental practice subjects did not make significantly fewer errors than subjects in the other groups. An analysis of the transfer task indicated that the faster tracing by the mental practice group might not have been the result of learning. Females tended to trace faster and make fewer errors than males. / M.S.

The topic of this work is motor imagery. Through 7 different studies, the process and proprieties of motor imagery and its impact on motor learning were assessed. It was found that this cognitive activity is closely linked to the motor system. This characteristic underlies the positive effects of motor imagery on motor learning. This can have many practical applications in the domain of sport and motor rehabilitation.<p><p> / Doctorat en Sciences Psychologiques et de l'éducation / info:eu-repo/semantics/nonPublished

published_or_final_version / Sports Science / Master / Master of Science in Sports Science

published_or_final_version / Speech and Hearing Sciences / Master / Master of Philosophy