Please read the FAQ and then contact me if you are interested in getting involved in any of these projects.
Natural Interactive Walking
The Natural Interactive Walking project aims to enable the experience of walking on natural ground surfaces of different types (e.g., on snow, sand, and through water) using techniques from virtual reality, with special emphasis on sound and haptic (touch) aspects. A number of research opportunities are available for highly motivated students, primarily related to mechanical device issues, multimodal interaction techniques, and HCI experiments on the above.
Specific examples include haptic device engineering: optimizing the mechanical and electronic configuration of the actuated floor component, with the aim of overcoming structural and actuator limitations on the fidelity of the display device; haptic rendering: development of haptic synthesis algorithms tailored to the interactive rendering of interaction types (e.g., compression, sliding friction) and materials (e.g., sand, pebbles) via a vibrotactile display device; and multimodal interaction: development of applications integrating immersive video projection, spatialized auditory display, and vibrotactile rendering using the existing actuated floor.
Three possible specific projects are described in further detail below.
First, a solution for a closed loop control should be studied for the vibration output previously designed. This strategy aims to improve the fidelity of the device (the proof of stability is not requested). The results obtained should be analyzed and compared with the current open loop strategy.
Second, some solutions for one degree of freedom (vertical) low-frequency force feedback (0 - 30 Hz) rendering should be investigated. These solutions include the investigation of the hardware (e.g. actuator type, the amplifier needed, etc.) and the software (adding acquisition and control of the actuator in the current controller).
Finally, a study of a smaller, more efficient voice coil actuator with different plate materials could be investigated for optimizing the device transparency (accurately reproducing a vibrotactile signal within the tactile frequency band addressed, from about 10 Hz to 1 kHz). An analysis of the transparency and the performance of the new device should be presented with some security aspects for the user.
Deliverables
Analyses of closed loop control strategies, force-feedback rendering, and rendering transparency across the 10 Hz to 1 kHz frequency range.
First, a literature review of virtual ground surface material rendering would be conducted. This review should include mathematical aspects of the interaction between the device model and the material model to render. A concise description of the experiments and the expected results should be presented based on hypotheses. Multiple rendering algorithms will be implemented, tested and compared.
Second, a complete immitance model (two-port network) should be developed for representing the final device, the user and the physical properties of the virtual object to render. From this model and the literature review, the synthesis algorithm could be designed and implemented inside the device controller in open-loop strategy.
Finally, an analysis of the influence of the device compliance on the realism of the rendering must be conducted. An evaluation of the different algorithms will be conducted in collaboration with the third project to confirm the hypotheses presented in the first part of the project.
Deliverables
Implementation and comparative evaluation of rendering algorithms.
First, some strategies should be conducted for the implementation of visual and audio effects for the water and the sand (or pebbles). The interactive scenario consists of a perceptually consistent display of the environmental and material properties of a virtual environment.
Second, a solution for the integration of at least two material models (e.g., sand and water) should be developed. One goal could be the development of a beach in a tropical environment.
Finally, in collaboration with the first and the second project, a solution for the integration of the haptic controller and the haptic synthesis algorithms with the virtual environment could be presented. This includes the study of the best software architecture for the haptic rendering hardware.
Deliverables
Simulated multimodal environment providing the experience of two or more material models.
Emotional Babelfish
As part of a Network Centres of Excellence dealing with Graphics, Animation, and New Media, our group is exploring the use of affect in improving computer-mediated human communication. One such possibility is the real-time manipulation of video content in a videoconference session based on automatic recognition of the participants' emotions. The intent is to use this ability as a universal "emotional translator", in particular, to assist groups from different cultures to recognize the mood of the other conversants. This project will build on our existing facial expression and biosignals-based emotion recognition technology, and experiment with various manipulations of the video stream to achieve the desired communicative effects, for example, through the use of colour (turning the face red when someone is angry) and halo effects (as if "radiating" happiness).
Enhanced Virtual Presence and Performance
As part of a Network Centres of Excellence dealing with Graphics, Animation, and New Media, our group is prototyping the next generation of telepresence technology. This is motivated by observation of the limitations of current videoconferencing systems with respect to gaze awareness, latency, and mobility of participants. The problems are not solely ones of bandwidth or video quality, but equally of the display architecture necessary to support complex, mobile, highly interactive group activity.
Several enabling technologies will be investigated in this project: video architectures, multimodal synthesis, support for haptics, and latency-reduction techniques. Beyond the prototyping of the architecture itself, a vital component of this project is evaluation of the fidelity and engagement resulting from the use of computer-generated representations of people and objects. In particular, this project will consider situations where participants are mobile and visibility is required simultaneously from multiple viewpoints, e.g., in an operatic performance setting in which a (remote) actor must be seen both by another actor and by the audience.
Last update: 8 January 2010