Novita Research Report - 2004 to 2007 - Novita Children's Services

64Project titleContributorsSummaryMethodResultsImplicationsStatusDeveloping a ‘no-touch’ switch and interface for a computer and specialisedcommunication deviceCristina Larenas (Flinders University), Anne-Louise Smith (Flinders Medical Centre), SharmilRandhawa (Flinders University), David Hobbs, Annabel Gregory, Jodie Hobbs (FlindersMedical Centre)For many people with a physical limitation, accessing and interacting with the environment isachieved through the use of a specialised switch. Assistive technology (AT) switches comein all shapes, sizes and functionality, depending on the needs of the user and the device orpiece of equipment that is being controlled. For example, a ‘wand’ switch provides singlelevel switch access for clients who have gross hand or arm function, whereas a ‘sip n puff’switch offers two levels of switching capability and is controlled by the user’s breath.Most switches require a physical press to be made to activate the switch, while a few offer‘light touch’ or ‘touch but no pressure’ activation when a sustained physical press is difficultfor the client to achieve. The aim of this project was to develop an easy to use ‘no-touch’switch and appropriate interface to allow a person with a disability to control a computer anda specialised communication device. A no-touch switch would enable someone with a highlevelspinal cord injury, or a condition like ‘locked in syndrome’, who is unable to control anyof their limbs or head, to be able to interact with other specialised devices.The project was jointly initiated by Flinders Biomedical Engineering and NovitaTech followingan enquiry about assistance for someone who acquired a brain-stem stroke and was‘locked in’. The project was submitted to the Flinders University School of Informatics andEngineering as a Final Year Honours Project for consideration, and was divided into fourstages:1. A literature review of current switches and switch practices, including reviewing knownresources and existing ‘no-touch’ switch methodologies, determining what is being usedand identifying the advantages, disadvantages and shortfalls of current systems2. Investigate current and new methodologies that would enable ‘no touch’ activation andselect a suitable technology3. Design and construct a system using the chosen technology4. Interface the system to a computer and existing communication device.After the initial literature review, the focus of the project shifted slightly to the important issueof communication at night-time, when the user may need to communicate with their carersabout being turned in bed, getting a drink or passing on important information.Eye blinking was seen as one method of alerting the system to the fact that the user needsattention, and this activity was targeted using a webcam with infrared (IR) capability incombination with a computer program called MATLAB to detect both the eyes and theaction of deliberate blinking. The prototype system that was developed used the fact that eyepupils have a high reflectance under IR illumination to identify the eyes, and detected blinksby comparing consecutive frames. The system successfully detected the eyes when testedunder varying conditions, resolutions, illuminations and distances from the camera, but wasless successful when detecting blinking events.The prototype system completed at the end of the academic year demonstrated that acamera with IR capability, when set up appropriately, was reliable at detecting a person’seyes at night-time. The robustness of the overall system, including how the system acquiresand processes real-time information to reliably and accurately use blinking as a controlmechanism still needs to be explored further and tested. This form of gesture recognitionmay prove to be more reliable than existing ‘no-touch’ switch methodologies.Completed