Recent, innovative developments in the field of gesture interfaces as input techniques have the potential to provide a basic, lower-cost, point-and-click function for graphic user interfaces (GUIs). Since these gesture interfaces are not yet widely used, indeed no tilt-based gesture interface is currently on the market, there is neither an international standard for the testing procedure nor a guideline for their ergonomic design and development. Hence, the research area demands more design case studies on a practical basis. The purpose of the research is to investigate the design factors of gesture interfaces for the point-andclick task in the desktop computer environment. The key function of gesture interfaces is to transfer the specific body movement into the cursor movement on the two-dimensional graphical user interface(2D GUI) on a real-time basis, based in particular on the arm movement. The initial literature review identified limitations related to the cursor movement behaviour with gesture interfaces. Since the cursor movement is the machine output of the gesture interfaces that need to be designed, a new accuracy measure based on the calculation of the cursor movement distance and an associated model was then proposed in order to validate the continuous cursor movement. Furthermore, a design guideline with detailed design requirements and specifications for the tilt-based gesture interfaces was suggested. In order to collect the human performance data and the cursor movement distance, a graphical measurement platform was designed and validated with the ordinary mouse. Since there are typically two types of gesture interface, i.e. the sweep-based and the tilt-based, and no commercial tilt-based gesture interface has yet been developed, a commercial sweep-based gesture interface, namely the P5 Glove, was studied and the causes and effects of the discrete cursor movement on the usability was investigated. According to the proposed design guideline, two versions of the tilt-based gesture 3 interface were designed and validated based on an iterative design process. Most of the phenomena and results from the trials undertaken, which are inter-related, were analyzed and discussed. The research has contributed new knowledge through design improvement of tilt-based gesture interfaces and the improvement of the discrete cursor movement by elimination of the manual error compensation. This research reveals that there is a relation between the cursor movement behaviour and the adjusted R 2 for the prediction of the movement time across models expanded from Fitts’ Law. In such a situation, the actual working area and the joint ranges are lengthy and appreciably different from those that had been planned. Further studies are suggested. The research was associated with the University Alliance Scheme technically supported by Freescale Semiconductor Co., U.S.
| Identifer | oai:union.ndltd.org:bl.uk/oai:ethos.bl.uk:516068 |
| Date | January 2009 |
| Creators | Chen, Tin Kai |
| Publisher | De Montfort University |
| Source Sets | Ethos UK |
| Detected Language | English |
| Type | Electronic Thesis or Dissertation |
| Source | http://hdl.handle.net/2086/2420 |
Page generated in 0.0016 seconds