Loading...
ZigBee-Based Tangible Input Devices
Khan, Waqqas Munir
Khan, Waqqas Munir
Description
A Master of Science thesis in Mechatronics Engineering by Waqqas Munir Khan entitled, "ZigBee-Based Tangible Input Devices," submitted in May 2015. Thesis advisor is Dr. Imran Zualkernan. Soft and hard copy available.
Abstract
ZigBee has emerged as a dominant low-power wireless standard for input devices. It has been used widely for building wireless input devices. However, the design of input devices in user environments has remained traditional and primarily consists of small keyboards and displays. The availability of cheap accelerometers and gyros has enabled the construction of gesture-based input devices. This thesis presents the design and implementation of gesture-based tangible input devices that are based on the ZigBee protocol. These ZigBee-based tangible input devices or ZTIDs are small, light-weight and graspable, and can roll, rub and tap to provide a surface gesturebased interface. This surface gesture-based interface is different from the traditional keypad-based devices because instead of a keypad, surface gestures will be used to communicate with the user environment. This type of interaction is not only natural but also creates a sense of intimacy between the user and their environment. Specialized hardware has been designed and implemented for this device. The ZTID hardware integrates the ZigBee platform, an accelerometer, a universal serial bus controller and a microcontroller with switches and an LED interface. The ZTID hardware has been tested for wireless link quality and range. It has been found that the wireless link quality is satisfactory and that ZTID can be used in star or mesh networking. The accelerometer, universal serial bus controller and a microcontroller have also been tested in the hardware. The device hardware is coupled with software implementing gesture-recognition using hidden Markov machines. Eight surface gestures have been defined for ZTID. A single Markov machine, one for each gesture, has been trained and then tested by using the HMM forward algorithm using the random sub-sampling approach. The resulting devices show an accuracy of 99.7% in gesture recognition. The ZTID gesture recognition is evaluated using receiver operating characteristics which show that the surface gestures that originate from ZTID can be used to control the user environment efficiently. Being based on ZigBee protocol, ZTID can be used as a building block for mechatronics design and also for integration with the existing products based on ZigBee.