Development of hand gesture recognition device

The design of the hand gesture recognition device for this project used an Arduino Uno microcontroller together with two MMA4851 triaxial accelerometers. One of the sensors was mounted on the back of the palm and the second was mounted on the index finger. While most designs typically use a resistive flex sensor on detect the motion of the fingers, in this project an accelerometer is used instead. The primary reason for this was due to the dimensional limitations of a resistive flex sensor. The flex sensor can only determine the curvature of the finger but is unable to detect movement in the adjacent plane. The accelerometer would not have this limitation. Thus, allowing a more accurate detection of the hand movement including the movement of the finger. Furthermore, the gesture recognition aspect can be done using machine learning algorithms. Data was collected from repeated movement of the hand for several gestures and stored in a data frame. This data was used to train a predictive model in machine learning that could recognize the gestures. However, the focus of this project was on the thought-process, fabrication, and design of the hand gesture recognition device along with the code used to collect the data. And lastly, some data exploration to validate and check the data collected for errors and feasibility. In this report, the details of the machine learning algorithms is mentioned but not explored in depth.Bachelor of Engineering (Mechanical Engineering