Micro electro mechanical systems (MEMS) enabled digital pen

• Main Author: Tan, Joel Qi Zhi
• Other Authors: Li King Ho Holden
• Format: Final Year Project
• Language: English
• Published: 2015
• Subjects: DRNTU::Engineering::Mechanical engineering::Assistive technology
• Online Access: http://hdl.handle.net/10356/63450
• Institution: Nanyang Technological University

Keyboards have been the key human-to-computer input device since they were first introduced to the commercial world in the 1970s. With keyboards, notes-taking became digitized and can be done in a much shorter time as compared to physical writing. However, when it comes to mobile note-taking, bringing a keyboard around can be a hassle; thus digital pens were invented to combine the advantages of both physical writing and keyboard input. In the market, there are many different types of digital pens being sold. However, most of them requires an additional item in complement to the pen to read the written notes. The aim of this project is to introduce a new generation of digital pen which incorporates Micro Electro Mechanical Systems (MEMS) technology. As MEMS tends to be small, it can be easily incorporated within the pen to compute its written path. In addition, using a MEMS enabled digital pen, users can take down their notes anywhere, without needing a surface to write on. The MEMS used in this project is STEVAL-MKI119V1 evaluation board from ST Microelectronics, which comprises of an accelerometer, LSM303DLHC, and a gyroscope, L3GD20. Firstly, Allan Variance was conducted to determine the initial sources of errors of the MEMS used for this project. Acceleration and angular rate data were collected from the MEMS evaluation board while it was being held in a static upright position on an optical table for a prolonged period of time. The runtime of this data collection lasted for 8 hours and 20 minutes, producing approximately three million data points. Upon completing the data collection, data were imported into AlaVar5 to compute the Allan standard deviation of each axis of the accelerometer and gyroscope. Secondly, experimental setup of the digital pen was fabricated. The fabricated experimental setup consists of a writing base and two adaptors, each serving a different purpose. The writing base was used to ensure the pen is in an upright position for each experiment. An adaptor was used to hold the pen at a specific location along the pen to ensure that it was rotated at the same pivot point throughout the experiments. Another adaptor was used to house the MEMS Inertia Measurement Units (IMUs) and to be attached to the pen. Thirdly, four letters were chosen to be written by the digital pen during the experiments. These four letters are ‘M’, ‘S’, ‘N’ and ‘C’. These letters were simple enough to be written with a continuous stroke of the digital pen, and therefore enable a continuous stream of data output from the MEMS. A total of 10 motions were recorded for each letter. The data from these 10 motions were then combined and averaged to eliminate any random characteristics of the human motion. The averaged data were then imported into MATLAB for further processing. Lastly, navigation algorithms were looked into to determine the best algorithm for use with the digital pen. In the end, Extended Kalman Filter (EKF) was chosen as it provides reasonable performance without too long of a computational time. MATLAB was then used to apply the EKF on the averaged data obtained from the writing simulation of each letters. A low pass filter was also applied after the EKF to eliminate any short-term fluctuations while keeping the long-term trend. The output of the filters was then used to plot out the letters in the digital world. It is found that the virtual letters plotted were similar to the actual letters written. The results of this project concludes that MEMS enabled digital pen is able to record and display written digital letters that were readable to users.