Development of pathological tremor cancelling for human-machine interaction

• Main Author: Fariadi, Yulius Setiadi
• Other Authors: Seet Gim Lee, Gerald
• Format: Final Year Project
• Language: English
• Published: 2010
• Subjects: DRNTU::Engineering::Mechanical engineering::Assistive technology
• Online Access: http://hdl.handle.net/10356/40124
• Institution: Nanyang Technological University

The project investigates the collaboration in human-machine interaction for the control system of robot with joystick as the user input interface. The purpose of this project is to implement an online pathological tremor cancelling method in hu-man-machine interaction. Current development of the project is the intention recognition system with filtering ability (IRS-F). The system can be divided into three main modules and one sub-module. The main modules are inferring module, decision making module, and behavior module. The sub-module is the feature recognition module. This project only focused in the inferring module of the system which uses the Bayesian theorem and Dynamic Bayesian Network as the basic architecture. In the previous system, the concept of fuzzy logic was implemented to the system by creating the overlapping regions as the tolerance for the tremor. Due to non-stationary process of pathological tremor, there is a major drawback in the fuzzy logic system when the tremor atte-nuates. Prior knowledge is required about the tremor. If a case is missed, the control-ler would not work properly. It is complicated to make so many different fuzzy states for every different case. The system is not able to compensate for the non-stationary process of pathological tremor. Hence, the author tried to overcome this problem by implementing a filter so that it is able to adapt itself to the tremor characteristics. Weighted-frequency Fourier linear combiner is well-suited for this application be-cause of its ability to estimate the tremor frequencies as well as amplitudes and then cancels the tremor. A simple 2D simulation program which requires the user to follow a narrow track was produced to measure the performance of the filter. When the simulation test had been successful, experiments were done to user with tremor to see the online cancelling of pathological tremor for computer input with WFLC filter.