Fall Detection in T-FLoW Humanoid Robot: V-REP Simulation

摘要

Research on T-FloW humanoid robot in EEPIS Robotics Research Center (ER2C) has entered the third generation. Previously, the robot already had primitive motions to mimic 'human-like' walking capabilities by making use of an open loop motion controller. Yet, a feedback control system is required in order to take corrective measures due to walking parameter variations. In this current stage of the work, the development is aimed to provide the robot with adequate number of sensors to detect the movement of the robot. This paper describes the utilisation of the sensors in our robot to analyze and detect its condition when the robot begins to fall down. An inertial measurement unit (IMU) sensor consists of 3-axes accelerometer and 3-axes gyrometer was in use. The analysis was done under V-REP simulation software which is equipped with Vortex dynamic engine. The simulation showed the detected characteristic of the walking parameters of the robot just before it begins to fall down. This simulation result is paramount for further T-FloW walking controller development.