Development of an Integrated Control System for a Humanoid Robot

Abstract

Humanoid robot is a huge field for modern robotics research. Many renowned organizations and intellectuals are still working on it. Some outcome of their long term research are ASIMO by Honda, NAO by Aldebaran, HUBO by KAIST, WABIAN by Waseda University etc. For the implementation of an idea, it is essential to have a full engineering system design, simulation and analysis. MISTBOY, is such a dream bot consist of 19 DoF, 45 cm height and 3.5 kg weight. Mechanical Engineering department of Military Institute of Science and Technology aims to design and built a humanoid that is capable of balancing, walking, turning, standing from a prostrate position and finally can play soccer autonomously. The first purpose, on the of making a soccer playing robot, is to develop a stable and universal humanoid platform on which can be implement various theories and algorithms such as forward kinematics, trajectory planning, dynamic walking, AI, Vision and Image recognition and navigation. Forward kinematics is helpful for balancing the robot and trajectory planning optimizes the motion of the robot. Forward kinematics calculates the position and orientation of the joint while trajectory planning is one of the fundamental issue in robot design and analysis which relates position, velocity and acceleration for each degree of freedom of manipulator with time. Soccer playing humanoid robot requires a minimum amount of energy to pass the ball to a certain distance which can be determined by analyzing the force and torque produced by the kick. This paper is focused on forward kinematics, trajectory planning, balancing, force and torque calculation and kick analysis. Machine vision (MV) is developed from the concept of human vision. It works on the basis of pattern and color matching like human. MV is nothing but digital image processing which involves extraction of information from an image. It is the young discipline of modern technology, widely used for industrial purposes, security purposes, medical diagnostics, weapon i

ii engineering and so on because of its significant level of accuracy and reliability. Now-a-days, in the field of robotics, MV is expanding its area of research like in ASIMO, HUBO, NAO. Intelligent vision algorithm is one of the most reliable and effective way to develop the control system of an autonomous robot as it can extracts maximum amount of real time data from the environment. Integration of vision data with other measurement unit, can make the control system more reliable, accurate and stable. Such kind of integrated control system has been implemented on MISTBOY, ongoing humanoid robot project. The vision system of MISTBOY is based on the NI LabVIEW Vision. National Instruments introduced a graphical programming platform LabVIEW for the engineers and scientist to solve problem easily, accurate the productivity and innovate continually. Various approach has been analyzed to find out the best method for faster and accurate tracking system of a robot. In this paper, the integrated vision based control system is also presented with analyzed result through mathematical derivation and simulation.