Realization of the Imitation Robot Based on Kinect and Kinematics

Abstract

In the field these days the robot application is extremely broad With the development of science and technology the robot has gradually replaced humans in repetitive dangerous and heavy work such as demolition explosives or seabed mining and so on If this kind of the works is replaced by a robot you can reduce or avoid unnecessary casualties Therefore the main purpose of this thesis is through simple human-machine interaction to achieve the robot control in order to facilitate the completion of the specified action to assist in the degree of difficulty or the relatively high risk of problem-solving work In the case of the current recession many industries also want to minimize the cost component so this thesis is also desirable to be able to maximize low costs to reach the system structure of imitation However Microsoft launched Xbox360 somatosensory peripheral Kinect sensor its low cost and a powerful image recognition library dramatically reduce the cost of image recognition applications This thesis will select Kinect sensor with 16 degree of freedom (DOF) humanoid robot RoboBuilder to construct a less expensive system of the imitation The content of the thesis is consisted of three parts including the Kinect sensor the software interface and calculating the joint angle by kinematics algorithms As for the software interface we use the C# of Microsoft Visual Studio 2012 with an official Kinect Software Development Kit (Kinect for Windows SDK) to develop the software interface In the part of algorithms inverse geometric kinematics is utilized to reduce the computational time to achieve the immediate imitation In this study we apply with Kinect sensor to capture human skeleton information through the proposed algorithm to calculate the angle of each joint and then the corresponding control signal to the robot servo motor to make the corresponding operation From the experimental results it shows that the proposed system can immediately achieve the visual effect of imitation

Date of Award	2016 Jul 28
Original language	English
Supervisor	Teh-Lu Liao (Supervisor)