Efficient Object Isolation In Complex Environment Using Manipulation Primitive On A Vision Based Mobile 6DOF Robotic Arm

• Main Authors: Mohd Annuar, Khalil Azha, Fahmi, Muhammad, Quah, Jit Shen
• Format: Article
• Language: English
• Published: IJENS Publisher 2018
• Subjects: T Technology (General); TJ Mechanical engineering and machinery
• Online Access: http://eprints.utem.edu.my/id/eprint/21983/2/23.%20Efficient%20Object%20Isolation%20in%20Complex%20Environment%20Using%20Manipulation%20Primitive%20on%20a%20Vision%20Based%20Mobile%206DOF%20Robotic%20Arm.pdf; http://eprints.utem.edu.my/id/eprint/21983/; http://ijens.org/Vol_18_I_01/171106-1801-8282-IJMME-IJENS.pdf; http://www.ijens.org/Vol_18_I_01/171106-1801-8282-IJMME-IJENS.pdf

This paper explores the idea of manipulation aided- perception in the context of isolating an object of interest from other small objects of varying degree of clusterization in order to obtain high quality training images.The robot utilizes a novel algorithm to plot out the position for each noise objects and its destined position as well as its trajectory and then utilizes manipulation primitives (pushing motion) to move said object along the planned trajectory.The method was demonstrated using Vrep simulation software which simulated a Kuka YouBot fitted with a camera on the gripper.We evaluated our approach by simulating the robot manipulators in an experiment which successfully isolate the object of interest from noise objects with at a rate of 77.46% at an average of 0.56 manipulations per object compared to others at 1.76 manipulations subsequently speeding up the time taken for manipulation from 12.58 minutes to 2.6 minutes however suffers from a tradeoff in terms of accuracy when comparing the similar works to our proposed method.