Susanto, Kevin and Rusyadi, Rusman (2016) Designing and Constructing an Autonomous Human-Following Robot With a Vision-Based System Using Several Object Tracking Algorithms. Bachelor thesis, Swiss German University.

Preview	Text Kevin Susanto 11112024 TOC.pdf Download (1MB) | Preview
	Text Kevin Susanto 11112024 1.pdf Restricted to Registered users only Download (580kB)
	Text Kevin Susanto 11112024 2.pdf Restricted to Registered users only Download (2MB)
	Text Kevin Susanto 11112024 3.pdf Restricted to Registered users only Download (7MB)
	Text Kevin Susanto 11112024 4.pdf Restricted to Registered users only Download (8MB)
	Text Kevin Susanto 11112024 5.pdf Restricted to Registered users only Download (362kB)
Preview	Text Kevin Susanto 11112024 Ref.pdf Download (679kB) | Preview

Abstract

The purpose of this thesis is to design and construct an autonomous human following robot with vision-based system using a camera in context of indoor environment. Practically, this robot is built to benefit the old or disabled people [13] [22] when they are shopping in the supermarket to reduce their physical effort. To make the robot follow the target human, there are some features that the system must be able to detect and track, such as upper-body, color of clothing or their appearance. To detect those features, several algorithms will be implemented and tested, including the CamShift Algorithm with Kalman Filter, Haar Cascade Classifier, HOG Person Detector and Condensation algorithm. Furthermore, the performance of these algorithms will be compared and choose the best one to be implemented. To maintain the safe distance between the robot and the human, distance measurement sensor will be implemented and will be processed by microcontroller board which will communicate with mini- PC through Firmata Communication Protocol. From the tests, the combination of Haar-Cascade Classifier and Camshift Algorithm with Kalman Filter is more suitable to be implemented in human-following robot in proper lighting condition indoor environment with 88% of success rate and infrared distance sensor is more accurate (90%) than ultrasonic sensor (55%) to measure the distance to the targeted human.

Item Type:	Thesis (Bachelor)
Uncontrolled Keywords:	CamShift Algorithm; Haar Cascade Classifier; Condensation Algorithm; HOG People Descriptor; Distance Measurement Sensor.
Subjects:	T Technology > TJ Mechanical engineering and machinery > TJ163.12 Mechatronics T Technology > TJ Mechanical engineering and machinery > TJ210 Robots (General)
Divisions:	Faculty of Engineering and Information Technology > Department of Mechatronics Engineering
Depositing User:	Atroridho Rizky
Date Deposited:	12 Nov 2020 13:16
Last Modified:	12 Nov 2020 13:16
URI:	http://repository.sgu.ac.id/id/eprint/1417

Actions (login required)

View Item