Obstacle Avoidance for Autonomous Navigation based on Context Awareness
- Author(s)
- Muhammad Tahir Rasheed
- Issued Date
- 2009
- Abstract
- 자율주행 로봇에 대한 연구는 인간이 이 세상에서 로봇을 꿈꾸던 때부터 시작 되었고 끊임없이 연구 되고 있다. 로봇이란 인공적인 인간이라고 할 수 있겠다. 이런 로봇에 자율성을 부여하고 스스로 장애물을 회피하면서 이동경로를 설정 움직이기란 쉬운 일이 아니다.
본 연구 논문은 상황인식 기반에서 자율 주행로봇을 위한 상황인식 기반 알고리즘을 보여주고 있으며 어떠한 장애물도 회피 할 수 있도록 디자인 되어 있다. 본 모델은 정지한 물체와 움직이는 물체를 다 고려해 디자인 되어 있다. 장애물을 회피하기 위해서 기본적으로 퍼지 시스템을 이용하였고 장애물 감지를 위해서는 적외선 센서를 스텝모터 위에 올려 패닝 스캔 방법을 사용하였다. 이동 물체에 대한 감지는 로봇의 중심체를 중심으로 원을 그려 반경을 구하고 패닝 스캔을 이용하여 감지 하였다. 그리고 움직이는 이동체의 거리 및 위치 반경을 예측하기 위해서 칼만 필터를 사용하였다. 실험을 위해서는 간단한 시뮬레이션과 실 로봇을 가지고 테스트 하였다. 많은 장애물들이 있었지만 거의 완벽하게 회피 하였다. 이동 물체에 대한 계산 량이 많아 위치 인식 및 로봇 이동 경로에 대한 문제가 있었지만 곧 해결될 것이다.|In this paper algorithm for obstacle avoidance in autonomous robot is proposed based on context awareness. The main object is to navigate the autonomous robot to the desired destination while detecting and avoiding any possible obstacles. A model theory of autonomous mechanism of robot navigation is showed. We have considered both stationary and moving obstacles. The stationary obstacles avoidance mechanism is based on fuzzy systems and infrared sensors mounted on step motors are used to detect the obstacles. The proposed algorithm detects the position and size of the obstacle and helps the robot to decide the movement action. For moving obstacles we consider the moving obstacle as a moving circle thus finding the center and radius of the circle becomes the primary concern. This is carried out by circularization. The velocity and position of the moving obstacle is estimated by using Kalman filter for future collision estimation. The collision is avoided by determining the angular velocity and direction of the moving obstacle. The test results show the navigation procedure and the action taken by the robot in certain situations.
Many other obstacle avoidance architectures are proposed consisting of resource, behavior, and controller with object-oriented approach structures showing good performance but their heavy calculation method makes it difficult to perform in real- time control. But in this paper, algorithms using geometrical methods have been introduced to ensure simple avoids with less computation aimed to improve real-time abilities. Moreover, the application of Kalman Filter led to the minimization of sensor and system errors.
- Alternative Title
- 상황 인식기반 자동 네비게이션을 위한 장애물 회피 기법
- Alternative Author(s)
- 무하마드 타히르 라쉬드
- Affiliation
- Chosun University, Graduate School of Chosun University
- Department
- 일반대학원 정보통신공학과
- Advisor
- Prof. Lee Joon
- Awarded Date
- 2010-02
- Table Of Contents
- Contents
Abstract
1.Introduction
1.1 Motivation
1.2 Current work
1.3 Thesis Overview
2.Obstacle Avoidance and Context Awareness
2.1 Obstacle Avoidance
2.2 Context Awareness
2.3 Ubiquitous Computing
2.4 Activity recognition
2.5 Challenges in Context-Aware Computing
2.6 Motivation to the Proposed Obstacle Avoidance Algorithm
3.Obstacle avoidance and autonomous navigation
3.1 Global and Local Obstacle Avoidance
3.2 Algorithm for autonomous navigation
3.3 Algorithm for panning scan by sensors for course driving
4.Moving obstacle avoidance using LRF sensor
4.1 Obstacle Detection
4.1.1 Obstacle Identification
4.1.2 Segmentation
4.1.3 Circularization
4.1.4 Estimation of future collision
4.2 Algorithm for Collision Avoidance
4.2.1 Obstacle's direction conversion and distance measured from the LRF sensor’s center to the robot’s center
4.2.2 Determination of moving direction
4.2.3 Velocity and angular velocity determination for obstacle avoidance
5.Test Results and Implementation
5.1 Panning scan System of Sensors
5.2 Implementation of Autonomous Navigation System using autonomous navigation algorithm
6.Conclusion
References
- Degree
- Master
- Publisher
- 조선대학교
- Citation
- Muhammad Tahir Rasheed. (2009). Obstacle Avoidance for Autonomous Navigation based on Context Awareness.
- Type
- Dissertation
- URI
- https://oak.chosun.ac.kr/handle/2020.oak/8401
http://chosun.dcollection.net/common/orgView/200000239176
-
Appears in Collections:
- General Graduate School > 3. Theses(Master)
- Authorize & License
-
- AuthorizeOpen
- Embargo2010-01-25
- Files in This Item:
-
Items in Repository are protected by copyright, with all rights reserved, unless otherwise indicated.