File(s) not publicly available

Reason: Restricted by author. A copy can be supplied under Section 51(2) of the Australian Copyright Act 1968 by submitting a document delivery request through your library or by emailing document.delivery@monash.edu

Improved mobile robot navigation using soft computing techniques

thesis
posted on 06.02.2017 by Soh , Chin Yun
Robotics is the science and technology of robots, their design, manufacture, and applications. The limitation and constrain of the existing mobile robot navigation techniques are overcome by proposing new mobile robot navigation techniques. Thus, the main focuses of this research are to investigate, propose, design and utilize the improved mobile robot navigation techniques. These techniques are applicable to mobile robots with built-in low cost ultrasonic sensors in a highly unstructured, unknown and uncertain environment. Some of the basic behaviors including Goal Seeking, Obstacle Avoidance and Wall Following are investigated using various sensor data received from the simple low cost ultrasonic sensors that are locally available on the mobile robot. Soft Computing techniques are used to improve mobile robot navigation. In this research, Fuzzy-Neural, Neural Q-Learning, Evolutionary Computation and Genetic-Fuzzy-Neural approaches are explored. In Fuzzy-Neural approach, comparative studies of the combination of Fuzzy Logic and Artificial Neural Network have been conducted. In Neural Q-Learning approach, Artificial Neural Network is used to enable the Q-Learning controller to learn from its environment and improve its perception of the environment through learning. In Evolutionary Computation approach, Goal Oriented Path Planning Algorithm (GOPPA) that is capable of obtaining an optimum collision free path in an unrecognized environment is proposed. The combination of three Soft Computing techniques including Genetic Algorithm (GA), Fuzzy Logic (FL) and Artificial Neural Network (ANN) is taken place in order to develop the proposed Genetic-Fuzzy-Neural Path Planning Algorithm (GFNPPA) for the specific problem domain i.e. near optimum path planning in goal seeking mobile robot navigation. The improved autonomous mobile robot navigation techniques are capable of guiding the mobile robot commencing from the starting point to the specified destination/target in the shortest collision free path. All of the four proposed algorithms are aimed to avoid any obstacles that are detected by the sonar sensors (ultrasonic sensors)of the mobile robot in the environment. Both simulation and real world implementations have been conducted. Their flexibility and robustness of the proposed algorithms have been shown. These proposed mobile robot navigation techniques have been compared with existing techniques which shown better performance.

History

Campus location

Australia

Principal supervisor

S. Parasuraman

Additional supervisor 1

Velappa Ganapathy

Additional supervisor 2

Bijan Shiruinzaden

Year of Award

2012

Department, School or Centre

Engineering (Monash University Malaysia)

Course

Doctor of Philosophy

Degree Type

DOCTORATE

Faculty

Faculty of Engineering

Exports