Intelligent Navigation for Autonomous Robots using Neural Network and Fuzzy Logic Techniques

Date
2013-01-21
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
This thesis proposes a simple navigation method for a mobile robot with limited computational power in uncertain obstacle-ridden environments. Rather than mapping the environment, a reactive method based on closed-loop control moves the robot around obstacles and ultimately to the target. Instead of relying on vision systems, simple sensors provide basic information on relative target position and the shape of straight-ahead obstacles. This sensor information is also provided to a neural-fuzzy system, allowing the robot to decide what sensor inputs truly represent an obstacle, based on learning experience. A Lyapunov-stable nonlinear control provides commanded linear and angular velocities to achieve target-seeking motion. In the control design, a filter is added to the robot dynamics, resulting in gently curved trajectories of motion. The information about the obstacle, after processing by the neural-fuzzy system, is input as a disturbance into the control signal. Simulations and experimental results with a simple Lego robot verify the expected behaviour of the robot.
Description
Keywords
Engineering--Electronics and Electrical, Robotics
Citation
Butt, T. (2013). Intelligent Navigation for Autonomous Robots using Neural Network and Fuzzy Logic Techniques (Master's thesis, University of Calgary, Calgary, Canada). Retrieved from https://prism.ucalgary.ca. doi:10.11575/PRISM/26593