Electronic Thesis and Dissertation Repository

Thesis Format

Monograph

Degree

Doctor of Philosophy

Program

Electrical and Computer Engineering

Supervisor

Tayebi, Abdelhamid

2nd Supervisor

Polushin, Ilia G.

Co-Supervisor

Abstract

This thesis provides some hybrid feedback based autonomous navigation algorithms. First, the problem of autonomous robot navigation in planar environments with arbitrarily-shaped convex obstacles is considered. The proposed navigation approach guarantees safe and global convergence to the target location through an appropriately designed switching strategy between two different modes, namely, the move-to-target mode and the obstacle-avoidance mode. A procedure for the implementation of the proposed hybrid feedback controller in a priori unknown environments is provided.

Subsequently, the problem of autonomous navigation in planar environments with non-convex obstacles is considered. An instrumental transformation that modifies (virtually) the non-convex obstacles, in a non-conservative manner, is introduced to facilitate the design of the obstacle-avoidance strategy. The proposed autonomous robot navigation scheme relies on a switching strategy between the move-to-target mode and the obstacle-avoidance mode. When initialized in the move-to-target mode, the proposed feedback control law guarantees safe and global convergence to the predefined target location in the modified obstacle-free workspace. The proposed controller has been successfully implemented in a priori unknown environments on the Turtlebot3 burger model in the Gazebo simulator using the ROS framework.

Finally, the problem of autonomous navigation in three-dimensional environments with arbitrarily-shaped convex obstacles is addressed. The proposed hybrid feedback control strategy, which consists in switching between the move-to-target and the obstacle-avoidance modes, guarantees safe and global autonomous robot navigation. A procedure for the implementation of the proposed autonomous navigation controller, in a priori unknown three-dimensional environments, is also provided.

Summary for Lay Audience

This thesis deals with the design of control algorithms allowing to safely guide a robot from any initial location to any target location in a given workspace. We consider three types of environments: two-dimensional environments with convex obstacles, two dimensional environments with arbitrarily shaped obstacles (possibly non-convex), and three-dimensional environments with arbitrarily shaped convex obstacles. The proposed control algorithms can also be implemented in a priori unknown environments, relying solely on information obtained from the sensors mounted on the robot.

Creative Commons License

Creative Commons Attribution 4.0 License
This work is licensed under a Creative Commons Attribution 4.0 License.

Download

Share

COinS