A neural network based approach for path planning on costmap

Author

Songjie Chen

Abstract

"In this thesis, a path planning method using a three-level hierarchical system model is proposed. In the top level of the hierarchy, the given location points to visit are taken as waypoints. The waypoint navigation process is formulated as a traveling salesman problem (TSP). The Lin-Kernighan algorithm is used to solve the TSP and transfer the solution to the lower level of the hierarchy. In the middle level, the path is represented by a grid-based costmap. A novel modified Grossberg neural network is designed to solve the point-to-point path planning. The bottom level of the hierarchy smoothens the path with kinematic constraints. The final results are simulated in a 3D virtual reality environment by using the MATLAB VR toolbox"--Abstract, page iii.

Advisor(s)

Balakrishnan, S. N.

Committee Member(s)

Dagli, Cihan H., 1949-
Wunsch, Donald C.

Department(s)

Mechanical and Aerospace Engineering

Degree Name

M.S. in Mechanical Engineering

Publisher

Missouri University of Science and Technology

Publication Date

Spring 2011

Pagination

vii, 44 pages

Note about bibliography

Includes bibliographical references (pages 41-43).

Rights

© 2011 Songjie Chen, All rights reserved.

Document Type

Thesis - Open Access

File Type

text

Language

English

Subject Headings

Traveling salesman problem
Neural networks (Computer science)
Intelligent control systems -- Mathematical models

Thesis Number

T 10227

Print OCLC #

863049523

Electronic OCLC #

909535007

Recommended Citation

Chen, Songjie, "A neural network based approach for path planning on costmap" (2011). Masters Theses. 4482.
https://scholarsmine.mst.edu/masters_theses/4482