Publication detail

Robot Navigation Using Genetic Algorithm and Case-Based Reasoning

ŠEDA, M., DVOŘÁK, J.

Original Title

Robot Navigation Using Genetic Algorithm and Case-Based Reasoning

Type

conference paper

Language

English

Original Abstract

The aim of the paper is to propose an integration of a genetic algorithm and case-based reasoning in robot motion planning in a partially known dynamic environment. The goal of the planning is to help find a path from a start to a goal position without collisions with known obstacles minimizing length and difficulty of the path. The environment is modelled by a grid in which known static obstacles or unattainable positions are defined. The robot motion is reduced in horizontal, vertical and diagonal directions. The paths realized are stored in a base of cases along with the degree of their traversability. When planning a path, first this base is searched so as to find the cases that are most similar to the given case and then they are adapted to it. If similar cases are not found in the base or adapted solutions are not good enough, a new path is searched for by a genetic algorithm.

Key words in English

motion planning, genetic algorithm, case-based reasoning

Authors

ŠEDA, M., DVOŘÁK, J.

RIV year

2003

Released

1. 5. 2003

Publisher

Association for Engineering Mechanics

Location

Svratka

ISBN

80-86246-18-3

Book

Book of Extended Abstracts of the National Conference with International Participation Engineering Mechanics 2003

Pages from

328

Pages to

329

Pages count

2

BibTex

@inproceedings{BUT13226,
  author="Miloš {Šeda} and Jiří {Dvořák}",
  title="Robot Navigation Using Genetic Algorithm and Case-Based Reasoning",
  booktitle="Book of Extended Abstracts of the National Conference with International Participation Engineering Mechanics 2003",
  year="2003",
  pages="2",
  publisher="Association for Engineering Mechanics",
  address="Svratka",
  isbn="80-86246-18-3"
}