間接的評価値推論を用いた遺伝的アルゴリズムによる自己位置推定用画像テンプレートの高速生成

概要

論文の詳細を見る
We have proposed a self-position estimation method for an autonomous mobile robot with variable processing time. In this method, the processing time for self-position estimation can be varied by changing the image size, and image templates are generated with Genetic Algorithm in order to realize efficient and stable self-position estimation for the size change. However, it takes enormous time to generate image templates even if the number of the image templates is very small. Considering the application of the proposed method to more practical problems for self-position estimation, it is necessary to generate image templates in shorter time. In order to overcome this problem, we apply Genetic Algorithm with a fitness inference system for template generation instead of SGA(Simple Genetic Algorithm) in this paper. Some simulation results show that the time for image template generation is reduced drastically with the proposed method.