Emergent Robot Navigation Integrating Unsupervised Word Discovery and Generative Model for Mapping and Localization
Project/Area Number |
16K12497
|
Research Category |
Grant-in-Aid for Challenging Exploratory Research
|
Allocation Type | Multi-year Fund |
Research Field |
Soft computing
|
Research Institution | Ritsumeikan University |
Principal Investigator |
|
Co-Investigator(Kenkyū-buntansha) |
萩原 良信 立命館大学, 情報理工学部, 講師 (20609416)
|
Project Period (FY) |
2016-04-01 – 2019-03-31
|
Project Status |
Completed (Fiscal Year 2018)
|
Budget Amount *help |
¥3,250,000 (Direct Cost: ¥2,500,000、Indirect Cost: ¥750,000)
Fiscal Year 2018: ¥1,170,000 (Direct Cost: ¥900,000、Indirect Cost: ¥270,000)
Fiscal Year 2017: ¥1,170,000 (Direct Cost: ¥900,000、Indirect Cost: ¥270,000)
Fiscal Year 2016: ¥910,000 (Direct Cost: ¥700,000、Indirect Cost: ¥210,000)
|
Keywords | ロボットナビゲーション / 確率モデル / ソフトコンピューティング / 地図生成 / 言語獲得 |
Outline of Final Research Achievements |
In this study, we aimed to construct an innovative space representation method for mobile robots that integrates vocabulary acquisition and map formation via a probabilistic generative model. Furthermore, we developed an emergent robot navigation method based on this space representation method. For this purpose, (1) we developed a probabilistic generative model integrating vocabulary acquisition and map generation and derived fast inference procedure, (2) we developed emergent robot navigation method based on probabilistic generation model of vocabulary and map, and (3) implemented the space representation method and demonstrated its effectiveness verification in a real environment.
|
Academic Significance and Societal Importance of the Research Achievements |
人間は移動や行動の指示を出す時に「玄関まで行って」「キッチンを掃除しておいて」などという言語的表現を自然と用いる.しかし,従来のロボットは場所に関する知識を客観的な座標系で管理しており,このような意味表現を持つわけではない.そこで,本研究ではロボットが言語的な知識で地図を学習し,表現し,そのような知識を用いて人間がロボットのナビゲーションを行える手法の研究を行った.具体的な成果としては語彙獲得と地図学習を融合的に行う手法や,言語的知識に基づく確率的な経路計画手法を開発することができた.
|
Report
(4 results)
Research Products
(14 results)