2022 Fiscal Year Final Research Report
Real-time SLAM for Dynamic Environments
| Project/Area Number |
19H02098
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (B)
|
| Allocation Type | Single-year Grants |
|---|
| Section | 一般 |
|---|
| Review Section |
Basic Section 20010:Mechanics and mechatronics-related
|
|---|
| Research Institution | Tokyo City University |
|---|
Principal Investigator |
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
関口 和真 東京都市大学, 理工学部, 准教授 (80593558)
大貝 晴俊 早稲田大学, 理工学術院(情報生産システム研究科・センター), 名誉教授 (80367169)
|
|---|
| Project Period (FY) |
2019-04-01 – 2022-03-31
|
| Keywords | SLAM / MHE / 実時間最適化 / 動的環境 / モデル予測制御 / 物体追跡 |
|---|
| Outline of Final Research Achievements |
This research investigated simultaneous localization and mapping (SLAM) for environments with dynamically moving objects. Although SLAM has become a commonly used method for mobile robots and self-driving vehicles, the conventional methods were sensitive to variation due to the moving objects. In this research, we built a novel SLAM that explicitly deals with moving objects in its model to estimate a map of both static and moving objects. To suppress the corruption of the map due to the assumption that it does not identify the moving objects, we introduced a moving horizon estimation (MHE) to devise an objective function that enhances the robustness of the map. In addition, a robust SLAM in singular environments was proposed. Finally, the simulation and experimental results show that the method generates a map in the real environment.
|
| Free Research Field |
制御工学
|
| Academic Significance and Societal Importance of the Research Achievements |
SLAMは地図生成の手法として広く使われていますが，一般的に建物や壁など静的な物体が対象でした．一方で移動体が含まれる場合は，静的物体と移動体の情報を分けて，別々に推定することが一般的でした．本研究ではMoving Horizon Estimation(MHE)を用いることで，動的物体を扱うモデルとそれに対応した評価関数を導入しました．これにより，動きのある物体が含まれた環境でも，移動物体の抽出などの処理を施すことなく，地図を生成し，移動物体を追跡する新しいSLAMの基礎を構築しました．これに加えて，遮蔽の生じる移動体の追跡や特異環境での推定などの成果もあげることができました．
|
