![Скачать презентацию 知能システム論1 12 移動ロボットのナビゲーション 2007 6 19 講義内容 1 はじめに 2 ベクトルの基礎 3 運動学 Kinematics Скачать презентацию 知能システム論1 12 移動ロボットのナビゲーション 2007 6 19 講義内容 1 はじめに 2 ベクトルの基礎 3 運動学 Kinematics](https://present5.com/wp-content/plugins/kama-clic-counter/icons/ppt.jpg)
dadc26566e00ee13b392ce704a088b83.ppt
- Количество слайдов: 40
知能システム論1(12) 移動ロボットのナビゲーション 2007.6.19
講義内容 1.はじめに 2.ベクトルの基礎 3.運動学(Kinematics) 4.動力学(Dynamics) 5.ロボットの腕の制御(Control)力制御 6.軌道計算(Trajectory) 7.移動ロボット(Mobile Robot)ナビゲーション
Indoor Navigation of Multiple Mobile Robots in a Dynamic Environment Using i. GPS Presented at the 2002 IEEE International Conference on Robotics and Automation
Contents • Objective • Approach to multi-robot navigation • Application of multi-robot navigation in a dynamic environment • Conclusions
Objective - Delivery Service Robots in Office -
Approach to Multi-robot Navigation • Car navigation using GPS • System configuration • Localization and map construction
Localization: i. GPS (indoor GPS) IR LED Unit [2] Y. Hada and K. Takase: ”Multiple Mobile Robot Navigation Using The Indoor Global Positioning System (i. GPS)”, Proc. of IROS 2001, pp. 1005 -1010, 2001.
Map construction Y X 14. 3[m] corridor desks Robot room 32. 2[m] We created a map using both building plan and a layout of furniture
Previous Work Multi-robot navigation in a static environment [2] Y. Hada and K. Takase: ”Multiple Mobile Robot Navigation Using The Indoor Global Positioning System (i. GPS)”, Proc. of IROS 2001, pp. 1005 -1010, 2001.
Multi-robot Navigation in a Dynamic Environment • We use the external robot control system for globally rational navigation • We need to introduce some methods to update environmental map by recognizing moving object – Using i. GPS – Using onboard sensor
Classification of obstacles • Labelable obstacles to which artificial marks can be attached – Chairs • Unlabelable obstacles which is difficult to bear artificial marks – People • Unmovable obstacles – Walls, Pillars, Desks, Shelves
Recognition of labelable Obstacles Using i. GPS • Mark-based vision [Y. Hada, K. Takase: 1997] – Object recognition can be replaced with simpler mark recognition We introduce mark-based vision into i. GPS to recognize labelable obstacles.
IR LED Unit y Start bits Object ID number bit End bit LED x LED Length: 134 mm Width : 36 mm
Object Model – pillar-like object v 4 x v 1 y Object ID=“1”…Chair v 2 v 3
Procedure of Object Recognition v 1 y v 4 x v 2 World Coordinate System v 3 Yw Xw V 1 V 4 V 2 V 3
Detection of unlable obstacles using an onboard sensor • Moving objects (people) are detected by onboard sensor • We assume that people temporarily impede a robot’s movement Laser range finder 30[degrees] Omnidirectional Mobile robot Direction of robot’s movement 60[cm]
Experimental setup
Experiment - Delivery Task -
Movable obstacle avoidance
Unload stowage
Detection of moving obstacle
Summary of Research • The multi-robot navigation system based on i. GPS was proposed. • The experiment of delivery task was carried out in the time-varying environment. • The system could successfully navigate multiple robots reliably and robustly, suggesting the practical usefulness.
Concluding Remarks • • Importance of robots for social application History of R/D on robot technology reviewed Social robots not realized by conventional R/D Intelligent Environment Supported Robot is proposed aiming at breakthrough in robotics • Mobile robot system for delivery in a building developed, demonstrating the feasibility and effectiveness of the proposed approach
Car navigation using GPS satellite • Planning route using map • Localization method Return
System Configuration External section • External sensing devices – Facilitate localization remarkably – Useful for indoor navigation Off-carrier sensors On-carrier sensors Internal section [3] R. Gutsche and F. M. Wahl: “A New Navigation Concept for Mobile vehicles”, Proc. Of IEEE IROS’ 92, pp. 215 -220, 1992. return
Mobile robot with IR LED Unit return
Experiment Goal of robot B Robot A Narrow space Bench with LED unit Static obstacles (2 tables, 3 benches) Stairs Goal of robot A Robot B
Future work • Introduce more effective sensor-based navigation scheme • Use “passive” marks like stealth barcode instead of IR LED unit.
Problem
全方向移動 ロボット • オムニホイル を用いた設計 • キネマティクス return
オムニホイル • 中心にモータ軸を 固定 • バレル部分が自由 に回転 • 60度ずらして2つ 組み合わせる モータの回転軸方向に自由に移動 120度ずらし3組配置、滑らかな全方向移動 return
全方向移動機構のキネマティクス V1 Vy r V2 Vx V3 対称型三輪機構 return
物体の実時間認識 • 物体にマークを貼付し、物体認識をマーク の認識に置き換える。 • 物体モデルとリンクすることで、その物体 が占める空間を認識する。
マークの認識 相関演算によるマークの同定 (トラッキングビジョン)
ベストマッチする場所とデストーション distortion 1 << distortion 2
情景中のマークとテンプレートとのdistortion値
物体形状モデルの作成
物体の認識 (輪郭の記述)
return
dadc26566e00ee13b392ce704a088b83.ppt