Improving Efficiency of Automatic Transportation in Dynamic Environments of Personal Intelligent Vehicles
Generally, it is difficult for machines to understand a dynamic environmentwith sensors which are installed on them. Therefore, vehicles often make inefficient movements when they move automatically in the dynamic environments. Forexample, it is not correctly predictable whether a passage which is scheduled torun is passable or impassable. If a passage is impassable because of obstacles suchas crowded people or pieces of baggage which are located on passage temporarily,vehicles have to change the route after they encountered with them. In that case,vehicle’s movement will become inefficient. It is necessary to detect changes ofstates of the route which is scheduled to run as earlier as possible based on information collected with the sensors not installed on the vehicle.
In this study,we aim at achievement of higher efficiency of automatic transportation in dynamicenvironments of personal intelligent vehicles. Our laboratory developed a personalintelligent vehicle called AT (Attentive Townvehicle). AT is a vehicle that adaptsto a person who is passenger and an environment that surrounds the vehicle, andis capable of omnidirectional movements and automatic transportation. AT canget environmental information and traveling route from servers which are installedin buildings. We also developed SUVs (Small Unmanned Vehicles), as new meansto capture environmental information.
SUV is a small vehicle capable of autonomous behavior, has a laser-range finder to get environmental information. The server can acquire changes of environmental information in almost real timeby using several SUVs that autonomously run in the environment to detect anyobstacles located on the ways.
The server can send appropriate action for ATswhich is moving automatically, according to changes of the environmental information. As a result, AT becomes possible to change the route dynamically beforeit reaches to a place where obstacles exist, if SUV discovers the obstacles in theplanned route which is scheduled to run, then the efficiency of automatic transportation of AT will improve.
To confirm usefulness of our mechanism, we checkedthe efficiency of AT’s automatic transportation by using a simulator. As a result,when SUVs were deployed in the dynamic environment, automatic transportationof AT became more efficient than the case that without SUVs. Therefore, we couldconfirm the effectiveness of our method.
1
2
2.1
Fugure1:
Fugure2:
Fugure3:
2.2
2.3
2.4
2.5
Fugure4:
2.6
2.7
2.7.1
Fugure5:
2.7.2
Fugure6:
2.8
2.8.1
2.8.2
Fugure7:
- 第1レイヤ：マニュアル走行 ・・・ 後述する操縦コントローラによって搭乗者の操作を受け付けるレイヤ
- 第2レイヤ：補助走行 ・・・ 一定距離以内に障害物が存在した場合，停止や進路を自動調節するレイヤ
- 第3レイヤ：障害物回避 ・・・ 障害物を検知しそれを回避するレイヤ
- 第4レイヤ：目的地誘導 ・・・ 地図に基づいて現在位置から目的地まで自動走行を行うレイヤ
- 第5レイヤ：連携走行 ・・・ 他の移動体と連携しながら走行するレイヤ
2.8.3
2.9
2.9.1
Fugure8:
Fugure9:
2.9.2
Fugure10:
Fugure11:
2.10
Fugure12:
Fugure13:
Fugure14:
Fugure15:
Fugure16:
2.11
2.11.1
Fugure17:
- 搭乗者が目的地を指定したときに目的地まで自動的に走行する機能
- 障害物を認識したらそれを自動的に回避する機能
- 他の移動体と連携・協調する機能
2.11.2
2.11.3
Fugure18:
Fugure19:
Fugure20:
Fugure21:
Fugure22:
Fugure23:
2.12
2.13
3
3.1
Fugure24:
3.2
3.2.1
Fugure25:
3.2.2
Fugure26:
3.2.3
Fugure27:
3.3
3.3.1
Fugure28:
3.3.2
Fugure29:
3.4
3.5
Fugure30:
Fugure31:
Fugure32:
Fugure33:
3.6
3.6.1
3.6.2
3.6.3
Fugure34:
Fugure35:
Fugure36:
3.6.4
Fugure37:
Fugure38:
3.6.5
Fugure39:
3.6.6
3.7
4
4.1
Fugure40:
4.1.1
Fugure41:
4.1.2
Fugure42:
4.1.3
Fugure43:
4.1.4
Fugure44:
4.1.5
4.1.6
Fugure45:
4.1.7
4.2
Fugure46:
4.3
Fugure47:
Fugure48:
Fugure49:
4.4
4.5
4.5.1
Fugure50:
Fugure51:
4.5.2
Fugure52:
Fugure53:
4.5.3
Fugure54:
Fugure55:
Fugure56:
4.5.4
4.6
5
5.1
- 経路変更の効率を上げること
- 環境の変化を知るタイミングの遅れを減らすこと
Fugure57:
Fugure58:
5.2
5.2.1
5.2.2
5.3
Fugure59:
5.4
- SUVが存在しない場合よりも存在する場合の方が自動走行の効率が良い
- SUVには台数効果があるが，台数が増えるといずれは効率性が頭打ちになり，最適な台数が決まる
5.4.1
Fugure60:
Fugure61:
Fugure62:
Fugure63:
Fugure64:
Fugure65:
Fugure66:
5.4.2
Fugure67:
5.4.3
Fugure68:
Fugure69:
Fugure70:
5.4.4
Fugure71:
6
6.1
6.1.1
- 従来のロボットは視覚センサが台車に固定されており，適切な視点からの視覚情報を得るのが難しい
- 単体のロボットでは，広範囲にわたる動的な環境の首尾一貫したモデルを維持するのが難しい
Fugure72:
6.1.2
6.2
6.2.1
6.2.2
7
7.1
7.2
7.2.1
7.2.2
7.2.3
7.2.4
Fugure73: