本文へジャンプ

成果報告書詳細
管理番号20190000000250
タイトル*2018年度中間年報 次世代人工知能・ロボットの中核となるインテグレート技術開発 人工知能技術の社会実装に向けた研究開発・実証 ロボット技術と人工知能を活用した地方中小建設現場の土砂運搬の自動化に関する研究開発
公開日2019/6/19
報告書年度2018 - 2018
委託先名国立大学法人東北大学 株式会社佐藤工務店 学校法人早稲田大学
プロジェクト番号P18002
部署名ロボット・AI部
和文要約
英文要約Title: Automation of earth and sand transporting dump truck system at medium and small sized company sites using robot technology and artificial intelligence
(FY2018-FY2019) FY2018 Annual Report
Tohoku University, Sato Komuten Co. Ltd., Waseda University

A. Robotization by simple equipment installation to existing construction vehicles
1) Development of small-sized and high-strength sensor boxes for measuring motion of construction vehicles
Small-size and high-strength sensor box was developed, which can measure motion of construction vehicle. It was easily attached to construction vehicles without any processing using magnetic force. Digging and loading motions of backhoe were measured by using the sensor boxes and were analyzed at offline.
2) Development of a new type of retrofit robot SAM for easy installation to dump truck
A new retrofit robot SAM was developed, which has a simple detachable handle driving mechanism that allows manual driving. For improving the SAM, new mechanism, which ca acquire the steering angular velocity, was designed and manufactured. Switching mechanism to change autonomous driving and teleoperation was also designed for safety.
3) Development of control method of autonomous driving on a rough road
A path following algorithm in forward/backward directions was developed and succeeded in demonstrations in the test field. Also, to detect obstacles on a path, an environmental laser scanner module was developed, and the function of the module was confirmed in the test field.

B. 3D mapping of construction sites and localization in mountainous areas
4) Development of localization method based on GNSS in mountainous areas
A GNSS based localization method was developed. It can be used in mountainous areas where GNSS multi-path signals affect GNSS positioning accuracy. By selecting the line-of-sight (LOS) satellite from pre-obtained 3D terrain map, the localization accuracy was successfully improved in comparison with the conventional GNSS positioning method.
5) Constructing 3D mapping system for construction site
A 3D mapping system was developed using an unmanned aerial vehicle (UAV). GNSS and LiDAR were used to generate the accurate 3D maps. Using this system, 3D map of a construction site was easily constructed. This map can be used for autonomous navigation of the dump truck.

C. Collaborative work with other construction machines by analysis of skilled workers
6) Development of a method to construct a prediction model of backhoe maneuvering behavior
A sensor box was added with a function to measure the backhoe operator’s horn cue. Cooperation works of backhoe and dump truck were measured using the sensor boxes. Data sets were built for modeling of backhoe operator.

D. Operation planning for dump tracks according to the situation in construction field
8) Development of a motion planning method for a six-wheeled dump truck that properly
A motion planner for dump track was developed based on Hybrid State A * algorithm. It generated a suitable trajectory for dump truck controlled by SAM. It was also confirmed that the dump truck integrated with the planner and SAM can autonomously move to the position specified by a human.
ダウンロード成果報告書データベース(ユーザ登録必須)から、ダウンロードしてください。

▲トップに戻る