| 研究実績の概要 |
This research aims at the automation of the hammering test for the inspection of concrete structures. Previous approaches based on supervised learning lacked practicability due to the high cost of training data. Thus this research focused on alleviating that cost, notably through domain adaptation, for the automation of the hammering test.
This year, attention was brought on the hardware side of the system, consisting of a UAV for hammering inspection. To solve the acoustic noise issue inherent to such robots, a noise suppression method based on prediction from propeller acceleration was proposed. Additionally, the use of a force sensor mounted on the tip of the hammer was explored and yielded highly promising results.
|
