| 片山 寛基 | M, 2回目発表 | ロボティクス | 小笠原 司, 浦岡 行治(MS), 中村 哲, 高松 淳, Gustavo Garcia |
|
title: Method Deveiopment of Surface Normal Measurement Using Sound for Roboto Manipulation
abstract: In robots, there are still problems in gripping in the dark and gripping transparent objects. As a solution, we propose a normal direction estimation method using sound waves.With reference to Shape-from-Shading represented by Photometric Stereo, we have developed a normal direction estimation method that has not been used in sound waves until now. In my method, it is used that the reflection characteristic that is a characteristic of a sound wave changes depending on the height of the frequency. language of the presentation: Japanesese 発表題目: 音波を用いたロボットの物体把持のための法線方向推定手法の開発 発表概要: ロボットにおいて、暗闇での把持や、透明物体の把持するにはまだ、問題がある。解決策として、音波を用いての法線方向推定手法を提案する。Photometric Stereoに代表される陰影解析を参考にして、これまで音波では使われてこなかった法線方向推定手法を開発した。その中では、音波の特徴となる反射特性が周波数の高さによって変化することを用いている。 | |||
| 三野 智貴 | M, 2回目発表 | ディペンダブルシステム学 | 井上 美智子, 浦岡 行治(MS), 大下 福仁, 新谷 道広 |
|
title: Efficient Challenge-Response Pairs Generation and Evaluation for PUF
Circuit Using BIST Circuit During Manufacturing Test
abstract: Device identification using challenge-response pairs (CRP) obtained from a physically unclonable function (PUF) circuit is expected as one of the countermeasures for the counterfeit ICs. However, conventional environment of the PUF circuit requires additional measurement and design cost to collect a large number of the CRPs This paper proposes a novel architecture to collect the CRP during production test by using logic build-in-self test (LBIST) circuit. Neither additional circuit implementation nor additional measurement is required for the CRP collection. In addition, the LBIST-PUF has a tolerance to machine learning (ML) attacks since the response generation of CRP becomes complicated by the LBIST. Through proof-of-concept implementation using field-programmable gate array (FPGA), we demonstrate that the PUF performance can be evaluated by the test pattern generated from the LBIST circuit and the vulnerability for ML attack is also improved. language of the presentation: Japanese | |||