• 전기전자학회논문지
• /
• 제15권1호
• /
• pp.10-14
• /
• 2011

In this paper, we present an efficient hardware architecture of unrolling image mapper of catadioptric omnidirectional imaging systems. The catadioptric omnidirectional imaging systems generate images of 360 degrees of view and need to be transformed into panorama images in rectangular coordinate. In most application, it has to perform the panorama unrolling in real-time and at low-cost, especially for high-resolution images. The proposed hardware architecture adopts a software/hardware cooperative structure and employs several optimization schemes using look-up-table(LUT) of coordinate conversion. To avoid the on-line division operation caused by the coordinate transformation algorithm, the proposed architecture has the LUT which has pre-computed division factors. And then, the amount of memory used by the LUT is reduced to 1/4 by using symmetrical characteristic compared with the conventional architecture. Experimental results show that the proposed hardware architecture achieves an effective real-time performance and lower implementation cost, and it can be applied to other kinds of catadioptric omnidirectional imaging systems.

• Journal of Semiconductor Engineering
• /
• 제2권2호
• /
• pp.130-135
• /
• 2021

Recently, deep neural networks (DNNs) are actively used for action control so that an autonomous system, such as the robot, can perform human-like behaviors and operations. Unlike recognition tasks, the real-time operation is essential in action control, and it is too slow to use remote learning on a server communicating through a network. New learning techniques, such as reinforcement learning (RL), are needed to determine and select the correct robot behavior locally. In this paper, we propose an energy-efficient DNN processor with a LUT-based processing engine and near-zero skipper. A CNN-based facial emotion recognition and an RNN-based emotional dialogue generation model is integrated for natural HRI system and tested with the proposed processor. It supports 1b to 16b variable weight bit precision with and 57.6% and 28.5% lower energy consumption than conventional MAC arithmetic units for 1b and 16b weight precision. Also, the near-zero skipper reduces 36% of MAC operation and consumes 28% lower energy consumption for facial emotion recognition tasks. Implemented in 65nm CMOS process, the proposed processor occupies 1784×1784 um2 areas and dissipates 0.28 mW and 34.4 mW at 1fps and 30fps facial emotion recognition tasks.

• 한국정보통신학회논문지
• /
• 제26권7호
• /
• pp.1071-1077
• /
• 2022

타원곡선 암호 (elliptic curve cryptography; ECC) 기반의 공개키 기반구조 구현에 사용될 수 있는 보안 SoC(system-on-chip)를 설계하였다. 보안 SoC는 타원곡선 디지털 서명 알고리듬 (elliptic curve digital signature algorithm; ECDSA)용 하드웨어 가속기가 AXI4-Lite 버스를 통해 Cortex-A53 CPU와 인터페이스된 구조를 갖는다. ECDSA 하드웨어 가속기는 고성능 ECC 프로세서, SHA3 (secure hash algorithm 3) 해시 코어, 난수 생성기, 모듈러 곱셈기, BRAM (block random access memory), 그리고 제어 FSM (finite state machine)으로 구성되며, 최소의 CPU 제어로 ECDSA 서명 생성과 서명 검증을 고성능으로 연산할 수 있도록 설계되었다. 보안 SoC를 Zynq UltraScale+ MPSoC 디바이스에 구현하여 하드웨어-소프트웨어 통합 검증을 하였으며, 150 MHz 클록 주파수로 동작하여 초당 약 1,000번의 ECDSA 서명 생성 또는 서명 검증 연산 성능을 갖는 것으로 평가되었다. ECDSA 하드웨어 가속기는 74,630개의 LUT (look-up table)와 23,356개의 플립플롭, 32kb BRAM 그리고 36개의 DSP (digital signal processing) 블록의 하드웨어 자원이 사용되었다.