네트워크

전기전자

아날로그집적회로시스템 연구실

 

Smart contact lens Controller IC

Related: Cheonhoo Jeon

This presents a controller IC for smart contact lens (SCL). For wireless telemetry, a transmitter supporting dual-mode operations with a single external loop antenna is proposed. In wireless-powered backscattering load-shift keying (LSK) mode, the transmitter is configured to use the antenna as an L for magnetic resonance coupling. The transmitter switches its configuration to be an active LC oscillator circuit in RF-radiated transmission mode. To suppress the average power consumption, a 0.2-percent duty-cycling is applied with a controlled timing schedule while minimizing static power consumption in sleep mode. To reduce the effect of instantaneous discharging, the period of data transmission is shortened by transmitting data with higher rate. The designed IC is fabricated using 0.18 μm CMOS. For in vitro verifications in both modes, a SCL is implemented by integrating the IC, a glucose sensor, a thin-film battery, a loop antenna and a supply-regulating capacitor. The IC achieves a power consumption of 143 nW with a input dynamic range of 89 dB.

 

 

Real-time object detection on Digital System

Related: Byung-Jun Kim

With Neural Network Algorithm, it's difficult to detect objection in real-time for HD video. Using Deformable Parts Model (DPM), Object detection on HD video is possible with efficient energy consumption.

 


Objectives
  • Implementation of real-time object detection for HD video on Digital System
  • Using DPM, designing power-efficient digital system

 


Neuromorphic Chip

Related: Jungho Lee

Widespread applications of artificial intelligence for internet-of-things (IoT) devices with endpoint computing have presented challenges for a breakthrough in the energy-efficient hardware implementation. Neuromorphic computing has been also studied as a promising candidate for energy-efficient hardware implementations in both learning tasks and inference tasks.

 


Objectives
  • Implementation of a neuromorphic system performing an on-chip deep-learning algorithm
  • Ultra-low-power operation of multiplication-and-addition in Analog domain

국가

대한민국

소속기관

포항공과대학교 (학교)

연락처

책임자

심재윤 jysim@postech.ac.kr

소속회원 0