RESEARCH


E-TUBE

All-electrical-domain wideband waveguide solution
E-TUBE is a cost-and-power-effective all-electrical-domain wideband waveguide solution. The E-TUBE achieves an unprecedented level of performance interms of bandwidth per carrier frequency, power, and density without requiring a precision manufacturing process unlike conventional optical/waveguide solutions.
*Lab spin-off startup: Point2 Technology (http://point2tech.com/)

more +


fNIRS

Near-infrared Spectroscopy
AI-based NIRS signal processing algorithm development
IC-based fNIRS solution
*Lab spin-off startup: OBELAB (http://www.obelab.com/)
more +


AI Accelerator

  • Energy-efficient Data Movement
  • Dataflow-aware Processing Core
  • Layer dependent Maximum Data Re-use

more +


AI Inspired
Medical Ultrasound Imaging System

  • Breast cancer diagnosis
  • Thyroid cancer diagnosis
  • Myocardial infarction diagnosis
  • Nonalcoholic fatty liver diagnosis
more +


AI Based
Bat-inspired
Ultrasound Sensor

In order for unmanned systems (UMS) to reach full autonomous level that does not require any human intervention, the collected data from sensors in UMS should suffice to manage the entire environmental scenarios. Therefore, UMS commonly employ a combination of sensors including RGB-D cameras, RADARs, LIDARs, and ultrasonic sensors that are complementary to each other.
Both RGB-D camera and LIDAR provide abundant high-resolution visual information, however, the visibility and accuracy can be severely compromised depending on environmental/weather conditions. On the contrary, RADAR and conventional ultrasonic sensor, measuring the time-of-flight of the reflected signal, are relatively less sensitive to operating circumstance but merely provide low-resolution ranging information. Consequently, a clear need exists for an imaging sensor that can precisely visualize 3D space irrespective of environmental conditions.
In this research, a high-resolution ultrasound 3D imaging system emulating the echolocation mechanism of a live bat is presented. (Published in NeurIPS 2019)

more +