DISPLAYING:

Events happening on (January 23, 2026)

• 23
  Jan

  EE FACULTY CANDIDACY TALK "FROM EEG BIOMARKERS TO SILICON: ULTRA-LOW-POWER INFERENCE PROCESSORS FOR BIOMEDICAL APPLICATIONS."

  Date: January 23, 2026
  Time: 11:00 AM
  Location: EE Reading Room

  Speaker: Dr. Abdul Rehman Aslam
  

  Emotion dysregulation and negative affectivity are among the most challenging behavioral traits associated with Autism Spectrum Disorder (ASD), yet existing diagnostic and assistive solutions rely heavily on software-based analysis that is power-hungry, latency-prone, and unsuitable for continuous real-time deployment. This seminar presents the design and validation of an ultra-low-power on-chip processor for emotion-aware assistance in individuals with ASD, using electroencephalography (EEG) signals as the primary sensing modality. The talk will cover the complete end-to-end hardware–algorithm co-design methodology, starting from large-scale EEG feature extraction and biomarker identification, followed by the development of patient-specific machine learning models optimized for hardware implementation. A custom neural-network–based inference architecture is then mapped onto a dedicated System-on-Chip (SoC), designed and taped-out in CMOS technology, achieving microjoule-level energy consumption per classification while maintaining clinically relevant accuracy. Key contributions include feature-aware hardware optimization, channel-efficient EEG processing, and a deeply pipelined low-power inference engine validated through FPGA prototyping and post-layout analysis. Building upon this foundation, the final part of the seminar will outline a future research vision centered on reconfigurable and scalable inference accelerators for long-term biomedical monitoring and next-generation embedded intelligence. This includes extending the presented architecture toward multi-modal physiological sensing, adaptive on-chip learning, and technology scaling to advanced nodes for further energy reduction. The vision also encompasses close integration of algorithmic robustness, circuit-level innovation, and system-level validation, enabling practical deployment in wearable and edge devices. Through this direction, the research aims to bridge fundamental VLSI design with real-world healthcare and embedded AI applications, while fostering interdisciplinary collaboration and externally funded research programs.

  • All Events
  • Public Events
  • Faculty and Staff Events
  • Departments
  • Electrical Engineering

  Read More