Ultra-Low-Power Neuromorphic Edge Accelerators for Surgical Workflow Intelligence and Rehabilitation Systems
Principal Investigator
Prof. Srinivasa Chakravarthy
Co-Investigator
Prof. Aniruddhan
Objective
- Spandan AI is developing Deep Oscillatory Neural Network (DONN) based analog neuromorphic AI accelerators that deliver ultra-low power, ultra-low latency edge intelligence. These continuous-time chips are uniquely suited for healthcare applications that demand real-time video-based movement tracking, privacy preservation, and offline operation.
Description
- This proposal requests Rs. 1,00,00,000 for 12 months to deliver two concrete outcomes: Two flagship healthcare edge applications and two DONN silicon prototype tape-outs. No clinical trials are planned in Year-1; the focus is application-ready hardware and pre-clinical technical validation, enabling clinical pilots in Year-2.
Impact
- Two fabricated DONN silicon revisions. Two fully functional healthcare edge applications. Video-based movement datasets (non-clinical). Energy/latency benchmarks vs digital AI. Partner-ready demos for hospitals and OEMs. Clear pathway to Year-2 clinical pilots.
Budget in Lakhs
100.00
Duration
1 Year