Cortical Processor
Abstract
The Cortical Processor program developed algorithms and hardware that can better handle the increasingly large and diverse sensor data streams used by battlefield systems. By leveraging advances in machine learning, the program yielded systems with the flexibility to understand and adapt to new contexts and new types of sensed data (e.g. new radio frequency or infrared signals). Current sensor platforms, conversely, are pre-programmed only to interpret specific data types and require a laborious coding effort to accommodate new types of data or contexts. Cortical Processor developed hardware implementations that gracefully handle multiple data streams and limit the programming burden required for sensing and interpreting a complex scenario. The program was enabled by bio-inspired algorithms that benefit from research into biological learning and data processing. Cortical Processor's applied research component investigated silicon circuit designs that are most suitable for high-performance, low-power, real-time sensing and data processing.
Document Details
- Document Type
- Accomplishment
- Publication Date
- Oct 01, 2019
- Source ID
- ca43838fba72aacb84f519589fb933d7