|
Briefing
|
- Automated Vehicle Identification – U.S. Air Force tested IBM’s brain-inspired neuromorphic TrueNorth chip against Nvidia’s Jetson TX-1 computer to identify visually vehicles such as tanks and anti-aircraft systems
- Test Results – While the TrueNorth chip and Jetson TX-1 both performed with 95% accuracy, identifying 10 classes of vehicles including Russian T-72 tanks, armored personnel carriers, and bulldozers, TrueNorth consumed less than one-twentieth the power
- Energy Efficiency Advantage – The TrueNorth chip’s neurons and synapses both store and process data, the secret to its energy efficiency advantage, unlike Nvidia’s Jetson TX-1 computer where calculations are separate from memory
- Other Applications – The TrueNorth chip can enhance machine vision in satellites, high-altitude aircraft, airbases reliant on generators and small drones in places where resources such as electricity and space are limited
|
|
Accelerator
|
|
|
Sector
|
Information Technology, Military and Security
|
|
Organization
|
Air Force Research Laboratory (AFRL), International Business Machines Corp.
|
|
Source
|
|
|
Original Publication Date
|
January 11, 2017
|
