CHANNELS

Aerospace

Automotive

Automated & Autonomous Vehicles

Battery & Electrification Technology

Defense

Drone TV

Electronics

Manufacturing & Prototyping

Materials

Medical

Motion Control

RF & Microwave Technology

Robotics & Automation

Sensors & IoT

Test & Measurement

Aerospace
Automotive
Automated & Autonomous Vehicles
Battery & Electrification Technology
Defense
Drone TV
Electronics
Manufacturing & Prototyping
Materials
Medical
Motion Control
RF & Microwave Technology
Robotics & Automation
Sensors & IoT
Test & Measurement
Aerospace & Defense
Search Results

Neurogrid: Super Fast & Efficent Circuit Board Based on the Human Brain

Stanford University scientists have developed a circuit board modeled on the human brain that is 9,000 times faster than a typical PC and could open up new frontiers in robotics and computing. Neurogrid is a circuit board consisting of 16 custom-designed 'Neurocore' chips. Together these 16 chips can simulate 1 million neurons and billions of synaptic connections. The team designed these chips with power efficiency in mind. Their strategy was to enable certain synapses to share hardware circuits. Neurogrid is about the size of an iPad, and can simulate orders of magnitude more neurons and synapses than other brain mimics on the power it takes to run a tablet computer. The scientists are working on developing prosthetic limbs that would be controlled by a Neurogrid-like chip.

Topics:
Automation Board-Level Electronics Computers Electronic Components Electronics & Computers Implants & Prosthetics Medical Motion Control Robotics Semiconductors & ICs

More From SAE Media Group

    Transcript

    00:00:00 [MUSIC] Stanford University. >> If we compute like the brain, we can get all of the technology that has the power of the brain, and so that allows us to develop a robot that uses much less power for his computation. You can develop much more autonomous robots. Your robots won't always be tethered to some large power supply. Neurogrid is a system that can simulate a million neurons in real time. And that kind of power, which is teraflops

    00:00:32 of computing power, is only available in super computers. And these super computers burn on the order of about a million watts. This is just to simulate a million neurons. Your brain has 100 billion and the brain is using about 20 watts. What we've done is we've taken the same transistors, instead of using them as digital logic, right? We use them actually as analog circuits. And we try to design a circuit, such that a level of voltage corresponds to the level of voltage inside a cell.

    00:01:06 The level of current corresponds to the synaptic input. So we are done with this you've saved a lot of power. >> So this is a Neurogrid board. So you get sixteen chips laid out in a tree network to give you one million neurons. The software is not restricted to be used as a simulation tool. So it's a wide spectrum of audience. Neuroscientists were interested in simulating, understanding the brain or engineers who are interested in using spiking neurons to do interesting applications.

    00:01:37 >> This becomes potentially clinically viable for say, quadriplegics, or people who have lost control over some limb for some reason or another. You can bypass wherever that damage is by taking signals directly from the cortex and then control some kind of robotic limb. The low power aspect, is required for the clinical viability, because it needs to be low power in order to be implantable. Neurogrid and the new morphic technology enables that. >> That's the promise of this kind of technology. That's the kind of inspiration that we get from the brain. >> [FOREIGN]

    00:02:14 >> For more, please visit us at www.stanford.edu.