NASA SBIR 2019-I Solicitation

Proposal Summary

 19-1- H6.22-3264
 Deep Neural Net and Neuromorphic Processors for In-Space Autonomy and Cognition
 Neuromorphic Spacecraft Fault Monitor
SMALL BUSINESS CONCERN (Firm Name, Mail Address, City/State/Zip, Phone)
Exploration Institute, LLC
Metro Office Park, Metro Parque 7, Street 1 Suite 204
Guaynab, PR 00968- 1111
(617) 599-0774

Principal Investigator (Name, E-mail, Mail Address, City/State/Zip, Phone)

Michael Mercury
Metro Office Park,Metro Parque 7,Street 1 Suite 204, GUAYNAB, PR 00968 - 1111
(626) 269-8248

Business Official (Name, E-mail, Mail Address, City/State/Zip, Phone)

Armin Ellis
Metro Office Park,Metro Parque 7, GUAYNABO, PR 00968 - 1111
(617) 599-0774
Estimated Technology Readiness Level (TRL) :
Begin: 4
End: 6
Technical Abstract (Limit 2000 characters, approximately 200 words)


Spacecraft are incredibly complicated, interdependent, somewhat autonomous systems operating in dynamic environments. When parameter limits are exceeded or watchdog timers are not reset, spacecraft automatically enter a “safe” mode where primary functionality is reduced or stopped completely. During safe mode the primary mission is put on hold while teams on the ground examine dozens to hundreds of parameters and compare them to archived historical data and the spacecraft design to determine the root cause and what corrective action to take. This is a difficult and time consuming task.

As human travel away from Earth, light travel time delays increase, increasing the time it takes for ground crews to respond to a safe event. A few astronauts will have a hard time replacing the brain power and experience of a team of experts on the ground. Therefore, a new approach is needed that augments existing capabilities to help the astronauts in key decision moments.

This proposal will develop a fault protection monitor that will better understand current faults to help diagnostic efforts, with the ultimate goal of presenting suggested corrective action and relevant collated information to onboard astronauts or predicting upcoming faults and deciding on corrective action.

This monitor enables a new kind of awareness within the spacecraft. This awareness is trained on past faults so that the sum total of understanding of the spacecraft’s test and development program is kept on board. This knowledge is stored in a low power neuromorphic chip whose architecture is modelled on the neural pathways of a neocortex to enable advanced pattern recognition and comparison to known patterns with minimal power consumption.

Phase I will demonstrate the simplest use of this fault monitor system, which will be to train a Spiked Neural Network (architected so it can work on currently available neuromorphic chips) to detect and report faults, learning from the environment and improving operations.

Potential NASA Applications (Limit 1500 characters, approximately 150 words)

This Neuromorphic Fault Monitor would be of use to all future spacecraft, human spaceflight and robotic. It would be useful in rovers and space suits. It would be useful both at the system level and, separately, within each subsystem, even at each sensor. Relevant NASA programs include: HEOMD Moon to Mars, STMD Robotic Exploration, ARMD Real-Time System-Wide Safety Assurance

Potential Non-NASA Applications (Limit 1500 characters, approximately 150 words)

Consistent and regulated safety across the UAV and autonomous car industries with government regulator supplied neuromorphic fault monitors. Such systems could also be used in any kind of safety monitoring from power plants to assembly lines, to commercial aircraft. The low power nature of this system makes it also applicable to Internet of Things devices, wearables, and medical devices.

Duration: 6

Form Generated on 06/16/2019 23:14:58