Hi David,
Thanks for the link.
I found this one which must have us a the leading contender although they do also suggest memristors which may be in Akida 2/3. It is interesting that they say:
"Additionally, this processing architecture shows promise for addressing the power requirements that traditional computing architectures now struggle to meet in space applications".
In other words "We've already had a look at thiis technology, and we like it." :
NASA SBIR 2021 Phase I Solicitation
Solicitation Period:09-Nov-2020 To 08-Jan-2021
H6.22 Deep Neural Net and Neuromorphic Processors for In-Space Autonomy and Cognition
https://sbir.nasa.gov/sbir_search?searchText=neural&searchTextType=any&searchType=topic&sort_solicit_program%5B%5D=SBIR&sort_solicit_program%5B%5D=STTR&sort_solicit_program_year%5B%5D=2021 [NB: use "neural" in the search box]
This subtopic specifically focuses on advances in signal and data processing.
Neuromorphic processing will enable NASA to meet growing demands for applying
artificial intelligence and machine learning algorithms onboard a spacecraft to optimize
and automate operations. This includes enabling cognitive systems to improve mission
communication and data-processing capabilities, enhance computing performance, and
reduce memory requirements. Neuromorphic processors can enable a spacecraft to
sense, adapt, act, and learn from its experiences and from the unknown environment
without necessitating involvement from a mission operations team. Additionally, this
processing architecture shows promise for addressing the power requirements that
traditional computing architectures now struggle to meet in space applications.
Additional areas of interest for research and/or technology development include:
(a) spiking algorithms that learn from the environment and improve operations,
(b) neuromorphic processing approaches to enhance data processing, computing
performance, and memory conservation, and
(c) new brain-inspired chips and breakthroughs in machine understanding/intelligence. Novel memristor approaches that
show promise for space applications are also sought.
This subtopic seeks innovations focusing on low-size, -weight, and -power (-SWaP)
applications suitable to lunar orbital or surface operations, thus enabling efficient
onboard processing at lunar distances. Focusing on SWaP-constrained platforms opens
up the potential for applying neuromorphic processors in spacecraft or robotic control
situations traditionally reserved for power-hungry general-purpose processors. This
technology will allow for increased speed, energy efficiency, and higher performance for
computing in unknown and uncharacterized space environments including the Moon and
Mars. Proposed innovations should justify their SWaP advantages and target metrics
over the comparable relevant state of the art.
Phase I will emphasize research aspects for technical feasibility and show a path toward a Phase II proposal.
Phase I deliverables include concept of operations of the research topic, simulations, and preliminary results. Early
development and delivery of prototype hardware/software is encouraged.
Phase II will emphasize hardware and/or software development with delivery of specific hardware and/or software
products for NASA, targeting demonstration operations on a low-SWaP platform. Phase II deliverables include a
working prototype of the proposed product and/or software, along with documentation and tools necessary for
NASA to use the product and/or modify and use the software. In order to enable mission deployment, proposed
prototypes should include a path, preferably demonstrated, for fault and mission tolerances. Phase II deliverables
should include hardware/software necessary to show how the advances made in the development can be applied
to a CubeSat, SmallSat, and rover flight demonstration.
The current SOA does not address the capabilities required for artificial intelligence and machine learning
applications in the space environment. These applications require significant amounts of multiply and accumulate operations, in addition to a substantial amount of memory to store data and retain intermediate states in a neural network computation. Terrestrially, these operations require general-purpose graphics processing units (GPGPUs), which are capable of teraflops (TFLOPS) each—approximately 3 orders of magnitude above the anticipated capabilities of the HPSC.
Neuromorphic processing offers the potential to bridge this gap through a novel hardware approach. Existing
research in the area shows neuromorphic processors to be up to 1,000 times more energy efficient than GP-GPUs
in artificial intelligence applications. Obviously, the true performance depends on the application, but nevertheless
the architecture has demonstrated characteristics that make it well-adapted to the space environment.
We may also be able to help with the next one on the list:
Solicitation:NASA SBIR 2021 Phase I SolicitationSolicitation Period:09-Nov-2020 To 08-Jan-2021Focus Area:5 Communications and NavigationSubtopic:H9.07 Cognitive Communication
"Learning-based Deep Neural Networks for Cognitive Space Communications"
(20min delay)