BRN Discussion Ongoing

Diogenese

Top 20
"Soaring Otter"... Who comes up with these names? 🤭





Posted Feb 19 2021, 11:11 am

Soaring Otter will be a one step, Closed BAA to advance, evaluate and mature Air Force autonomy capabilities, leveraging the latest advancements in both the fundamental science of autonomy and Machine Learning (ML) and the most modern computing technologies designed to support them.

The Air Force (AF) is increasingly employing the science of autonomy to solve complex problems related to global persistent awareness, resilient information sharing, and rapid decision making. Current autonomy approaches include a growing spectrum of techniques ranging from the well-understood to the truly novel: ML, Artificial Intelligence (AI), many varieties of Neural Networks, Neuromorphic Computing, Data Exploitation and others. Together with the rapid progress of autonomy algorithms and methodologies is the equally rapid progress of hardware and software designed to support the efficient execution of autonomy. These computing solutions bring new capabilities, but also new challenges, including how best to develop applications with them, and how to integrate them into larger systems. The application space for autonomy is rapidly growing, with critical technologies like target identification and recognition, Positioning, Navigation and Timing (PNT) and Unmanned Aircraft Systems (UAS) route planning. Finally, how best to integrate and test these new solutions within reasonable constraints of cost and risk is still not well understood, and there is need for a well-defined progression from lab prototype, through realistic System Integration Lab (SIL) testing, finally through field and flight testing for Technology Readiness Level (TRL) increase.

The scope includes the following seven main topic areas:

Autonomy Development and Testing: Develop novel approaches to solving autonomy problems using the latest techniques in ML, neural networks, AI and other fields. Constantly seek to leverage the newest developments from both government and industry; mature existing approaches toward greater levels of robustness and determine early what is required for the eventual successful transition of these autonomy technologies to the warfighter.

Evaluation of Autonomy Capabilities: Provide neutral 3rd party evaluation of algorithms from Government, Academia and Industry. Provide unbiased analysis of alternatives for algorithms being produced by the Government, Industry and Academia to provide actionable information to AFRL about which algorithms are performing best against objective criteria, as well as determine which solutions are most ready for maturation and integration into systems. Design and perform trade studies to identify best-of-breed solutions and make recommendations to the Government for their application and further maturation.

Novel Computing Approaches: This area will focus on compact computing solutions that push processing to the edge for real or near real time solutions to support the warfighter. Assess the latest emerging computing architectures from government and industry, together with the latest approaches to efficiently developing applications using these technologies.

New Application Spaces: Evaluate emerging Air Force priorities and user requirements, to determine where autonomy can bring the greatest benefit, focusing on Intelligence, Surveillance, and Reconnaissance (ISR).

Open System Architectures for Autonomy: Assess existing and emerging Open System Architectures (OSAs) as fundamental elements of future autonomous systems.

Autonomy Technology Integration and Testing: Plan and execute paths by which new autonomy technologies can be rapidly integrated into larger systems for lab, SIL and field/flight testing.

Maturing System Support: Plan and execute technology transition and system transition activities for operational partners. System deployment support and participation, system integration, testing, and assessment support activities.

Additionally, this BAA will address the issues both individually and collectively. No R&D conducted under this program will be done in isolation, but rather in full consideration of how the new technologies can progress toward full integration with large, complex systems, ready to transition to support of the warfighter.

I'm sure @Yak52 will correct me, but there was a de Havilland Twin Otter propeller aircraft used in regional airlines 40 years or so ago.
 
  • Like
  • Love
  • Fire
Reactions: 5 users

jtardif999

Regular
I tend to agree. All he says is that Edge impulse have matched Brainchips akida to a Prophesee Camera. That does not directly link us with Prophesee in my eyes.
Look as far as I’m aware the Synsense tech is not multimodal - it is not capable of sensor fusion, therefore Akida will be what everyone will eventually want because if you listen to Zach Shelby in that presentation sensor fusion makes Edge Ai more accurate and reliable. I think also PVDM may have previously discounted the Synsense tech as too slow when compared to Akida. AIMO. Akida Ubiquitous. 😎
 
  • Like
  • Fire
  • Love
Reactions: 15 users

VictorG

Member
Your wrong I like my definition better I even have one where the Ogre represents repressed homosexuality which I share with Professors of English Literature. 😂🤣😎

FF



AKIDA BALLISTA
Wham, straight to the upper cut, not even a jab.
 
  • Haha
  • Like
  • Love
Reactions: 12 users
Wham, straight to the upper cut, not even a jab.
With @Diogenese you only get one chance and you better make it a good one or he will come back with a five punch combination and it’s all over red rover. 😎

FF

AKIDA BALLISTA
 
  • Haha
  • Like
Reactions: 18 users

Bravo

If ARM was an arm, BRN would be its biceps💪!
One step closer for "Seeing Machines"to become involved with BrainChip for in cabin monitoring, if they aren't already.

I think I can hear a bottle of champers calling my name. 🍾🍾🍾


NVISO + SEEING MACHINES
NVISO + BRAINCHIP
SEEING MACHINES + HONDA

  • 16 Dec 2019 NVISO and Seeing Machines to bring Human Beviour AI to the Automotive World “Our vision is to help create safer and more personalized driver and passenger experiences by deeply embedding our AI in every car in the world. Working with industry leader Seeing Machines is a great first step to realizing our vision.” said Tim Llewellynn, CEO and Co-founder of NVISO.
  • 19 April 2022 BrainCHip and NVISO Partner on Human Behavious Analytics
  • 10 May 2022 Seeing Machines has won its first Japanese car manufacturer, rumoured to be Honda, in a deal initially worth A$21m (US$14.6m). The cars will go into production in 2025.


 
Last edited:
  • Like
  • Fire
  • Love
Reactions: 35 users

RobjHunt

Regular
I've just subscribed, thank you Zeebot your forum is phenomenal, a totally different league to that other hell hole HCrapper thank you so much. Also a huge thank you to all the great posters here, you guys and gal's are awesome and so generous of your precious time. 🥰
We are all hoping to have wonderful Xmas present this year in the shape of a very positive share price. If Xmas day is your birthday you get to have two wonderful bites at the cheery ;)
 
  • Like
  • Love
Reactions: 8 users

Bravo

If ARM was an arm, BRN would be its biceps💪!
One step closer for "Seeing Machines"to become involved with BrainChip for in cab monitoring, if they aren't already.

I think I can hear a bottle of champers calling my name. 🍾🍾🍾


NVISO + SEEING MACHINES
NVISO + BRAINCHIP
SEEING MACHINES + HONDA

  • 16 Dec 2019 NVISO and Seeing Machines to bring Human Beviour AI to the Automotive World “Our vision is to help create safer and more personalized driver and passenger experiences by deeply embedding our AI in every car in the world. Working with industry leader Seeing Machines is a great first step to realizing our vision.” said Tim Llewellynn, CEO and Co-founder of NVISO.
  • 19 April 2022 BrainCHip and NVISO Partner on Human Behavious Analytics
  • 10 May 2022 Seeing Machines has won its first Japanese car manufacturer, rumoured to be Honda, in a deal initially worth A$21m (US$14.6m). The cars will go into production in 2025.




Bummer... Had to put the cork back inside the bottle because Seeing Machines appear to have teamed up with Ambarella.😔

Bit strange because Seeing Machines is a Canberra-based company. Maybe they can be convinced to support another Aussie company,what d'ya reckon mate?

@Dhm please warm up the typewriter.🤭

 
Last edited:
  • Like
  • Fire
  • Love
Reactions: 7 users
According to PvdM (or maybe Sean?) we have 100 this year I believe ? (same patents in different countries was my impression)
Oh f***
 
  • Like
Reactions: 7 users

TheFunkMachine

seeds have the potential to become trees.
Hi TFM,

Renesas do have their own in-house brand of AI called DRP-AI (Dynamically Reconfigurable Processor -AI) which they have worked on for a decade.

View attachment 8549 The recent Renesas announcement did disclose that they were using Akida in very low power products (battery operated?), but it seems they are set on using their DRP-AI wherever they can, even though Akida would give better power consumption and speed.

That's why I don't see Renesas using Akida with a 64-bit processor any time soon, even though it could be used to advantage as an image data accelerator as an input for the 64-bit von Neumann processor. However, competitive pressures may alter that.
Cheers! Thanks for clarifying.
 
  • Like
  • Love
Reactions: 10 users

M_C

Founding Member
For anyone wondering if Mercedes is going it alone with AKIDA and thinking we have missed out on BMW, I believe QUALCOMM will have us covered................pure speculation






BMW1.PNG






In addition to this -


Old mate Leendert Van Doorn who I posted about leaving Microsoft last week has turned up at...............QUALCOMM..............check out his comments on the subject AND WHO LIKED IT....................I believe this further backs up my theory that QUALCOMM acquired NUVIA as a smokescreen for hiding the fact they are using our ip..........



Capture.PNG
Capture1.PNG
 
  • Like
  • Fire
  • Love
Reactions: 54 users
Bummer... Had to put the cork back inside the bottle because Seeing Machines appear to have teamed up with Ambarella.😔

Bit strange because Seeing Machines is a Canberra-based company. Maybe they can be convinced to support another Aussie company,what d'ya reckon mate?

@Dhm please warm up the typewriter.🤭

They have an accelerator called Occula that is fabricated in 28nm & uses 2.5 watts running at 71 fps for head tracking. The chip can also track limbs at up to 169 fps but that exhausts its capacity.

Nviso driver monitoring apps running AKIDA on 28nm can do head tracking, eye tracking and passenger monitoring at up to 1,000 fps but demonstrates all these functions at 60 fps and if needed to run at 1,000 fps would use less than 1 watt.

In addition the AKIDA neural fabric is not exhausted and can have further apps added for driver health monitoring including blood pressure and other systems.

Tim Llewellyn stated that AKIDA was the only semiconductor powerful enough to accommodate all Nviso’s apps.

My opinion only DYOR
FF

AKIDA BALLISTA
 
  • Like
  • Fire
  • Love
Reactions: 53 users

uiux

Regular
They have an accelerator called Occula that is fabricated in 28nm & uses 2.5 watts running at 71 fps for head tracking. The chip can also track limbs at up to 169 fps but that exhausts its capacity.

Nviso driver monitoring apps running AKIDA on 28nm can do head tracking, eye tracking and passenger monitoring at up to 1,000 fps but demonstrates all these functions at 60 fps and if needed to run at 1,000 fps would use less than 1 watt.

In addition the AKIDA neural fabric is not exhausted and can have further apps added for driver health monitoring including blood pressure and other systems.

Tim Llewellyn stated that AKIDA was the only semiconductor powerful enough to accommodate all Nviso’s apps.

My opinion only DYOR
FF

AKIDA BALLISTA

Hey FF,

I was chatting to Tim @ the NVISO demo they had running in the Rocks. Apparently, the demonstration in the car was only running the emotion recognition network on the Akida chip. It only required 5 nodes to run it on the Akida chip. The other networks/features will still need to be ported over to the Akida device but it was stated it was going to be a very quick process.

So the 1000 FPS is simply the emotion recognition functionality and you can understand why it was able to reach that extreme FPS due to the smallish footprint (5 nodes) on the device.

I will be interested to see the performance and node requirements when the full gamut of apps are operating on the device
 
  • Like
  • Fire
  • Love
Reactions: 59 users
I guess this has been discussed. The Renesas guy on the video suggests Akida will be use on the new Prophesee camera.
Is he talking about this mob:


Sorry don’t have my laptop. I have fallen behind a bit.
Great find but isn't that guy Zac Shelby?
 
  • Like
  • Love
Reactions: 8 users

Dolci

Regular
MC - I think it was 50 to be lodged this year across different countries. Still a good number though - aggressively pursuing patents. When you combine that with the lead we have and the price we are charging it seems inevitable that we'll be ubiquitous.
... if that's the case from 2015 till now, we have lodged so far only eight or ten in the past seven years, so I would say if we see another ten that would be great not fifty, as some are on cloud nine in the Co...:eek:
 
  • Like
  • Thinking
Reactions: 5 users

uiux

Regular
... if that's the case from 2015 till now, we have lodged so far only eight or ten in the past seven years, so I would say if we see another ten that would be great not fifty, as some are on cloud nine in the Co...:eek:

Dolci, I understand the 100+ number to include international variants of our patents

So 1 US patent may end up being 15 international patents etc
 
  • Like
  • Love
  • Fire
Reactions: 47 users
“In order to cool the system, HPE pumps 6,000 gallons of water through Frontier's cabinets every minute using four 350-horsepower pumps.”


USA has the number one super computer in the world with the first true Exascale system called

“The HPE Frontier system can perform over 1 quintillion calculations per second”​


It is apparently more powerful than AKIDA but does not disclose the power requirements so probably uses more watts than AKIDA. 😂🤣

FF

AKIDA BALLISTA
 
  • Like
  • Haha
  • Wow
Reactions: 25 users
Dolci, I understand the 100+ number to include international variants of our patents

So 1 US patent may end up being 15 international patents etc
I suppose it is a question of whether or not to believe Peter van der Made when he said in answer to a question during the AGM that they had 50 patent applications almost ready for filing.

As it was a statement made during the AGM I suppose those who do not believe this to be possible could seek written clarification from the company.

My opinion only DYOR
FF

AKIDA BALLISTA
 
Last edited:
  • Like
  • Love
  • Fire
Reactions: 39 users
D

Deleted member 118

Guest
  • Like
  • Love
  • Fire
Reactions: 13 users

Diogenese

Top 20
... if that's the case from 2015 till now, we have lodged so far only eight or ten in the past seven years, so I would say if we see another ten that would be great not fifty, as some are on cloud nine in the Co...:eek:
From the sound of that, it could be entering into the PCT National Phase for, eg, 5 patents in 10 countries. So the PCT applications would have been filed 30 months earlier.
 
  • Like
  • Love
  • Wow
Reactions: 10 users
Top Bottom