BRN Discussion Ongoing

[rgupta]

Primax have been busy​

​

Primax announced reference design​

partnership with Qualcomm in IoT gateway platform​

Staff reporter, DIGITIMES Asia, TaipeiThursday 30 March 20230

Credit: DIGITIMES

Following Qualcomm's Embedded World 2023 announcement of new IoT processors designed to support four major operating systems, Primax is announcing a reference design partnership for a QCS6490 based IoT gateway platform specifically designed for Fleet and Industrial IoT applications. High compute capabilities and a strong feature set make the QCS6490 a versatile solution for next generation edge processing devices such as cameras, gateways, industrial automation equipment (IPC, PLC), autonomous mobile robotics, and more.
The Primax developed QCS6490 reference platform will support multiple wireless interfaces including Wi-Fi 6, BLE 5.2, Thread, and LoRa, and allow for integrating up to eight cameras, connected via AHD, PoE, USB(UVC), or Wi-Fi. The platform is capable of supporting multiple wired or wireless sensors and will offer integration with the vehicle's on-board diagnostics system, and 5G or LTE network connectivity. Primax's own design not only encompasses dual-mic echo cancellation and noise suppression, ELD compliance, but also precision camera image via spectra DSP pipeline tuning.
"As a global leader in edge computing and vision technologies, Primax Electronics is committed to serving our clients with the most advanced technologies for machine vision and Industrial IoT applications. We are proud to continue our close collaboration with Qualcomm Technologies to help widen support for the management of physical assets and further enable our clients to bring cutting-edge products to market. Our QCS6490 reference design will not only accelerate product time to market, but also empower developers to drive innovations in AI and edge computing." said Jason Hsu, Vice President of Primax Connected Mobility Business Unit.
"Qualcomm Technologies is proud to power the Primax reference platform and to advance Industrial IoT application use," stated Dev Singh, vice president, business development, and head of building, enterprise and industrial automation, Qualcomm Technologies, Inc. "Collaborations between Qualcomm Technologies and Primax Electronics continue to accelerate the future of fleet management with state-of-the-art technology behind it."
Founded in 1984, Primax is a global leader in Vision, Audio, and Human Machine Interface technologies. Headquartered in Taipei with operations across Asia, North America, and Europe, Primax clients include some of the world's most respected consumer, industrial, and automotive brands.

Another link where brn is in question.
MB and Qualcomm
Prophesee and Qualcomm
Primax and qualcomm
So either Qualcomm is an IP partner or have a similarly technology to us
Fingers crossed but in either circumstances 47 cents a share is too cheap.
Dyor

 

[Diogenese]

Primax Electronics Ltd (Taiwan) have just released a smart door lock

Ambarella must be using Akida IP to achieve the functionality described below? It’s got BrainChip written all over it

View attachment 33245
View attachment 33246


View attachment 33244


View attachment 33250
View attachment 33249
View attachment 33251


View attachment 33252
View attachment 33253

Ambarella Inc on LinkedIn: #iscwest #aienvisioned #ambarella #primaxelectronics #iscwest #security…

#ISCWest - Primax Electronics Ltd. announced a new high-precision 3D artificial intelligence (AI) facial recognition smart door lock based on our low-power…

www.linkedin.com

https://www.primax.com.tw/en/press-center/company-news?view=article&id=252:primax-introduces-smart-door-lock-based-on-ambarella-s-low-power-edge-ai-vision-soc&catid=15

Hi tls,

Ambarella have their own CNN doorbell SoC:

US11563927B2 Mounting calibration of structured light projector in mono camera stereo system

[0089] The processor/SoC 802 may be configured to execute computer readable code and/or process information. In various embodiments, the computer readable code may be stored within the processor/SoC 802 (e.g., microcode, etc.) and/or in the memory 806 . In an example, the processor/SoC 802 may be configured to execute one or more artificial neural network models (e.g., facial recognition CNN, object detection CNN, object classification CNN, etc.) stored in the memory 806 . In an example, the memory 806 may store one or more directed acyclic graphs (DAGs) and one or more sets of weights defining the one or more artificial neural network models. The processor/SoC 802 may be configured to receive input from and/or present output to the memory 806 . The processor/SoC 802 may be configured to present and/or receive other signals (not shown). The number and/or types of inputs and/or outputs of the processor/SoC 802 may be varied according to the design criteria of a particular implementation. The processor/SoC 802 may be configured for low power (e.g., battery) operation.


[0112] One of the hardware modules 809 a - 809 n (e.g., 809 b ) may implement a convolutional neural network (CNN) module. The CNN module 809 b may be configured to perform the computer vision operations on the video frames. The CNN module 809 b may be configured to implement recognition of the objects and/or events through multiple layers of feature detection. The CNN module 809 b may be configured to calculate descriptors based on the feature detection performed. The descriptors may enable the processor 802 to determine a likelihood that pixels of the video frames correspond to particular objects (e.g., the people, pets, items, text, etc.).

[0113] The CNN module 809 b may be configured to implement convolutional neural network capabilities. The CNN module 809 b may be configured to implement computer vision using deep learning techniques. The CNN module 809 b may be configured to implement pattern and/or image recognition using a training process through multiple layers of feature-detection. The CNN module 809 b may be configured to conduct inferences against a machine learning model.

 

[Diogenese]

Valeo announces 2 new major contracts for its third-generation LiDAR​

Valeo, the global leader in ADAS, is proud to announce that its SCALA 3 LiDAR has been chosen by a leading Asian manufacturer and a leading American robotaxi company. Valeo has now registered orders worth more than 1 billion euros for SCALA 3.



Valeo SCALA 1 and 2 made conditional autonomous driving in traffic jams a reality. SCALA 3 dramatically increases the domain of operation – and therefore the customer value – of private vehicles thanks to a wider scope of use and higher speed support. Valeo is also positioned as a leader in robotaxis equipment as SCALA 3 enables our clients to make robotaxis a reality at scale, using a proven, best-in-class technology in mass production.
Valeo was the first, and remains the leading company in the production of automotive LiDAR sensors and perception software on an industrial scale. Our clients recognize our ability to deliver on our ambitions.
Marc Vrecko, President of Valeo’s Comfort and Driving Assistance Business Group said: “With over 10 years of experience in development and 5 years of production, Valeo’s LiDAR technology is proven to be reliable and protected by more than 575 patents filed. These new orders show that our technology meets the expectations of different markets, being individual cars or robotaxis. But it also confirms our world leader position in the domain with clients in Europe, Asia and North America. We are excited to be working with new partners to bring safer and more automated mobility around the world.“
Valeo SCALA 3 LiDAR is the perception centerpiece of the next generation of highly automated systems. It will expand the domain of operation of conditionally automated vehicles (level 3) to wider areas and higher speed (up to 130 kph), and make robotaxi (level 4) deployment at scale possible.
Thanks to its laser-based system, Valeo SCALA 3 generates a 3D image of the vehicle’s surroundings delivering a point cloud of a yet unparalleled resolution for an automotive system. Boasting more than 12 million pixels per second (x16 compared with the SCALA 2) and a detection range of 200 meters for low reflectivity objects and 300 meters for high reflectivity objects, it sees almost everything, especially objects that are invisible to the human eye.
In addition to its hardware capabilities, SCALA 3 comes with a suite of software modules, including perception and artificial intelligence based algorithms. Its LiDAR functions guarantee unparalleled safety and reliability through blockage detection, rain & spray detection, online calibration or misalignment detection. It can identify objects – such as a tire – left on an unlit black asphalt road more than 150 meters ahead that neither the driver, cameras nor radars can reliably detect. It finally identifies, classifies and tracks objects, giving the car fusion system the full 3D mapping around the car, and finally making autonomous driving possible. These software modules can easily be embedded on major SoC platforms and run on dedicated ECUs or domain controllers.
The ADAS market is expected to grow by 17% per year to reach 60 billion euros in 2030, driven by enhancements in safety and comfort. Valeo develops software and the most complete suite of sensors (ultrasonic sensors, cameras, radars, LiDARS) and computing units (domain controllers) that enable the vehicle to make the right decisions.

Valeo announces 2 new major contracts for its third-generation LiDAR

Valeo, the global leader in ADAS, is proud to announce that its SCALA™ 3 LiDAR has been chosen by a leading Asian manufacturer and a leading American robotaxi company. Valeo has now registered orders worth more than 1 billion euros for SCALA™ 3.

www.valeo.com

Holy cow!

1 billion euros in orders for Scala 3 ...

... and BrainChip is in a Joint Development with Valeo!

https://smallcaps.com.au/brainchip-joint-development-agreement-akida-neuromorphic-chip-valeo/

BrainChip signs joint development agreement for Akida neuromorphic chip with Valeo​

By George Tchetvertakov - June 9, 2020

In a JD, it is likely that there would be no licence, just a share of receipts based on relative contribution.

LdN said we had a sweet spot for LiDaR, sort of like Prophesee, nViso, SiFive ...

It's time one filled one's boots ...

 

[Diogenese]

Holy cow!

1 billion euros in orders for Scala 3 ...

... and BrainChip is in a Joint Development with Valeo!

https://smallcaps.com.au/brainchip-joint-development-agreement-akida-neuromorphic-chip-valeo/

BrainChip signs joint development agreement for Akida neuromorphic chip with Valeo​

By George Tchetvertakov - June 9, 2020

In a JD, it is likely that there would be no licence, just a share of receipts based on relative contribution.

LdN said we had a sweet spot for LiDaR, sort of like Prophesee, nViso, SiFive ...

It's time one filled one's boots ...

Obviously Valeo has to make the equipment and deliver it to the customers before we would see the money.

Also, we do not have any official confirmation that Scala 3 has Akida inside, but ...

"Boasting more than 12 million pixels per second (x16 compared with the SCALA 2)" - Valeo will need a bigger NN processor.

 

[Tothemoon24]

 

[DingoBorat]

Press a button for engine sound and even sporty engine sound?

Bit of a joke really..
What's the mental age, of the people they want to sell these to ..

 

[Diogenese]

View attachment 33271

Finally there's an advantage to having size 12s.

 

[Tothemoon24]

Finally there's an advantage to having size 12s.

That’s what she said

 

[Diogenese]

Obviously Valeo has to make the equipment and deliver it to the customers before we would see the money.

Also, we do not have any official confirmation that Scala 3 has Akida inside, but ...

"Boasting more than 12 million pixels per second (x16 compared with the SCALA 2)" - Valeo will need a bigger NN processor.

On the up side, the LiDaRs will be delivered before the cars hit the market ...

 

[Frangipani]

On the up side, the LiDaRs will be delivered before the cars hit the market ...

… and hopefully those cars won‘t hit anything except the market, otherwise BRN‘s SP will be hitting even lower lows…

Edit: (provided Brainchip is collaborating with Valeo on its SCALA 3 LiDAR, that is)

 

[Kachoo]

On the up side, the LiDaRs will be delivered before the cars hit the market ...

What's the likely hood that Akida is in the Scala 3.

Are there any patent changes that would in someway incorporate Akida applied by Valeo.

I do remember when Louis or ould CEO said Akida well I believe he used the word Brainchip had a sweet spot for Lidar.

 

[stockduck]

This is a quite interesting read....but I don`t know if it is related to Akida technolgy:

Adopting neural and automated supply chain for intelligent operations

Read how digital transformation of supply chain and modern technologies will lead to the creation of adaptable and purpose-driven neural manufacturing firms.

www.tcs.com

"...The neural self-acting stage is the pinnacle of capabilities. It supports real-time data collection, alerting, monitoring and actioning, multi-level personalization, seamless and secured data sharing, cognitive abilities. Further, this stage fully leverages AI, ML, mobile, and cloud-based technologies.
..."

".....
THE NEURAL FUTURE HAS JUST STARTED

Currently, technological enablers such as AI, ML, self-learning algorithms, data science, and internet of things (IoT) are in various stages of development and implementation maturity. Manufacturing supply chains are still fully connected to one another. Thus, creating a neural self-acting supply chain is a work in progress, which is taking place at a rapid pace.

However, in the not-so-distant future, the distinct elements of a supply chain will align and manifest themselves as an army of AI bots, each specializing in a specific functional aspect with collaborative capabilities.

Bots from procurement will collaborate with those from forecasting, planning, inventory, manufacturing, transportation, logistics, and more. They will gain knowledge about an organization’s supply chain by drawing upon existing data models, structured and unstructured inputs, historical data, and will reinforce that learning by using self-learning algorithms. This will create a true self-acting neural supply chain that is connected, intelligent, collaborative, and personalized.
..."

TCS launches Autoscape™ solution suite

Tata Consultancy Services (TCS) has launched the Autoscape™ solution suite to enable connected and autonomous vehicle ecosystem players,...

www.telematicswire.net

 

[dippY22]

Attached find a short video on a robotic product that smells to me like it could incorporate Akida tech. Maybe not. I am trying to get the poop on it's specs but alas the website seems to be down. So I'm left holding an empty bag which stinks. However, if you are a dog owner and turn your nose up at following Rover around your yard, this product may be for you.

 

[Boab]

What's the likely hood that Akida is in the Scala 3.

Are there any patent changes that would in someway incorporate Akida applied by Valeo.

I do remember when Louis or ould CEO said Akida well I believe he used the word Brainchip had a sweet spot for Lidar.

Fingers crossed but.....
From the Valeo website referring to Scala3
Valeo designs and manufactures the entire system, from the hardware to the software and the associated artificial intelligence, the “brain” that collates all the information and enables the vehicle to make the right decision instantly. The software automatically adapts to the environment and improves its performance through progressive over-the-air updates.

 

[Diogenese]

On the up side, the LiDaRs will be delivered before the cars hit the market ...

Scala 3 is scheduled to hit the market in 2024:

Valeo SCALA™ LiDAR

Discover Valeo’s third generation LiDAR sensors for autonomous vehicles: detects objects around the car, adapts to light conditions…

www.valeo.com

“This third- generation LiDAR from Valeo is a major technological step forward for autonomous vehicles. Valeo is currently the only player on the market mass producing an automotive LiDAR system. This advance strengthens Valeo’s technological and industrial leadership position in that market. The primary objective of this system is to save lives on the road,” says Geoffrey Bouquot,
...
Valeo’s third-generation laser LiDAR technology, which is scheduled to hit the market in 2024, will take autonomous driving even further, making it possible to delegate driving to the vehicle in many situations, including at speeds of up to 130 km/h on the highway. Even at high speeds on the highway, autonomous vehicles equipped with this system are able to manage emergency situation autonomously."

 

[Diogenese]

Attached find a short video on a robotic product that smells to me like it could incorporate Akida tech. Maybe not. I am trying to get the poop on it's specs but alas the website seems to be down. So I'm left holding an empty bag which stinks. However, if you are a dog owner and turn your nose up at following Rover around your yard, this product may be for you.

Back in the last millennium, someone invented a superdooper pooper scooper.

 

[Boab]

Technology and hardware for neuromorphic computing​

This was a European Commission project in 2020.
There were 19 participants but one notably withdrew. VALEO
Was it because it was made aware of the secret sauce?? June 9 2020 JD with Brainchip

Technology and hardware for neuromorphic computing | TEMPO Project | Fact Sheet | H2020 | CORDIS | European Commission

Massive adoption of computing in all aspects of human activity has led to unprecedented growth in the amount of data generated. Machine learning has been employed to classify and infer patterns from this abundance of raw data, at various levels of abstraction. Among the...

cordis.europa.eu

 

[Kachoo]

Fingers crossed but.....
From the Valeo website referring to Scala3
Valeo designs and manufactures the entire system, from the hardware to the software and the associated artificial intelligence, the “brain” that collates all the information and enables the vehicle to make the right decision instantly. The software automatically adapts to the environment and improves its performance through progressive over-the-air updates.

Well regardless this shows how big the markets are for Brainchip. I went on the BRN website which partner pops up first Valeo.

Now maybe they have not finished building the system and they are not in Scala 3 maybe they are we don't know. We now that this partnershipnwas announce but Valeo has been very secretive Yet they are on the BRN site.
So they have to be working on something but that's what we don't know.

 

[Diogenese]

What's the likely hood that Akida is in the Scala 3.

Are there any patent changes that would in someway incorporate Akida applied by Valeo.

I do remember when Louis or ould CEO said Akida well I believe he used the word Brainchip had a sweet spot for Lidar.

This wording suggests that Valeo was new to NNs possibly as late as December 2019:

WO2020127151A1 METHOD FOR IMPROVED OBJECT DETECTION

"A powerful possibility to process 2-dimensional array of data points is to use a neural network. State of the Art applications of neural networks for image processing are typically based on deep neural networks. Usage of such network types has shown promising results at an affordable price. Neural networks comprise an input and an output layer, as well as one or multiple hidden layers. The hidden layers of a deep neural network typically consist of convolutional layers, pooling layers, fully connected layers and normalization layers. The convolutional layers apply a convolution operation to the input, passing the result to the next layer. The convolution emulates the response of an individual neuron to visual stimuli."

Similarly, the only reference to NNs in the claims is in claim 9, but there is no defining description of NNs at a circuit layout level in the specification.

"9. Method according to any preceding claim, characterized in that
the step of identifying one or more troublemakers (24) in the 2-dimensional array of data points (26)
comprises identifying one or more troublemakers (24) in the 2-dimensional array of data points (26) by an image recognition algorithm and/or by a neural network."

So my considered guess is that until 2020, Valeo were not particularly familiar with NNs, and were even considering a software algorithm as a viable alternative.

The fact that Valeo entered into a Joint Development with BrainChip in mid-2020 strongly suggests to me that they are using Akida. I don't recall whether or not there was an ASX announcement of the JD, but, as @Fact Finder reminds us, if there were an announcement, then there would be an obligation to notify ASX if the JD was terminated.

 

[Evermont]

This wording suggests that Valeo was new to NNs possibly as late as December 2019:

WO2020127151A1 METHOD FOR IMPROVED OBJECT DETECTION

"A powerful possibility to process 2-dimensional array of data points is to use a neural network. State of the Art applications of neural networks for image processing are typically based on deep neural networks. Usage of such network types has shown promising results at an affordable price. Neural networks comprise an input and an output layer, as well as one or multiple hidden layers. The hidden layers of a deep neural network typically consist of convolutional layers, pooling layers, fully connected layers and normalization layers. The convolutional layers apply a convolution operation to the input, passing the result to the next layer. The convolution emulates the response of an individual neuron to visual stimuli."

Similarly, the only reference to NNs in the claims is in claim 9, but there is no defining description of NNs at a circuit layout level in the specification.

"9. Method according to any preceding claim, characterized in that
the step of identifying one or more troublemakers (24) in the 2-dimensional array of data points (26)
comprises identifying one or more troublemakers (24) in the 2-dimensional array of data points (26) by an image recognition algorithm and/or by a neural network."

So my considered guess is that until 2020, Valeo were not particularly familiar with NNs, and were even considering a software algorithm as a viable alternative.

The fact that Valeo entered into a Joint Development with BrainChip in mid-2020 strongly suggests to me that they are using Akida. I don't recall whether or not there was an ASX announcement of the JD, but, as @Fact Finder reminds us, if there were an announcement, then there would be an obligation to notify ASX if the JD was terminated.

08 June 2020 @Diogenese