Embedded Computing Design

Compute in the automobile is (again) moving toward consolidation and away from the discrete nature of multiple processors around the vehicle. Is this a trend that’s going to stick? I guess it depends on who you ask. 

In this case, I asked Robert Moran, the GM & VP for Automotive Processors at NXP Semiconductors. As we add AI and autonomous drive to the software-defined vehicle (SDV), a move like this is pretty much a requirement. Hear it right from the horse’s mouth on this week’s Embedded Executives podcast.

The Cyber Resilience Act (CRA) is a regulation designed to enhance cybersecurity for products mainly sold in Europe. It establishes common cybersecurity standards for hardware and software, requiring manufacturers to build security into their products from design through the products’ complete lifecycle. It’s this latter concept that can be troubling, as sometimes devices stay in use for a long, long time. 

So what’s a manufacturer to do? That’s the concept that was on the table for this week’s Embedded Executives podcast with Eystein Stenberg, founder and CTO of Northern.tech.

While most autonomous vehicle focus is on the automobile, Mozee is going in a slightly different direction, no pun intended. Mozee is focused on autonomous buses, or vehicles that transport large numbers of people. As you might expect, such a system has its challenges, in addition to those typically associated with an automobile. What makes the company’s solution different is that it operates in real-time, without the need for “pre-mapping” the city in which the vehicle travels. It does this through its AI engine. Hear more directly from Shawn Taikratoke, Mozee’s CEO, on this week’s Embedded Executives podcast.

Two technologies seem to be barreling ahead at a rapid pace, and there is likely a convergence coming shortly, if it hasn’t already occurred. I’m referring to space travel and artificial intelligence. But in my limited knowledge, there appears to be a disconnect in this marriage, whereby how could AI make decisions on something it’s never encountered before? 

To boost my limited knowledge, I invited Sakya Dasgupta, the Founder and CEO, EdgeCortix, to be my guest on this week’s Embedded Executives podcast. He has experience in this space (pun intended) and we discussed what’s happening today and what we can expect in the near future.

Designing wearables with integrated AI /ML just got one step easier with Nordic Semiconductor’s acquisition of Neuton.AI. While the two companies had already been working together, you can expect a lot more integration between Nordic’s MCUs and the very low-power AI IP that Neuton.AI has become famous for. 

On this week’s Embedded Executives podcast. I spoke to Øyvind Strom, the EVP of Nordic’s Short-Range Business Unit. He explains why a developer should sit up and take notice, and what the design community can expect from this powerful combination down the road.

The line between a maker and a maker pro is quite fuzzy. In my opinion, the maker wants to build one of something, and likely won’t want to build more than one. The maker pro starts out building one, but knows somewhere down the line that the project could turn into a far larger volume. 

If you fall into that latter group, does that change some of your design decisions, even where you begin? 

Do you start with a Pi, or do you use something that’s more “professional?” 

This is the conversation I had with Okan Saracoglu, the Vice President of Growth at Sixfab. Hear what he had to say on the subject forward, on this week’s Embedded Executives podcast.

If you believe all the hype, it seems that AI is being integrated into nearly every embedded computing platform. It may not be everyone, but the percentage is fairly high and growing. 

So, as a developer, do you know what your starting point should be for your project? 

What Vin and I discovered, thanks to Avijit Sinha, Senior VP of Strategy and Global Business Development at Wind River, is that a whole lot of questions need to be answered before you even get to the starting point. Hear it for yourself on this episode of the DevTalk with Rich and Vin podcast.

ISO 8800 is a standard aimed at functional safety in automotive applications, pertaining to artificial intelligence. It provides the guidelines for integrating AI systems into any safety-related functions of road vehicles. In full disclosure, I wasn’t aware of this standard, and I was frankly surprised I wasn’t. 

To remedy the situation, I invited Jill Britton, the Director of Compliance at Perforce, onto this week’s Embedded Executives podcast. Jill explained what the standard is all about and which parts of it are key for embedded developers to understand.

Here’s a walk down memory lane. The Eclipse Foundation dates back more than 20 years. Mike Milinkovich, the Foundation’s Executive Director, has held this role since its inception. The role of the foundation and the embedded industry in general has evolved more than anyone could have conceived. 

We talk about AI a lot here at Embedded Computing Design, and that’s a huge understatement. What we don’t talk about often is the power required to keep the data centers going, specifically the data centers that form the backbone of all this AI activity. 

When you talk to experts in this space, they’ll tell you that the amount of power needed, specifically the power density, is staggering. One of those experts, Tim Phillips, President, CEO, and Founder of Empower Semiconductor, joined me on this week’s Embedded Executives podcast. He gave some insight into what we need to do to handle the power (and associated heat) issue.

The best part about writing Tear Down articles is that I get to try out the latest/coolest products. The worst part is that I have to take the products apart, and only having a hammer in my toolbox, the products never work again.

Nordic Semiconductor is certainly one of the leaders in the embedded space regarding microcontroller technology. Frankly, I feel like the company is one of the best-kept secrets. When I had the opportunity to speak with their (sort of) new CEO, I jumped at it. 

Hear my discussion with Vegard Wollan, Nordic’s President and CEO, where he explains some of the moves he has made in his tenure, and what can be expected going forward, on this week’s Embedded Executives podcast.

Some “experts” in the embedded space will swear that systems based on open-source software or hardware can never be secure. At the same time, others will say that such a notion is pure nonsense.

To get a better feel for the argument and understand what is accurate and what is not, I spoke to Jan Pleskac, the Co-Founder and CTO of Tropic Square, on this week’s Embedded Executives podcast.

The company is a fabless semiconductor vendor that specializes in secure elements. Jan and I went through the process behind staying secure and the role that open source plays in it.

In a typical SoC, you handle all your processing functionality within the processor core. Makes sense, right? What if you can handle your processing within the SoC’s memory subsystem? If I understand the concept correctly, the time and power to handle those processing functions would be reduced considerably, which is needed in an AI application. So why hasn’t someone done this? 

Someone has, and that would be PIMIC. To learn what this means and how it works, I spoke to Subi Krishnamurthy, the Founder and CEO of PIMIC, on this week’s Embedded Executives podcast.

Functional safety is no longer a “nice to have” or “should have” in many applications, having moved to “must have” in spaces like automotive, industrial automation, medical devices, and aviation. Passing the functional safety testing process can be rigorous, especially if you’re going through it for the first time. But even the best experts take at least six months to get software certified. 

To understand exactly what functional safety is, what the testing process is like, and how and when a developer should get started, I spoke to Bill Lamie, the President and CEO of PX5 RTOS, and someone who’s been through the rigors of functional safety testing on this week’s Embedded Executives podcast.

You likely have heard of the Rust programming language, known for its safety features. But, like me, you may not know its details, and frankly, why you should be considering it for your next design. For those reasons, Rich and Vin invited Jonathan Pallant, a Senior Engineer with Ferrous Systems, to join them on this episode of DevTalk with Rich and Vin.

In programming jargon, the term “real time” is not used in its literal sense, which can be confusing. The literal definition means that when something happens in real time, it happens instantaneously. But in practice, something that occurs very quickly is mistitled as being real-time. Why does the misconception continue to occur, and what problems does it cause? This is the beginning of the discussion I had with Steve Graves, the Founder and CEO of McObject, on this week’s Embedded Executives podcast. Steve has to deal with this issue regularly, so his opinions are quite valid.

When I attended CES in Las Vegas back in January, I came across Morse Micro, a provider of Wi-Fi HaLow ICs. I was unfamiliar with that specification, so I attempted to learn more. Apparently, I wasn’t the only one looking for more insight into this specification, because the company’s tabletop area was jam-packed, and being that I was already late for another appointment, I never hooked up with the Morse team. That changed when I attended Embedded World in Nuremberg last month. I got the full download on the spec, the company, and its latest SoC/transceiver.

If system developers had unlimited power for their designs, they’d likely find a way to take advantage of it. And the same goes for performance. However, neither of those features is available in infinite quantities. And in just about every case, one becomes a tradeoff for the other. 

So, how do you manage that tradeoff, and when you do, what other factors come into play? 

To better handle this topic, I spoke to Praveen Vaidyanathan, the Vice President and General Manager of Micron Technology’s Data Center Business Unit, on this week’s Embedded Executives podcast.

Just about everybody in the engineering space has reverse-engineered a product at some point. It might have been for professional reasons, like seeing how a competitor is attacking a problem or simply seeing how they engineered a product. Or it could have been for personal reasons, because the product stopped working, or simply because they wanted to. 

Note: If you search for “Tear Down” on Embedded Computing Design, you can see the products we’ve reverse-engineered.

Then there are those people, like the folks at emproof, who get paid to reverse-engineer products, particularly on the software side. Tim Blazytko, the company’s Chief Scientist, Head of Engineering, and Co-Founder, joined us on this episode of DevTalk with Rich and Vin to share some of his experiences with reverse engineering.