Home Artificial Intelligence Dr. Mehdi Asghari, President & CEO of SiLC Technologies – Interview Series

Dr. Mehdi Asghari, President & CEO of SiLC Technologies – Interview Series

by admin
mm

Mehdi Asghari is currently the President & Chief Executive Officer at SiLC Technologies, Inc. Prior to this, he worked as the CTO & SVP-Research & Development at Kotura, Inc. from 2006 to 2013. He also held positions as Vice President-Silicon Photonics at Mellanox Technologies Ltd. and Vice President-Research & Development at Bookham, Inc. Asghari holds a doctorate degree from the University of Bath, an undergraduate degree from the University of Cambridge, and graduate degrees from St. Andrews Presbyterian College and Heriot-Watt University.

SiLC Technologies is a silicon photonics innovator providing coherent vision and chip-scale FMCW LiDAR solutions that enable machines to see with human-like vision. Leveraging its extensive expertise, the company is advancing the market deployment of coherent 4D imaging solutions across a variety of industries, including mobility, industrial machine vision, AI robotics, augmented reality, and consumer applications.

Dr. Asghari, you have an extensive background in Silicon Photonics and have been involved in multiple startups in this space. Could you share what first sparked your interest in this field?

I went into photonics as I wanted to be in the closest branch of engineering to physics that I could. The idea was to be able to develop products and viable businesses while playing at the front line of science and technology. At that time, around 30 years ago, being in photonics meant that you either did passive devices in glass, or active devices (for light emission, modulation or detection) in III/V materials (compound of multiple elements such as In, P, Ga, As). Both industries were migrating to integration for wafer scale manufacturing. Progress for both was very slow, primarily due to material properties and a lack of well-established fabrication process capabilities and infrastructure.

I was in the III/V camp and came across a small startup called Bookham which was using silicon to make optical devices. This new idea offered the major advantage of being able to use mature silicon wafer fabrication processes to make a highly scalable and cost-effective platform. I felt this could transform the photonics industry and decided to join the company.

With over 25 years of experience and over 50 patents, you’ve had a significant impact on the industry. What do you see as the most transformative developments in Silicon Photonics during your career?

Bookham was the first company ever to try to commercialize silicon photonics, which meant there was no existing infrastructure to use. This included all aspects of the development process, from design to fabrication to test, assembly and packaging. On design, there was no simulation tool that was adapted to the large index steps we were using. On the fab side, we had to develop all the fabrication processes needed, and since there was no fab ready to process wafers for us, we had to build wafer fabs from scratch. On assembly and packaging, there was virtually nothing there.

Today, we take all of these for granted. There are fabs that offer design kits with semi-mature libraries of devices and many of them even offer assembly and packaging. While these remain far from the maturity level offered by the IC industry, life is so much easier today for people who want to do silicon photonics.

SiLC is your third Silicon Photonics startup. What motivated you to launch SiLC, and what challenges did you set out to address when founding the company in 2018?

Throughout my career, I felt that we were always chasing applications that more mature micro-optics technologies could address. Our target applications lacked the level of complexity (e.g. number of functions) to truly justify deployment of such a powerful integration platform and the associated investment level. I also felt that most of these applications were borderline viable in terms of the volume they offered to make a thriving silicon-based business. Our platform was by now mature and did not need much investment, but I still wanted to address these challenges by finding an application that offered both complexity and volume to find a true long-lasting home for this amazing technology.

When you founded SiLC, what was the primary problem you aimed to solve with coherent vision and 4D imaging? How did this evolve into the company’s current focus on machine vision and LiDAR technology?

COVID-19 has shown us how vulnerable our logistics and distribution infrastructure are. At the same time, almost all developed countries have been experiencing a significant drop in working age population (~1% year on year for a couple of decades now) resulting in labor shortages. These are the underlying major trends driving AI and Robotic technologies today, both of which drive enablement of machine autonomy. To achieve this autonomy, the missing technology piece is vision. We need machines to see like we do If we want them to be unchained from the controlled environment of the factories, where they do highly repetitive pre-orchestrated work, to join our society, co-exist with humans and contribute to our economic growth. For this, humanlike vision is critical, to allow them to be efficient and effective at their job, while keeping us safe.

The eye is one of the most complex optical systems that I could imagine making, and if we were to put our product on even a small portion of AI driven robots and mobility devices out there, the volume was certainly going to be huge. This would then achieve both the need for complexity and volume that I was seeking for SiLC to be successful.

SiLC’s mission is to enable machines to see like humans. What inspired this vision, and how do your solutions like the Eyeonic Vision System help bring this to life?

I saw our technology as enabling AI to assume a physical incarnation and get actual physical work done. AI is wonderful, but how do you get it to do your chores or build houses? Vision is critical to our interactions with the physical world and if AI and Robotics technologies wanted to come together to enable true machine autonomy, these machines need a similar capability to see and interact with the world.

Now, there is a major difference between how we humans see the world and how existing machine vision solutions work. The existing 2D and 3D cameras or TOF (Time of Flight) based solutions enable storage of stationary images. These then have to be processed by heavy computing to extract additional information such as movement or motion. This motion information is key to enabling hand-eye coordination and our ability to perform complex, prediction-based tasks. Detection of motion is so critical to us, that evolution has devoted >90% of our eye’s resources to that task. Our technology enables direct detection of motion as well as accurate depth perception, thus enabling machines to see the world as we do, but with much higher levels of precision and range.

Your team has developed the industry’s first fully integrated coherent LiDAR chip. What sets SiLC’s LiDAR technology apart from other solutions on the market, and how do you foresee it disrupting industries like robotics, C-UAS and autonomous vehicles?

SiLC has a unique integration platform that enables it to integrate all the key optical functions needed into a single chip on silicon, while achieving very high-performance levels that are not attainable by competing technologies (>10X better). For the robotics industry, our ability to provide very high-precision depth information in micrometer to millimeter at long distances is critical. We achieve this while remaining eye-safe and independent of ambient lighting, which is unique and critical to enabling widespread use of the technology. For C-UAS applications, we enable multi-kilometer range for early detection while our ability to detect velocity and micro-doppler motion signatures together with polarimetric imaging enables reliable classification and identification. Early detection and classification are critical to keeping our people and critical infrastructure safe while allowing peaceful usage of the technology for commercial applications. For mobility, our technology detects objects hundreds of meters away while using motion to enable prediction-based algorithms for early reactions with immunity to multi-user interference. Here, our integration platform facilitates the ruggedized, robust solution needed by automotive/mobility applications, as well as the cost and volume scaling that is needed for its ubiquitous usage.

FMCW technology plays a pivotal role in your LiDAR systems. Can you explain why Frequency Modulated Continuous Wave (FMCW) technology is critical for the next generation of AI-based machine vision?

FMCW technology enables direct and instantaneous detection of motion on a per pixel basis in the images we create. This is achieved by measuring the frequency shift in a beam of light when it reflects off of moving objects. We generate this light on our chip and hence know its exact frequency. Also, since we have very high-performance optical components on our chip, we are able to measure very small frequency shifts and can measure movements very accurately even for objects far away.  This motion information enables AI to empower machines that have the same level of dexterity and hand-eye coordination as humans. Furthermore, velocity information enables rule-based perception algorithms that can reduce the amount of time and computational resources needed, as well as the associated cost, power dissipation and latency (delay) to perform actions and reactions. Think of this as similar to the hardwired, learning and reaction-based activities we perform like driving, playing sports or shooting ahead of a duck. We can perform these much faster than the electro-chemical processes of conscious thinking would allow if everything had to go through our brain to be processed fully first.

Your collaboration with companies like Dexterity shows a growing integration of SiLC technology in warehouse automation and robotics. How do you see SiLC furthering the adoption of LiDAR in the broader robotics industry?

Yes, we see a growing need for our technology in warehouse automation and industrial robotics. These are the less cost-sensitive, and more performance-driven applications. As we ramp up production and mature our manufacturing and supply chain eco-system, we will be able to offer lower cost solutions to address the higher volume markets, such as commercial and consumer robotics.

You recently announced an investment from Honda. What is the impact of this partnership with Honda and what does it mean for the future of mobility?

Honda’s investment is a major event for SiLC, and it is a very important testament to our technology. A company like Honda does not make investments without understanding the technology and performing in-depth competitive analysis. We see Honda as not just one of the top automotive and truck manufacturers but also as a super gateway for potential deployment of our technology in so many other applications. In addition to motor bikes, Honda makes powersports vehicles, power gardening equipment, small jets, marine engines/equipment and mobility robotics. Honda is the largest manufacturer of mobility products in the world. We believe our technology, guided by Honda and their potential deployment, can enable mobility to reach higher levels of safety and autonomy at a cost and power efficiency that could enable widespread usage.

Looking forward, what is your long-term vision for SiLC Technologies, and how do you plan to continue driving innovation in the field of AI machine vision and automation?

SiLC has only just begun. We are here with a long-term vision to transform the industry. We have spent the better part of the past 6 years creating the technology and knowledge base needed to fuel our future commercial growth. We insisted on dealing with the long pole of integration head-on from day one. All of our products use our integration platform and not components sourced from other players. On top of this, we have added full system simulation capabilities, developed our own analog ICs, and invented highly innovative system architectures. Added together, these capabilities allow us to offer solutions that are highly differentiated and end-to-end optimized. I believe this has given us the foundation necessary to build a highly successful business that will play a dominant role in multiple large markets.

One area where we have focused more attention is how our solutions interface with AI. We are now working to make this simpler and faster such that everyone can use our solutions without the need to develop complex software solutions.

As for driving future innovation, we have a long list of wonderful advancements we would like to make to our technology. I believe that the best way to prioritize implementation of these as we grow is to listen carefully to our customers, and then find the simplest and smartest way to offer them a highly differentiated solution that builds on our technological strengths. It is only when you make clever use of your strengths that you can deliver something truly exceptional.

Thank you for the great interview, readers who wish to learn more should visit SiLC Technologies.

Source Link

Related Posts

Leave a Comment