Sharan brought a complementary skill set to the table. “Shashwath is naturally more creative while I am naturally more systematic. Real-world experience has since trained us to be creative and systematic respectively. Over the years, that particular skill combination has allowed us to hold each other up and bring out innovative solutions as a team,” he says.

Sharan had picked up an undergraduate degree in computer science from Chennai-based Meenakshi Sundararajan Engineering College before proceeding to ETH Zürich in Switzerland — famous as the institute where Albert Einstein received his education and later taught — to study computational science. Right after, he joined Shashwath at Lucid.

They began to put their heads together. “We kept talking about how to build hardware and high-performance computing and related things,” Shashwath says.

Sharan enrolled into a Doctor of Philosophy (PhD) programme at IIT Madras. His thesis adviser is Professor Veezhinathan Kamakoti, who leads the efforts of the Reconfigurable Intelligent Systems Engineering (RISE) group at IIT Madras to build the indigenous open-source Shakti processor ecosystem.

“And then the pandemic hit and I hadn’t seen the inside of an office since March 2020,” Shashwath tells me.

Sharan adds: “I had spent the calendar year 2019 shuttling between IIT-M for coursework and Lucid for product development work literally every working day. Then spent all day during the weekends catching up on assignments. I had just completed my comprehensive examinations at IIT-M when the first lockdown hit.”

In an instant, Sharan went from “being a total workaholic to nearly idle”. He says, “It was nice to catch up on some sleep initially. And while there were still deliveries to be made in an official capacity, this was still the uncharted waters of the first wave and workload was low and slow. I then began to talk to Shashwath about the possibility of doing something.”

By the end of the year, Shashwath and Sharan got started on Mindgrove.

When an opportunity came knocking from the Ministry of Electronics and Information Technology to make good use of the Shakti microprocessor, through the Swadeshi Microprocessor Challenge, Shashwath and Sharan decided to have a crack at it.

“We thought of everything from a baby monitor up,” Shashwath says. “Because I had a child, I thought, ‘Okay, a baby monitor is something I would definitely buy’.”

They also thought of building secure communication systems, but soon enough, Shashwath knew which path he had to take. “My entire focus in my career had been imaging — industrial imaging, image processing, signal processing. So I wanted to do something along those lines, do something which I know,” he says.

“I adore learning and I am always up for an engineering challenge. The 2010s was a decade where I explored domains ranging from neuroscience to additive manufacturing while naturally converging on vision and allied systems. I was quite happy to follow along,” Sharan adds.

After zeroing in on imaging, and taking good stock of their capabilities and limitations imposed by the pandemic, the Shashwath-Sharan duo took off on building a ground penetrating radar.

Ground Penetrating Radar

A ground penetrating radar or GPR is a tool primarily used for inspection and mapping in the infrastructure space. Roads, buildings, and bridges can be scanned quickly and cost-effectively using GPR so that the “health” of the structure can be ascertained.

An early look, courtesy the GPR, allows for faster decision making that can improve, among various things, public safety.

Shashwath wasn’t new to GPR. It was a technology he had engaged with in 2011. The project back then had involved scanning concrete via multiple platforms like ultrasound and GPR so that the inside of the concrete could be imaged. It would reveal details such as the presence of a crack and changes in the density of concrete.

Impressed by the technology, he had sought to keep up with GPR and, over time, picked up some of the essential know-how. He realised it was a better alternative to ultrasound — a tool he used predominantly over the years — because of the challenges with procuring and working with the latter.

“A radar doesn’t require so much of that stuff (compared to ultrasound). All it needs is a radio transmitter and a receiver of the correct frequency and a few other components like an antenna,” Shashwath says.

Sharan explains: “In my ETH days, I experienced first-hand how seemingly tiny optimisations in software can add up to huge performance improvements at scale. So when done right, I knew that a good portion of computations done today, for example, in the cloud, can be done on the system right then and there. This is known in the industry today as ‘edge computing’. So the idea was to take the radio signal and write the software to process said signal right there, on the Shakti processor.”

And thus, it came to be — Mindgrove would work with GPR.

They began the groundwork for the technology in 2020, kicked off official operations in March this year, and are now very close to a prototype.

Shashwath and Sharan have done much of the heavy lifting during this period, with useful help coming from project interns — picked from the Master of Technology (MTech) programme at IIT Madras — who are working on components that will be integrated into the GPR.

Even as Mindgrove is getting the GPR ready, they have their eyes set on extended applications. “From GPR, we plan to move to other forms of radar, such as the millimetre-wave radar, which is used in the automotive industry,” Shashwath tells me.

The situational awareness of autonomous vehicles such as that made by Tesla, for instance, interestingly comes from the radar they have on board.

The millimetre-wave radar provides position and velocity with reasonable accuracy. Therefore, Mindgrove’s frequency modulated continuous wave radar, which is still in the works, could be used to detect, say, unmanned aerial vehicles.

They are also planning radar applications using drones, which, Shashwath says, has strategic importance for India. “Right now all of these devices are being imported even though the drones can be made in India. The more we make these kinds of things within India, the more we can reduce the cost and actually turn a profit,” he says.

An important byproduct, of course, is greater self-reliance within the Indian industry.

Mindgrove’s radar uses new electronics that cost less while also being energy-efficient enough to be mounted on a drone so that a larger area can be scanned faster than before.

In addition, the GPR prototype can be extended to synthetic aperture radar (SAR) to serve purposes such as remote sensing. SAR uses a radar that is mounted on a moving platform to create 2D or 3D images of a target.

SAR is, in fact, the basis for a NASA-ISRO Earth-observing space mission that is proposed to be launched by India in early 2023.

The joint satellite mission will use radar to take a good, close look at many Earth processes — earthquakes, volcanoes, landslides, sea ice and glacier movement, among others — over all the land masses, including the polar cryosphere and the Indian Ocean region.

Such impressive applications are a possibility for Mindgrove in the future.