Aiming to build a future laced with autonomy and human-machine co-existence, Grene Robotics started with new combinations of Nature-Inspired Tech in 2008. The company, through its rigorous research, developed a full spectrum hyper-automation platform, greneOS, which unifies people, machines and systems. Director-Emerging technologies at Grene Robotics, Wing Commander MVN Sai (Retd, IAF), shared some of the ideas, achievements and future plans of Grene Robotics. Here are some excerpts from his conversation with EFY
Q.We would like to know the problem that you’re trying to solve.
A. Good question to start with. In nature, everything is autonomous, right? The whole process is set, it will look after itself, and it heals itself. The solution gets into the framework. The evolution of every single species on Earth has become a part of its growth story. But it is not true for the entire software stack that we have been building over the last 40-50 years. A typical program is not evolutionary in nature. But, like nature, we are inspired to solve the autonomy puzzle.
Q. How exactly do you intend to solve this autonomy challenge?
A. Firstly, technology is an enabler only. It is not an end by itself. All the technologies we speak about, like AI, machine learning, and real-time operating systems are all sub-terrain issues. We are not banking on learning how to use GIS to solve your problem. At Grene Robotics, we bring all the technologies there, at a level below and all the algorithms further below the layer of technology, thus keeping it simple for the user.
We ask not what it takes to automate, but what it takes to make the process autonomous. We don’t want to solve a piece of the puzzle. You want to make the whole thing move together.
Q. How do you derive inspiration from nature?
A. Needless to say, nature has its best autonomy collaborations. For example, let’s take the coordination between bees in building a beehive, or a colony of ants going in a particular direction seeking food. There is a lot of communication and shared intelligence between all the elements. But, if a vehicle knows that there is a pothole, the information is not shared with the next vehicle.
Similarly, in the case of software deployment that we do, there is no shared awareness. We want to apply collaborative AI in a manner that the learning gets faster. You don’t rediscover the same problem, but you figure out a solution using collective knowledge.
To sum up, we’re trying to make everything autonomous. That’s an unsolved challenge. And we are pushing the barriers.
“We ask not what it takes to automate, but what it takes to make the process autonomous.”
Q. Every company has a unique story as to how its idea first came up. So what is the story behind Grene Robotics?
A. While the various verticals have come up over time, the central idea emerged when the founder of the company, Kiran Penumacha was experimenting with autonomous vehicles at Carnegie Mellon (CMU), during his post-graduation. He realized that most of the systems are prescriptive in nature. That means, they will define the problem and leave the choice to the operator. This way, the efficiency of the operator dictates the efficiency of the system as a whole. But autonomy can be applied to the entire stack of human existence, from enterprises to defence, to every other business we do. So, the inspiration is the curiosity to make systems autonomous the way nature can.
Q. Since you are the director of the defence arm of Grene Robotics, we would like to know about the challenges that you faced particularly in this sector.
A. It relates to the problem statement in the defence space. ‘Make in India’ and ‘Aatmanirbhar Bharat’ are hot topics today, and everyone is aware, but in the software space, our defence has been self-reliant for more than 50 years. We typically don’t import any operational software. We make it because we don’t want to give away our rulebook to somebody else to write our story.
So, the story for defence is that humans in the loop are slowing down the process significantly and a battlefield is a very dynamic space. You will have 200-300 influences, in the form of hearing something, some radar illuminating you, etc. in every mission. The issue here is whatever influences you had in your first sortie, have you learned from that for your next one? This is a very complex challenge.
Q. What solution do you propose for this sort of trouble?
A. The collection of data in the safest possible manner is the most important part of the story. Unless you go to the enemy territory, you will not experience the influences of the enemy territory. You can’t simulate them in your own space. But if you come back for the next mission, you should know everything that was previously there and it should become a part of your solution.
So, the challenge for this particular area of work, for example, is not only to improve efficiency but also to continuously learn information that is not otherwise available to you. You can write code for it, and it should get into the process of your decision framework.
Q. How was the first prototype (or MVP) made?
A. It’s not a hardware system per se, it’s an ecosystem of solutions. The first prototype starts with an evolutionary interface between systems. So if you understand sensors and get the data on the sensors into a single framework, then you can do something with the data. For example, an Internet of Things (IoT), or Internet of Military Things is the beginning point. That means the data collection from sensors and systems without humans getting the data to you. Human Interface is not only slow but also inaccurate. On the other hand, a sensor interface is the most efficient and realistic, both temporally and spatially, way of getting the data.
Q. What kind of electronics or technology did you use in your very first prototype?
A. The first thing we did was a huge IoT project which was far ahead of its time, nearly 18 years ago. The project was about what we know today, smart metering, which was not so common a decade and a half ago. Thereafter, we started building an entire solution called an autonomous solution with low-code/no-code for the greneOS. It can be deployed across enterprises in a simple ‘drag-and-drop’ fashion. Over time, the platform has been enabled with RPAs (Robotic Process Automation) and bots. And above the bots, it’s the artificial intelligence engine, which helps it make decisions like a human being.
Q. Is there any niche that you have developed over time, concerning sensors or any other industry-specific niche?
A. We are comfortable with every kind of sensor, whether it’s military or civilian or banking interface, a phone or computer or thermal imager or an aircraft mission computer.
Data collected at the source has to come into the process. So, the entire ability to collect the data is one part of the story. The second part of the story is about how you build processes efficiently without getting into the elaborate project development process.
We aren’t into knowledge acquisition. We go from sensor Interface, collection of data, and process automation, and then show the MVP to the customer within 15 to 45 days. This helps us concentrate more on the segment where we’re solving the customer’s problem rather than understanding his/her requirements.
Q. Tell us about your drone defence system, Indrajaal.
A. Since conventional defences will be overwhelmed in a swarm attack scenario, an AI-enabled autonomous dome with its own ecosystem of sensors and processing is the way forward. Indrajaal is the first such autonomous drone defence dome system in India. It can protect a large area of 1000-2000 sq. km per system against threats such as UAVs, incoming weapons, loitering munitions, and low-RCS targets autonomously.
There is a need to consider an integrated and comprehensive approach against all the UAVs, loiter ammunition swarm drones and low RCS threats with indigenously developed autonomous defence systems.
Q. When it comes to sensor interfaces, how do you ensure cybersecurity and that the data goes to the right channels?
A. Cybersecurity doesn’t come into the picture because we don’t use an open public network. Typically, our deployments are cloud deployments. And the cloud itself comes with a whole stack of security devices. But in case we are doing an on-prem deployment, we build our own cloud, and then all the protocols are followed and the whole apparatus is in place. So, the data, per se, will not be seen even by us.
Q. What is the USP of Grene Robotics?
A. At Grene Robotics, we offer a full spectrum hyper-automation platform. Our key operation system, greneOS is designed to get into the skin of the problem and help organizations with data unification and hyper-automation. We break the process down into a set of deliverables and metrics.
While automation is a smaller process, we make sure that we try to give autonomy to the system and the OS is our benchmark.
“A sensor interface is the most efficient and realistic, both temporally and spatially, way of getting the data.”
Q. Where in India is your R&D Hub majorly located?
A. Our team sits everywhere around the globe. There is no single centre where you can come and see what we do. As we said, we have been a completely virtual company for the last 5-6 years, much before the pandemic forced most of us to move there. We have the frameworks to ensure data security. We believe that creative minds cannot sit in physical offices.
Q. Being a hyper-automation partner to so many MNCs, what has been your commercialization story?
A. The 13-year R&D journey has a very strong foundation. We didn’t acquire a customer to start building our tech on his/her problem statement. After ample time was spent on building the platform, our commercialization started only in 2019.
One part of the commercialization strategy is to stay in the business for long and secondly, to acquire more sectors and industries. And obviously more mission-critical systems. The strategy is to expand the sector and the customer-reach within each sector. We intend to go and acquire markets as a whole, rather than organically try to find solutions.
The other part is definitely the business strategy of ‘land and expand’. We ask for additional segments to be added to it. So, that’s how the expansion takes place and we enter the market.
Q. How long does it take to work on a client’s project, and how do you go about it?
A. Our projects are generally not limited-period projects. When we sign a contract, it goes on for a couple of years, depending on the client’s needs. Once they see the solutions that the platform provides, the clients like to bring every aspect of their business into the system. So, that’s how it goes on for a minimum of 5-7 years. We take care of the platform for the lifecycle of the contract.
Q. What has been your hiring trend?
A. We are an open organization and there is full freedom to work. Every individual is a key stakeholder and has an equal say. We like people who are self-driven and highly committed, whom we hire as per the requirements. The numbers are not very significant, but we have around 75-80 people at this point, with a typical batch of three or four.
Q. Do you hire freshers as well?
A. Of course, we have people coming straight out of college. Even in the second year, for example, one kid just wrote to us one day that he wanted to join. He’s now heading one of the robotics arms, just two years after passing out. We also fund candidates for certain projects And also offer internship opportunities.
Q. What are your expansion plans in India and abroad?
A. We currently operate in the Indian and US markets. We have identified growth markets and we are working with partners in Australia, Israel, Singapore and Europe. So, there are partners that we have onboarded through whom we strategise to expand. However, we are open to working in all markets.
Q. As you said that you have partners in the US and India, are there further opportunities for other firms also to partner with you, in enabling you to expand your customer base?
A. Yeah, we have a business development vertical. So internally, we are adding segments of work. A few years back, we got into the defence space. If there is a worthy partner, why not collaborate? That’s the only way we can expand. We do not white-label anything. We become the unification and automation partner of our clients.
Q. What would you like to tell the budding entrepreneurs in this field?
A. Defence is a slow sector. Everything happens over time. Nothing can be announced before the execution of the actual contracts. And even then, 90 per cent of the time you are not mentioned.
So, if you are there, that’s enough validation. We have consciously spent 10 years building our platform and the commercialization started only in 2019. Along with defence, we have expanded our capabilities in telecom, banking, healthcare, pharma, retail, and hospitality. The only thing that I’d like to tell budding entrepreneurs is that there’ll be initial hiccups but as long as you are solving a relevant problem, there’ll be an initial partner who will come along and trust. So don’t give up if you believe in your concept.