From having a hundred horses lug a few kilograms of load to a single autonomous bot carrying a tonne of payload, humans have surely come a long way. Fortunately, this isn’t the end as the road ahead is even longer. Ati Motors’ Founder and CEO, Saurabh Chandra, shares his wisdom with Yashasvini Razdan from Electronics For You on the future of bots in the material movement automation industry
Q. Could you tell us about Ati Motors’ journey?
A. We started our journey six years ago with the dream of making autonomous robots for material movement, inspired by self-driving cars, and not by traditional robotics.
Since 2021, we have been able to bring out a product in the market, which matches the industry requirements, and has the right specifications, price point, and functionality, allowing us now to gain traction with many good customers.
Q. What solutions do you currently provide?
A. We have two products right now—the Sherpa Tug and Sherpa RollerTop. One moves trolleys autonomously inside factories, while the other moves plastic bins in factories.
Q. Could you explain your technology to our engineering audiences?
A. To put that simply, an autonomous vehicle needs to ask three questions in a loop—where is it, what is around it, and then what will its next step be that will take it towards its goal? As long as you can answer these questions successfully in a tight loop, you will get to your location. That at a very high level is what we do. We need to ask these questions every 100 milliseconds and we do that. Our product is ISO 3691-4 certified. That is the best safety standard for such vehicles available. Safety is very important, especially in automotive companies
Q. Why do we need this technology?
A. Generally, some material movement in factories is a repetitive, boring, hazardous, and backbreaking activity. The fundamental thing is that human beings are not supposed to do this repetitive activity which involves high fatigue. We cannot do it at the right accuracy level as people will get bored. Also, the number of people required to do these things with time is not increasing. This becomes the genesis of the demand for such things. The other reason is the volume of production. It is not possible to get the right volume and quality of production, without the correct automation. One cannot achieve this by just putting more people on the process. After some time, you have to automate many of these things to get the required throughput and volume. We are a part of the overall journey of automation, which is a natural journey for any industry as it matures, scales up and gains volume. This becomes the core need for the kind of product that we are making.
Q. We see bots acting as delivery agents in the US in universities. Will we ever see that in India and how soon can that be?
Frankly, bots doing delivery is very feasible. The challenge is not the technology but good footpaths. In India, unfortunately, we don’t have good footpaths. When India gets good footpaths, this technology won’t be too complicated to be ready for deployment. I would say the limiting part is not the bot technology but the footpath technology!
Q. What is the need for such systems in developing countries such as India, where labour is easily available?
A. That is a good question. It is a misconception that labour is very easily available in India. You talk to any business and there’s always a shortage of skilled manpower. Also, the Indian average ages look very good on paper. People will often tell you that the average age in India is just 26 years or 27 years, but the average age in Tamil Nadu is 35 years. The average age in Karnataka is also very similar now. These are the more industrialised areas of our country. That is one part. The other part is that youngsters who are coming out of college, and for a good reason, aspire for better jobs. Today’s youngsters don’t want to do a mind-numbing job of just moving material from one part to another where there is no application of the brain. This is not really very dignified work for human beings and over time people should not be doing this kind of work. So it’s not a question of ‘whether’ people are available or not to do work. It’s a question of ‘should’ people be doing this kind of work or not.
Q. You’re talking about skilled labour. What about unskilled labour? We do have plenty of that in India.
A. Yeah, that is true, but in a factory, you have to be skilled. You can’t operate in a modern factory unless you have some skills. You have to understand discipline. When we say skilled and unskilled, it’s not just knowing how to operate complex machinery and tools. It is accompanied by the discipline of working in a factory. If you have to do something every 15 minutes it cannot be 16 or 14. These kinds of work habits and discipline come as a process of becoming skilled. Unskilled people don’t have these kinds of work delivery habits, which are mandatory things today for modern manufacturing. What we need to do is increase our ITI training output from 2.5 million per annum to 10 million per annum on a mission mode. The more skilled people we have in the country across areas, the wealthier we are as a nation.
Q. Amazon and Walmart have started drone deliveries in certain parts of the US, which are supervised by skilled employees. Earlier deliveries were conducted by unskilled labour. In India, these transportation jobs are done by people whose livelihood depends on it. Don’t you think this is eating into their job market?
A. For the US, the answer is even clearer. They just don’t have enough people. For India, we had people doing the transportation job earlier. That’s the evolution of the human journey. People would keep moving towards higher-value addition jobs, and we’d need to keep automating our lower-value addition jobs. The economy yields better gains by adding more productivity per person. The only magic humankind knows is technology. It has happened before as well—somebody invented the wheel; whatever we could do manually in a field, we could do better with a bullock cart, and even better with the tractor. Horses were the technology for transportation, and there was so much technology which went into the horse—somebody had to invent the saddle, the spur, and the horseshoe—all of which were connected with the horse to really make it into a transportation machine. There has been the automotive revolution and we will at some point see the autonomy revolution as well.
Automation is not always about labour-saving. Think about welding robots. Today, almost all welding for making the car is done by robotic arms. Initially, some people did it, but you can’t manufacture many cars if you weld manually. It’s beyond human capability to weld with that accuracy repeatedly, at that speed. So, the only option is to have a robotic arm which does it. One could argue why there is a conveyor belt when somebody could lift the packet and run. The person will never be able to meet the speed of the conveyor moving something. Similarly, sometimes a robot is required because conveyors are not practical or you have large material that humans cannot carry. You will not get the repeatability, accuracy, and reliability with humans that you get with robots, especially in the factory.
When you come to public areas, or deliveries, the challenge is of availability now. There aren’t enough young people who want to do this work, especially in the West. So it’s just not viable anymore. Even in India, there aren’t enough people available in the country to work in factories. The total manufacturing workforce is 50 million people typically, in India right now. We produce only 2.5 million people annually from ITIs. That is hardly 5%. So we’d just be replenishing the attrition whereas this 50 million should become 100 million over time. To increase this figure we need to increase our manufacturing output, which is not going to happen unless we add enough skilled people to that pool. The scale of operations is also increasing. We can’t reach that scale at all without automation. There’s just no way. On the assembly line, you have a car coming out every 30 seconds or one minute. A full car! How do you do that if you have not automated a lot of stuff!
Q. So what are the challenges that you faced in building this product?
A. For any startup, life is only full of challenges. It’s difficult to name very few things. Since we’re doing multidisciplinary engineering we do everything. We make the mechanical vehicle, handle the electronics, and cover the software stack as well. So any multidisciplinary engineering system, by definition, becomes a little challenging. Apart from that, supply-chain challenges have been hard for us, like for many people in the industry. These challenges exist in any startup’s journey, especially when it comes to hardware startups. Running a product in the lab is reasonably easy, but taking it to a customer’s production site is very challenging. Crossing the chasm from lab to production is difficult. Another challenge lies in increasing the production capacity, while ensuring the same quality repeatedly. At different stages, one has to keep crossing these challenges, and they never stop. Customers will keep coming in. We have to support our old products while developing new products. Product management is also very complex and challenging, especially with hardware. You can just update the software in so many places, but supporting old hardware is not that easy. We have to do it because once a customer buys a product, it is with the expectation that we will support them for the next 5-10 years, and we have to live by that commitment.
Q. What are the future trends and technologies that you expect to see in this space?
A. Firstly, we are still in the very early days of this space. You’ll see all the line-following robots in many of the factories. But the penetration of these modern, without-line, flexible, mobile robots is still very low. It’s minuscule. I would share another interesting statistic. There is a number which people track which is called robot density, which calculates the number of robots per 10,000 workers. The world average of that number is around 126. The US is at 250. South Korea is in the 950 range, which is a benchmark even for the US which has a long way to go. China has gone from 50 to 250 in just five years. India is still at 4. There is so much that we need to do to catch up not just with the world, but just with our neighbours. So, I would say forget future trends, let’s catch up with the present first. We have to do that at double the speed. We have to compete globally as we are trying to attract global manufacturing into India. We cannot do that if we are not at the level of our global competitors. Such automation is important to produce things at the same quality scale and cost.
Q. How do you see the global competition in this space?
A. We are quite unique globally also in our space. The approach of self-driving car like technology inside factories is still novel. There are demos but significant production deployment as we have done across more than 25 locations is rare. People have tried it in other use cases. Outrider will move full trucks to the yard of the warehouse or the factory. Einride is doing teleoperated automation where somebody is remotely driving; they’re not trying to do full autonomy. Different people have different approaches and everybody has some market because there are so many different things that customers want. No one company will do everything for them. I would like to see more companies in India doing this.