CXOToday has engaged in an exclusive interview with Gokul NA, Founder- Design, Product & Brand at CynLr and Nikhil Ramaswamy, Co-Founder and CEO at CynLr.
- Can you give a brief about the company?
Today, most industrial robots perform “dumb” tasks like tracking along a predefined set of coordinates. Any task that requires the robotic arm to be versatile needs human intervention.
We at CynLr want to transform these mindless machines into mindful robots. To that end, we are building a visually intelligent robotic system that is capable of understanding, picking, and manipulating (orienting, placing, moving) any physical object placed in a cluttered environment or presented in random arrangements and orientations. While this sounds simple (a human child can intuitively do this), this problem has perplexed the world of robotics for more than 40 years now and is often touted as the Holy Grail of Robotics.
CynLr is currently engaged in research collaborations with manufacturing lighthouses & giants from automotive and machine tool industries, to build the next generation of autonomous and intelligent robots. We envision this visual object technology to change the way the factories’ function.
2. How is CynLr using innovation to contribute towards visual object intelligence for industrial robots?
Today, modern approaches in AI and machine vision technology used in robotics heavily rely on predicting object positions based on static images, regardless of their format. However, this approach faces challenges when dealing with varying object orientations, lighting conditions, and reflective surfaces like mirrors. Traditional methods relying on color and shape for object identification lack universality. To achieve true automation with flexibility and reliability, robots need human-like perception, understanding, and interaction abilities.
At CynLr, we design fundamental components of visual intelligence for robotic arms, integrating them with object awareness. Our goal is to enable robotic arms that can manipulate objects with exceptional precision, adapting seamlessly to variations in shape, orientation, and weight. This adaptability allows them to meet future requirements. Our groundbreaking visual object intelligence platform empowers industrial robotic arms to see, comprehend, and manipulate objects in random and unstructured environments, offering a comprehensive solution for automation and object handling.
Our highly dynamic and adaptive product renders customised solutions obsolete. CynLr’s vision dynamic robots effortlessly handle any object in any environment without requiring training. Whether it’s a packet of chips, a metal part, or a mirrored spoon, the same robot can handle them all. These future-proof robots are capable of performing a wide range of tasks, paving a way for the futuristic universal factories.
(You can refer to our demo videos – Untrained Dynamic Tracking & Grasping of Random Objects and Oriented Grasps and Picks of Untrained Random Objects)
3. What is your take on India’s evolving industrial robotics landscape?
Industrial automation is at a nascent stage in India. However, with the recent push from the government to produce locally and reduce dependencies on China or Taiwan, India is bound to have accelerated use of industrial robots.
India will become a true manufacturing hub when the perception problem surrounding robots is solved. Currently, robots are viewed as cost-cutting tools rather than value creators. Manufacturers should realise that robots will not replace humans but will add to their productivity by performing repetitive tasks. A simple case in point is the use of a power screwdriver. A person with a power screwdriver can deliver 10 times more output than someone with a regular screwdriver. The manufacturing ecosystem is slowly realising the benefits of industrial automation with a few entrepreneurs and companies leading the market.
4. What are some of the current challenges/roadblocks you are facing?
As a deep-tech startup company, which is working on a fundamental technology such as object manipulation, we confront several technological and operational challenges that limit our capacity to scale our solution:
The investment landscape is a huge impediment. CynLr is subject to what venture capitalists call the “India Discount,” in which investors have this assumption that an Indian company can only build solutions relevant for the India market. This assumption, however, ignores the fact that CynLr sources from over 400 suppliers in over 20 countries, resulting in input prices equivalent to those of US or European enterprises. It is critical to bridge this perceived gap and connect investor expectations with CynLr’s global requirements.
Furthermore, there is a gap between the investment ecosystem and the manufacturing industry. Most investors, including venture capitalists, associates, and analysts, have backgrounds in software or business, which limits their grasp of the complex difficulties confronting the manufacturing industry. So, every time we approach VCs, we have to prove our concept over and over again even in early development phases.
On the talent front, we find it hard to find the right people for the job. CynLr’s vision stack is built on a solid foundation of fundamental disciplines such as evolutionary biology, neuroscience, and physics. This involves hiring of employees with a solid basis in these domains, an understanding of the company’s complicated objectives, and the patience to handle long-term difficulties. However, such candidates are in short supply, resulting in an extraordinarily low selection rate despite receiving a large number of applications.
Nonetheless, we are predicting a shift in opinions as we continue to execute pilots with customers in the United States and EU region. We are also partnering up with academic institutions and R&D labs in Europe. CynLr’s machine vision stack is a foundational layer for vision-based automation We bring together the ecosystem to reach a universal solution. Some of the academic institutions we have engaged with so far include RSL, ETH Zurich, Karlsruhe University of Applied Sciences, HM Hochschule München University of Applied Sciences Munich and EPFL, Switzerland.
5. Which industries are you servicing currently and where are your customers located?
We primarily target discrete manufacturing industries, including automotive, electronics, warehousing white goods, aerospace, and so on. Our visual intelligence technology enables customers to automate previously non-automatable operations, specifically focusing on tasks involving the placement and tightening of fasteners like bolts and screws.
The ability to direct robots in inserting bolts accurately, aligning them with screw holes, and ensuring proper tightening with minimal slippage has been an unsolved problem in automation. Considering that a substantial portion of manufacturing efforts—approximately 70%—is dedicated to manual fastening, our visual intelligence technology has the potential to significantly streamline and automate these processes.
While we are already working with leading OEM players in the European Union (EU) and the United States, we also have plans to further expand our business in India, to serve a broader market and potentially collaborate with more Indian manufacturers.
6. Could you please emphasise the importance of universal factories and how they would aid in simplifying factories and logistics in the manufacturing industry?
Universal factories will play a crucial role in simplifying factories and logistics in the manufacturing industry, and CynLr’s technology can contribute to achieving this vision. By creating universal factories, manufacturers can establish flexible and adaptable production environments that streamline operations, enhance efficiency, and optimize logistics.
One of the key advantages of universal factories is their ability to accommodate diverse product lines without the need for significant reconfiguration or specialized infrastructure. CynLr’s visual intelligence technology aligns with this concept by providing a platform that simplifies deployment and eliminates the need for hardware adaptations for different tasks or objects. This versatility allows manufacturers to rapidly deploy robots and adapt their operations to new products or changing demands, saving time and resources.
Moreover, the automation of previously non-automatable processes, such as the placement and tightening of fasteners, through CynLr’s visual intelligence technology can contribute to the development of universal factories. These processes often require intricate coordination of visual direction, tactile feedback, and specialised knowledge. By automating these tasks, CynLr enables manufacturers to eliminate human-intensive processes, reduce labour costs, and enhance overall productivity.
In the context of logistics, universal factories simplify operations by reducing the complexity associated with managing multiple production lines or specific configurations for different products. The streamlined deployment and adaptability of CynLr’s technologies allow for faster implementation and reconfiguration of robot cells, making logistics more efficient and responsive to changing demands. This flexibility leads to shorter lead times, increased production throughput, and improved overall supply chain management.
By embracing universal factories and leveraging CynLr’s technology, manufacturers can transform their operations into agile, adaptable, and optimized systems. The simplification of factories and logistics will result in reduced costs, increased productivity, improved product quality, and enhanced competitiveness in the manufacturing industry.
7. In your opinion, what is more important, hardware or software?
Both hardware and software are essential components in the development and functioning of technology systems, and their importance can vary depending on the specific context and application. In some cases, hardware may take precedence, such as in systems where real-time processing, high computational power, or specialized sensors are crucial. In other cases, software may be the key differentiator, particularly in systems that require sophisticated algorithms, machine learning, or complex decision-making capabilities.
For example, in CynLr’s case, hardware takes precedence. Software cannot do the magic alone. It’s physical elements and physical objects that we’re actually dealing with. So, without structuring the system to be able to pick or capture information about an object, we cannot create a holistic model of the object or intelligence about the object. Also, information about an object cannot be gotten just from a camera. For example, we think autofocus just brings objects into focus, but actually, it is the first layer of depth construction that we do. There are a lot of missing features in the way we use these physical devices like cameras. So, our inference algorithm begins with the hardware itself
8. What are your future plans going forward?
Our vision is to make ‘Universal Factories’ a reality. We aim to add our machine vision technology stack to any computing system performing a task. We are starting off with manufacturing plants which have tasks that are not completely automable with the current technology. In future, we will see how this technology can be further used for different kinds of applications apart from manufacturing. It turns out that this capability can indeed have use-cases beyond object manipulation alone, such as in advancing navigation capabilities of ground robots/humanoid robots and for Advanced Driver Assistance (ADAS) and autonomous driving. CynLr will be investing in building more products to cater to such adjacent spaces. In a gist, wherever the human eye aids humans excel at a job, CynLr’s vision stack can eventually find a place in helping automate those tasks.
Our strategic plans include establishing new offices in both the US and EU, expanding our workforce to exceed 50 employees, and enhancing our capabilities to deliver 100 robots annually to our valued customers.