In Part 1 , we highlighted why while robot automation has advanced, we still have a long way to go. Here, we discover why we cannot keep building robot solutions the traditional way. We need a paradigm shift in robot system design to achieve hyper-automation.
Robots Have “Force”
Seek Jedi Master.
The industrial robot industry is a whopping $16 billion. The market is worth double that when factoring in all the components and services that feed into this industry (International Federation for Robotics).
Industrial robots can be re-programmed and used for various applications. The most versatile of industrial robots today are articulate robot arms — sophisticated electro-mechanical machines but with no “intelligence” associated. These robot arms can operate with three to even seven degrees of freedom. An significant evolution from the early days of rigid automation for every type of motion.
Robots today can repeat their movements with high precision. They can move away from a position and then return to it with as little as 0.01 mm difference from the previous time it occupied the place. This is known as position repeatability.
Additionally, they completely dwarf human ability by achieving speeds upto 3 m/s, and payload capacities of several hundred kilograms.
However, they are limited in situations where they need to adapt to dynamic conditions.
What are some complex jobs for robots?
1 -Bin Picking & Placing— Identifying items from a bin, picking the right one irrespective of orientation and accurately orienting the item as required
Working with screws — Picking up screws, placing them in the right hole, and holding it right while making the first few turns. Think about it. Almost every single object manufactured needs screws!
Dealing with wires,cables and flexible objects — changing geometries each time
Ability of robots to dynamically handle wires is still a distant pipe dream
Bin picking, orienting and placing is so simple for adults; we forget how complex it could be for a robot. It involves many steps –
Industrial robots cannot “look” and “figure out” that a specific part has been moved a few millimeters away, or the orientation is a little askew, or even recognize an object that may have a shadow cast on it. They then become incapacitated and unreliable.
The real issue is that having achieved the original goals of automation, we stopped asking the right questions.
Historically, we asked our robots to -
1 - Perform a specific task predictably and repeatedly.
2 - Complete their tasks without “seeing,” “thinking,” or “understanding” what they were doing.
Ultimately, a vast majority of tasks in any industry require flexibility in the face of changing parts, orientations, or locations, but our robots are not equipped to accommodate for that.
In an attempt to fix this, a large part of the robotics world is devoted to building costly customized solutions with very tiny tolerances into standard automation processes so their robots can pick items “blindly.”
However, products have gotten more sophisticated, greater productivity is the need of the hour, and flexibility in an ever-changing world is paramount.
It is no longer feasible for customized mechanisms to be the norm.
In other words…
It’s Not You, It’s Us.
In this age of hyper-consumerism, intense competition from across the globe, and rising customer demands for the next generation of products, sometimes within a few months of the original launch, the old automation model cannot be sustained.
It already costs large manufacturers trillions of dollars in wasted opportunities, high labor costs, and losses in productivity.
Take the example of Adidas recently shutting down its “Speedfactories.” Adidas had factories in Germany and the United States designed to decentralize its operations and bring products to customers in Europe and North America much faster.
Speedfactories were considered to be futuristic and highly automated with cutting-edge robotics. They were also meant to be an alternative to outsourcing to lower-cost labor markets like China. Alas, this too remained a damp squib.
We need a paradigm shift in how we envision robot automation.
Sophisticated robot arms need advanced robot eyes and brains to see better, think, understand what they are doing, and adapt to shifting products, orientations, and situations.
Robot system designers need to aim for a new goal and, in the words of Adidas’s arch nemesis, we are here to “Just Do It.”
I am Nikhil Ramaswamy, CEO and Co-Founder of Cynlr, with my business partner, CTO and Co-Founder, Gokul N A. We are a deep-tech and venture-funded startup looking to advance the field of robotics through next-gen visual object intelligence.
Whether a curious investor, robot-aficionado, software and algorithm whiz or a combination of the above, discover more at www.cynlr.com