Quo vadis, robotics?
Robotics will change the world! In combination with artificial intelligence (AI), it will be shaping the next 50 to 100 years to a similar extent as the emergence of computer technology, the internet, and smartphones have affected life in the last half-century. In various forms, intelligent automation solutions will permeate the world we live in and – if humanity uses them correctly – comprehensively enhance it. In that world, our grandchildren will be growing up as the first Generation ‘R’ of Robotic Natives – surrounded by autonomous, intelligent assistants supporting them in both their personal and professional lives. What will that world of tomorrow look like, though? What shape will robots in manufacturing, the service domain, entertainment, and our households take? Will they continue to be large, highly specialized, stationary automation solutions – similar to those in the early days of computer technology – or will “robot apprentices” and “robot housekeepers” soon become a reality across the board?
The author
Dominik Bösl is a technology geek, professor of informatics, father of two, and convinced that technology can really make the world a better place to live – if we get it right! He investigates the “how to” for the UN, the European Commission, the government, and others. His heart beats for technology ethics and the collection of Iron Man action figures. Those are optimum prerequisites for writing about humanoid robots because he’d be their first customer!
Robotic revolutions
In its triumphant history, computer technology went through four evolutionary stages: huge, room-filling mainframe computers made the leap onto our desks due to miniaturization of crucial components and were made available to the public as desktop computers. Thanks to the resulting falling prices and continued downsizing, they evolved into mobile systems – i.e., laptops and notebooks. In the next stage, they took shape as smartphones and tablets becoming omnipresent in the human world. And today, our lives are permeated by digital systems and services most of which are no longer even visible.
The four "robotic revolutions"
Robotics is developing in a very analog manner in four disruptive stages, also known as “robotic revolutions.” The first stage saw the creation of classic industrial and manufacturing robotics as we’ve known it for several decades, for instance in the automotive industry. These highly efficient but potentially dangerous machines could only be used inside safety enclosures and behind fences – like in a “robot zoo.” In the second stage, the robots were released from their cages and can now interact directly with humans in the form of safe, sensitive cobots. The third revolution involves the mobilization of systems: a workpiece no longer needs to be taken to the robot but the machine moves to where it will work – an important prerequisite for flexible, adaptive service robotics. The fourth and highest evolutionary stage is currently being created by the collaboration of robotics and artificial intelligence. Only when these safe, mobile systems become autonomous and “intelligent” can they proactively support a wide variety of tasks in manufacturing and everyday life. This last and decisive step is the necessary basis for science fiction to become a reality.
The potential of legs
What crucial role will legs play for robotics in the future? When life left water, the ability to crawl, scuttle, hop, and, later, to run was essential to conquering new habitats. Primates – and ultimately humans – proved to be particularly adaptable. Not because humanoids were best-adapted to any environment or habitat but because they lacked exactly that level of specialization: humans are not particularly good at anything but good enough to adapt to new conditions very quickly. Its upright, two-legged gait is one of the key skills that have made the “human” model so successful: free hands enable the use of tools and movement in and on a wide variety of terrains. So, it supposedly makes sense to rely on legs when mobilizing (service) robots and to create machines in our image. Especially in unstructured, constantly changing environments – e.g., offices, retail settings, and households – legs are a necessity. However, the two-legged gait has disadvantages too: constant balancing of the systems requires high computing power because they cannot stand upright and maintain stability on their own. Plus, regular motion, as well, requires more energy than a system with wheels would, for example. The same applies to the use of humanoid robots: the gain in flexibility is offset by a loss of efficiency and effectiveness.
The history of bipeds
The desire for artificial, human- or animal-like assistants is thousands of years old. There’s evidence from ancient Egypt and Greece that resourceful engineers constructed so-called “automatons” in the form of figures of gods distributing small objects to worshippers in temples after having inserted a coin. The golem, known from Jewish mythology, was a figure made of clay that served its master. Leonardo da Vinci tried to give his mechanical lion anatomically correct movements, thus creating an artificial living being. The alchemists of the Middle Ages followed the same idea on their mission to create “humunculi” – laboratory-bred, human-like figures – as servants. Even today’s concept of the robot originates from a similar idea. In 1920, novelist Karel Čapek described artificial humans in his drama “R.U.R – Rossum’s Universal Robots” – (today, they would most likely be called androids) engaged in cheap labor without any rights. However, they all have one thing in common: they tried to create legged beings that were supposed to move around “normally” in the human environment.
Motivation for humanoid robots
A look at the media reveals that there have been frequent reports in recent weeks and months about the (supposedly) imminent use of humanoid robots in industrial manufacturing. In its current forecast for the robotics market, the International Federation of Robotics (IFR) names androids as a defining technology trend for 2024. Companies such as Daimler, BMW, and Tesla are talking publicly about related field tests, technology partnerships, or even proprietary developments. On the one hand, these reports raise hopes for imminent introduction of humanoid robots across the board; on the other hand, they underscore the urgent, underlying problem: the shortage of skilled labor. The Volkswagen Group predicts that by 2035, 15% of the current workforce will retire – and that they will no longer be found on the global labor market. The United Nations and the OECD predict, as well, that by the 2050s at the latest, there will be a global labor shortage that can only be compensated for by the widespread use of flexible, autonomous automation technologies. By now, it’s clear that we will have to automate! Humanoid systems have the potential to automate areas of manufacturing and everyday life that were previously considered not to be fit for automation: fields of application with high complexity and small batch sizes, inadequate efficiency for stationary automation solutions, or areas with rapidly changing environmental conditions.
18
years. That’s the age the humanoid robot Justin (below) of the German Aerospace Center has just reached. Initially just a torso, Justin now has sensory and motor abilities that at least come close to those of humans. It made its major development leaps especially in the most recent past – thanks to AI.
State of the art
Research has long been engaged in the design of bipedal and multi-legged robots. Every major German university involved in mechanical engineering and mechatronics is dedicated to at least one of the aspects necessary for the construction of humanoids. Particularly at the Karlsruhe Institute of Technology (KIT), in the robotics focus group of the German Aerospace Center (DLR), the various Fraunhofer Institutes, and at the Technical University of Munich, attempts have long been made to have household robots load dishwashers, design lightweight actuators for robot legs and arms, coat robots with sensitive sensor skins to increase human safety, or to make robots fit for use in domestic and nursing settings. The list of international academic institutions would even be longer including Stanford University, MIT, the University of Washington, and Georgia Tech University, as well as the leading institutes in Italy, France, and Japan. Some of the best-known demonstrators and concepts come from this source: Boston Dynamics has no doubt been setting the standard for two-legged robots for years with Atlas, and the videos on the internet in which the android does somersaults, is nimbly running and jumping a parkour, or indulging in a dance with other robots are well known to most people.
“Machines that at best only need to have tasks shown or explained to them and that – just like human employees – can move freely on the factory floor are the dream of many companies.”
But even these humanoids are far from being as autonomous as we’d like them to be for use at home as robot housekeepers or robot butlers. So, what has changed – beyond the impending skilled labor shortage – prompting global corporations such as BMW, Tesla, and Daimler to publicly talk about the use of humanoid machines? The disruptive changes in the field of artificial intelligence have no doubt made a significant contribution to that discussion. The rapid development of technologies such as ChatGPT and company have reignited hopes for fast progress in the field of autonomous, reactive, and “intelligent” systems. Manufacturers of initial systems such as Apptronik, Unitree, Agility Robotics, Figure 01, or NEO are entering into collaborative partnerships with well-known customers to test the maturity of their technology and systems under realistic conditions.
As magnificent and impressive as these technology demonstrators may be, there’s one thing they’re definitely not (yet): products. And they’re certainly not suitable for industrial-scale or reliable use in processes. For a long time, the number of available systems was very limited because they’re expensive to build, maintenance-intensive, limited in their functionality, and could be operated only for a few minutes on a single battery charge. It seems downright grotesque that the funny and stunning demonstration videos do not show the key aspects: the five to nine scientists and engineers in the background remotely controlling the system, manually adjusting parameters, or standing by to catch the androids when the energy supply diminishes.
No fear of the robocalyps
One thing is certain, though: The prospect of flexible, autonomous robot workers is tempting. Machines that at best only need to have tasks shown or explained to them and that – just like human employees – can move freely on the factory floor are the dream of many companies. Mechanically, a lot has happened in this area over the last 35 years – if it just weren’t for the continuation of the well-known issues: inadequate energy density for storage technologies, inadequate payload-bearing load ratio for actuators (the human arm can lift and carry many times its own weight, robotics is still struggling with a ratio of 2:1), and inadequate cognitive abilities. Only when these problems have been sufficiently solved and the AI has progressed far enough to be able to recognize, “understand,” and then grasp, handle, and discard a previously unknown object, will the hoped-for “General Purpose Robotic” come within reach.
The good news? None of us need to be afraid of a “robocalypse” – i. e. an uprising of robots – now (or in the next half-century!). Change will happen, automation of the world will progress. But it will happen at a slow yet steady pace, similar to the development of autonomous vehicles: First, car manufacturers’ top-end models will be equipped with increasingly intelligent and autonomous assistants; subsequently, the technologies will migrate to the lower-priced segments. And one day, autonomous driving will be “with us” as well – but not as a revolution overnight, but as an evolution over years. The same applies to AI and autonomous, humanoid (service) robotics. Unfortunately, this also means that household robots that do the laundry, tidy, clean, and cook will be long in coming as well. But come they will!
Artificial intelligence and robotics are here to stay – if we assume responsibility now, we can shape this innovation and develop it into a valuable, sustainable support technology for humanity. The introduction of the washing machine has freed up as much as 30% of the human time that was previously tied up for household chores (unevenly distributed between women and men, of course) – the establishment of automation technologies such as artificial intelligence, robotics, autonomous driving, etc. across the board will have a similar impact on society. Let’s hope that it will be used for the benefit and growth of humanity – and not for cat videos on the internet.
