3

There have been many attempts to replicate the hand in the form of a machine. While some have succeeded in imitating specific tasks, all of them have failed to capture the true dexterity of the hand.

If hands are so difficult to render robotic, why bother at all? Motivations for developing robot hands include:
– Safety: Keeping human hands away from hazardous jobs
– Productivity: Robotic hands can work almost non-stop
– Precision: Improving repeatable standards

Hands are far from perfect. Each finger has a limited range of motion; the two upper joints always move together in each finger; and human gestural habits create bodily limitations in our range of motion.
Could a superior hand be invented?

In reality, a robot hand cannot really be like a hand, it is a constellation of many interlinked components, working together to bring the hand to life. The hand works together with, and as extension of the arm.

Controllers – Controllers are the robotic arms’ main processors and operate as their brains. They can be set to behave automatically, or they can be manually operated by receiving instructions from a specialist.

Arms – The arm, which is made up of three major parts: the shoulder, elbow, and wrist, is the major portion of the robotic arm. The shoulder can move forward, backward, or spin. The elbow allows the upper part to move forward and backwards independently of the bottom section. The wrist connects to the end effector at the very end of the upper arm.

End effector – The end effector serves as the robotic arm’s hand. In many cases, it has two claws, but it can also have three or more which can open and close on demand. It may also be spun on the wrist, making it simple to move materials and equipment.

Drives – Drives are the motors that regulate the movement and manoeuvring between the joints. Some models employ belts that are comparable to those seen in automobile engines.

Sensors – Sensors allow the robots and their parts to detect and respond to their surroundings. They can, for example, avoid accidents between two robots operating nearby or enable a robot to modify its hold on a delicate object to avoid injuring it.

Efforts to render our material world into life-like automatic creations has existed, at least in thought, for over a millenia.

1st century AD – Automata
The ancient Greeks used the term “automatons” to describe inanimate objects that seemed to borrow the defining feature of living creatures, self-motion. As early as the 1st century AD, engineer Hero of Alexandria described lots of automata but never made them.

11th Century – Su Song’s Clock
The polymath Su Song designed a clock relating water worlds to outer planets. His monumental clock sat on a river that ran across vast fields of agriculture. This four-floor high clock had a half-automaton cosmologist on top who studied the earth between his two hands.

12th century – Ismail al-Jazari’s Automatons
The Arab inventor Ismail Al-Jaziri designed several automatons in his lifetime and has been described as “the father of robotics”. These included a peacock fountain, a humanoid automata that could serve drinks, and a musical robot band.

1495 – “Leonardo’s Robot”
Leonard Da Vinci designed a robot that could open and close its anatomically correct jaw, wave its arms, and move its head. This robot consisted of two independent systems: the legs and the arms. The orientation of the arms showed that it could only whole-arm grasp with the joints moving in unison, so the hands were not really independently moving, but worked as extension of arms.

1598 – Waterworks for the Royal Palace at Saint-Germain-en-Laye
Tomasso and Alessandro Francini were commissioned to build elaborate and theatrical waterworks for the royal palace at Saint-Germain-en-Laye. Their hydraulic grottoes included automaton blacksmiths, weavers, millers, carpenters, knife-grinders, fishermen, and farriers conducting the obligatory watery attacks on spectators.

1738 – Automaton Flute Player
In 1738, Jacques de Vaucanson designed and built an automaton flute player, the “androide”. His advancement was in the detail of the mechanic arm and hand. The Abbé Desfontaines, who was agape about the human-like characteristics of the flautist, described the inside as containing an “infinity of wires and steel chains…” that made the androide’s hands work to play the flute.
1770s – Androids by the Jaquet-Droz family
In 1774, Swiss clockmakers Pierre Jaquet-Droz and his son Henri-Louis executed three life-sized automata with particular emphasis on their human-like capabilities. It is likely that the village surgeon helped with the development of the arms and hands of these androids. They created an artist, a writer, and a musician automaton. The “Writer” wrote 40 characters; the “Draughtsman” sketched four pictures in charcoal; and the “Musician” played several airs on a harpsichord.

1769 – Wolfgang von Kempelen’s Mechanical Turk
Wolfgang von Kempelen designed “the Turk”, a life-sized chess model. The automaton enacted the chess player’s left hand. This chess game was actually an elaborate hoax with a human operator concealed inside the cabinetry underneath the chessboard. Many knew it was a hoax, but they were fascinated anyway, because it performed the ultimate question of machine intelligence, i.e. can a hand act autonomously?

1920s – Karel Čapek’s Robot
Karel Čapek coined the term “robot” in his 1920 play RUR, Rossum’s Universal Robots. The play is set in a factory where humans and robots work together. At the end of the play, robot workers revolt and bring the end of humans.

1939 – Elektro at the New York World’s Fair
World’s Fairs were a product of the industrial revolution and its desire to churn out ever-more novel modes of production. Many robotic advancements featured in these fairs. Westinghouse introduced the mechanical man Elektro at the New York World’s Fair in 1939. Elektro used a 78- rpm record player to simulate conversation, and had a vocabulary of more than 700 words. Later on, Elektro appeared in the movie Beauty and the Robot (1960).

1962 – The Unimate Robot
The first modern industrial robot, the Unimate, was introduced into factories in the 1960s. George Devol invented Unimate’s first robotic arm in 1962. It was installed at the General Motors plant in Ternstedt, New Jersey, for automated die casting. Approximately 8,500 units were sold.

1963 – The Rancho Arm
The Rancho Los Amigos Hospital in Downey, California introduced the “Rancho Arm”, an artificial robotic arm initially designed for the disabled. It was one of the first robotic arms controlled by a computer. It had six joints which brought it closest to human movement.

1968 – The Minsky-Bennet arm
The Minsky-Bennet arm was biologically influenced by crayfish claws and nerves. It was able to perform a variety of features, like being able to pick up and put down a pencil or tube.

1973 – The Famulus
In 1973, Kuka, one of the world’s largest robotic manufacturers today, launched their first robot, the Famulus – a robotic arm that worked by a revolutionary use of six electromagnetic axles.

1981 – The Canadarm
In space, robot hands work as the remote extensions of the astronauts. The Canadarm series is an example of the multi-degree freedom of robotic arms. These arms inspected the Space Shuttle using a specially deployed boom with cameras and sensors attached at the end effector.

2000 – ASIMO
In 2000, Honda Motors unveiled a spaceman-like robot called ASIMO, which featured a couple of dexterous humanoid hands, which could open the cover of a cup. Development of the robot ceased in 2018, as Honda stated it would focus on more practical applications using ASIMO-derived technology. Design details of these robots are still kept confidential.

2000 – da Vinci
The US Department of Health and Human Services (FDA) approved the use of the first surgical robot, the da Vinci.

2005 – Aichi World’s Fair
At the Aichi World’s Fair in 2005 a section was called “We Live in a Robot Age”, which included robots that aided sanitation, garbage collection, childcare and duties for the disabled. The section “Robot Station” was to interact with a whole host of robot-based venues.

2009 – i-Cub
The i-Cub by the Italian Institute of Technology and later with RobotCub Consortium, was developed to test embodied cognition and artificial intelligence. The robot hands developed at the consortium are thus capable of ‘learning’ motor behaviour over time, such as grasping small objects, archery and crawling.

2011 – Inmoov
Inmoov was invented as the first open-source 3D-printable life-size robot. It was originally made as an art project by French sculptor Gaël Langevin. It is controlled by Arduino microcontrollers and open-source robotics software. Because of the open-sourced nature of its construction, there are potentially limitless uses and design variations for the robot.

2013 – The Atlas Robot
Atlas robot was developed by Boston Robotics for outdoor search and rescue. Several diverse, powerful non-humanoid robotic hands can be linked to its arm one at a time for use in different scenarios. Boston Robotics, bought by Hyundai in 2022, aims to focus further on creating the perfect automated robot hand.

Robots are not able to copy or imitate the human hand, at least not fully, yet, because the human hand is far more complex and articulated to cover all activities that humans want to perform. Hands work with tendons and many hinge joints that make it possible to handle weight and operate flexibly. They can repeat tasks, but also behave autonomously and ad-hoc.

The internet is awash with examples of robotic-shaming hubris, when it comes to our foolish attempts to automate the hand: botched haircuts from errant automated scissors; poorly constructed ceramics from robotic arms colliding with pottery wheels, the clumsiness of a robot unable to catch their fall with an outstretched hand.

Even in the world of AI-imagery, hands throw a proverbial spanner in the works. We see awe-inspiringly realistic photographs being churned out by increasingly sophisticated AI-bots, with one glaringly obvious tell that they’re not real. The hands are rubbish. We see portraits with beautiful faces, but three hands, seventeen fingers, all contorted like sausages in floating ether. The communal AI brain, feeding off of countless terabytes of data, sourced by millions of people, still cannot fathom the basics of a human hand.

Automation intensifies the use of robot hands for the sake of safety, productivity, and precision, however it is still hard to justify this investment, humans are still the most adaptive and resilient.

Rather than shrink back in terror, perhaps there is something to celebrate here. In our gradual march towards a fully automated future, one stubborn truism persists: the hand remains a mystery. Once shunned, as something to be tamed and ashamed of, the hand is due for a comeback, as our last defense against a world that wants to render humankind secondary to machines.