THE 1940S: ASIMOV’S LAWS OF ROBOTICS AND THE FIRST ARTIFICIAL NEURAL NETWORKS
While the 1920s saw the prelude of the term “robot”, it wasn’t until Isaac Asimov’s 1942 novella “Runaround” that the phrase “robotics” appeared. In this novel, Asimov set out his well known three laws of robotics — that robots must not harm humans, that they must obey orders from humans, and that they must protect themselves from threats provided their self-preservation doesn’t break either of the first two laws. Though written in fiction, these laws gave the foundation for many of the ethical questions covering robots and autonomous technologies, and are still applied to today.
Isaac Asimov’s “The Three Laws of Robotics”
The 1940s also witnessed the creation of the very first artificial neural networks. In 1943 Warren McCulloch and Walter Pitts created a fundamental neural network applying electrical circuits to better realize how neurons function in the brain. Their studies led the way for the first autonomous robots that could showcase complex behavior, due to making use of artificial neural networks.
In 1948 and 1949 William Grey Walter developed two such robots — Elmer and Elsie. Labeled “tortoises”, the automatic robots could react to and move towards light, directing themselves to their re-charging stations when their electric batteries were low. Walter named his machines tortoises, after the tortoise in Alice in Wonderland. He’s work preceded for BEAM (Biology, Electronics, Aesthetics and Mechanics) robotics, which do not demand the computing power of a microprocessor.
The original neural network was developed in 1951 by Marvin Minsky and Dean Edmonds. Named SNARC (the Stochastic Neural Analog Reinforcement Computer), it was generated from vacuum tubes, motors and clutches and was tasked with assisting a virtual rat resolve a puzzle.
THE 1950S: THE TURING TEST AND THE UNIMATE
A further turning point moment in the evolution of robotics happened in 1950, when Alan Turing summarized his assessment of a machine’s artificial intelligence. The Turing Test has become the benchmark of AI, since it measures to which degree a machine’s intelligence is equal to or equivalent from that of a human. In its most basic form, the objective of the test is to determine whether a machine can think. In spite of the test being criticized for its simplicity, it has become authoritative in the way we think of artificial intelligence.
Statue of Alan Turing
His project created a needed framework for the establishment of the discipline of Artificial Intelligence in Dartmouth College in 1956.
The 1950s also saw the conception of the premier industrial robot — the Unimate. The patent for the Unimate was filed by George Devol in 1954, and displayed a robotic arm capable of transporting die-cast components and welding them into place. The revolutionary tool has led the way for industrial robots to complete recurring, challenging or hazardous tasks and transformed the face of the manufacturing industry permanently.
THE 1960S: THE INDUSTRIAL ROBOT REVOLUTION
In 1968, MIT’s AI Laboratory co-founder Marvin Minsky put together a “tentacle arm” — a robotic 12-jointed arm that was powered by hydraulics and could be managed via a joystick. Minsky’s robotic tentacle was robust enough to lift a person, and could reach around obstacles easily. His research ushered for a number of the soft robotics innovations arising today.
‘Backpropagation’ may not seem like much, but it’s actually one of the most important algorithm in the evolution of AI. In layperson’s terms, backpropagation enables a neural network to adjust its layers in the event that the outcome obtained by its computations isn’t the objective it’s aiming for. This means that the system can learn through trial and error to proceed to refine its outcome in a similar way to a child learning ways to do something. It works utilizing statistics — each time the system results in an error, the likelihoods are readjusted and a new procedure is undertaken. Despite the fact that the idea which became the groundwork of backprop was initially proposed in 1969, it wasn’t regularly integrated into machine learning until the mid ‘80s.
“Shakey” the robot was designed by the Stanford Research Institute between 1966 and 1972 and was a turning point in robotics as a result of its blending of hardware and software so that it could perceive its natural environments. Shakey took robots into the general public awareness after receiving extensive media attention.
Right after Devol was granted his patent for the Unimate in 1961, the application of robots in industrial environments advanced rapidly. That same year, General Motors applied Unimate on their production line in Ewing, New Jersey. After the triumph of Unimate at General Motors, it entered comprehensive production in 1966.
In 1969 Victor Scheinman designed the Stanford Arm, a robotic arm that is considered as being among the first robots being overseen entirely from a computer. This was a massive breakthrough, as at that time Unimate maneuvered from a magnetic disk. It spotlighted six points of articulation and was assembled entirely in Stanford’s Artificial Intelligence Lab. Though applied mainly for educational purposes, the Stanford Arm indicated a major breakthrough for industrial machines that could be regulated via computers.
The 1960s saw a quantity of innovations and growths on the central idea of Devol’s robotic, industrial arm. At present, we find conversing to natural language processing programs like Alexa and Siri a part of everyday lifestyle. But back in the ’60s, voice assistants were a little more revolutionary. ELIZA was among the first of these, developed by a professor at the MIT Artificial Intelligence Laboratory between the years 1964 and 1966. The software could undertake an interaction via text by following a ‘script’ that ordained it on the ways to respond.
NASA used cutting edge computing and robotics technology to get human beings on the moon for the first time in 1969.
THE 1970S: THE WABOT-1, INDUSTRIAL INNOVATIONS, AND ROBOTS IN SPACE
The early 1970s saw the introduction of the world’s first comprehensive humanoid robot — the WABOT-1. The WABOT-1 was a sequent to 1967’s WABOT, and was created by Ichiro Kato in Tokyo’s Waseda University. The WABOT-1 had a vision and arms and legs control system, empowering it to maneuver itself and move freely. It could even estimate distances between objects. Its hands came with tactile sensors, meaning it could pick up and transport objects. It had an approximated intelligence equivalent to that of an 18-month-old human, and marked a significant advancement in humanoid robotics.
In 1978 SCARA — the Selective Compliance Assembly Robotic Arm — was constructed. Created by University of Yamanashi professor, Hiroshi Makino, the arm could slide 4 axis and became a common fixture in production line in the early 1980s.
The first of all robots to come down on Mars were Viking 1 and Viking 2, who visited the red planet in 1976. Both robots were powered by radioisotope thermo electric dynamos, which produced power from the heat radiated by decaying plutonium. While the records gathered by both Vikings was inconclusive, they were the legitimate leaders of the Mars rovers we know today.
Carl Sagan with a model of a Viking lander
The 1970s also saw the progress of industrial robotics when, in 1973, German company KUKA launched the FAMULUS — the 1st industrial robot with six electro mechanically driven axes. The next year, Richard Hohn developed the first industrial computer being powered by a minicomputer — The Tomorrow Tool, or T3.
George Lucas popularized humanoid robots with the production of Star Wars in 1977 — A New Hope and the procreation of two of the most well-known robots to this day: R2-D2 and C-3PO.
THE 1980S: ROBOTS IN THE HOME, THE CANADARM, AND GENGHIS
Frequently considered one of the most significant robots in modern history, 1989’s Genghis was a hexagonal robot made by scientists in MIT. As a result of its small size and inexpensive materials, Genghis is credited with reducing production time and cost for future space robot concepts. It was assembled with 12 servo motors and 22 sensors, and could pass through rocky landscapes.
Self-driving cars are still not regular at present, so it may be surprising to learn that the first self-governing car took itself for a drive in 1986. The Mercedes-Benz van incorporated mirrors and sensors and had the ability to drive safely and effectively on empty streets. The project was accomplished by researchers at Germany’s Bundeswehr University. Ever since then the technology used in self-driving cars has continued to improve.
It was in the 1980s that robots formally entered into the conventional consumer market, though mainly as simple toys. Among the most prominent of these robotic toys was the Omnibot 2000 by TOMY. The Omnibot 2000 was remote-controlled, and came complete with a holder for offering drinks and snacks.
The ’80s saw further advancements in the field of industrial robots, with Ford incorporating numerous robots to their production line around the world. The Ford Fiesta was noteworthy for being among the first cars globally whose anti-corrosion sealers were injected by robots.
Robots continued their adventures through the universes in the ’80s further, with the release of the Canadarm on the Space Shuttle Columbia in 1981. The Canadian-made automatic arm was 50 feet (15.2 meters) long and had six points of juncture. It could be governed by one crew member in the command station, and conducted 90 successful missions during its period in service.
THE 1990S: THE CYBERKNIFE, THE SOJOURNER, AND AIBO
The late ’90s saw the launch of one of the most legendary robots of the 20th Century — Sony’s AIBO robotic dog. Introduced in 1999, AIBO was one of the first robotic pets to hit the buyer market. AIBO could react to voice commands and chase after a pink ball which came along with the purchase of the robot. Earlier this year Sony revealed a new, renewed AIBO for the 21st Century market which came equipped with two cameras and space-mapping abilities.
In 1996 the Sojourner became the first rover being sent to Mars. The small, lightweight robot was taken to Mars by the Pathfinder, and with success settled on the planet’s surface area in July 1997. Throughout its time on Mars, Sojourner examined 2,691 square feet (250 square meters) of land and made 550 images. As a result of the information collected by Sojourner, scientists had the chance to ascertain that Mars once had a warm, wet climate. The quest marked the starting point of several more NASA rover assignments to Mars.
The early 1990s saw robots come into the operation auditorium with the Cyberknife — a radiosurgery equipment that could surgically address tumors. Designed by Stanford University neurology professor, John R. Adler, the Cyberknife was a non-invasive surgical instrument which traced and intended tumors with narrowly-focused light beams of radiation. By 2010, the Cyberknife was used in 5% of all Stanford Cancer Center’s procedures.
In May 1997 — IBM’s robot Deep Blue beat the world chess champion Garry Kasparov in a match. It had beaten Kasparov in a single game in 1996.
THE 21ST CENTURY: THE STATE OF ROBOTS IN TODAY’S TIMES
Although we’re a mere 18 years into this century, robotics have readily improved and formed a lot of our technological landscape. Many homes now have their personal Roombas — robotic vacuum cleaners that are able to clean your flooring’s autonomously. We’ve seen the application of drones throughout from the military to home distributions. There have been numerous landmark breakthroughs over the last few years, that they would require their own post. When reviewing the robotic achievements of recent years, nonetheless, it would neglect not to point out two robots particularly — Sophia and the Boston Dynamics Dog.
Sophia made headings last year when she eventually became the first robot to be granted citizenship to a nation. The Android robot, made by Hanson Robotics, was awarded Saudi Arabian citizenship in October 2017. The following month, she came the very first non-human to be given a United Nations title when she was labelled the UN Development Programme’s Innovation Champion. Sophia’s AI is cloud-based which enables deep learning, and she can identify and reproduce a wide array of human facial expressions.
In 2000 — Honda unleashes ASIMO (Advanced Step in Innovative Mobility), its famous humanoid robot designed to become a personal assistant that can recognize voice commands, motions and interact with its natural environments.
2004 — IBM begins work with Watson, its next version on Deep Blue. Watson made headlines in 2008 when it defeated human beings at the quiz show Jeopardy! which demands complex knowledge of natural language. Watson beat down former champions Ken Jennings and Brad Rutter in 2011.
Boston Dynamics has been proclaimed in the press as leading the charge in modern robotics, as a result of their autonomous creations. One of the most famous of these is the Boston Dynamics Dog, or BigDog, which grabbed global attention upon its introduction in 2005. It was designed to become a robotic pack horse for military purpose, and contained 50 sensors over its body system. It was capable of transporting weights of unto 340 lbs (150 kg) and could run at a stunning 4 mph (6.4 km/h).
Just Recently, Boston Dynamics showed two more headline-grabbing robots — the MiniSpot and Atlas. MiniSpot is a sovereign robot dog that can open doors on its own, whilst Atlas is a state-of-the-art humanoid robot capable of jogging and jumping over barriers.
On March 15 2016 — AlphaGo, an AI program developed by UK company DeepMind, defeated the world champion Lee Sedol at the ancient board game Go. This was an important breakthrough for DeepMind’s research into creating artificial intelligence that can ‘learn’ ways to resolve problems no matter the context, as opposed to Deep Blue that is configured for a specific use scenario.
DeepMind’s Artificial Intelligence (AI) AlphaGo won the final match in the five-game Go series against South Korean champion Lee Se-dol
In 2017 — a new model of AlphaGo, AlphaGo Zero, learnt how to play Go by itself in merely three days after only being explained the rules. Prior versions have learnt the ways to play the game by training themselves up on 1000’s of games taken by professionals. However this variation simply gamed itself repeatedly countless times, beginning by positioning the stones on the board randomly but rapidly learning winning strategies. This is fantastic because it illustrates that AI can develop knowledge by itself with hardly any human direction. And this has far more valuable purposes than beating mortals at board games. At this moment, the software is tasked with figuring out how proteins fold, a discovery that could revolutionize drug innovation.
If Boston Dynamics is any example to move onward, robotic inventions are arising on a near-weekly basis within this 21st Century. Robotics have enjoyed a lengthy and fabled past history, and it appears that we have much more to look forward to.