Where seconds count: A challenge in space exploration
A new atomic clock is in Earth orbit promising unprecedented accuracy in time keeping. On the ground, artificial intelligence has solved the famous Rubik’s Cube puzzle in just over a second, learning all by itself, setting a world record. The implications of these mind-boggling developments for future space exploration are fascinating.
As Chandrayaan-2 is well on its course, the Moon’s orbit is shifting away from the Earth at 3.8 centimetres a year. This is not a new development and it does not pose any problem for the lunar probe or communication with it. A signal from the Earth will still reach the Moon at the speed of light viz., 1.27 seconds, as the average distance to be covered would be 384, 000km or about 30 Earth diameters.
It has become possible to observe the small increase in the distance of the Moon’s orbit accurately because the astronauts of Apollo 11, 14 and 15 missions had left retro- reflectors on the lunar surface. The reflectors are still in good shape, as there is no air or pollution there. The panels reflect the incoming laser light sent from the Earth in the same direction they came from. Each of these panels has 100 corner cubes placed in a 46cm square aluminium base. It is a key to measure the distance to the Moon, as the time taken by light to reach it is known. It has been observed that the light beam widens to 20 km by the time it returns from the Moon. The measurement has been fine-tuned over the years thanks to advances in computer technology. Originally developed by the Jet Propulsion Laboratory in California, several researchers led by Dell Agnello of Italy have contributed to its improvement.
Successful lunar laser range measurements to the retro-reflectors were first reported by the 3.1 m telescope at Lick Observatory, (U.S.) Air Force Cambridge Research Laboratories, Lunar Ranging Observatory in Arizona, thePic du Midi Observatory in France, the Tokyo Astronomical Observatory, and McDonald Observatory in Texas. Similar reflectors are proposed to be placed on the Moon and Mars by NASA and the European Space Agency in 2020.
Chandrayaan-2 too has taken a passive laser retro-reflector to be placed on the Moon, but it is accessible from the lunar orbit. The object is to better understand the dynamics of the Earth- Moon system and derive clues on the lunar interior, including possible presence of a liquid core.
As exploratory satellites go beyond the Moon, the signals will take longer and longer time to reach and return. A round trip signal to Mars will take anywhere between 8 and 48 minutes, depending on the distance of the red planet from the Earth. A signal to Jupiter will take one and half hours and to Saturn, two and a half hours. A signal from NASA’s Voyager-2, (launched in 1977) now some 18 billion (yes, billion!) kilometres away, takes 16.5 hours to travel to the Earth. In contrast, light from the Sun reaches us in just about 8 minutes!
Accurate timing is essential to reach planets. A clock that is off by even a single second could mean the difference between landing on Mars and missing it by miles. As the distance between the Earth and exploratory satellites increases, real- time communication with them would come to a halt. They would not be able to know where they are in space and where they are moving to, as signals sent from the Earth would become irrelevant by the time they reach. It would be impossible to deal with unexpected situations and do even simple course corrections. They do not have a ring of satellites in space to provide them coordinates, as given to us by the Earth -orbiting satellites. One basic way of addressing the problem is to provide the satellites with hyper accurate atomic clocks to help them get their coordinates in real-time, without depending on Earth-bound clocks.
On the Earth, navigation satellites like those of the Global Positioning System (GPS) provide exact data that tell us where we are and where we should go to reach our destination. India too has its own navigation satellite system consisting of seven satellites. They all need corrections every day, because the signals which take slightly different time to reach the destination and are also distorted by the atmosphere. Their success basically depends on the accuracy of the atomic clocks they carry. Imported rubidium-based-atomic clocks used on the Indian Navigation satellites malfunctioned some time ago. India has since developed and made its own atomic clocks.
Why atomic clocks, one may ask. The reason is simple. They are super accurate as they are based on the unchanging fluctuations inside atoms. Even the second on the Earth is defined (since 1967) as exactly 91 92 631 770 periods of radiation corresponding to the transition between two hyperfine levels of the ground state of caesium-133 atom. A clock based on such accuracy will be far more stable than a quartz watch, which will be off by a millisecond (a thousandth of a second) in just six weeks. A rubidium-based clock ticks 385 billion times a second! Still better, a clock based on mercury ions will lose a second once in 400 million years. Such a clock is now in Earth orbit. It is called Deep Space Atomic Clock, developed by JPL and its collaborators.
The demonstration clock was launched in June last by the American super rocket, Falcon X. The test flight of the clock is planned to last one year. The researchers say that the new atomic clock is 50 times more uniform than others now in use. They hope to confirm that the clock will show a delay of just 0.2 nano seconds (less than one part in a billion). Unimaginable!
Moreover, the size of the clock is that of a toaster. It weighs only 16 kg and needs 47W of power. Further reductions are planned in future models, so that they can easily be taken by satellites.
Hyper accuracy of clocks will enable satellites in deep space exploration to function as autonomous units. They would then be able to have artificial intelligence (AI) to guide them. Is the world getting ready to send such robots into space? The answer is a definite yes.
AI Solves Rubik’s Cube puzzle in record time
An AI system has recently created a world record, when it solved the Rubik’s Cube puzzle in just over a second, smashing the current human record by more than two seconds. It is a unique achievement because it learnt all by itself and did not mimic the human brain.
The six--sided Rubik’s Cube puzzle was invented in 1974 by Ernő Rubik, a Hungarian Professor, to teach his students spatial relations in geometry. With six sides representing nine blocks of a single color—orange, yellow, green, red, white, and blue—a Rubik’s is said to hold 43 quintillion potential configurations. That’s 43,000,000,000,000,000,000 possible ways of combinations! The challenge is to combine the blocks in such a way that each of the cube’s six sides displays a single colour, in the shortest time.
The AI system, developed by the University of California at Irvine solved the puzzle without prior knowledge of the game or coaching from human handlers. Billions of potential moves were offered to the AI system, requiring it to reach the goal. The goal was: each of the cube’s six sides should display a single colour within 30 moves. It reached the goal in 20 moves. Human records in the 1980s were less than 15 seconds and now it is down to 3.5 seconds , according to World Cube Association.
The California algorithm is not the first or the fastest non-human to solve the puzzle. That honour goes to a system devised at the Massachusetts Institute of Technology, dubbed the min2phase algorithm, which solved the puzzle three times faster. In 2018, researchers built a robot that solved the puzzle in 0.38 seconds. But that system did not use a neural network - which mimics how the human brain works - or machine learning techniques. It was programmed just to solve the puzzle.
The significance of the latest record is that it marks the creation of a system that teaches itself to move beyond games to solve real-world problems.
Pierre Baldi, a professor of computer science, says “The solution to the Rubik’s Cube involves more symbolic, mathematical and abstract thinking, so a deep learning machine that can crack such a puzzle is getting closer to becoming a system that can think, reason and plan.”
With such an AI system and an improved atomic clock that gives coordinates on board, the dawn of a new era of space exploration by intelligent robots will not remain science fiction for long
(The writer is the author of Space Today published by NBT.)