The HAL 9000 supercomputer in the science fiction movie "2001: Space Odyssey" is best remembered for the cooling the way space explorers are killed. Researchers are currently working on AI as HAL to help space travelers without killing them and their model effectively controls a stimulated planetary base for quite some time.
In the masterpiece, the HAL 9000 is an AI controlling the Discovery One spacecraft. In the novel of Sci-Fi legend Arthur C. Clark, written alongside the film, HAL is described as suitable for conversation with space travelers "in the ideal idiomatic English he has learned in the short weeks of his childhood."
AI researcher Pete Bonasso at TRACLabs Inc. in Webster, Texas, for the first time observed "2001: Space Odyssey" in his senior year in West Point, where he programmed the institute's computer to play a virtual version of the pool.
"When I saw 2001, I realized I had to make the computer into another creature that resembles HAL 9000," he said in a study published online in Science Robotics.
The AI software, which Bonasso and its partners have already designed with the designation "Cognitive Architecture for Space Agents" or CASE, is made up of three key layers. The first is a sequential control layer that connects to and works with hardware such as robotic hands and eyes. In particular CASE controls the stimulation of a planetary base.
The second layer completes the procedures underlying the routine exercises. These include connecting battery power, controlling oxygen generation and carbon dioxide removal systems, and charging and sending rover to extract planetary scale samples.
The last layer consists of programmed planning software that chooses how to achieve CASE's programmed goals for the day and the order in which to execute them. It can also automatically re-adjust activities when problems arise, such as gas leaks, broken engines or planetary dust storms, Bonasso said.
Each of the three layers is linked to an ontology server that is a thorough and comprehensive database that can understand its information. For example, if someone moved a toolbox from the hardware cabinet to the neighborhoods of the team, the ontology server indicated that most of the tools in the cell would also change their location, the researchers said.
The software may display data, such as those relating to the maintenance of the robot's life and status, but may also speak with individuals so that they can ask questions, send commands, and be warned of whatever approaching problems. In order to maintain a circumstance like the one where the HAL has killed its space passengers, CASE is meant to make plans just after sharing them with individuals and getting consent to action.
"Our NASA partners and colleagues do not worry that our HAL can get out of power," said Bonasso at Space.com. "This is because it can not do anything that is not programmed to do."
In experiments, CASE is engaged in a stimulated planetary base for about 4 hours. Nonetheless, analysts have focused on more work that is needed before it can run a real base.
Also, "CASE is unbelievable, but not fully recognized HAL 2001: Cosmic Odyssey," nor is it a leader on Star Trek: the next generation, said Bonasso. "Its capacity is extremely limited, focusing on events taking place on a planetary basis, although it can keep the functioning of life support systems, there is no idea who will win the last presidential election."
Currently, researchers are working with what NASA calls analogues – places where volunteers pretend to live in a remote world. In the long run, the goal is to include CASE in the analogues to see how the software can help better future space expeditions.
Sometimes Bonas forgets how this work intends to create a real version of HAL for distant space travelers.
"When you're somewhere in the software, you forget from time to time that you really imagine what it looks like to live on Mars or on the moon," Bonasso said. "Now and again we have to retreat and say," Hello! This is absolutely amazing. "
Posts for posts: