News classification
Contact us
- Add: No. 9, North Fourth Ring Road, Haidian District, Beijing. It mainly includes face recognition, living detection, ID card recognition, bank card recognition, business card recognition, license plate recognition, OCR recognition, and intelligent recognition technology.
- Tel: 13146317170 廖经理
- Fax:
- Email: 398017534@qq.com
Artificial intelligence changes space exploration, plans to Mars
Artificial intelligence changes space exploration, plans to Mars
Mars 2020 is an ambitious plan. NASA plans to use 1.25 years or 28 Mars, Mars Earth month collects 20 cores and soil samples without artificial intelligence, this task can not be completed, because the probe will waste too much time waiting for the instruction.
At present, NASA's jet propulsion laboratory team planning "Mars Science Laboratory Rover Mars daily activities, and then through the NASA" deep space network "to send instructions to spend about eight hours. The project leader will tell when the shuttle car should get up, how long it takes to warm up its equipment, and how to avoid the rocks that damage its already battered wheels.
Mars 2020 will need it to have more autonomy. "The mission is based on the ground cycle," said the head of the Mars Science Laboratory, the Mission Director of the Mars Science laboratory." The more it can do it, the better it will do."
Beyond the HAL that never makes mistakes
The $2 billion 400 million Mars 2020 mission is just one example of the growing need for artificial intelligence in NASA, although the term itself is unsettling for some. Many NASA scientists and engineers prefer to call it machine learning, not artificial intelligence. In space, this broad term is sometimes reminiscent of Arthur C. Clarke, the fictional computer HAL9000 in Space Odyssey 2001, which claims to have a perfect record and never makes mistakes.
It's important to note that NASA is not trying to create HAL. Engineers are developing software and algorithms to meet the specific requirements of the task.
Senior scientist, the field of automatic system of the NASA Ames research center of intelligent robot team leader Kelly Fong said: "what we do today is not the focus of attention to intelligence issues, but to make the system more independent, more automatic and more independent."
For manned spaceflight, it provides software for astronauts to help them cope with emergencies such as equipment failures and emergency medical treatment. For example, a combination of reasoning and learning algorithm with data mining of medical aid, help astronauts to perform multiple tasks in a month on Mars, processing from the daily care of illness or injury to all sorts of things - "no longer need to talk to a lot of flight control staff to negotiate," Fong said.
With robotic missions to Mars, NASA is showing its increasingly powerful detector. While NASA's Mars probe, the spirit and opportunity, landed on Mars in 2004, it was still very limited despite its autonomy through software upgrades. By contrast, "curiosity" is obviously much better.
Last year, curiosity began to use a technique called "scientific knowledge collection independent exploration" software, computer vision and machine learning together, according to the scientists to determine the criteria for selection of rock and soil samples. The Mars vehicle can shoot its target with a laser of its chemical camera, analyze the burning gas, pack images and data, and send it back to earth.
"Scientists are responsible for this task are excited, because in the past they had to watch video, select the target, and then wait for the data to send commands," researcher Jiri Wagstaff JPL of machine learning and independent instrument group said.
Although data can only be sent back from Mars to earth in 10 to 30 minutes, mission controllers can send and receive data only on the time they are allocated on the deep space network.
"Even if the detector can talk to us all day, we shouldn't be answering it all day long, can we?"" Wagstaff says. "We have only two times a day, it's time to talk to a 10 minute, because of the deep space network is busy contact signal receiving Cassini, Voyager, pioneer, new horizons and all other tasks."
The goal of the Mars 2020 project's probe design is to do more alone to better communicate with mission managers. In the process to send to its task, and if there is surplus electricity to do some chores before it woke myself up, the instrument is heated to a suitable temperature.
"Ideally, that's good for us.". What we want is the image of the target and the data from the instrument. Gather these and contact us, and then we'll use the information to guide it to get a sample." Tero J Per says.
NASA has yet to achieve this goal, but the Mars 2020 project prompted them to improve the software in this direction, so that the detector on the Martian surface can move from one point to another point, while avoiding obstacles. "It's like the basic skills of toddlers, don't hit obstacles." He said. This autonomy is increasingly being added to our space program. Looking to the future, I think we will use these smart technologies more and more.
Future missions, such as NASA's Europa spacecraft (Europa Clipper), artificial intelligence will need more powerful, to find the plumes rising from underground oceans, and caused by hydrothermal vent ice cracks on the surface of the moon. When scientists are unable to predict when or where they can make new discoveries, they need artificial intelligence to "watch, watch, collect data, and send data back," Wagstaff says.