Current Research on a New Spacecraft Autonomy Architecture named Remote Agent

Remote Agent Part 1: Introduction of Spacecraft Autonomy In Part 1 of this two-part series, we will introduce current research on a new spacecraft autonomous architecture called remote agent applying artificial intelligence technology. It is for the control function of the spacecraft. The need for autonomous technology far exceeds that of deep space missions. Remote Agent is a model-based autonomous system developed by Jet Propulsion Laboratory and Ames Research Center of National Aeronautics and Space Administration.

"NASA's remote agent is the first onboard master plan to control the spacecraft's runtime schedule.The remote agent generates a plan based on the upper goal specified on the ground, I will monitor the operation of the spacecraft and recover. "Artificial intelligence gave birth to early enthusiasm and expectation for realism. It first responds to some of the most basic questions about human existence by understanding the nature of intelligence. But it has evolved into the field of science and technology that affect many aspects of business, business and society. As Russell and Novig state, the commercial success of applications such as expert systems is driving the development of the artificial intelligence industry. "Overall, the industry has grown from several millions of dollars in 1980 to tens of dollars in 1988. Billions of dollars"

Recent research on landing spacecraft focuses on developing algorithms that increase the level of autonomy in aerospace systems. Some of the main problems of spacecraft or rocket landings include vacuum stage, software error, guidance, and sensor problems. Machine learning and computer vision are core optimization and evaluation methods for successful landing. SpaceX Falcon 9 successfully landed at the Cape Canaveral Air Force Base in 2015 and demonstrated the machine learning and computer vision ability to change the space exploration. SpaceX uses the convex optimization algorithm to determine the best way to land a rocket and real - time computer vision data facilitates route prediction. These sophisticated machine learning applications enable the first reusable rocket in space exploration history - scientists believe this is a prerequisite for developing deep space exploration