Title:
Small Body Reconnaissance by Multiple Spacecraft via Deep Reinforcement Learning
Small Body Reconnaissance by Multiple Spacecraft via Deep Reinforcement Learning
Files
Author(s)
Tomita, Kento
Shimane, Yuri
Ho, Koki
Shimane, Yuri
Ho, Koki
Advisor(s)
Editor(s)
Collections
Supplementary to
Permanent Link
Abstract
Small-body investigations by spacecraft are one of the most scientifically important
space exploration missions. Due to the strong uncertainty of the dynamics
around the body, geological surface features, and scientific values of candidate target
sites, these missions require dedicated planning and execution from the ground.
As a study of automated operations for asteroid investigation, this paper investigates
how small-body reconnaissance operations could be performed by multiple
spacecraft. By comparing baseline policies with different model parameters and a
policy trained via deep reinforcement learning, we discuss the optimal balance of
exploration and exploitation for our science model.
Sponsor
Date Issued
2022-08
Extent
Resource Type
Text
Resource Subtype
Paper
Rights Statement
Unless otherwise noted, all materials are protected under U.S. Copyright Law and all rights are reserved