Method for training aircraft control agent

An example includes a method (200) for training an agent (115) to control an aircraft (10A). The method includes: selecting (202), by the agent (115), first actions (Aa) for the aircraft to perform within a first environment (112A) respectively during first time intervals (Ta) based on first states...

Ausführliche Beschreibung

Gespeichert in:

Bibliographische Detailangaben
Hauptverfasser:	HUNG, Fan, KHOSLA, Deepak, SOLEYMAN, Sean, FADAIE, Joshua G, CHEN, Yang
Format:	Patent
Sprache:	eng
Schlagworte:	ADVERTISING APPLIANCES FOR TEACHING, OR COMMUNICATING WITH, THE BLIND,DEAF OR MUTE CRYPTOGRAPHY DIAGRAMS DISPLAY EDUCATION EDUCATIONAL OR DEMONSTRATION APPLIANCES GLOBES PHYSICS PLANETARIA SEALS
Online-Zugang:	Volltext bestellen
Tags:	Tag hinzufügen Keine Tags, Fügen Sie den ersten Tag hinzu!

Beschreibung
Zusammenfassung:	An example includes a method (200) for training an agent (115) to control an aircraft (10A). The method includes: selecting (202), by the agent (115), first actions (Aa) for the aircraft to perform within a first environment (112A) respectively during first time intervals (Ta) based on first states (Sa) of the first environment (112A) during the first time intervals (Ta), updating (204) the agent (115) based on first rewards (Ra) that correspond respectively to the first states (Sa), selecting (206), by the agent (115), second actions (AP) for the aircraft (10A) to perform within a second environment (112B) respectively during second time intervals (T) based on second states (SP) of the second environment (112B) during the second time intervals (T), and updating (208) the agent (115) based on second rewards (RP) that correspond respectively to the second states (SP). At least one first rule of the first environment (112A) is different from at least one rule of the second environment (112B). z zo