Computational investigation of environment learning in guidance and navigation

Abhishek Verma, Berenice F Mettler May

Research output: Contribution to journalArticle

2 Citations (Scopus)

Abstract

This paper presents a computational framework for learning autonomous guidance behavior in unknown or partially known environments. The learning framework uses a receding horizon trajectory optimization associated with a spatial value function. The function describes optimal (for example, minimum time) guidance behavior represented as cost and velocity at any point in geographical space to reach a specified goal state. For guidance in unknown environments, a local spatial value function based on the current vehicle state is updated online using environment data from onboard exteroceptive sensors. The proposed learning framework has the advantage, in that it learns information directly relevant to the optimal guidance and control behavior, enabling optimal trajectory planning in unknown or partially known environments. The learning framework is evaluated by measuring performance over successive runs in three-dimensional indoor flight simulations. The test vehicle in the simulations is a Blade-Cx2 coaxial miniature helicopter. The environment is a priori unknown to the learning system. The paper investigates changes in performance, dynamic behavior, spatial value function, and control behavior in the body frame as a result of learning over successive runs.

Original languageEnglish (US)
Pages (from-to)371-389
Number of pages19
JournalJournal of Guidance, Control, and Dynamics
Volume40
Issue number2
DOIs
StatePublished - Jan 1 2017

Fingerprint

Learning Environment
navigation
learning
Guidance
Navigation
Value Function
Unknown
Trajectories
Trajectory Optimization
Flight Simulation
Trajectory Planning
flight simulation
trajectory
Helicopters
Optimal Trajectory
trajectory optimization
Coaxial
test vehicles
Helicopter
trajectory planning

Cite this

Computational investigation of environment learning in guidance and navigation. / Verma, Abhishek; Mettler May, Berenice F.

In: Journal of Guidance, Control, and Dynamics, Vol. 40, No. 2, 01.01.2017, p. 371-389.

Research output: Contribution to journalArticle

@article{b8b46dd4ddcd4baeb56e2b221dce40ca,
title = "Computational investigation of environment learning in guidance and navigation",
abstract = "This paper presents a computational framework for learning autonomous guidance behavior in unknown or partially known environments. The learning framework uses a receding horizon trajectory optimization associated with a spatial value function. The function describes optimal (for example, minimum time) guidance behavior represented as cost and velocity at any point in geographical space to reach a specified goal state. For guidance in unknown environments, a local spatial value function based on the current vehicle state is updated online using environment data from onboard exteroceptive sensors. The proposed learning framework has the advantage, in that it learns information directly relevant to the optimal guidance and control behavior, enabling optimal trajectory planning in unknown or partially known environments. The learning framework is evaluated by measuring performance over successive runs in three-dimensional indoor flight simulations. The test vehicle in the simulations is a Blade-Cx2 coaxial miniature helicopter. The environment is a priori unknown to the learning system. The paper investigates changes in performance, dynamic behavior, spatial value function, and control behavior in the body frame as a result of learning over successive runs.",
author = "Abhishek Verma and {Mettler May}, {Berenice F}",
year = "2017",
month = "1",
day = "1",
doi = "10.2514/1.G001889",
language = "English (US)",
volume = "40",
pages = "371--389",
journal = "Journal of Guidance, Control, and Dynamics",
issn = "0731-5090",
publisher = "American Institute of Aeronautics and Astronautics Inc. (AIAA)",
number = "2",

}

TY - JOUR

T1 - Computational investigation of environment learning in guidance and navigation

AU - Verma, Abhishek

AU - Mettler May, Berenice F

PY - 2017/1/1

Y1 - 2017/1/1

N2 - This paper presents a computational framework for learning autonomous guidance behavior in unknown or partially known environments. The learning framework uses a receding horizon trajectory optimization associated with a spatial value function. The function describes optimal (for example, minimum time) guidance behavior represented as cost and velocity at any point in geographical space to reach a specified goal state. For guidance in unknown environments, a local spatial value function based on the current vehicle state is updated online using environment data from onboard exteroceptive sensors. The proposed learning framework has the advantage, in that it learns information directly relevant to the optimal guidance and control behavior, enabling optimal trajectory planning in unknown or partially known environments. The learning framework is evaluated by measuring performance over successive runs in three-dimensional indoor flight simulations. The test vehicle in the simulations is a Blade-Cx2 coaxial miniature helicopter. The environment is a priori unknown to the learning system. The paper investigates changes in performance, dynamic behavior, spatial value function, and control behavior in the body frame as a result of learning over successive runs.

AB - This paper presents a computational framework for learning autonomous guidance behavior in unknown or partially known environments. The learning framework uses a receding horizon trajectory optimization associated with a spatial value function. The function describes optimal (for example, minimum time) guidance behavior represented as cost and velocity at any point in geographical space to reach a specified goal state. For guidance in unknown environments, a local spatial value function based on the current vehicle state is updated online using environment data from onboard exteroceptive sensors. The proposed learning framework has the advantage, in that it learns information directly relevant to the optimal guidance and control behavior, enabling optimal trajectory planning in unknown or partially known environments. The learning framework is evaluated by measuring performance over successive runs in three-dimensional indoor flight simulations. The test vehicle in the simulations is a Blade-Cx2 coaxial miniature helicopter. The environment is a priori unknown to the learning system. The paper investigates changes in performance, dynamic behavior, spatial value function, and control behavior in the body frame as a result of learning over successive runs.

UR - http://www.scopus.com/inward/record.url?scp=85014181879&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85014181879&partnerID=8YFLogxK

U2 - 10.2514/1.G001889

DO - 10.2514/1.G001889

M3 - Article

AN - SCOPUS:85014181879

VL - 40

SP - 371

EP - 389

JO - Journal of Guidance, Control, and Dynamics

JF - Journal of Guidance, Control, and Dynamics

SN - 0731-5090

IS - 2

ER -