東華大學圖書館 |

Language: English

Help

回圖書館首頁

手機版館藏查詢

Back to Search results for [ null ]

Switch To: Labeled | MARC Mode | ISBD

Evolving Rule Based Explainable Arti...

Keneni, Blen M.

Linked to FindBook

Google Book

Amazon

博客來

Evolving Rule Based Explainable Artificial Intelligence for Decision Support System of Unmanned Aerial Vehicles.

Record Type:	Electronic resources : Monograph/item
Title/Author:	Evolving Rule Based Explainable Artificial Intelligence for Decision Support System of Unmanned Aerial Vehicles./
Author:	Keneni, Blen M.
Published:	Ann Arbor : ProQuest Dissertations & Theses, : 2018,
Description:	113 p.
Notes:	Source: Masters Abstracts International, Volume: 80-07.
Contained By:	Masters Abstracts International80-07.
Subject:	Computer Engineering. -
Online resource:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=13819496
ISBN:	9780438773196

Evolving Rule Based Explainable Artificial Intelligence for Decision Support System of Unmanned Aerial Vehicles.
Keneni, Blen M.

Evolving Rule Based Explainable Artificial Intelligence for Decision Support System of Unmanned Aerial Vehicles. - Ann Arbor : ProQuest Dissertations & Theses, 2018 - 113 p.

Source: Masters Abstracts International, Volume: 80-07.

Thesis (M.S.)--The University of Toledo, 2018.

This item must not be sold to any third party vendors.

UAVs are used for many purposes including agriculture, industry, law enforcement, and defense. These autonomous systems have several advantages over manned aerial vehicles as not only they reduce expenses by avoiding human error, but they also save the lives of fighter jet pilots. Nowadays black-box machine learning algorithms are used to train unmanned vehicles to make decisions on their own. However, while these techniques give good predictive abilities, they fail to provide the reasoning behind decisions, thus rendering them untrustworthy. To address that concern, in this thesis, an intelligent rule based model that explains the logic behind the decisions of a UAV while it is on a predefined mission, has been developed. An effective XAI should be able to deliver explanation with high level of accuracy, handle uncertainty, and learn from experience. To address these points and provide meticulous explanation, this thesis utilizes a hybrid learning technique that combines explanation ability of Fuzzy logic which incorporates uncertainty with learning abilities of nature inspired Artificial Neural Networks. Before developing an explainable artificial intelligence (XAI), first model of UAV missions are created using Mamdani fuzzy inference system (FIS). Various patterns of paths for UAV mission are defined. On each path, weather conditions and enemies are placed at random locations. During a mission, UAV navigates through these predefined paths taking into consideration adverse weather patterns and its distance from a nearby enemy. UAV deviates from the predefined path and engages in attacking an enemy when the conditions demand. Data is gathered regarding the actual route the UAV took under those weather and enemy conditions and the actions it engaged in while traversing the planned route. The data gathered from UAV missions is used to create a reverse model. The model is Sugeno type fuzzy inference system based on subtractive clustering. It has seven inputs (time, x-coordinate, y-coordinate, heading direction, engage in attack, continue mission, steer UAV); and two outputs (weather conditions and distance from enemy). Then ANFIS is used to train the Sugeno fuzzy model. Fuzzy rules of Sugeno type in rule view window provide the XAI in a visual format. The rules provide explanation by illustrating the episodes that led to why UAV deviated from the planned route, engaged in attacking an enemy, or continued mission even though it has detected a nearby enemy. The predictive accuracy of the model is computed in terms of Root Mean Square Error (RMSE) of actual weather pattern and predicted weather pattern as well as the actual distance from enemy and predicted distance from enemy. In addition, the accuracy percentage is calculated by defining a threshold RMSE to calculate percentage error. Furthermore, to check the robustness of the model, Gaussian random noise is added to a UAV path and the prediction accuracy is validated. The validity of XAI is cross checked by visualizing the UAV mission data in parallel coordinates.

ISBN: 9780438773196Subjects--Topical Terms:

1567821
Computer Engineering.
Subjects--Index Terms:

ANFIS

Evolving Rule Based Explainable Artificial Intelligence for Decision Support System of Unmanned Aerial Vehicles.
LDR:04477nmm a2200409 4500 001 2271136
005 20201007134525.5
008 220629s2018 ||||||||||||||||| ||eng d
020 $a 9780438773196
035 $a (MiAaPQ)AAI13819496
035 $a (MiAaPQ)OhioLINK:toledo1525094091882295
035 $a AAI13819496
040 $a MiAaPQ $c MiAaPQ
100 1 $a Keneni, Blen M. $3 3548546
245 1 0 $a Evolving Rule Based Explainable Artificial Intelligence for Decision Support System of Unmanned Aerial Vehicles.
260 1 $a Ann Arbor : $b ProQuest Dissertations & Theses, $c 2018
300 $a 113 p.
500 $a Source: Masters Abstracts International, Volume: 80-07.
500 $a Publisher info.: Dissertation/Thesis.
500 $a Advisor: Kaur, Devinder.
502 $a Thesis (M.S.)--The University of Toledo, 2018.
506 $a This item must not be sold to any third party vendors.
506 $a This item must not be added to any third party search indexes.
520 $a UAVs are used for many purposes including agriculture, industry, law enforcement, and defense. These autonomous systems have several advantages over manned aerial vehicles as not only they reduce expenses by avoiding human error, but they also save the lives of fighter jet pilots. Nowadays black-box machine learning algorithms are used to train unmanned vehicles to make decisions on their own. However, while these techniques give good predictive abilities, they fail to provide the reasoning behind decisions, thus rendering them untrustworthy. To address that concern, in this thesis, an intelligent rule based model that explains the logic behind the decisions of a UAV while it is on a predefined mission, has been developed. An effective XAI should be able to deliver explanation with high level of accuracy, handle uncertainty, and learn from experience. To address these points and provide meticulous explanation, this thesis utilizes a hybrid learning technique that combines explanation ability of Fuzzy logic which incorporates uncertainty with learning abilities of nature inspired Artificial Neural Networks. Before developing an explainable artificial intelligence (XAI), first model of UAV missions are created using Mamdani fuzzy inference system (FIS). Various patterns of paths for UAV mission are defined. On each path, weather conditions and enemies are placed at random locations. During a mission, UAV navigates through these predefined paths taking into consideration adverse weather patterns and its distance from a nearby enemy. UAV deviates from the predefined path and engages in attacking an enemy when the conditions demand. Data is gathered regarding the actual route the UAV took under those weather and enemy conditions and the actions it engaged in while traversing the planned route. The data gathered from UAV missions is used to create a reverse model. The model is Sugeno type fuzzy inference system based on subtractive clustering. It has seven inputs (time, x-coordinate, y-coordinate, heading direction, engage in attack, continue mission, steer UAV); and two outputs (weather conditions and distance from enemy). Then ANFIS is used to train the Sugeno fuzzy model. Fuzzy rules of Sugeno type in rule view window provide the XAI in a visual format. The rules provide explanation by illustrating the episodes that led to why UAV deviated from the planned route, engaged in attacking an enemy, or continued mission even though it has detected a nearby enemy. The predictive accuracy of the model is computed in terms of Root Mean Square Error (RMSE) of actual weather pattern and predicted weather pattern as well as the actual distance from enemy and predicted distance from enemy. In addition, the accuracy percentage is calculated by defining a threshold RMSE to calculate percentage error. Furthermore, to check the robustness of the model, Gaussian random noise is added to a UAV path and the prediction accuracy is validated. The validity of XAI is cross checked by visualizing the UAV mission data in parallel coordinates.
590 $a School code: 0232.
650 4 $a Computer Engineering. $3 1567821
650 4 $a Electrical engineering. $3 649834
653 $a ANFIS
653 $a Explainable artificial intelligence
653 $a Fuzzy logic
653 $a Parallel coordinates
653 $a Subtractive clustering
653 $a Unmanned aerial vehicle
690 $a 0464
690 $a 0544
710 2 $a The University of Toledo. $b Electrical Engineering. $3 3170575
773 0 $t Masters Abstracts International $g 80-07.
790 $a 0232
791 $a M.S.
792 $a 2018
793 $a English
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=13819496