Skip to Main Content
 

Global Search Box

 
 
 
 

Files

File List


Final_Thesis.pdf (2.96 MB)

ETD Abstract Container

Abstract Header

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

Keneni, Blen M, Keneni
http://rave.ohiolink.edu/etdc/view?acc_num=toledo1525094091882295

Abstract Details

2018, Master of Science, University of Toledo, Electrical Engineering.
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.
Devinder Kaur (Committee Chair)
Vijay Devabhaktuni (Committee Co-Chair)
Ahmad Javaid (Committee Member)
Richard Molyet (Committee Member)
Computer Engineering; Electrical Engineering
Explainable Artificial Intelligence, fuzzy logic, ANFIS, Unmanned Aerial Vehicle, Subtractive Clustering, Parallel Coordinates

Recommended Citations

Citations

  • Keneni, Keneni, B. M. (2018). Evolving Rule Based Explainable Artificial Intelligence for Decision Support System of Unmanned Aerial Vehicles [Master's thesis, University of Toledo]. OhioLINK Electronic Theses and Dissertations Center. http://rave.ohiolink.edu/etdc/view?acc_num=toledo1525094091882295

    APA Style (7th edition)

  • Keneni, Keneni, Blen. Evolving Rule Based Explainable Artificial Intelligence for Decision Support System of Unmanned Aerial Vehicles. 2018. University of Toledo, Master's thesis. OhioLINK Electronic Theses and Dissertations Center, http://rave.ohiolink.edu/etdc/view?acc_num=toledo1525094091882295.

    MLA Style (8th edition)

  • Keneni, Keneni, Blen. "Evolving Rule Based Explainable Artificial Intelligence for Decision Support System of Unmanned Aerial Vehicles." Master's thesis, University of Toledo, 2018. http://rave.ohiolink.edu/etdc/view?acc_num=toledo1525094091882295

    Chicago Manual of Style (17th edition)

toledo1525094091882295
2,472
© , some rights reserved.Creative Commons License Evolving Rule Based Explainable Artificial Intelligence for Decision Support System of Unmanned Aerial Vehicles by Blen M Keneni Keneni is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 3.0 Unported License. Based on a work at etd.ohiolink.edu.
This open access ETD is published by University of Toledo and OhioLINK.