Organizations in many domains generate a considerable amount of heterogeneous data every day. Such data can be processed to enhance these organizations’ decisions in real…
Organizations in many domains generate a considerable amount of heterogeneous data every day. Such data can be processed to enhance these organizations’ decisions in real time. However, storing and processing large and varied datasets (known as big data) is challenging to do in real time. In machine learning, streaming feature selection has always been considered a superior technique for selecting the relevant subset features from highly dimensional data and thus reducing learning complexity. In the relevant literature, streaming feature selection refers to the features that arrive consecutively over time; despite a lack of exact figure on the number of features, numbers of instances are well-established. Many scholars in the field have proposed streaming-feature-selection algorithms in attempts to find the proper solution to this problem. This paper presents an exhaustive and methodological introduction of these techniques. This study provides a review of the traditional feature-selection algorithms and then scrutinizes the current algorithms that use streaming feature selection to determine their strengths and weaknesses. The survey also sheds light on the ongoing challenges in big-data research.
An artificial immune system (AIS) for the detection and identification of abnormal operational conditions affecting an unmanned air vehicle (UAV) is developed using the…
An artificial immune system (AIS) for the detection and identification of abnormal operational conditions affecting an unmanned air vehicle (UAV) is developed using the partition of the universe approach. The performance of the proposed methodology is assessed through simulation within the West Virginia University (WVU) unmanned aerial system (UAS) simulation environment.
An AIS is designed and generated for a fixed wing UAV using data from the WVU UAS simulator. A novel partition of the universe approach augmented with the hierarchical multiself strategy is used to define the self, within the AIS paradigm. Several 2-dimensional and 3-dimensional commanded trajectories are simulated under normal and abnormal conditions affecting actuators and sensors. Data recorded are used to build the AIS and develop an abnormal condition detection and identification scheme for the two categories of subsystems. The performance of the methodology is evaluated in terms of detection and identification rates, false alarms and decision times.
The proposed methodology for UAV abnormal condition detection and identification has the potential to support a comprehensive and integrated solution to the problem of aircraft subsystem health management. The novel partition of the universe approach has been proven to be a promising alternative to the previously investigated clustering methods by providing similar or better performance for the cases investigated.
The promising results obtained within this research effort motivate further investigation and extension of the proposed methodology toward a complete system health management process, including abnormal condition evaluation and accommodation.
The use of the partition of the universe approach for AIS generation may potentially represent a valuable alternative to current clustering methods within the AIS paradigm. It can facilitate a simpler and faster implementation of abnormal condition detection and identification schemes.
In this paper, a novel method for AIS generation, the partition of the universe approach, is formulated and applied for the first time for the development of abnormal condition detection and identification schemes for UAVs. This approach is computationally less expensive and mitigates some of the issues related to the typical clustering approaches. The implementation of the proposed approach can potentially enhance the robustness of UAS for safety purposes.