Carnegie Mellon AI Collaborates With Pentagon To Improve Helicopter Reliability

       The researchers team of Carnegie Mellon University (CMU) with the Pentagon’s Joint Artificial

Intelligence Center (JAIC) are utilizing artificial intelligence procedures in improving the dependability and

availability of helicopters used by the U.S. Army's 160th Special Operations Aviation Regiment (SOAR), also

known as Night Stalkers.

‍

        On January 26, the Predictive

Maintenance (PMx) was considered

during an Armed Services Committee

meeting. Artur Dubrawski, Alumni

Research Professor of Computer Science

at CMU and director of the Auton Lab, said that this project aimed to employ

machine learning to recognize

circumstances which manifest the

remaining power within a few flight

hours.

‍

        For instance, when the engine overheats during power operation , the indicator shows the early

impending failure. However, Dubrawski, Kyle Miller and other lab members have evolved a variety of

consideration models, including engine pressures, temperatures and acceleration. Thus, in this mountain of

data to find some effective factors, machine learning algorithms can identify designs that can be effective as

early warning signs.

        
        According to an amount of analyzing flight data, including maintenance records and other references,

Dubrawski and his team have studied for identifying certain factors that can lead to problems. This analysis

has been evaluated at more than 100$ million to keep the maintenance staff away from all crises.

‍

       Furthermore, the Predictive Maintenance (PMx), a research team of Carnegie Mellon University,

investigated large amounts of data which contained relatively fewerneedles. With the huge flight and

maintenance report manifested healthy working conditions, the airplanes are carefully preserved.

‍

        Therefore, the finding of impending failure was a challenge in which machine learning techniques

require a lot of information, Dubrawski said. Moreover, this project requested many solutions, such as model

construction that incorporated physical principles of the power engine.

‍

      Although the JAIC project had ended in September, the Auton Lab is continuing work on the predictive

maintenance problem as a part of the Army AI Task Force headquartered at Carnegie Mellon University.

Source: https://www.cs.cmu.edu/news/carnegie-mellon-ai-collaborates-pentagon-improve-helicopter-reliability