No abstract

Deep learning has seen tremendous growth over the past decade. It has set new performance limits for a wide range of applications, including computer vision, speech recognition, and machinery health monitoring. With the abundance of instrumentation data and the availability of high computational power, deep learning continues to prove itself as an efficient tool for the extraction of micropatterns from machinery big data repositories. This study presents a comparative study for feature extraction capabilities using stacked autoencoders considering the use of expert domain knowledge. Case Western Reserve University bearing dataset was used for the study, and a classifier was trained and tested to extract and visualize features from 12 different failure classes. Based on the raw data preprocessing, four different deep neural network structures were studied. Results indicated that integrating domain knowledge with deep learning techniques improved feature extraction capabilities and reduced the deep neural networks size and computational requirements without the need for exhaustive deep neural networks architecture tuning and modification.

No abstract

No abstract

Aircraft prototyping and modeling is usually associated with resource expensive techniques and significant post-flight analysis. The NASA Learn-To-Fly concept targets the replacement of the conventional ground-based aircraft development and prototyping approaches with an efficient real-time paradigm. The work presented herein describes a learning paradigm of a quadcopter unmanned aircraft that utilizes real-time flight data. Closed-loop parameter estimation of a highly collinear model terms such as those found on a quadrotor is challenging. Using phase optimized orthogonal multisine input maneuvers, collinearity of flight data decreases leading to fast and accurate convergence of the Fourier transform regression estimator. The generated models are utilized to reconfigure a nonlinear dynamic inversion controller in normal, failure, and learning testing conditions. Results show highly accurate model estimation in different testing scenarios. Additionally, the nonlinear dynamic inversion controller easily integrates the identified model parameters without any need for gain scheduling or computationally expensive methods. Overall, the proposed technique introduces an efficient integration between real-time modeling and control adaptation utilizing the limited computational power of the quadcopter’s microcomputer.