Classifying Human Body Acceleration Patterns Using a Hierarchical Temporal Memory

Sassi, Federico; Ascari, Luca; Cagnoni, Stefano

doi:10.1007/978-3-642-10291-2_50

Cited by 4 publications

(1 citation statement)

References 7 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…HTM is a recently developed machine learning technology that aims to capture the structural and algorithmic properties of the neocortex in human brain. 27 It has been applied to classify human body acceleration patterns, 28 vision-based hand shape, 29 remote gaze gesture, 30 and sign language. 31 In comparison with traditional ANNs, HTM not only has better self-adaptability, higher learning efficiency, and lower requirements for the number of samples but also can recognize complicated patterns with strong noise.…”

Section: Hierarchical Temporal Memorymentioning

confidence: 99%

Integrative intrinsic time-scale decomposition and hierarchical temporal memory approach to gearbox diagnosis under variable operating conditions

Duan

Yao

Wang

et al. 2016

Advances in Mechanical Engineering

View full text Add to dashboard Cite

Gearbox diagnosis under stationary operating conditions has been extensively investigated; however, variable operating conditions such as load and speed changes play important roles in affecting the accuracy of gearbox diagnosis. This article presents an integrative approach of intrinsic time-scale decomposition and hierarchical temporal memory for gearbox diagnosis under variable operating conditions. A total of two modules are emphasized including a feature extraction method and an integrative feature fusion and classification model. Intrinsic time-scale decomposition method is investigated to extract the gearbox features which are insensitive to variable operating conditions, and its performance overcomes the commonly used empirical mode decomposition in terms of decomposition result and computational efficiency. Hierarchical temporal memory integrates feature fusion and pattern classification in one model to autonomously diagnose gearbox defect. Performance comparison among the presented method, back-propagation neural network, support vector machine, and fuzzy c-means clustering using experimental data demonstrate the effectiveness of the presented method.

show abstract