A wavelet tensor fuzzy clustering scheme for multi-sensor human activity recognition

He, Hong; Tan, Yonghong; Zhang, Wuxiong

doi:10.1016/j.engappai.2018.01.004

Cited by 34 publications

(30 citation statements)

References 37 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The distance function in (12) of NNCHTC can be transformed into the inner product calculation of the tensors.…”

Section: B the Proposed Nnchtc Methodsmentioning

confidence: 99%

“…Although WPT has been widely used in the decomposition and feature extraction of gear vibration signals, it is only used to extract the feature from single channel vibration signals or the signals of a triaxial accelerometer. He et al [12] have successfully used the DWPT method to extract features of multi-sensor signals, and then used the features represented as tensors to realize human activity recognition. Therefore, in the field of gear intelligent fault diagnosis, it is of great significance to further apply the WPT method to extract feature tensor from multi-channel signals of diverse sensors.…”

Section: Related Work a Feature Tensor Expressionmentioning

confidence: 99%

“…) is easily affected by the outliers. From (12), it can be found that d 2 h (X , CH(X h )) is mainly affected by the coefficient α i of the outliers. Therefore, when there are outliers in the training samples, in order to improve the robustness of the NNCHTC method, reduction factor µ is introduced into the NNCHTC method.…”

Section: Tensor Training Sample Set and C H (Cmentioning

confidence: 99%

“…However, when the raw signals are from diverse sensors or multi-channels of a sensor, the tensor expression of the features can well reflect the inherent structural information of the raw signals. In [12], He et al proved that the feature tensors extracted from the multi-sensor can be more effectively used for classification. Therefore, if the raw signals in gear faults intelligent diagnosis are from multi-sensor, the tensor expression of features from the raw signals can better contain the fault information of the gear.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Nearest Neighbor Convex Hull Tensor Classification for Gear Intelligent Fault Diagnosis Based on Multi-Sensor Signals

Cheng

Wang

2019

IEEE Access

View full text Add to dashboard Cite

For the feature tensor of multi-sensor signals classification problem in gear intelligent fault diagnosis, a new tensor classifier named nearest neighbor convex hull tensor classification (NNCHTC) is proposed in this paper. First, the convex hull distance from a test tensor sample to the convex hull is taken as the similarity measure for classification. Then, the convex hull distance calculation is transformed into the feature tensor inner product, and CANDECOMP/PARAFAC (CP) decomposition is applied to the calculation process to capture the intrinsic information of the feature tensor. Furthermore, the reduction factor is introduced into NNCHTC to enhance its robustness. Finally, feature tensors are obtained from multi-sensor signals by wavelet packet transform (WPT) and used to identify gear working condition by NNCHTC. The experimental results show that NNCHTC not only can be effectively applied to the gear intelligent fault diagnosis based on multi-sensor signals but also has better robustness.INDEX TERMS Gear intelligent fault diagnosis, feature tensor, multi-sensor signals, nearest neighbor convex hull tensor classification, reduction factor.

show abstract

“…The distance function in (12) of NNCHTC can be transformed into the inner product calculation of the tensors.…”

Section: B the Proposed Nnchtc Methodsmentioning

confidence: 99%

Section: Related Work a Feature Tensor Expressionmentioning

confidence: 99%

Section: Tensor Training Sample Set and C H (Cmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Nearest Neighbor Convex Hull Tensor Classification for Gear Intelligent Fault Diagnosis Based on Multi-Sensor Signals

Cheng

Wang

2019

IEEE Access

View full text Add to dashboard Cite

show abstract

“…HAR is essentially a pattern recognition problem, to which many traditional machine learning based approaches have been applied [7]. These traditional approaches rely on either expert-crafted features [8], [27], [28] or some self-learned shallow features to train the classifier. However, such features cannot sufficiently represent the latent characteristics of complex human activities [7] and usually become the performance bottleneck.…”

Section: Related Workmentioning

confidence: 99%

Baseline Model Training in Sensor-Based Human Activity Recognition: An Incremental Learning Approach

Xiao

Chen

et al. 2021

IEEE Access

View full text Add to dashboard Cite

Human activity recognition (HAR) based on wearable sensors has attracted significant research attention in recent years due to its advantages in availability, accuracy, and privacy-friendliness. HAR baseline model is essentially a general-purpose classifier trained to recognized multiple activity patterns of most user types. It provides the input for subsequent steps of model personalization. Training a good baseline model is of fundamental importance because it has significant impacts on the ultimate HAR accuracy. In practice, baseline model training in HAR is a non-trivial problem that faces two challenges: insufficient training data and biased training data. This paper proposes a novel baseline model training scheme to tackle the two challenges using Deep InfoMax (DIM)-based unsupervised feature extraction and Broad Learning System (BLS)-based incremental learning, respectively. Experimental results demonstrate that the proposed scheme outperform conventional methods in terms of overall accuracy, computational efficiency, and the ability to adapt to dynamic scenarios with changing data characteristics. INDEX TERMS Human activity recognition, baseline model, broad learning system, DIM, incremental learning.

show abstract

Smart Sensors for Activity Recognition

Rayan

Zafar

Rafique

et al. 2022

Intelligent Systems for Rehabilitation Engineering

View full text Add to dashboard Cite

A wavelet tensor fuzzy clustering scheme for multi-sensor human activity recognition

Cited by 34 publications

References 37 publications

Nearest Neighbor Convex Hull Tensor Classification for Gear Intelligent Fault Diagnosis Based on Multi-Sensor Signals

Nearest Neighbor Convex Hull Tensor Classification for Gear Intelligent Fault Diagnosis Based on Multi-Sensor Signals

Baseline Model Training in Sensor-Based Human Activity Recognition: An Incremental Learning Approach

Smart Sensors for Activity Recognition

Contact Info

Product

Resources

About