A Stacked Human Activity Recognition Model Based on Parallel Recurrent Network and Time Series Evidence Theory
As the foundation of Posture Analysis, recognizing human activity accurately in real time assists in using machines to intellectualize living condition and monitor health status. In this paper, we focus on recognition based on raw time series data, which are continuously sampled by wearable sensors,...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2020-07-01
|
| Series: | Sensors |
| Subjects: | |
| Online Access: | https://www.mdpi.com/1424-8220/20/14/4016 |
| id |
doaj-eb9f9423cb0b4579af085e3900d1bf82 |
|---|---|
| record_format |
Article |
| spelling |
doaj-eb9f9423cb0b4579af085e3900d1bf822020-11-25T03:43:05ZengMDPI AGSensors1424-82202020-07-01204016401610.3390/s20144016A Stacked Human Activity Recognition Model Based on Parallel Recurrent Network and Time Series Evidence TheoryPeng Zhang0Zhenjiang Zhang1Han-Chieh Chao2Department of Electronic and Information Engineering, Key Laboratory of Communication and Information Systems, Beijing Municipal Commission of Education, Beijing Jiaotong University, Beijing 100044, ChinaDepartment of Electronic and Information Engineering, Key Laboratory of Communication and Information Systems, Beijing Municipal Commission of Education, Beijing Jiaotong University, Beijing 100044, ChinaDepartment of Electrical Engineering, National Dong Hwa University, Hualien 97401, TaiwanAs the foundation of Posture Analysis, recognizing human activity accurately in real time assists in using machines to intellectualize living condition and monitor health status. In this paper, we focus on recognition based on raw time series data, which are continuously sampled by wearable sensors, and a fine-grained evidence reasoning approach has been proposed to produce a timely and reliable result. First, the basic time unit of input data is selected by finding a tradeoff between accuracy and time cost. Then, the approach uses Long Short Term Memory to extract features and project raw multidimensional data into probability assignments, followed by trainable evidence combination and inference network that reduce uncertainly to improve the classification accuracy. Experiments validate the effectiveness of fine granularity and evidence reasoning while the final results indicate that the recognition accuracy of this approach can reach 96.4% with no additional complexity in training.https://www.mdpi.com/1424-8220/20/14/4016human activity recognitionrecurrent neural networktime series evidence theory |
| collection |
DOAJ |
| language |
English |
| format |
Article |
| sources |
DOAJ |
| author |
Peng Zhang Zhenjiang Zhang Han-Chieh Chao |
| spellingShingle |
Peng Zhang Zhenjiang Zhang Han-Chieh Chao A Stacked Human Activity Recognition Model Based on Parallel Recurrent Network and Time Series Evidence Theory Sensors human activity recognition recurrent neural network time series evidence theory |
| author_facet |
Peng Zhang Zhenjiang Zhang Han-Chieh Chao |
| author_sort |
Peng Zhang |
| title |
A Stacked Human Activity Recognition Model Based on Parallel Recurrent Network and Time Series Evidence Theory |
| title_short |
A Stacked Human Activity Recognition Model Based on Parallel Recurrent Network and Time Series Evidence Theory |
| title_full |
A Stacked Human Activity Recognition Model Based on Parallel Recurrent Network and Time Series Evidence Theory |
| title_fullStr |
A Stacked Human Activity Recognition Model Based on Parallel Recurrent Network and Time Series Evidence Theory |
| title_full_unstemmed |
A Stacked Human Activity Recognition Model Based on Parallel Recurrent Network and Time Series Evidence Theory |
| title_sort |
stacked human activity recognition model based on parallel recurrent network and time series evidence theory |
| publisher |
MDPI AG |
| series |
Sensors |
| issn |
1424-8220 |
| publishDate |
2020-07-01 |
| description |
As the foundation of Posture Analysis, recognizing human activity accurately in real time assists in using machines to intellectualize living condition and monitor health status. In this paper, we focus on recognition based on raw time series data, which are continuously sampled by wearable sensors, and a fine-grained evidence reasoning approach has been proposed to produce a timely and reliable result. First, the basic time unit of input data is selected by finding a tradeoff between accuracy and time cost. Then, the approach uses Long Short Term Memory to extract features and project raw multidimensional data into probability assignments, followed by trainable evidence combination and inference network that reduce uncertainly to improve the classification accuracy. Experiments validate the effectiveness of fine granularity and evidence reasoning while the final results indicate that the recognition accuracy of this approach can reach 96.4% with no additional complexity in training. |
| topic |
human activity recognition recurrent neural network time series evidence theory |
| url |
https://www.mdpi.com/1424-8220/20/14/4016 |
| work_keys_str_mv |
AT pengzhang astackedhumanactivityrecognitionmodelbasedonparallelrecurrentnetworkandtimeseriesevidencetheory AT zhenjiangzhang astackedhumanactivityrecognitionmodelbasedonparallelrecurrentnetworkandtimeseriesevidencetheory AT hanchiehchao astackedhumanactivityrecognitionmodelbasedonparallelrecurrentnetworkandtimeseriesevidencetheory AT pengzhang stackedhumanactivityrecognitionmodelbasedonparallelrecurrentnetworkandtimeseriesevidencetheory AT zhenjiangzhang stackedhumanactivityrecognitionmodelbasedonparallelrecurrentnetworkandtimeseriesevidencetheory AT hanchiehchao stackedhumanactivityrecognitionmodelbasedonparallelrecurrentnetworkandtimeseriesevidencetheory |
| _version_ |
1724521331377045504 |