Self-Taught Learning for Activity Spotting in On-body Motion Sensor Data

2011 ◽  
Author(s):  
Oliver Amft
Keyword(s):  
Body Motion ◽  
Sensor Data ◽  
Motion Sensor

Data processing techniques of hydraulic bed sensor and doppler radar for elder care

2017 ◽  
Author(s):  
◽  
Bo-Yu Su
Keyword(s):  
Heart Rate ◽  
Doppler Radar ◽  
Detection Method ◽  
Blood Pressure Measurement ◽  
Fall Detection ◽  
The United States ◽  
Heart Beat ◽  
Body Motion ◽  
Sensor Data ◽  
Heart Rate Monitoring

[ACCESS RESTRICTED TO THE UNIVERSITY OF MISSOURI AT AUTHOR'S REQUEST.] Population aging is a common phenomenon in a society. The developed country like the United States, eldercare is becoming an important issue nowadays. There are many aspects we need to address for eldercare, including - circulatory system, alimentary system, nervous system and so on. In this research study, we focus on the heart rate monitoring and estimation using a hydraulic bed sensor. In addition, we also develop the fall detection technique using a Doppler radar. The hydraulic bed sensor for heart rate monitoring is placed under the mattress. The sensor system contains four tubes filled with water and uses the pressure sensor to obtain the Ballistocardiogram (BCG) signal. The BCG signal contains the information of heart beat, respiratory rate and body motion. Two algorithms are developed to process the bed sensor data. One uses the Hilbert transform and the other is based on the energy. By using the algorithms we developed, we can extract the heart beat information to estimate the heart rate. The system has been validated in a well controlled lab environment and a nursing house. In addition to the heart rate, the relative blood pressure measurement by using two features extracted from the bed sensor signal has also been developed and validated with 48 people data. The results show high correlation coefficient with the groundtruth. The Doppler radar for human fall detection is mounted in the ceiling. The radar senses the motion of an object and produces outputs based on the Doppler shift effect. We propose an effective method based on Wavelet Transform (WT) for fall vs. nonfall classification. The proposed fall detection classi er can distinguish between the fall and daily activities. The good performance of the proposed detection method has been validated through the data from the lab and in-home environments, with the falls from stunt actors and senior residents. To further improve the performance, we introduce an additional radar mounted on the wall. Based on the same detection method as when using one radar, we extract and concatenate the features from two radars for classification. The result shows outstanding improvement.

scholarly journals A Hierarchical Deep Fusion Framework for Egocentric Activity Recognition Using a Wearable Hybrid Sensor System

Sensors ◽  
2019 ◽  
Vol 19 (3) ◽  
pp. 546 ◽  
Author(s):  
Haibin Yu ◽  
Guoxiong Pan ◽  
Mian Pan ◽  
Chong Li ◽  
Wenyan Jia ◽  
...  
Keyword(s):  
Activity Recognition ◽  
Short Term Memory ◽  
Specific Activity ◽  
Negative Influence ◽  
Sensor System ◽  
Sensor Data ◽  
Motion Sensor ◽  
Motion Sensors ◽  
Motion State ◽  
Fusion Framework

Recently, egocentric activity recognition has attracted considerable attention in the pattern recognition and artificial intelligence communities because of its wide applicability in medical care, smart homes, and security monitoring. In this study, we developed and implemented a deep-learning-based hierarchical fusion framework for the recognition of egocentric activities of daily living (ADLs) in a wearable hybrid sensor system comprising motion sensors and cameras. Long short-term memory (LSTM) and a convolutional neural network are used to perform egocentric ADL recognition based on motion sensor data and photo streaming in different layers, respectively. The motion sensor data are used solely for activity classification according to motion state, while the photo stream is used for further specific activity recognition in the motion state groups. Thus, both motion sensor data and photo stream work in their most suitable classification mode to significantly reduce the negative influence of sensor differences on the fusion results. Experimental results show that the proposed method not only is more accurate than the existing direct fusion method (by up to 6%) but also avoids the time-consuming computation of optical flow in the existing method, which makes the proposed algorithm less complex and more suitable for practical application.

scholarly journals Perancangan Sistem Efisiensi Penggunaan Energi Listrik Menggunakan Sensor Gerak dan Sensor Arus

2020 ◽  
Vol 4 (3) ◽  
pp. 626
Author(s):  
Sarmayanta Sembiring ◽  
Hadir Kaban ◽  
Rido Zulfahmi
Keyword(s):  
Human Movement ◽  
Electrical Energy ◽  
Electronic Equipment ◽  
Sensor Data ◽  
Current Sensor ◽  
Motion Sensor ◽  
Air Conditioners ◽  
Average Current ◽  
Sensor Current ◽  
Predetermined Time

Efficiency system in using electrical energy has been designed using a PIR motion sensor, current sensor SCT-013-030, infrared LED and relay with a controller using Arduino Uno. The system is designed to turn off electronic equipment such as air conditioners, projectors and lights automatically as a solution from users forgetting to turn off electronic equipment when it is no longer in use. The experimental results show that the system has been running well, where the system can detect no movement for a predetermined time by using a PIR motion sensor. Detection of electronic equipment using sensors SCT-013-030 has been able to distinguish the state of the equipment whether it is ON or OFF based on differences in sensor output data that is read by the Arduino analog port. Sensor data when detecting the lamp when OFF is average = 1.333 while the mini projector and TV when off the average sensor data value = 1.667. The average current sensor data when detecting lights when ON = 5,333, mini projector = 8,333 and TV = 11,333. Overall the system designed has been able to turn off the equipment that is still active when the sensor does not detect any human movement during a predetermined time

scholarly journals Scalable Polyhedral Verification of Recurrent Neural Networks

2021 ◽  
pp. 225-248
Author(s):  
Wonryong Ryou ◽  
Jiayu Chen ◽  
Mislav Balunovic ◽  
Gagandeep Singh ◽  
Andrei Dan ◽  
...  
Keyword(s):  
Neural Networks ◽  
Computer Vision ◽  
Recurrent Neural Networks ◽  
Sensor Data ◽  
Motion Sensor ◽  
Use Case ◽  
Prior Work ◽  
Non Linear ◽  
Speech Classification ◽  
Sampling Optimization

AbstractWe present a scalable and precise verifier for recurrent neural networks, called Prover based on two novel ideas: (i) a method to compute a set of polyhedral abstractions for the non-convex and non-linear recurrent update functions by combining sampling, optimization, and Fermat’s theorem, and (ii) a gradient descent based algorithm for abstraction refinement guided by the certification problem that combines multiple abstractions for each neuron. Using Prover, we present the first study of certifying a non-trivial use case of recurrent neural networks, namely speech classification. To achieve this, we additionally develop custom abstractions for the non-linear speech preprocessing pipeline. Our evaluation shows that Prover successfully verifies several challenging recurrent models in computer vision, speech, and motion sensor data classification beyond the reach of prior work.

Sign in / Sign up

Export Citation Format

Share Document