Accuracy Of Activity Detection Research Articles

State-of-the-art in human activity recognition based on inertial measurement unit sensors: survey and applications

Human activity recognition systems using wearable sensors is an important issue in pervasive computing, which applies to various domains related to healthcare, context aware and pervasive computing, sports, surveillance and monitoring, and the military. Three approaches can be considered for activity recognition: video sensor-based, physical sensor-based, and environmental sensor-based. This paper investigates the related work regarding the physical sensor-based approaches to motion processing. In this paper, a wide range of artificial intelligence models, from single classifications to methods based on deep learning, have been reviewed. The human activity detection accuracy of different methods, under natural and experimental conditions poses several challenges. These challenges cause problems regarding the accuracy and applicability of the proposed methods. This paper analyzes the methods, challenges, approaches, and future work. The goal of this paper is to establish a clear distinction in the field of motion detection using inertial sensors.

Attention-Based Multihead Deep Learning Framework for Online Activity Monitoring With Smartwatch Sensors

The expeditious propagation of Internet of Things (IoT) technologies implanted in different smart devices such as smartphones and smartwatches have a ubiquitous consequence on the modern population. These devices are employed to collect data and to aid in tracking and analyzing the users’ daily activities using various human activity monitoring and recognition (HAR) techniques. However, most current HAMR approaches rely on exploratory case-based shallow feature learning architectures, which endeavor to recognize activities correctly in real-world situations. To address this issue, we offer a unique strategy for HAMR that leverages the attention mechanism with multi-head convolutional neural networks (CNNs) and Long-Short-Term-Memory (LSTM). The accuracy of activity detection is improved in the presented method by integrating attention into multi-head CNNs followed by LSTM for better feature extraction and selection. Verification investigations are carried out using data from the University of California (UCI) repository, which is publicly available. The results show that our proposed framework is more accurate than current frameworks using both the 10-fold and leave-one-subject-out cross-validation. Finally, the proposed method can recognize human activity in real-time, regardless of the type of smart device.

Message Passing-Based Joint User Activity Detection and Channel Estimation for Temporally-Correlated Massive Access

This paper studies the user activity detection and channel estimation problem in a temporally-correlated massive access system where a very large number of users communicate with a base station sporadically and each user once activated can transmit with a large probability over multiple consecutive frames. We formulate the problem as a dynamic compressed sensing (DCS) problem to exploit both the sparsity and the temporal correlation of user activity. By leveraging the hybrid generalized approximate message passing (HyGAMP) framework, we design a computationally efficient algorithm, HyGAMP-DCS, to solve this problem. In contrast to only exploiting the historical estimations, the proposed algorithm performs bidirectional message passing between the neighboring frames for activity likelihood update to fully exploit the temporally-correlated user activities. Furthermore, we develop an expectation maximization HyGAMP-DCS (EM-HyGAMP-DCS) algorithm to adaptively learn the hyperparameters during the estimation procedure when the system statistics are unknown. In particular, we propose to utilize the analysis tool of state evolution to find the appropriate hyperparameter initialization of EM-HyGAMP-DCS. Simulation results demonstrate that our proposed algorithms can significantly improve the user activity detection accuracy and reduce the channel estimation error.

Deep Learning-Based Activity Detection for Grant-Free Random Access

The cellular Internet of Things wireless network is a promising solution to provide massive connectivity for machine-type devices. However, designing grant-free random access (GF-RA) protocols to manage such connections is challenging, since they must operate in interference-aware scenarios with sporadic device activation patterns and a shortage of mutually orthogonal resources. Supervised machine learning models have provided efficient solutions for activity detection, noncoherent data detection, and nonorthogonal preamble design in scenarios with massive connectivity. In this article, we develop two deep learning (DL) sparse support recovery algorithms to detect active devices in massive machine-type communication random access. The DL algorithms, developed to deploy GF-RA protocols, are based on the deep multilayer perceptron and the convolutional neural network models. Unlike previous works, we investigate the impact of the type of sequences for preamble design on the activity detection accuracy. Our results reveal that preambles based on the Zadoff–Chu sequences, which present good correlation properties, achieve better activity detection accuracy with the proposed algorithms than random sequences. Besides, we demonstrate that our DL algorithms achieve activity detection accuracy comparable to state-of-the-art techniques with extremely low computational complexity.

Deep Transferable Intelligence for Spatial Variability Characterization and Data-Efficient Learning in Biomechanical Measurement

Biomechanical measurement is of promising value for rehabilitation, assisted-living, and lifestyle management applications. Nevertheless, the understanding is still limited on the spatial variability of biomechanical dynamics that is essential for optimal motion sensor configuration. Besides, training physical activity detectors is usually data-heavy and time-consuming. Targeting these two challenges, in this study, we propose a novel deep transfer intelligence framework, which leverages deep learning to characterize spatial variability of different motion sensors on diverse body locations, and further leverages inter-subject transfer learning to maximize data-efficiency in challenging scarce data learning. More specifically, to characterize the spatial variability, we propose deep convolutional neural networks to investigate capabilities of both different sensor locations and channels, on physical activity measurement. The characterization determines both optimal sensor configuration and optimal channel configuration. Further, we propose a transfer learning approach to mine inter-subject similarity and then share learned knowledge among subjects, thereby minimizing the training effort and maximizing the data-efficiency in the wearable scarce data learning scenario. Our evaluation experiments have determined the optimal sensor location from seven options as thigh, and the optimal sensor&channel configuration from 42 options as thigh-accelerometer-axis-Y. Our experiments have further demonstrated that, with transfer learning under the optimal sensor&channel configuration, only 10% of data from the target subject for model fine-tuning can yield a physical activity detection accuracy up to 91.6%, with a performance boosting of 9% compared with direct learning without transfer learning. Therefore, the deep transferable learning framework will greatly advance spatial variability characterization for optimal sensor and channel configuration, and efficient scarce data learning in biomedical measurement.

Abstract 13824: Lifestyle Management for Cardiac Health With Deep Learning of Biomechanical Dynamics

Physical activities are of great values and, with appropriate principles, are reported to be clinically effective for rehabilitation from cardiac diseases. The physical activity-based lifestyle stimulation is thus highly promising to foster cardiac health management. Target the challenge that, the understanding of how different body locations contribute to the physical activity detection accuracy is still limited, we propose to leverage deep learning to investigate comprehensively seven body locations and six signal channels for each location. They are the chest, forearm, head, shin, thigh, upper-arm, and waist locations, and for each location, there are tri-axis acceleration signals and tri-axis gyroscope signals. In total there are forty-two combinations. The proposed convolutional neural network, when applied on each of the forty-two cases, yields a comprehensive comparison of these configurations in terms of physical activity type detection accuracy. Evaluated on fifteen subjects and six different physical activities, including climbing downstairs, climbing upstairs, jumping, lying, running/jogging, and walking, the determined optimal configuration is the chest location and the accelerometer axis-Y channel, with an accuracy of 95.1%. This study has not only maximized the physical activity detection accuracy through optimal configuration determination, but also deepened our understanding of how biomechanical dynamics are generated through different body locations and sensor channels. The research findings will greatly benefit cardiac health management.

Spatial Correlation Aware Compressed Sensing for User Activity Detection and Channel Estimation in Massive MTC

Grant-free access is considered as a key enabler for massive machine-type communications (mMTC) as it promotes energy-efficiency and small signalling overhead. Due to the sporadic user activity in mMTC, joint user identification and channel estimation (JUICE) is a main challenge. This paper addresses the JUICE in single-cell mMTC with single-antenna users and a multi-antenna base station (BS) under spatially correlated fading channels. In particular, by leveraging the sporadic user activity, we solve the JUICE in a multi measurement vector compressed sensing (CS) framework under two different cases, with and without the knowledge of prior channel distribution information (CDI) at the BS. First, for the case without prior information, we formulate the JUICE as an iterative reweighted <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\ell _{2,1}$ </tex-math></inline-formula> -norm minimization problem. Second, when the CDI is known to the BS, we exploit the available information and formulate the JUICE from a Bayesian estimation perspective as a maximum <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">a posteriori</i> probability (MAP) estimation problem. For both JUICE formulations, we derive efficient iterative solutions based on the alternating direction method of multipliers (ADMM). The numerical experiments show that the proposed solutions achieve higher channel estimation quality and activity detection accuracy with shorter pilot sequences compared to existing algorithms.

Data Transmission in Wearable Sensor Network for Human Activity Monitoring Using Embedded Classifier Technique

Abstract: The recent development of wireless wearable sensor networks has opened up a slew of new possibilities in industries as diverse as healthcare, medicine, activity monitoring, sports, safety, human-machine interface, and more. The battery-powered sensor nodes' longevity is critical to the technology's success. This research proposes a new strategy for increasing the lifetime of wearable sensor networks by eliminating redundant data transmissions. The proposed solution is based on embedded classifiers that allow sensor nodes to determine whether current sensor readings should be sent to the cluster head. A strategy was developed to train the classifiers, which takes into account the impact of data selection on the accuracy of a recognition system. This method was used to create a wearable sensor network prototype for human monitoring of activity Experiments were carried out in the real world to assess the novel method in terms of network lifetime, energy usage, and human activity recognition accuracy. The proposed strategy allows for a large increase in network lifetime while maintaining excellent activity detection accuracy, according to the results of the experimental evaluation. Experiments have also demonstrated that the technology has advantages over state-of-the-art data transmission reduction strategies. Keywords: wireless sensor network; wearable sensors; activity recognition; lifetime; energy con- sumption; transmission suppression; embedded machine learning

Data Transmission in Wearable Sensor Network for Human Activity Monitoring using Embedded Classifier technique

The recent development of wireless wearable sensor networks has opened up a slew of new possibilities in industries as diverse as healthcare, medicine, activity monitoring, sports, safety, human-machine interface, and more. The battery-powered sensor nodes' longevity is critical to the technology's success. This research proposes a new strategy for increasing the lifetime of wearable sensor networks by eliminating redundant data transmissions. The proposed solution is based on embedded classifiers that allow sensor nodes to determine whether current sensor readings should be sent to the cluster head. A strategy was developed to train the classifiers, which takes into account the impact of data selection on the accuracy of a recognition system. This method was used to create a wearable sensor network prototype for human monitoring of activity Experiments were carried out in the real world to assess the novel method in terms of network lifetime, energy usage, and human activity recognition accuracy. The proposed strategy allows for a large increase in network lifetime while maintaining excellent activity detection accuracy, according to the results of the experimental evaluation. Experiments have also demonstrated that the technology has advantages over state-of-the-art data transmission reduction strategies.

A coupled deep learning-based internal heat gains detection and prediction method for energy-efficient office building operation

Occupants' behaviour and the use of electrical equipment can significantly impact the building energy demand. Accurate occupancy and equipment usage information are key to improving the performance of demand-driven control, which can automatically adjust the heating, cooling and ventilation system operation. Employing static schedules is commonly used for the operation of heating, ventilation and air-conditioning systems, while it cannot satisfy the actual requirements due to the dynamic variations within the conditioned spaces. This study introduces a coupled real-time occupancy and equipment usage detection and recognition approach using deep learning and computer vision techniques for efficient building energy controls. The experimental results presented an overall equipment detection and occupancy activity detection accuracy of 78.39% and 93.60%. To investigate the influence of the implementation of the approach on building energy demand, a case study office building was selected to conduct experimental tests and modeled using a building energy simulation tool. Four scenarios with different occupancy and equipment profiles were defined and evaluated. The simulation results showed that heat gains, when employing static profiles were larger than the heat gains predicted when using the deep learning influenced profiles. Up to 53.95% lower heat gains were estimated when using both occupancy and equipment detection approaches than static schedules solely. The results highlighted the importance of monitoring real-time occupancy and electrical equipment usage and the advantages of using deep learning detection techniques to provide data for demand-driven controls, optimising building energy efficiency while maintaining a comfortable indoor environment.

A Personalized Classifier for Human Motion Activities With Semi-Supervised Learning

The motion activities performed by a mobile or wearable device user are characteristic of the user and the accuracy of detection of motion activity context can be improved by incorporating personalized samples of motion activity. But obtaining personalized samples of motion activity from device user is not practically feasible. In this article, a semi-supervised method is presented for improving the classification accuracy of the mobile or wearable device user’s motion activities by deriving a classifier that is personalized to the user starting from a factory-set generalized classifier. The novelty lies in using information theoretic criterion to select personalized data samples from all the personalized data samples collected from the target user. This is useful in improving the motion activity detection accuracy, since each device user has a distinct gait personality as compared to the motion activity data patterns available in a generalized database that are used to train a generalized classifier. The average accuracy for 11 target users of a generalized classifier is 93.11% that increases to 96.50% using the proposed method when the device is used in in-hand mode. It is observed that the accuracy improvement with a personalized classifier over a generalized classifier is greater for those subjects whose accuracy is lower than others with the generalized classifier.

Scheduling Sensor Duty Cycling Based on Event Detection Using Bi-Directional Long Short-Term Memory and Reinforcement Learning.

A smart home provides a facilitated environment for the detection of human activity with appropriate Deep Learning algorithms to manipulate data collected from numerous sensors attached to various smart things in a smart home environment. Human activities comprise expected and unexpected behavior events; therefore, detecting these events consisting of mutual dependent activities poses a key challenge in the activities detection paradigm. Besides, the battery-powered sensor ubiquitously and extensively monitors activities, disputes, and sensor energy depletion. Therefore, to address these challenges, we propose an Energy and Event Aware-Sensor Duty Cycling scheme. The proposed model predicts the future expected event using the Bi-Directional Long-Short Term Memory model and allocates Predictive Sensors to the predicted event. To detect the unexpected events, the proposed model localizes a Monitor Sensor within a cluster of Hibernate Sensors using the Jaccard Similarity Index. Finally, we optimize the performance of our proposed scheme by employing the Q-Learning algorithm to track the missed or undetected events. The simulation is executed against the conventional Machine Learning algorithms for the sensor duty cycle, scheduling to reduce the sensor energy consumption and improve the activity detection accuracy. The experimental evaluation of our proposed scheme shows significant improvement in activity detection accuracy from 94.12% to 96.12%. Besides, the effective rotation of the Monitor Sensor significantly improves the energy consumption of each sensor with the entire network lifetime.

Non-Employment Activity Type Imputation from Points of Interest and Mobility Data at an Individual Level: How Accurate Can We Get?

Human activity type inference has long been the focus for applications ranging from managing transportation demand to monitoring changes in land use patterns. Today’s ever increasing volume of mobility data allow researchers to explore a wide range of methodological approaches for this task. Such data, however, lack reference observations that would allow the validation of methodological approaches. This research proposes a methodological framework for urban activity type inference using a Dirichlet multinomial dynamic Bayesian network with an empirical Bayes prior that can be applied to mobility data of low spatiotemporal resolution. The method was validated using open source Foursquare data under different isochrone configurations. The results provide evidence of the limits of activity detection accuracy using such data as determined by the Area Under Receiving Operating Curve (AUROC), log-loss, and accuracy metrics. At the same time, results demonstrate that a hierarchical modeling framework can provide some flexibility against the challenges related to the nature of unsupervised activity classification using trajectory variables and POIs as input.

TSFS: An Integrated Approach for Event Segmentation and ADL Detection in IoT Enabled Smarthomes

With the recent advances in Internet of Things, smarthomes containing a variety of sensors are expected to monitor and react to the activities of daily living (ADL) of the residents with the goal of improving convenience, comfort, and safety. However, a single human activity may trigger multiple sensors, and each sensor is triggered by multiple activities. Also, the very human nature of interleaving activities along with the presence of multiple inhabitants makes ADL detection a complex problem. This paper tackles the problem with a data driven de-multiplexing approach termed TSFS (Temporal-Sensor-Frequency-Stitch) that disentangles each individual activity from the sensor stream and thus simplifies the ADL recognition problem. TSFS leverages temporal($T$T), sensory($S$S), and frequency($F$F) information to discriminate between two consecutive/parallel activities, followed by a stitch($S$S) to enhance segmentation. The multi-faceted approach helps us in correctly identifying segments; extensive evaluation on 10 public datasets reveals that TSFS is better by 185 percent when compared to a simple temporal segmentation approach. The activity detection accuracy yields within 93 percent of the best possible (an oracle) approach.

Activity-Aware Sensor Cycling for Human Activity Monitoring in Smart Homes

Smart homes are one of the Internet of Things domains intended to support and aid the residents through various smart services. These services require accurate context inferences using daily activity patterns and environmental properties. To satisfy such a need with battery-powered sensors, various duty cycling schemes were introduced. In this letter, we propose an activity-aware sensor cycling approach that makes the best tradeoff for duty cycle adjustments by exploiting the predictable behavior of residents, thereby significantly improving the activity detection accuracy at a marginal increase of the energy consumption. Evaluation results demonstrate that it achieves up to 99% accuracy of activity detection and extends the network lifetime by supporting balanced energy consumption among sensors.

Super-Audible Voice Activity Detection

In this paper, reflected sound of frequency just above the audible range is used to detect speech activity. The active signal used is inaudible to humans, readily generated by the typical audio circuitry and components found in mobile telephones, and is robust to background sounds such as nearby voices. In use, the system relies upon a wideband excitation signal emitted from a loudspeaker located near the lips, which reflects from the mouth region and is then captured by a nearby microphone. The state of the lip opening is evaluated periodically by tracking the resonance patterns in the reflected excitation signal. When the lips are open, deep and complex resonances are formed as energy propagates into and then reflects out from the open mouth and vocal tract, with resonance depth being related to the open lip area. When the lips are closed, these resonance patterns are absent. The presence of the resonances can thus serve as a low complexity detection measure. The technique is evaluated for multiple users in terms of sensitivity to source placement and sensor placement. Voice activity detection performance using this measure is further evaluated in the presence of realistic wideband acoustic background noise, as well as artificially added noise. The system is shown to be relatively insensitive to sensor placement, highly insensitive to background noise, and able to achieve greater than 90% voice activity detection accuracy. The technique is even suitable when a subject is whispering in the presence of much louder multi-speaker babble. The technique has potential for speech-based systems operating in high noise environments as well as in silent speech interfaces, whisper-input systems and voice prostheses for speech-impaired users.

Optimal Placement of Accelerometers for the Detection of Everyday Activities

This article describes an investigation to determine the optimal placement of accelerometers for the purpose of detecting a range of everyday activities. The paper investigates the effect of combining data from accelerometers placed at various bodily locations on the accuracy of activity detection. Eight healthy males participated within the study. Data were collected from six wireless tri-axial accelerometers placed at the chest, wrist, lower back, hip, thigh and foot. Activities included walking, running on a motorized treadmill, sitting, lying, standing and walking up and down stairs. The Support Vector Machine provided the most accurate detection of activities of all the machine learning algorithms investigated. Although data from all locations provided similar levels of accuracy, the hip was the best single location to record data for activity detection using a Support Vector Machine, providing small but significantly better accuracy than the other investigated locations. Increasing the number of sensing locations from one to two or more statistically increased the accuracy of classification. There was no significant difference in accuracy when using two or more sensors. It was noted, however, that the difference in activity detection using single or multiple accelerometers may be more pronounced when trying to detect finer grain activities. Future work shall therefore investigate the effects of accelerometer placement on a larger range of these activities.

Energy-aware Activity Classification using Wearable Sensor Networks.

This paper presents implementation details, system characterization, and the performance of a wearable sensor network that was designed for human activity analysis. Specific machine learning mechanisms are implemented for recognizing a target set of activities with both out-of-body and on-body processing arrangements. Impacts of energy consumption by the on-body sensors are analyzed in terms of activity detection accuracy for out-of-body processing. Impacts of limited processing abilities for the on-body scenario are also characterized in terms of detection accuracy, by varying the background processing load in the sensor units. Impacts of varying number of sensors in terms of activity classification accuracy are also evaluated. Through a rigorous systems study, it is shown that an efficient human activity analytics system can be designed and operated even under energy and processing constraints of tiny on-body wearable sensors.

Wearable Networked Sensing for Human Mobility and Activity Analytics: A Systems Study.

This paper presents implementation details, system characterization, and the performance of a wearable sensor network that was designed for human activity analysis. Specific machine learning mechanisms are implemented for recognizing a target set of activities with both out-of-body and on-body processing arrangements. Impacts of energy consumption by the on-body sensors are analyzed in terms of activity detection accuracy for out-of-body processing. Impacts of limited processing abilities in the on-body scenario are also characterized in terms of detection accuracy, by varying the background processing load in the sensor units. Through a rigorous systems study, it is shown that an efficient human activity analytics system can be designed and operated even under energy and processing constraints of tiny on-body wearable sensors.

Noise Robust Voice Activity Detection Based on Switching Kalman Filter

This paper addresses the problem of voice activity detection (VAD) in noisy environments. The VAD method proposed in this paper is based on a statistical model approach, and estimates statistical models sequentially without a priori knowledge of noise. Namely, the proposed method constructs a clean speech / silence state transition model beforehand, and sequentially adapts the model to the noisy environment by using a switching Kalman filter when a signal is observed. In this paper, we carried out two evaluations. In the first, we observed that the proposed method significantly outperforms conventional methods as regards voice activity detection accuracy in simulated noise environments. Second, we evaluated the proposed method on a VAD evaluation framework, CENSREC-1-C. The evaluation results revealed that the proposed method significantly outperforms the baseline results of CENSREC-1-C as regards VAD accuracy in real environments. In addition, we confirmed that the proposed method helps to improve the accuracy of concatenated speech recognition in real environments.