Virtual sensors model the sensing operation of physical sensors deployed in an area of interest by generating sensory data with accuracy and precision close to those collected by physical sensors. their use in applica...
详细信息
ISBN:
(纸本)9781665495127
Virtual sensors model the sensing operation of physical sensors deployed in an area of interest by generating sensory data with accuracy and precision close to those collected by physical sensors. their use in applications such as augmenting the infrastructure of IoT facilities and test beds, monitoring and calibrating the operation of physical sensors, and developing Digital Twins of physical systems have led virtual sensors to attract research attention. Machine learning provides methods for modelling patterns in complex and big data generated by IoT sensing devices, allowing to model the behaviour of these devices. In this work, we investigate ML methods as means of implementation for virtual sensors. In particular, we evaluate the performance of six ML methods in terms of their effectiveness, accuracy and precision in generating sensory data based on data from physical sensors. In our study, we use a multi-modal dataset comprising IoT sensory data for temperature, humidity and illumination collected over a period of two years in an office space at University of Geneva. Our results show that the best performing model at predicting an output of a missing sensor is the Random Forest method, achieving MAPE error below 3%, 5% and 18% respectively for temperature, humidity and illuminance. the worst performing models were the linear radial basis function neural network and linear regression. In future research, we plan to deploy the best performing models natively on IoT devices, making use of tinyML and extreme edge computing methods.
the article considers a methodology for the synthesis of adaptive distributed control systems. By the nature of changes in the control device, the systems considered in the article belong to the class of self-tuning (...
详细信息
the low-voltage distribution network is facing a situation of blowout, small capacity and decentralized distributed new energy access, withthe emergence of large-scale power reverse transmission and frequent voltage ...
详细信息
In Storm systems, an efficient resource placement strategy is key to ensure application performance. However, the Storm platform uses a polled resource placement strategy, which leads to large resource and communicati...
详细信息
the proceedings contain 359 papers. the topics discussed include: classification and structuring of DL-based method for finding Alzheimer's disease through MRI images;impact of social media on the professional dev...
ISBN:
(纸本)9798350383522
the proceedings contain 359 papers. the topics discussed include: classification and structuring of DL-based method for finding Alzheimer's disease through MRI images;impact of social media on the professional development of health professionals;use of sensor networks with self-organizing algorithm to increase the agricultural productivity;cyberbullying : research challenges and opportunities;improving brain tumor detection and classification using intelligent methods;rotten and fresh fruits classification using deep learning;structured implementation of ML algorithms for cardiovascular disease detection;enhanced retail shopper behavioral analysis using human machine interaction and model validation;java-powered digital healthcare management: innovating medical administration systems;and applicability of machine learning in waste water quality detection.
the pressure-sensing diaphragms of micro-electromechanical systems (MEMS) pressure sensors are prone to fracture under high pressure and eventually result in sensor performance failure. To address this issue, a techni...
详细信息
ISBN:
(纸本)9798350363272;9798350363265
the pressure-sensing diaphragms of micro-electromechanical systems (MEMS) pressure sensors are prone to fracture under high pressure and eventually result in sensor performance failure. To address this issue, a technique is proposed in this study that can improve the overload capacity of MEMS polycrystalline silicon nano-diaphragm pressure sensors. A polycrystalline silicon diaphragm with a thickness of 80 nm is utilized for constructing the pressure-sensing diaphragm structure. Furthermore, the cavity height of the membrane for bottoming protection is analyzed using the static large deformation and contact nonlinear finite element analysis methods. Accordingly, the association of diaphragm sizes withsensor overload capacity is deduced, and the influence of diaphragm size variation on the overload capacity under different sensor ranges is discussed. the simulation results reveal that the MEMS polycrystalline silicon nanomembrane pressure sensor designed for a 1 MPa range exhibits an overload capacity of 8 MPa, which is eight times the sensor range. this study provides novel ideas that can contribute to the preparation of MEMS pressure sensors in real-time production.
the proceedings contain 25 papers. the topics discussed include: machine learning-based task allocation in computational networks;reinforcement learning based edge microservice migration strategy under resource constr...
ISBN:
(纸本)9798350369007
the proceedings contain 25 papers. the topics discussed include: machine learning-based task allocation in computational networks;reinforcement learning based edge microservice migration strategy under resource constraints;distributedsensor-assisted DNN-based massive-MIMO channel reconstruction;modeling and evaluation of low-frequency noise influence in a MIMO communication system;an anonymous authentication scheme based on blind signatures for the FIDO protocol;MANETs connectivity management for age of information decreasing in real-time applications;the equivalence principle for adaptive load balancing in networks and systems;how useful is communication scheduling for distributed training?;and HeavyCache: a generic sketch for summarizing data streams.
the rapid deployment of IoT networks in different industrial services has caused the emanation of a huge volume of data from sensors and monitors. the efficient analysis and compact representation of the big data gene...
详细信息
We consider large scale Peer-to-Peer sensor Networks, which try to calculate and distribute the mean value of all sensor inputs. For this we design, simulate and evaluate distributed approximation algorithms which red...
详细信息
ISBN:
(纸本)9781665495127
We consider large scale Peer-to-Peer sensor Networks, which try to calculate and distribute the mean value of all sensor inputs. For this we design, simulate and evaluate distributed approximation algorithms which reduce the number of messages. the main difference of these algorithms is the underlying communication protocol which all use the random call model, where in discrete round model each node can call a random sensor node with uniform probability. the amount of data exchanged between sensor nodes and used in the calculation process affects the accuracy of the aggregation results leading to a trade-off situation. the key idea of our algorithms is to limit the sample size using the Finite Population Correction (FPC) method and collect the data using a distribution aggregation using Push-Pull Sampling, Pull Sampling, and Push Sampling communication protocols. It turns out that all methods show exponential improvement of Mean Squared Error (MSE) withthe number of messages and rounds.
this paper proposes an autoencoder-based approach to effectively extract sensor features by leveraging an autoencoder as a data preprocessing method. the autoencoder constrains the hidden units in a bottleneck structu...
详细信息
ISBN:
(纸本)9798350307627
this paper proposes an autoencoder-based approach to effectively extract sensor features by leveraging an autoencoder as a data preprocessing method. the autoencoder constrains the hidden units in a bottleneck structure, resulting in a compressed knowledge representation of sensor readings. In the latent space representation, the encoded data learns and describes the most prominent latent attributes of sensor readings. the algorithm is experimentally validated in a real-world setting, demonstrating its effectiveness in accurately extracting relevant features from sensor data. Nine flexible bending sensors are utilized for posture sensing of a bellow-shaped fluidic elastomer actuator. Compared to previous studies, the results demonstrate that valuable features can be extracted without employing a large dropout rate for overfitting prevention, while maintaining prediction accuracy and reducing the entire sensor signals to half. Additionally, the training time is reduced by 7.2%. By providing a reduced and featured input to the regression neural network, the proposed approach not only prevents overfitting but also alleviates the computational redundancy and complexity brought by an increasing number of sensors.
暂无评论