Pub Date : 2020-08-01DOI: 10.1109/COINS49042.2020.9191402
Philippos Barabas, Emanuel Regnath, S. Steinhorst
Vehicular applications heavily rely on location information to improve road safety and efficiency as well as to provide a personalized driving experience through a variety of location-based services. To determine their position, vehicles depend on different technologies like GPS, which might be unreliable or vulnerable to interference or spoofing. In the safety-critical vehicular world, a secure mechanism must be in place which guarantees the accuracy and trustworthiness of location information to the service that requires it. In this work we propose COLAW, a COoperative Location proof Architecture based on Witnessing that leverages the distributed nature of vehicular ad-hoc networks to create verifiable and secure location proofs. The evaluation of COLAW shows that it is possible for a group of neighboring vehicles to generate secure location proofs for each other with a significantly lower message overhead than previously proposed approaches and that the protocol’s performance can be further improved, by taking certain environmental parameters and road conditions into consideration.
{"title":"COLAW: Cooperative Location Proof Architecture for VANETs based on Witnessing","authors":"Philippos Barabas, Emanuel Regnath, S. Steinhorst","doi":"10.1109/COINS49042.2020.9191402","DOIUrl":"https://doi.org/10.1109/COINS49042.2020.9191402","url":null,"abstract":"Vehicular applications heavily rely on location information to improve road safety and efficiency as well as to provide a personalized driving experience through a variety of location-based services. To determine their position, vehicles depend on different technologies like GPS, which might be unreliable or vulnerable to interference or spoofing. In the safety-critical vehicular world, a secure mechanism must be in place which guarantees the accuracy and trustworthiness of location information to the service that requires it. In this work we propose COLAW, a COoperative Location proof Architecture based on Witnessing that leverages the distributed nature of vehicular ad-hoc networks to create verifiable and secure location proofs. The evaluation of COLAW shows that it is possible for a group of neighboring vehicles to generate secure location proofs for each other with a significantly lower message overhead than previously proposed approaches and that the protocol’s performance can be further improved, by taking certain environmental parameters and road conditions into consideration.","PeriodicalId":350108,"journal":{"name":"2020 International Conference on Omni-layer Intelligent Systems (COINS)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2020-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130997851","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2020-08-01DOI: 10.1109/COINS49042.2020.9191409
Christoph Lehnert, Grischan Engel, Thomas Greiner
Security and traceability of smart sensor data in centrally organized IoT-architectures require a third party of trust. In order to overcome this issue, Distributed Ledger Technologies (DLT) apply consensus mechanisms. Current approaches suggest DLT-based IoT-architectures which are static and only provide limited data precision in specific applications. Thus, they rely on custom tokens and additional technologies such as SQL databases. In addition, the design of the applied smart contracts (sc) allow unauthorized access. In contrast, in this paper an adaptable, scalable and purely DLT-based IoT-architecture for secure and decentral software services is proposed. It employs sc for the secure and decentralized interaction between users, software services and IoT devices, such as smart sensors. Thereby, sc are adjustable and their access is controlled by an address comparison of authorized wallets. Finally, a case-study on a sc based software service for an industrial smart temperature sensor demonstrates applicability and benefits of the proposed approach.
{"title":"Distributed Ledger and Smart Contract Based Approach for IoT Sensor Applications","authors":"Christoph Lehnert, Grischan Engel, Thomas Greiner","doi":"10.1109/COINS49042.2020.9191409","DOIUrl":"https://doi.org/10.1109/COINS49042.2020.9191409","url":null,"abstract":"Security and traceability of smart sensor data in centrally organized IoT-architectures require a third party of trust. In order to overcome this issue, Distributed Ledger Technologies (DLT) apply consensus mechanisms. Current approaches suggest DLT-based IoT-architectures which are static and only provide limited data precision in specific applications. Thus, they rely on custom tokens and additional technologies such as SQL databases. In addition, the design of the applied smart contracts (sc) allow unauthorized access. In contrast, in this paper an adaptable, scalable and purely DLT-based IoT-architecture for secure and decentral software services is proposed. It employs sc for the secure and decentralized interaction between users, software services and IoT devices, such as smart sensors. Thereby, sc are adjustable and their access is controlled by an address comparison of authorized wallets. Finally, a case-study on a sc based software service for an industrial smart temperature sensor demonstrates applicability and benefits of the proposed approach.","PeriodicalId":350108,"journal":{"name":"2020 International Conference on Omni-layer Intelligent Systems (COINS)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2020-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125072531","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2020-08-01DOI: 10.1109/COINS49042.2020.9191418
M. Hashemi, Bahareh J. Farahani, F. Firouzi
The human factor has been documented as the primary contributor to road accidents bringing outrageous costs, such as property damage, disabling injuries, and loss of life. To mitigate accident-related costs and to enhance driver safety, particularly during unfavorable driving conditions, the transportation industry strives to integrate IoT and Deep Learning technologies. In this work, we propose a holistic IoT-based multimodal technique to monitor driver fatigue by exploiting the facial and physiological information of the driver. A novel deep neural network is designed to classify the eye and mouth states. The results of the classification are fed into the cloud to be fused with other data sources (e.g., health records) in order to assess the corresponding driver risk accurately. Experimental results on various datasets show that the proposed mouth classification and eye state detection solution results in 99.5% and 99.01% accuracy, respectively.
{"title":"Towards Safer Roads: A Deep Learning-Based Multimodal Fatigue Monitoring System","authors":"M. Hashemi, Bahareh J. Farahani, F. Firouzi","doi":"10.1109/COINS49042.2020.9191418","DOIUrl":"https://doi.org/10.1109/COINS49042.2020.9191418","url":null,"abstract":"The human factor has been documented as the primary contributor to road accidents bringing outrageous costs, such as property damage, disabling injuries, and loss of life. To mitigate accident-related costs and to enhance driver safety, particularly during unfavorable driving conditions, the transportation industry strives to integrate IoT and Deep Learning technologies. In this work, we propose a holistic IoT-based multimodal technique to monitor driver fatigue by exploiting the facial and physiological information of the driver. A novel deep neural network is designed to classify the eye and mouth states. The results of the classification are fed into the cloud to be fused with other data sources (e.g., health records) in order to assess the corresponding driver risk accurately. Experimental results on various datasets show that the proposed mouth classification and eye state detection solution results in 99.5% and 99.01% accuracy, respectively.","PeriodicalId":350108,"journal":{"name":"2020 International Conference on Omni-layer Intelligent Systems (COINS)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2020-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132061239","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2020-08-01DOI: 10.1109/COINS49042.2020.9191423
Joao Falcao, Paulo Menezes, R. Rocha
In recent years, Internet of Things has been gaining popularity due to its capabilities and flexible implementation. Current developments make use of several sensor types building large wireless sensor networks, where each sensor can have a degree of connection over the others. It is usually more perceptible with the use of motion sensors in different rooms where physical paths taken by a subject are strongly correlated to temporal sequences detected in the nodes. This study presents two methods for the detection of these correlations between nodes, one requiring the user to perform a path across every sensor and another method that tries to infer information without any explicit human intervention, by analysing the first events of each day where entropy is low. The results show that the latter method, which does not require explicit human intervention, presents some degradation if a low number of sensors is used in the network and these sensors have a high periodic activation. The former method is in general more accurate for small to medium sized networks, but can be problematic in large networks where passing across every sensor can be a tedious or unpractical requirement.
{"title":"Automatic Identification of Wireless Sensor Network Topology in a IoT Domestic Setup and Discovery of User Routines","authors":"Joao Falcao, Paulo Menezes, R. Rocha","doi":"10.1109/COINS49042.2020.9191423","DOIUrl":"https://doi.org/10.1109/COINS49042.2020.9191423","url":null,"abstract":"In recent years, Internet of Things has been gaining popularity due to its capabilities and flexible implementation. Current developments make use of several sensor types building large wireless sensor networks, where each sensor can have a degree of connection over the others. It is usually more perceptible with the use of motion sensors in different rooms where physical paths taken by a subject are strongly correlated to temporal sequences detected in the nodes. This study presents two methods for the detection of these correlations between nodes, one requiring the user to perform a path across every sensor and another method that tries to infer information without any explicit human intervention, by analysing the first events of each day where entropy is low. The results show that the latter method, which does not require explicit human intervention, presents some degradation if a low number of sensors is used in the network and these sensors have a high periodic activation. The former method is in general more accurate for small to medium sized networks, but can be problematic in large networks where passing across every sensor can be a tedious or unpractical requirement.","PeriodicalId":350108,"journal":{"name":"2020 International Conference on Omni-layer Intelligent Systems (COINS)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2020-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128804073","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2020-08-01DOI: 10.1109/COINS49042.2020.9191674
Gilles Callebaut, Geoffrey Ottoy, L. Perre
Many IoT applications require long range yet low power connectivity in dynamic environments. We assess and optimize their energy efficiency in a Long Range Wide Area Network (LoRaWAN) network following a cross-layer approach. The analysis demonstrates that the channel variation may significantly impact the quality of the transmission and the energy consumption of the nodes. A proactive adjustment strategy of the Adaptive Data Rate settings allows for optimization of the transmit energy. In addition, we show how trade-offs between robustness, energy efficiency and throughput can be made.
{"title":"optimizing Transmission of IoT Nodes in Dynamic Environments","authors":"Gilles Callebaut, Geoffrey Ottoy, L. Perre","doi":"10.1109/COINS49042.2020.9191674","DOIUrl":"https://doi.org/10.1109/COINS49042.2020.9191674","url":null,"abstract":"Many IoT applications require long range yet low power connectivity in dynamic environments. We assess and optimize their energy efficiency in a Long Range Wide Area Network (LoRaWAN) network following a cross-layer approach. The analysis demonstrates that the channel variation may significantly impact the quality of the transmission and the energy consumption of the nodes. A proactive adjustment strategy of the Adaptive Data Rate settings allows for optimization of the transmit energy. In addition, we show how trade-offs between robustness, energy efficiency and throughput can be made.","PeriodicalId":350108,"journal":{"name":"2020 International Conference on Omni-layer Intelligent Systems (COINS)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2020-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129473429","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2020-03-07DOI: 10.1109/COINS49042.2020.9191371
Rishab Sharma, Rahul Deora, Anirudha Vishvakarma
In this paper, we propose an end to end solution for image matting i.e high-precision extraction of foreground objects from natural images. Image matting and background detection can be achieved easily through chroma keying in a studio setting when the background is either pure green or blue. Nonetheless, image matting in natural scenes with complex and uneven depth backgrounds remains a tedious task that requires human intervention. To achieve complete automatic foreground extraction in natural scenes, we propose a method that assimilates semantic segmentation and deep image matting processes into a single network to generate detailed semantic mattes for image composition task. The contribution of our proposed method is two-fold, firstly it can be interpreted as a fully automated semantic image matting method and secondly as a refinement of existing semantic segmentation models.We propose a novel model architecture as a combination of segmentation and matting that unifies the function of upsampling and downsampling operators with the notion of attention. As shown in our work, attention guided downsampling and upsampling can extract high-quality boundary details, unlike other normal downsampling and upsampling techniques. For achieving the same, we utilized an attention guided encoder-decoder framework which does unsupervised learning for generating an attention map adaptively from the data to serve and direct the upsampling and downsampling operators. We also construct a fashion e-commerce focused dataset with high-quality alpha mattes to facilitate the training and evaluation for image matting.
{"title":"AlphaNet: An Attention Guided Deep Network for Automatic Image Matting","authors":"Rishab Sharma, Rahul Deora, Anirudha Vishvakarma","doi":"10.1109/COINS49042.2020.9191371","DOIUrl":"https://doi.org/10.1109/COINS49042.2020.9191371","url":null,"abstract":"In this paper, we propose an end to end solution for image matting i.e high-precision extraction of foreground objects from natural images. Image matting and background detection can be achieved easily through chroma keying in a studio setting when the background is either pure green or blue. Nonetheless, image matting in natural scenes with complex and uneven depth backgrounds remains a tedious task that requires human intervention. To achieve complete automatic foreground extraction in natural scenes, we propose a method that assimilates semantic segmentation and deep image matting processes into a single network to generate detailed semantic mattes for image composition task. The contribution of our proposed method is two-fold, firstly it can be interpreted as a fully automated semantic image matting method and secondly as a refinement of existing semantic segmentation models.We propose a novel model architecture as a combination of segmentation and matting that unifies the function of upsampling and downsampling operators with the notion of attention. As shown in our work, attention guided downsampling and upsampling can extract high-quality boundary details, unlike other normal downsampling and upsampling techniques. For achieving the same, we utilized an attention guided encoder-decoder framework which does unsupervised learning for generating an attention map adaptively from the data to serve and direct the upsampling and downsampling operators. We also construct a fashion e-commerce focused dataset with high-quality alpha mattes to facilitate the training and evaluation for image matting.","PeriodicalId":350108,"journal":{"name":"2020 International Conference on Omni-layer Intelligent Systems (COINS)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2020-03-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114699388","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2019-08-08DOI: 10.1109/COINS49042.2020.9191636
Boyu Zhang, A. Davoodi, Y. Hu
To deploy a CNN on resource-constrained edge platforms, channel pruning techniques promise a significant reduction of implementation costs including memory, computation, and energy consumption without special hardware or software libraries. This paper proposes CHaPR, a novel pruning technique to structurally prune the redundant channels in a trained deep Convolutional Neural Network. CHaPR utilizes a proposed subset selection problem formulation for pruning which it solves using pivoted QR factorization. CHaPR also includes an additional pruning technique for ResNet-like architectures which resolves the issue encountered by some existing channel pruning methods that not all the layers can be pruned. Experimental results on VGG-16 and ResNet-50 models show 4.29X and 2.84X reduction, respectively in computation cost while incurring 2.50% top-1 and 1.40% top-5 accuracy losses. Compared to many existing works, CHaPR performs better when considering an Overall Score metric which accounts for both computation and accuracy.
{"title":"CHaPR: Efficient Inference of CNNs via Channel Pruning","authors":"Boyu Zhang, A. Davoodi, Y. Hu","doi":"10.1109/COINS49042.2020.9191636","DOIUrl":"https://doi.org/10.1109/COINS49042.2020.9191636","url":null,"abstract":"To deploy a CNN on resource-constrained edge platforms, channel pruning techniques promise a significant reduction of implementation costs including memory, computation, and energy consumption without special hardware or software libraries. This paper proposes CHaPR, a novel pruning technique to structurally prune the redundant channels in a trained deep Convolutional Neural Network. CHaPR utilizes a proposed subset selection problem formulation for pruning which it solves using pivoted QR factorization. CHaPR also includes an additional pruning technique for ResNet-like architectures which resolves the issue encountered by some existing channel pruning methods that not all the layers can be pruned. Experimental results on VGG-16 and ResNet-50 models show 4.29X and 2.84X reduction, respectively in computation cost while incurring 2.50% top-1 and 1.40% top-5 accuracy losses. Compared to many existing works, CHaPR performs better when considering an Overall Score metric which accounts for both computation and accuracy.","PeriodicalId":350108,"journal":{"name":"2020 International Conference on Omni-layer Intelligent Systems (COINS)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2019-08-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121257592","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：