{"title":"Task Offloading in Autonomous IoT Systems using Deep Reinforcement Learning and ns3-gym","authors":"Abdel Kader Chabi Sika Boni, H. Hassan, K. Drira","doi":"10.1145/3494322.3494325","DOIUrl":"https://doi.org/10.1145/3494322.3494325","url":null,"abstract":"IoT systems grow quickly and are massively present in urban areas. Their successful deployment requires autonomy that can be built on automated learning technologies such as Deep Learning. The IoT applications require important computational resources, rarely available on devices. Autonomous IoT systems require the computation power available on the edge and cloud servers in order to offload some tasks related to the supported applications and the underlying platforms. Task offloading constitutes a big challenge in autonomous IoT systems due to the huge number of IoT devices for scenarios of the family of smart cities. Managing task offloading in such contexts requires adaptive strategies capable of taking into consideration the rapid evolution of available resources and proposing efficient offloading solutions to all received requests. In this paper we use a Deep Reinforcement Learning (DRL) approach capable of handling large state spaces, and resolve the optimization problem in this context, where other techniques can not scale efficiently. Our solution is based on a DRL agent that was developed in the ns3-gym framework and was tested on IoT system scenario implemented in the NS3 simulator. The results obtained show that the DRL agent can adapt quickly to resource evolution in the IoT system and can handle big number of demands fulfilling scalabilty requirements of autonomous IoT systems.","PeriodicalId":348063,"journal":{"name":"Proceedings of the 11th International Conference on the Internet of Things","volume":"21 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-11-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126587464","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}
{"title":"An Expert Digital Companion for Working Environments","authors":"Janick Spirig, Kimberly García, S. Mayer","doi":"10.1145/3494322.3494326","DOIUrl":"https://doi.org/10.1145/3494322.3494326","url":null,"abstract":"Research in proactive agents capable of anticipating users’ needs has been conducted in different application areas, from agents aiming to help users accomplish their next task while using a text editor to digital assistants that provide support to medical professionals in hospitals. Considering these works and following the rapid development in machine learning, the growing adoption of the IoT, the prevalence of pervasive computing, and the various innovative user interfaces that are becoming mainstream, we propose the creation of expert digital companions. Expert digital companions are agents that know the specifics of an environment and the available physical equipment in a space as well as software components reachable in that environment. On their side, personal digital companions know the preferences of a specific user. The separation of concerns between expert and personal companions enables them to evolve independently. Thus, expert and personal digital companions collaborate to provide customized assistance to a user given her environment. In this paper, we present two expert digital companions for employees working in an office and a shopfloor environment that takes advantages of: the decoupling that the W3C WoT Thing Description provides to interact with and control devices and other tools; a Knowledge Graph to provide richer descriptions of elements in the environment; computer vision to perceive the physical world; and Mixed Reality as the medium to deliver assistance to workers.","PeriodicalId":348063,"journal":{"name":"Proceedings of the 11th International Conference on the Internet of Things","volume":"13 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-11-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121657935","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}
{"title":"Poster: An Exploration Into Tracking the Use of Consumer Goods with RFID","authors":"G. Berumen, Anthony Brown, M. Baumers, J. Fischer","doi":"10.1145/3494322.3494344","DOIUrl":"https://doi.org/10.1145/3494322.3494344","url":null,"abstract":"Smart consumer goods have the potential to positively impact consumers by providing support in areas of great concern, such as helping to reduce waste and promote healthy eating, so tracking their use is essential for the development of their enhanced versions. To explore how we could track the use of these products, we tested a prototype of a radio-frequency identification system (RFID) that detects interactions by recording obstructions to the signals of sensors attached to consumer goods. We tested the system using items associated with the preparation of a cup of tea. Our system was highly reliable in detecting interactions; however, it was imprecise in estimating their durations. We discuss how further development of the proposed RFID system, including improvements in the collection and classification of the signal, could lead to be a viable method for tracking the use of consumer goods and help incorporate a layer of ‘smartness’.","PeriodicalId":348063,"journal":{"name":"Proceedings of the 11th International Conference on the Internet of Things","volume":"10 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-11-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116524492","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}
Dario Salvi, C. Olsson, Gent Ymeri, Carmen Carrasco-López, K. C. Tsang, Syed Ahmar Shah
{"title":"Mobistudy: mobile-based, platform-independent, multi-dimensional data collection for clinical studies","authors":"Dario Salvi, C. Olsson, Gent Ymeri, Carmen Carrasco-López, K. C. Tsang, Syed Ahmar Shah","doi":"10.1145/3494322.3494363","DOIUrl":"https://doi.org/10.1145/3494322.3494363","url":null,"abstract":"Internet of Things (IoT) can work as a useful tool for clinical research. We developed a software platform that allows researchers to publish clinical studies and volunteers to participate into them using an app and connected IoT devices. The platform includes a REST API, a web interface for researchers and an app that collects data during tasks volunteers are invited to contribute. Nine tasks have been developed: Forms, Positioning, Finger tapping, Pulse-oximetry, Peak Flow measurement, Activity tracking, Data query, Queen’s College step test and Six-minute walk test. These leverage sensors embedded in the phone, connected Bluetooth devices and additional APIs like HealthKit and Google Fit. Currently, the platform is used in two clinical studies by 25 patients: an asthma management study in the United Kingdom, and a neuropathic pain management study in Spain.","PeriodicalId":348063,"journal":{"name":"Proceedings of the 11th International Conference on the Internet of Things","volume":"3 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-11-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124833311","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}
{"title":"Business Process Models for Analysis of Industrial IoT Applications","authors":"A. Krishna, Gwen Salaün","doi":"10.1145/3494322.3494336","DOIUrl":"https://doi.org/10.1145/3494322.3494336","url":null,"abstract":"Industrial automation is a complex process involving various stakeholders. The two important aspects to consider during the automation system development are its business production goals and its technical implementation. The international standard IEC 61499 helps to specify distributed automation using a generic architectural model, targeting the technical development of the automation. However, it is not easy to analyse whether these IEC 61499 models satisfy production goals due to their informal semantics and inherent complexities of distributed logic. In this paper, we propose to use Business Process Model Notation (BPMN) to express the underlying IEC 61499 specification. We propose a transformation from an IEC 61499 specification to the BPMN model. This model presents the automation from a business point-of-view, and it also enables quantitative analysis of process models. Specifically, it allows business analysts to perform analysis related to cost, resource allocation, and time in automation. This analysis is achieved by transforming the business processes to formal model in Maude rewriting logic. The viability of our proposals is illustrated using an industrial case study of a packaging system.","PeriodicalId":348063,"journal":{"name":"Proceedings of the 11th International Conference on the Internet of Things","volume":"72 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-11-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121099816","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}
Jaroslava Huber, Michael Haslgrübler, Martin Schobesberger, A. Ferscha, Viktorijo Malisa, Georg Effenberger
{"title":"Addressing Worker Safety and Accident Prevention with AI","authors":"Jaroslava Huber, Michael Haslgrübler, Martin Schobesberger, A. Ferscha, Viktorijo Malisa, Georg Effenberger","doi":"10.1145/3494322.3494342","DOIUrl":"https://doi.org/10.1145/3494322.3494342","url":null,"abstract":"The following document presents a conceptual and technical framework for a multimodal cognitive HumanAI system for worker safety and accident prevention. The authors incorporate two common approaches to safety (Safety-I and Safety-II), focusing primarily on the implementation of the latter. The framework of the system, which is envisioned to serve industrial workers, integrates a work assistant with an emergency break assistant. Purposeful interaction between the workers and the system is based on the system’s context awareness.","PeriodicalId":348063,"journal":{"name":"Proceedings of the 11th International Conference on the Internet of Things","volume":"43 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-11-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124536306","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}
{"title":"HYDRO: Hybrid Orchestration of In-Network Computations for the Internet of Things","authors":"Uthra Ambalavanan, Dennis Grewe, N. Nayak, J. Ott","doi":"10.1145/3494322.3494331","DOIUrl":"https://doi.org/10.1145/3494322.3494331","url":null,"abstract":"For some of the Internet-of-Things (IoT) applications, there is a need for processing huge amounts of data stemming from the network edge in near real-time manner. To this end several in-network computation frameworks have been developed which deploy functions in the computing infrastructure from the edge to the cloud. Some of these frameworks take a centralized approach to the orchestration problem with a central entity making deployment decisions of the functions. Others decentralize such decisions. In this paper, we present HYdrO, which is a hybrid approach to orchestration of functions, i.e., a centralized instance gathers the state of the network and provides hints to the individual computing nodes for an efficient resolution of computation requests. We show with a concrete example of Named Function Networking (NFN), a computing framework based on Information Centric Networking, that HYdrO can improve completion times of requests and distribute the load of the infrastructure.","PeriodicalId":348063,"journal":{"name":"Proceedings of the 11th International Conference on the Internet of Things","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-11-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123180602","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}
Jakob Hviid, Aslak Johansen, Gabe Fierro, M. Kjærgaard
{"title":"Service Portability and Information Discovery in Building Operating Systems using Semantic Modeling","authors":"Jakob Hviid, Aslak Johansen, Gabe Fierro, M. Kjærgaard","doi":"10.1145/3494322.3494337","DOIUrl":"https://doi.org/10.1145/3494322.3494337","url":null,"abstract":"To achieve cost-efficient IoT based Building Operating Systems (BOS), portable building services are needed. Most previous work has gone into hardware abstraction, while service abstraction has been neglected. This paper presents an information discovery mechanism for BOSs that is based on an ontology which integrates with other semantic models used in the space. The built environment is characterized by extreme heterogeneity; no buildings are entirely alike. Equipment is replaced or updated, control systems and building functionality evolve, as applications, models, forecasters, and controllers improve. For services deployed in such settings to operate at a scale, they must be robust to change. Describing service interfaces using a semantic model, together with the physical context in a building, enables applications to query for their service dependencies. Applications then depend on an abstract query, instead of specific services. For evaluation, nine services running on models of the service ecosystems of three concrete buildings, are implemented, demonstrated, and discussed. Results show services have successfully been made portable and adapts to the changing environments of buildings. Merging and splitting services without code changes to depending services also work as intended, as well as increasing system resilience, by arbitrating similar services.","PeriodicalId":348063,"journal":{"name":"Proceedings of the 11th International Conference on the Internet of Things","volume":"15 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-11-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116269835","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}
Georgios Sopidis, Abdelrahman Ahmad, Michael Haslgrübler, A. Ferscha, Martin Baresch
{"title":"Micro Activities Recognition and Macro Worksteps Classification for Industrial IoT Processes","authors":"Georgios Sopidis, Abdelrahman Ahmad, Michael Haslgrübler, A. Ferscha, Martin Baresch","doi":"10.1145/3494322.3494356","DOIUrl":"https://doi.org/10.1145/3494322.3494356","url":null,"abstract":"Automated understanding of worksteps in industrial assembly work is important for IoT-based assistive guidance technologies in employee-machine collaboration and for industrial IoT (IIoT) environments. Our aim is to identify macro worksteps using depth images and micro activities of employees during the assembly of ATM machines for auxiliary purposes in their daily complex tasks using hand-operated tools. Due to the advance of inertial measurement unit (IMU) technologies and pattern recognition systems [11], IMU based sensing together with machine learning have gained momentum on workstep recognition and were selected for this study in combination with a depth camera sensor which is mounted on a ceiling with a top-down angle. In this work the focus is at a seamless embedding of non-impeding body-worn IMUs or their integration into smart tools and IoT devices, and the depth sensor ensuring the privacy of the operators[5][7][8], allowing for unobstructed monitoring of tools’ usage pattern and thus assembly workstep recognition. The results for this study are evidenced with empirical observations of assembly workstep executions by (i) hand screwing, (ii) screw driver screwing, (iii) machine screwing, (iv) wrench screwing, with the null class being disproportionally dominant in the data set. Deep Learning models including LSTM, Temporal Convolutional Networks (TCN) and CNN architectures are proposed for the detection of micro activities and macro worksteps and the identification of the current workstep which will be beneficial also for the recognition of the transition between each two consecutive macro worksteps. A sophisticated counting mechanism of the classified activities is recognized as the next research challenge, with the extraction of features from each IMU sensor and the temporal information from the depth sensor, integrated in an IoT system.","PeriodicalId":348063,"journal":{"name":"Proceedings of the 11th International Conference on the Internet of Things","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-11-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122940860","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}
{"title":"Is it Nemo or Dory? Fast and accurate object detection for IoT and edge devices","authors":"Sudhanshu Agarwal, R. Vuduc","doi":"10.1145/3494322.3494335","DOIUrl":"https://doi.org/10.1145/3494322.3494335","url":null,"abstract":"Current state-of-the-art object detection neural networks, such as YOLO and SSD, are trained and developed on server-class GPUs. These neural networks do not scale down well to resource-constrained devices, with both accuracy and precision taking a significant hit at the expense of speed. This is particularly concerning as object detection algorithms are often used for low-powered devices, such as surveillance and smart-home cameras, where accuracy is critical. Therefore, these devices generally tend to forward data to servers for processing, which adds network latency. In other cases, algorithms developed for these devices shrink neural networks to reduce computation at the expense of accuracy, which is often not acceptable. We create an alternative object detection scheme for static-camera systems such as those used in surveillance and smart-home settings. Our model does not require positions of objects in training data, enabling us to retain only the relevant parts of a video frame for training data and reduce data storage cost while preserving privacy of other subjects in the video. When evaluated on static-camera video feeds using an NVIDIA Jetson Nano (a hybrid Arm and GPU IoT embedded device platform), our method increases throughput compared to Tiny YOLOv3. Further, it takes half of the time that Tiny YOLOv3 takes when run on OpenCV’s CUDA-optimized DNN framework per frame while having high accuracy of bounding box prediction and classification. The lessons learned from this work also suggest other strategies for tailoring the deployment of deep learning algorithms at the edge.","PeriodicalId":348063,"journal":{"name":"Proceedings of the 11th International Conference on the Internet of Things","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-11-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114661770","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}