Gurkan Solmaz, Pankaj Baranwal, Flavio Cirillo: “CountMeIn: Adaptive Crowd Estimation with Wi-Fi in Smart Cities”, IEEE PerCom 2022 (accepted)
The widespread use of pervasive sensing technologies such as wireless sensors and street cameras allows easy deployment of crowd estimation solutions in smart cities. However, existing Wi-Fi-based systems do not provide highly accurate crowd size estimation. Furthermore, these systems do not adapt to the dynamic changes in-the-wild, such as unexpected crowd gatherings. To address the crowd estimation problem, this paper presents a new adaptive machine learning system, called CountMeIn, using polynomial regression and neural networks. The approach transfers the calibration task from cameras to machine learning after a short training with people counting from stereoscopic cameras, Wi-Fi probe packets, and temporal features. After the training, CountMeIn calibrates Wi-Fi using the trained model and maintains high accuracy for longer duration without cameras. We test the approach in our pilot study in Gold Coast city, Australia for about 5-month time period. Compared to the state-of-the-art approach, CountMeIn achieves 44% and 72% error reductions in minutely and hourly crowd estimations.
Presented at: IEEE PerCom 2022
Flavio Cirillo, Bin Cheng, Raffaele Porcellana, Marco Russo, Gurkan Solmaz, Hisashi Sakamoto, Simon Pietro Romano: “IntentKeeper: Intent-oriented Data Usage Control for Federated Data Analytics”, IEEE LCN 2020
Data usage control is of utmost importance for federated data analytics across multiple business domains. However, the existing data usage control approaches are limited due to their complexity and inefficiency. This paper proposes an intentoriented data usage control system for federated data analytics, called IntentKeeper. The system allows users to specify intents for data usage policies and services easily. Thus, it reduces the data sharing complexity for data providers and consumers. Moreover, IntentKeeper enforces preventive and proactive data usage control for better security and efficiency through joint decisions based on policy enforcement and service orchestration. The use case validations for the automotive industry scenario show that IntentKeeper significantly reduces the complexity of policy specification (up to 75% for moderately complex scenarios) compared to the state-of-the-art flow-based approach. Lastly, the experimental results show that the IntentKeeper system provides sufficiently short response times (less than 40ms) with minimal overhead (less than 10ms).
Presented at: IEEE Conference on Local Computer Networks (LCN), 2020
Available at: IEEE Xplore
Jonathan Fuerst, Mauricio Fadel Argerich, Bin Cheng, Ernö Kovacs: “Knowledge Infusion for Robust and Transferable Machine Learning in IoT”, OJIOT, 2020
Machine learning (ML) applications in Internet of Things (IoT) scenarios face the issue that supervision signals,such as labeled data, are scarce and expensive to obtain. For example, it often requires a human to manually labelevents in a data stream by observing the same events in the real world. In addition, the performance of trainedmodels usually depends on a specific context: (1) location, (2) time and (3) data quality. This context is not staticin reality, making it hard to achieve robust and transferable machine learning for IoT systems in practice. In thispaper, we address these challenges with an envisioned method that we name Knowledge Infusion. First, we presenttwo past case studies in which we combined external knowledge with traditional data-driven machine learning inIoT scenarios to ease the supervision effort: (1) a weak-supervision approach for the IoT domain to auto-generatelabels based on external knowledge (e.g., domain knowledge) encoded in simple labeling functions. Our evaluationfor transport mode classification achieves a micro-F1 score of80.2%, with only seven labeling functions, on parwith a fully supervised model that relies on hand-labeled data. (2) We introduce guiding functions to ReinforcementLearning (RL) to guide the agents’ decisions and experience. In initial experiments, ourguided reinforcementlearningachieves more than three times higher reward in the beginning of its training than an agent with no externalknowledge. We use the lessons learned from these experiences to develop our vision ofknowledge infusion. Inknowledge infusion, we aim to automate the inclusion of knowledge from existing knowledge bases and domainexperts to combine it with traditional data-driven machine learning techniques during setup/training phase, butalso during the execution phase.
Presented at: The International Workshop on Very Large Internet of Things (VLIoT 2020)
Gurkan Solmaz: “Tutorial Proposal: Combining IoT and ML for situation classification”, ECAI 2020
Presented at: The 24th European Conference on Artificial Intelligence (ECAI 2020)
G. Solmaz, J. Fürst, S. Aytac, F.-J. Wu: “Group-In: Group Inference from Wireless Traces of Mobile Devices”, ACM/IEEE IPSN’20, April 2020
This paper proposes Group-In, a wireless scanning system to detect static or mobile people groups in indoor or outdoor environments. Group-In collects only wireless traces from the Bluetooth-enabled mobile devices for group inference. The key problem addressed in this work is to detect not only static groups but also moving groups with a multi-phased approach based only noisy wireless Received Signal Strength Indicator (RSSIs) observed by multiple wireless scanners without localization support. We propose new centralized and decentralized schemes to process the sparse and noisy wireless data, and leverage graph-based clustering techniques for group detection from short-term and long-term aspects. Group-In provides two outcomes: 1) group detection in short time intervals such as two minutes and 2) long-term linkages such as a month. To verify the performance, we conduct two experimental studies. On consists of 27 controlled scenarios in the lab environments. The other is a real-world scenario where we place Bluetooth scanners in an office environment, and employees carry beacons for more than one month. Both the controlled and real-world experiments result in high accuracy group detection in short time intervals and sampling liberties in terms of the Jaccard index and pairwise similarity coefficient.
Full paper download: Group-In_Group_Inference_from_Wireless_Traces_of_Mobile_Devices_2005.12848.pdf
F. Cirillo, D. Gomez, L. Diez, I. Elicegui Maestro, T. Barrie Juel Gilbert, R. Akhavan, “Smart City IoT Services Creation through Large Scale Collaboration”, IEEE IoT Journal, March
K. Shankari, J. Fuerst, M. Fadel Argerich, E. Avramidis, J. Zhang. (2020). “MobilityNet: Towards a Public Dataset for Multimodal Mobility Research”, Climate Change AI 2020, February
F. Cirillo, G. Solmaz, E. L. Berz, M. Bauer, B. Cheng and E. Kovacs “A Standard-based Open Source IoT Platform: FIWARE.” IEEE Internet of Things Magazine (IoTM), 2020.
Flavio Cirillo, Bin Cheng, Raffaele Porcellana, Marco Russo,Gurkan Solmaz, Hisashi Sakamoto, Simon Pietro Romano: “IntentKeeper: Goal-Oriented Data Usage Control for Federated Data Analytics”, IEEE LCN 2020
Data usage control is of utmost importance for fed-erated data analytics across multiple business domains. However,the existing data usage control approaches are limited due totheir complexity and inefficiency. This paper proposes anintent-orienteddata usage control system for federated data analytics,calledIntentKeeper. The system allows users to specify intentsfor data usage policies and services easily. Thus, it reducesthe data sharing complexity for data providers and consumers.Moreover, IntentKeeper enforces preventive and proactive datausage control for better security and efficiency through jointdecisions based on policy enforcement and service orchestration.The use case validations for the automotive industry scenarioshow that IntentKeeper significantly reduces the complexity ofpolicy specification (up to 75% for moderately complex scenarios)compared to the state-of-the-art flow-based approach. Lastly, theexperimental results show that the IntentKeeper system providessufficiently short response times (less than 40ms) with minimaloverhead (less than 10ms).
Index Terms—IoT platforms, federation, privacy and security,data usage control
M.F. Argerich, J. Fuerst, B. Cheng: "Tutor4RL: Guiding Reinforcement Learning with External Knowledge", AAAI Spring Symposium: Combining Machine Learning with Knowledge Engineering, December 2019
J. Fuerst, M.F. Argerich, K. Shankari, G. Solmaz, B. Cheng: "Applying Weak Supervision to Mobile Sensor Data: Experiences with Transport Mode Detection", Artificial Intelligence of Things (AIoT) Workshop at AAAI 2020, November 2019
G. Solmaz and D. Turgut. "A Survey of Human Mobility Models", in IEEE Access, vol. 7, pp. 125711-125731, September 2019.
B. Cheng, G. Solmaz, F. Cirillo, “Intent-based Fog Computing with FogFlow”, demo paper accepted by the 44th IEEE Conference on Local Computer Networks (LCN’19). August 2019
G. Solmaz, E. Luis Berz, M. Dolatabadi Farahani, S. Aytac, J. Fuerst, B. Cheng, J. den Ouden: "Learn from IoT: Pedestrian Detection and Intention Prediction for Autonomous Driving", ACM Mobicom’19, SMAS workshop, 2019. July 2019
B. Cheng, J. Fürst, G. Solmaz, T. Sanada, “Fog Function: Serverless Fog Computing for Data Intensive IoT Services” in the proceedings of 2019 IEEE International Conference on Services Computing (SCC), pp. 28-35, Milan, July, 2019
G. Romano, M. Falcitelli, S. Noto, P. Pagano, M. Djurica, G. Karagiannis, G. Solmaz, “Autonomous driving progressed by oneM2M, the experience of Autopilot Project”, Accepted to EuCNC’19, June 2019
F. Cirillo, N. Capuano, E. Kovacs, S. P. Romano: "LIoTS: League of IoT Sovereignties. A Scalable-approach for a Transparent Privacy-safe Federation of Secured IoT Platforms", IEEE LCN 2019
Jose Manuel Cantera, Martin Bauer, Abdullah Abbas: “Position Statement from ETSI ISG CIM”, W3C Workshop on Web Standardization for Graph Data, March 2019
F. Cirillo, D. Straeten, D. Gomez, J. Gato, L. Diez, I. Elicegui Maestro, R. Akhavan, “Atomic Services: sustainable ecosystem of smart city services through pan-European collaboration”, IEEE Global IoT Summit 2019. March 2019
J. G. An, F. Le Gall, J. Kim, J. Yun, J. Hwang, M. Bauer, M. Zhao, J. Song, "Towards Global IoTenabled Smart Cities Interworking using Adaptive Semantic Adapter" in IEEE Internet of Things Journal. Accepted Date March 2019
M. F. Argerich, B. Cheng, J. Fuerst, “Reinforcement Learning based Orchestration for Elastic Services”. Accepted to WF-IoT’19. Accepted Date Feb 2019
G. Solmaz, et al.: Towards Understanding Crowd Mobility in Smart Cities through Internet of Things, Accepted to IEEE Communications Magazine, Feature Topic on Crowd Management. Jan 2019
F. Cirillo, FJ. Wu, G. Solmaz, E. Kovacs: Embracing the Future of Internet of Things, Best Paper Award at GioTS 2018, Accepted to MDPI Sensors, Special Issue "Selected Papers from the 2nd Global IoT Summit: IoT Technologies and Applications for the Benefit of Society". Jan 2019
Jorge Lanza, Luis Sanchez, Juan Ramón Santana, Rachit Agarwal, Nikos Kefalakis, Paul Grace, Tarek Elsaleh, Mengxuan Zhao, Elias Tragos, Hung Nguyen, Flavio Cirillo, Ronald Steinke, John Soldatos: “Experimentation as a Service over Semantically Interoperable Internet of Things Testbeds”, IEEE Access, September 2018
Infrastructures enabling experimental assessment of Internet of Things (IoT) solutions arescarce. Moreover, such infrastructures are typically bound to a specific application domain, thus, notfacilitating the testing of solutions with a horizontal approach. This paper presents a platform that supportsExperimentation as s Service (EaaS) over a federation of IoT testbeds. This platform brings two majoradvances. First, it leverages semantic web technologies to enable interoperability so that testbed agnosticaccess to the underlying facilities is allowed. Second, a set of tools ease both the experimentation workflowand the federation of other IoT deployments, independently of their domain of interest. Apart from theplatform specification, this paper presents how this design has been actually instantiated into a cloud-basedEaaS platform that has been used for supporting a wide variety of novel experiments targeting differentresearch and innovation challenges. In this respect, this paper summarizes some of the experiences fromthese experiments and the key performance metrics that this instance of the platform has exhibited duringthe experimentation.