F. Zonzini, Letizia Burioli, Andi Gashi, Nicola Francesco Mancini, L. Marchi
{"title":"A Tiny Convolutional Neural Network driven by System Identification for Vibration Anomaly Detection at the Extreme Edge","authors":"F. Zonzini, Letizia Burioli, Andi Gashi, Nicola Francesco Mancini, L. Marchi","doi":"10.1109/COINS57856.2023.10189195","DOIUrl":"https://doi.org/10.1109/COINS57856.2023.10189195","url":null,"abstract":"Vibration data analysis is the driving tool for the Structural Health Monitoring (SHM) of structures in the dynamic regime, i.e., structures showing important oscillatory behaviours, which largely dominate the transportation back-bone: from terrestrial/aerial vehicles (e.g., trains, aircraft, etc.) to the supporting infrastructures (e.g., bridges, viaducts, etc.). Outstanding opportunities have recently been disclosed in the field of Intelligent Transportation Systems (ITS) by the advent of sensor-near processing functionalities, eventually empowered by Artificial Intelligence (AI). The latter allow for the extraction of damage-sensitive features at the extreme edge, without the need of transmitting long time series over the monitoring network. In this work, we explore for the first time a novel anomaly detection workflow for on-sensor vibration diagnostics, which combines the unique advantages of embedded System Identification (eSysId) as a data compression strategy with the computational/energy advantages of Tiny Machine Learning (TinyML). Experimental results conducted on a representative SHM dataset demonstrate that the proposed pipeline can achieve high classification scores (above 90%) for the health assessment of the well-known Z24 bridge. In particular, the minimal inference time (less than 44 ms) and power consumption performed while running on three different general-purpose microprocessors make it a promising solution for the development of the next generation of SHM-oriented ITS.","PeriodicalId":328385,"journal":{"name":"2023 IEEE International Conference on Omni-layer Intelligent Systems (COINS)","volume":"29 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-07-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128215256","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":"NormSoftmax: Normalizing the Input of Softmax to Accelerate and Stabilize Training","authors":"Zixuan Jiang, Jiaqi Gu, D. Pan","doi":"10.1109/COINS57856.2023.10189242","DOIUrl":"https://doi.org/10.1109/COINS57856.2023.10189242","url":null,"abstract":"Softmax is a basic function that normalizes a vector to a probability distribution and is widely used in machine learning, most notably in cross-entropy loss function and dot product attention operations. However, the optimization of softmax-based models is sensitive to the input statistics change. We observe that the input of softmax changes significantly during the initial training stage, causing slow and unstable convergence when training the model from scratch. To remedy the optimization difficulty of softmax, we propose a simple yet effective substitution, named NormSoftmax, where the input vector is first normalized to unit variance and then fed to the standard softmax function. Similar to other existing normalization layers in machine learning models, NormSoftmax can stabilize and accelerate the training process, and also increase the robustness of the training procedure against hyperparameters. Experiments on Transformer-based models and convolutional neural networks validate that our proposed NormSoftmax is an effective plug-and-play module to stabilize and speed up the optimization of neural networks with cross-entropy loss or dot-product attention operations.","PeriodicalId":328385,"journal":{"name":"2023 IEEE International Conference on Omni-layer Intelligent Systems (COINS)","volume":"2 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-07-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125787538","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}
O. Ilina, Ksenia Tagirova, Artem Barger, V. Gorgadze
{"title":"SoK: How Blockchain and Tokenization Will Transform the Energy Sector","authors":"O. Ilina, Ksenia Tagirova, Artem Barger, V. Gorgadze","doi":"10.1109/COINS57856.2023.10189196","DOIUrl":"https://doi.org/10.1109/COINS57856.2023.10189196","url":null,"abstract":"Tokenization and blockchain technology might significantly improve the energy industry by making electricity production, transmission, and consumption more open, safe, and effective. Tokenization enables the establishment of digital assets that reflect ownership or access to energy resources and enables blockchain to trace the origin and flow of energy produced. By incorporating these elements, we can build a more distributed and egalitarian energy infrastructure and make it easier to use renewable energy. However, legal barriers, integration with current systems, and scalability remain obstacles to the widespread use of these technologies in the energy industry. Therefore, there is a need for more study and development before the energy industry can ultimately reap the benefits of blockchain and tokenization. Accordingly, in this work, we suggest a systematization of knowledge to capture recent development and summarize research in the blockchain domain and its application to the energy sector.","PeriodicalId":328385,"journal":{"name":"2023 IEEE International Conference on Omni-layer Intelligent Systems (COINS)","volume":"190 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-07-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121406625","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":"Similarity and Location-based Real-time Loop Closure: SNAPS for SLAM† in Unexplored-Environments","authors":"Kiavash Fathi, Alireza Darvishy, H. W. V. D. Venn","doi":"10.1109/COINS57856.2023.10189300","DOIUrl":"https://doi.org/10.1109/COINS57856.2023.10189300","url":null,"abstract":"Loop closure is an inseparable part of any accurate and reliable visual simultaneous localization and mapping (SLAM) algorithm for autonomous vehicles and mobile robots. Loop closure potentially decreases the impact of the cumulative drift while generating the map of the traversed environment. In this paper, a heuristic similarity and location-based approach for loop closure in unexplored environments is introduced. The current SLAM implementation on average requires 0.295 seconds per frame from which only 0.0270 seconds are the runtime latencies of the similarity and location-based real-time loop closure (SNAPS), which includes trajectory correction. The proposed approach results in a 65% decrease in the mean deviation from the ground truth. In the conducted study, neither conventional bag-of-words models, nor computationally expensive deep neural networks have been used to detect and perform loop closure, which makes the proposed approach both interpretable and efficient. In fact, we propose a method which tries to find loop closure candidates based on the location and also an interpretable similarity score attained from the generated thumbnails of the read frames instead of the local descriptors. Additionally, the employed discount factor applied on the pose trajectory update rule guarantees a consistent and accurate map. Lastly, the KITTI dataset is used to demonstrate the efficiency and accuracy of SNAPS for SLAM.","PeriodicalId":328385,"journal":{"name":"2023 IEEE International Conference on Omni-layer Intelligent Systems (COINS)","volume":"111 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-07-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"117240632","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}
Umuralp Kaytaz, Saba Ahmadian, F. Sivrikaya, S. Albayrak
{"title":"Graph Neural Network for Digital Twin-enabled Intelligent Transportation System Reliability","authors":"Umuralp Kaytaz, Saba Ahmadian, F. Sivrikaya, S. Albayrak","doi":"10.1109/COINS57856.2023.10189200","DOIUrl":"https://doi.org/10.1109/COINS57856.2023.10189200","url":null,"abstract":"Digital Twin (DT) is a promising technology for executing real-time monitoring and diagnostic operations in Intelligent Transportation Systems (ITSs). Guaranteeing safety and functionality in DT-enabled ITS domain requires reliable data streams among vehicular edge and cloud network resources. Therefore, in this paper, we propose Graph Neural Network based Anomaly Detection (GNN-AD) as a generic end-to-end framework for achieving reliability in multi-dimensional data streams of ITS platforms. Leveraging unsupervised and supervised machine learning algorithms, namely BiDirectional Generative Adversarial Network (BiGAN), affinity propagation and Graph Convolutional Network (GCN), we semantically model the complex distributions of ITS data streams and perform anomaly detection utilizing graph representations. Employing DT applications and diverse sensory measurements from the BeIntelli smart mobility platform of TU Berlin, we compare the performance of our approach with traditional and deep learning based methods in terms of detection accuracy and Receiver Operating Characteristics (ROC). Experimental results show that GNN-AD results in higher accuracy and larger Area Under the Curve (AUC) values compared to other supervised and unsupervised multi-stage end-to-end machine learning techniques.","PeriodicalId":328385,"journal":{"name":"2023 IEEE International Conference on Omni-layer Intelligent Systems (COINS)","volume":"416 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-07-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"117307067","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":"Use Case Examples of Ethereum Non-Fungible Tokens Tied to Assets Using ERC-4519","authors":"Javier Arcenegui, Rosario Arjona, I. Baturone","doi":"10.1109/COINS57856.2023.10189330","DOIUrl":"https://doi.org/10.1109/COINS57856.2023.10189330","url":null,"abstract":"Ethereum is a dynamic blockchain that grows every day thanks to a community that creates and decides on various of its issues. The community participates through Ethereum Improvement Proposals (EIPs) at many levels, from proposals that describe new standards for the creation of new tokens to proposals that define new ways for the Ethereum main network to generate new blocks. This paper describes how an Ethereum Request for Comments (ERC)-type EIP, the ERC-4519, was proposed last year to standardize the way to define non-fungible tokens (NFTs) representing assets that can generate their own Ethereum addresses and obey users and owners. Advantages provided by ERC-4519 in several use cases are illustrated. The examples show the facilities and the security improvements introduced in the management of both physical and digital assets by their owners and users. Particularly, use cases with physical assets such as IoT devices are illustrated.","PeriodicalId":328385,"journal":{"name":"2023 IEEE International Conference on Omni-layer Intelligent Systems (COINS)","volume":"46 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-07-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"117324399","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":"Reputation-based User Vehicle Assignment in Intelligent and Connected Vehicle Platoons","authors":"Subham Datta, Panagiota Nikolaou, M. Michael","doi":"10.1109/COINS57856.2023.10189279","DOIUrl":"https://doi.org/10.1109/COINS57856.2023.10189279","url":null,"abstract":"Vehicle platooning is a promising and emerging framework in intelligent transportation systems. Recent works consider reputation-based approaches for head selection in single platoons, in order to optimize safety and security. When a large number of user vehicles having the same destination want to benefit from platooning, several platoons need to be formed. To this end, user vehicles are allocated to different platoons, with each platoon being led by a platoon head. To ensure necessary network bandwidth and latency, the number of user vehicles between the newly formed platoons must be balanced. However, an arbitrary assignment of user vehicles in platoons can bias the future selection of the platoon heads in such reputation-based approaches. This work considers reputation-based platooning systems and proposes an optimal approach to balance the number of user vehicles in platoons while at the same time ensures fairness in the reputation score of platoon heads. A mixed-linear integer programming formulation is proposed, which provides an optimal allocation of the user vehicles in platoons based on the above objectives. The approach is validated using multiple synthetically generated datasets using 14 different input parameters. The obtained results demonstrate the optimality of the proposed method while achieving significant time performance speedups $(sim 140mathrm{x}$ on average) when compared to a brute-force exhaustive method.","PeriodicalId":328385,"journal":{"name":"2023 IEEE International Conference on Omni-layer Intelligent Systems (COINS)","volume":"5 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-07-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115521929","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}
Sajjad Parvin, Sallar Ahmadi-Pour, C. Jha, F. Sill, R. Drechsler
{"title":"Lo-RISK: Design of a Low Optical Leakage and High Performance RISC-V Core","authors":"Sajjad Parvin, Sallar Ahmadi-Pour, C. Jha, F. Sill, R. Drechsler","doi":"10.1109/COINS57856.2023.10189332","DOIUrl":"https://doi.org/10.1109/COINS57856.2023.10189332","url":null,"abstract":"Optical Probing Attacks (OPA) have been shown as an effective solution to bypass several known protection schemes of integrated circuits and to read-out sensitive information, like security keys or Intellectual Property (IP). As countermeasure, we propose a method for designing high-performance OPA-hardened digital circuits. Several existing solutions for OPA-hardened designs require changes in the fabrication process, resulting in high cost increase. Other approaches suffer from notable performance reductions and require significant changes of the employed gate libraries. In this work, we alleviate these limitations and propose a methodology to design high-performance OPA-hardened circuits. We achieve this by using a two-step methodology. First, we design a high-performance, and Low optical Leakage Dual-Rail Logic (LoL-DRL) gate library based on a standard CMOS gate library. That means, no complete redesign of the layout required, unlike comparable approaches. Second, we propose a lightweight synthesis technique to synthesize OPA-hardened circuits from conventional circuits. Furthermore, we applied our methodology on a RISC-V core to design the first OPA-hardened RISC-V core named Lo-RISK. Our method ensures a negligible performance penalty, however at the notable costs in terms of area and power.","PeriodicalId":328385,"journal":{"name":"2023 IEEE International Conference on Omni-layer Intelligent Systems (COINS)","volume":"9 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-07-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130681721","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":"Analyzing ModuloNET Against Transition Effects","authors":"Martin Schmid, E. Kavun","doi":"10.1109/COINS57856.2023.10189305","DOIUrl":"https://doi.org/10.1109/COINS57856.2023.10189305","url":null,"abstract":"While becoming an indispensable part of our lives, artificial intelligence comes also with some problems. For instance, the theft of trained Machine Learning (ML) models through side-channel attacks on their inference engines has started to be a significant threat. Recently, as an approach to protect such inference hardware, the protected ML model ModuloNET was proposed. ModuloNET is based on binarized neural networks and uses modular arithmetic throughout the calculations, making it compatible with usual masking techniques. In the original work, ModuloNET is proven to be first-order glitch-extended probing secure. However, recent research has shown that, in addition to glitches, transitions in memory cells have a significant impact on leakage of a circuit. Hence, in this paper, we analyse the security of ModuloNET under the first-order glitch-and transition-extended probing model. We start with discussing the original work on ModuloNET and later extend the proofs presented in the previous study to a stronger threat model. We also identify potential next steps as well as future directions for further research.","PeriodicalId":328385,"journal":{"name":"2023 IEEE International Conference on Omni-layer Intelligent Systems (COINS)","volume":"16 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-07-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122075290","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}
Dylan Wright, C. Mourning, Justin Murray, Treyce Albin, Charles Young, Ethan Krimins
{"title":"Weather Visibility Sensor Network: Platform Characterization","authors":"Dylan Wright, C. Mourning, Justin Murray, Treyce Albin, Charles Young, Ethan Krimins","doi":"10.1109/COINS57856.2023.10189260","DOIUrl":"https://doi.org/10.1109/COINS57856.2023.10189260","url":null,"abstract":"This paper presents the performance characterization of an Internet of things (IoT) based sub-GHz radio platform to be utilized as a sensor node in a Wireless Sensor Network (WSN). A CC1312Rx series microcontroller is the heart of each sensor node; it is battery powered and recharged by a small solar array affixed to each unit. The implemented sensor network will operate to transmit images from an in-unit electro-optical camera to a base station to perform machine learning assisted estimation of node localized atmospheric visibility. The main design characteristics of the proposed system are long range transmission capability and long term independent battery operation, while remaining low cost and easy to deploy. As such, this paper will focus on the performance characteristics of the individual sensor node platform in regards to transmission signal strength. This is analyzed from the measured Received Signal Strength Indicator (RSSI) in real world conditions. To validate the utilized radio subsystem long range capabilities, the measured data is compared against computer simulated data using a third party radio frequency (RF) software suite.","PeriodicalId":328385,"journal":{"name":"2023 IEEE International Conference on Omni-layer Intelligent Systems (COINS)","volume":"26 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-07-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122523435","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}