arXiv - CS - Systems and Control最新文献_第3页

TV White Space and LTE Network Optimization towards Energy Efficiency in Suburban and Rural Scenarios 电视白区和 LTE 网络优化，实现郊区和农村场景中的能源效率

arXiv - CS - Systems and Control Pub Date : 2024-05-04 DOI: arxiv-2405.02693

Rodney Martinez Alonso, David Plets, Margot Deruyck, Luc Martens, Wout Joseph

引用次数: 0

Unsupervised machine learning for data-driven classification of rock mass using drilling data: How can a data-driven system handle limitations in existing rock mass classification systems? 利用钻探数据进行岩体数据驱动分类的无监督机器学习：数据驱动系统如何处理现有岩体分类系统的局限性？

arXiv - CS - Systems and Control Pub Date : 2024-05-04 DOI: arxiv-2405.02631

T. F. Hansen, A. Aarset

{"title":"Unsupervised machine learning for data-driven classification of rock mass using drilling data: How can a data-driven system handle limitations in existing rock mass classification systems?","authors":"T. F. Hansen, A. Aarset","doi":"arxiv-2405.02631","DOIUrl":"https://doi.org/arxiv-2405.02631","url":null,"abstract":"Rock mass classification systems are crucial for assessing stability and risk\u0000in underground construction globally and guiding support and excavation design.\u0000However, systems developed primarily in the 1970s lack access to modern\u0000high-resolution data and advanced statistical techniques, limiting their\u0000effectiveness as decision-support systems. Initially, we outline the\u0000limitations observed in this context and later describe how a data-driven\u0000system, based on drilling data as detailed in this study, can overcome these\u0000limitations. Using extracted statistical information from thousands of MWD-data\u0000values in one-meter sections of a full tunnel profile, thus working as a\u0000signature of the rock mass, we have demonstrated that it is possible to form\u0000well-defined clusters that can act as a foundational basis for various rock\u0000mass classification systems. We reduced the dimensionality of 48-value vectors\u0000using nonlinear manifold learning techniques (UMAP) and linear principal\u0000component analysis (PCA) to enhance clustering. Unsupervised machine learning\u0000methods (HDBSCAN, Agglomerative Clustering, K-means) were employed to cluster\u0000the data, with hyperparameters optimised through multi-objective Bayesian\u0000optimisation for effective clustering. Using domain knowledge, we experienced\u0000improved clustering and system tuning opportunities in adding extra features to\u0000core clusters of MWD-data. We structured and correlated these clusters with\u0000physical rock mass properties, including labels of rock type and rock quality,\u0000and analysed cumulative distributions of key MWD-parameters for rock mass\u0000assessment to determine if clusters meaningfully differentiate rock masses. The\u0000ability of MWD data to form distinct rock mass clusters suggests substantial\u0000potential for future classification systems grounded in this objective,\u0000data-driven methodology, free from human bias.","PeriodicalId":501062,"journal":{"name":"arXiv - CS - Systems and Control","volume":"16 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-05-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"140882491","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}

引用次数: 0

Risk Assessment for Nonlinear Cyber-Physical Systems under Stealth Attacks 隐形攻击下非线性网络物理系统的风险评估

arXiv - CS - Systems and Control Pub Date : 2024-05-04 DOI: arxiv-2405.02633

Guang Chen, Zhicong Sun, Yulong Ding, Shuang-hua Yang

引用次数: 0

A Mathematical Model of the Hidden Feedback Loop Effect in Machine Learning Systems 机器学习系统中隐藏反馈回路效应的数学模型

arXiv - CS - Systems and Control Pub Date : 2024-05-04 DOI: arxiv-2405.02726

Andrey Veprikov, Alexander Afanasiev, Anton Khritankov

{"title":"A Mathematical Model of the Hidden Feedback Loop Effect in Machine Learning Systems","authors":"Andrey Veprikov, Alexander Afanasiev, Anton Khritankov","doi":"arxiv-2405.02726","DOIUrl":"https://doi.org/arxiv-2405.02726","url":null,"abstract":"Widespread deployment of societal-scale machine learning systems necessitates\u0000a thorough understanding of the resulting long-term effects these systems have\u0000on their environment, including loss of trustworthiness, bias amplification,\u0000and violation of AI safety requirements. We introduce a repeated learning\u0000process to jointly describe several phenomena attributed to unintended hidden\u0000feedback loops, such as error amplification, induced concept drift, echo\u0000chambers and others. The process comprises the entire cycle of obtaining the\u0000data, training the predictive model, and delivering predictions to end-users\u0000within a single mathematical model. A distinctive feature of such repeated\u0000learning setting is that the state of the environment becomes causally\u0000dependent on the learner itself over time, thus violating the usual assumptions\u0000about the data distribution. We present a novel dynamical systems model of the\u0000repeated learning process and prove the limiting set of probability\u0000distributions for positive and negative feedback loop modes of the system\u0000operation. We conduct a series of computational experiments using an exemplary\u0000supervised learning problem on two synthetic data sets. The results of the\u0000experiments correspond to the theoretical predictions derived from the\u0000dynamical model. Our results demonstrate the feasibility of the proposed\u0000approach for studying the repeated learning processes in machine learning\u0000systems and open a range of opportunities for further research in the area.","PeriodicalId":501062,"journal":{"name":"arXiv - CS - Systems and Control","volume":"16 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-05-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"140882512","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}

引用次数: 0

A Robust Data-Driven Iterative Control Method for Linear Systems with Bounded Disturbances 有界扰动线性系统的稳健数据驱动迭代控制方法

arXiv - CS - Systems and Control Pub Date : 2024-05-04 DOI: arxiv-2405.02537

Kaijian Hu, Tao Liu

{"title":"A Robust Data-Driven Iterative Control Method for Linear Systems with Bounded Disturbances","authors":"Kaijian Hu, Tao Liu","doi":"arxiv-2405.02537","DOIUrl":"https://doi.org/arxiv-2405.02537","url":null,"abstract":"This paper proposes a new robust data-driven control method for linear\u0000systems with bounded disturbances, where the system model and disturbances are\u0000unknown. Due to disturbances, accurately determining the true system becomes\u0000challenging using the collected dataset. Therefore, instead of designing\u0000controllers directly for the unknown true system, an available approach is to\u0000design controllers for all systems compatible with the dataset. To overcome the\u0000limitations of using a single dataset and benefit from collecting more data,\u0000multiple datasets are employed in this paper. Furthermore, a new iterative\u0000method is developed to address the challenges of using multiple datasets. Based\u0000on this method, this paper develops an offline and online robust data-driven\u0000iterative control method, respectively. Compared to the existing robust\u0000data-driven controller method, both proposed control methods iteratively\u0000utilize multiple datasets in the controller design process. This allows for the\u0000incorporation of numerous datasets, potentially reducing the conservativeness\u0000of the designed controller. Particularly, the online controller is iteratively\u0000designed by continuously incorporating online collected data into the\u0000historical data to construct new datasets. Lastly, the effectiveness of the\u0000proposed methods is demonstrated using a batch reactor.","PeriodicalId":501062,"journal":{"name":"arXiv - CS - Systems and Control","volume":"1 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-05-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"140882143","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}

引用次数: 0

Stable Distributed Online Feedback Optimization for Distribution System Voltage Regulation 配电系统电压调节的稳定分布式在线反馈优化

arXiv - CS - Systems and Control Pub Date : 2024-05-03 DOI: arxiv-2405.02487

Sen Zhan, Nikolaos G. Paterakis, Wouter van den Akker, Anne van der Molen, Johan Morren, J. G. Slootweg

引用次数: 0

New design of smooth PSO-IPF navigator with kinematic constraints 具有运动学约束条件的平滑 PSO-IPF 导航器新设计

arXiv - CS - Systems and Control Pub Date : 2024-05-03 DOI: arxiv-2405.01794

Mahsa Mohaghegh, Hedieh Jafarpourdavatgar, Samaneh Alsadat Saeedinia

{"title":"New design of smooth PSO-IPF navigator with kinematic constraints","authors":"Mahsa Mohaghegh, Hedieh Jafarpourdavatgar, Samaneh Alsadat Saeedinia","doi":"arxiv-2405.01794","DOIUrl":"https://doi.org/arxiv-2405.01794","url":null,"abstract":"Robotic applications across industries demand advanced navigation for safe\u0000and smooth movement. Smooth path planning is crucial for mobile robots to\u0000ensure stable and efficient navigation, as it minimizes jerky movements and\u0000enhances overall performance Achieving this requires smooth collision-free\u0000paths. Partial Swarm Optimization (PSO) and Potential Field (PF) are notable\u0000path-planning techniques, however, they may struggle to produce smooth paths\u0000due to their inherent algorithms, potentially leading to suboptimal robot\u0000motion and increased energy consumption. In addition, while PSO efficiently\u0000explores solution spaces, it generates long paths and has limited global\u0000search. On the contrary, PF methods offer concise paths but struggle with\u0000distant targets or obstacles. To address this, we propose Smoothed Partial\u0000Swarm Optimization with Improved Potential Field (SPSO-IPF), combining both\u0000approaches and it is capable of generating a smooth and safe path. Our research\u0000demonstrates SPSO-IPF's superiority, proving its effectiveness in static and\u0000dynamic environments compared to a mere PSO or a mere PF approach.","PeriodicalId":501062,"journal":{"name":"arXiv - CS - Systems and Control","volume":"13 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-05-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"140882337","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}

引用次数: 0

A Flow-Based Model for Conditional and Probabilistic Electricity Consumption Profile Generation and Prediction 基于流量的条件和概率用电曲线生成与预测模型

arXiv - CS - Systems and Control Pub Date : 2024-05-03 DOI: arxiv-2405.02180

Weijie Xia, Chenguang Wang, Peter Palensky, Pedro P. Vergara

引用次数: 0

Incremental Volt/Var Control for Distribution Networks via Chance-Constrained Optimization 通过机会约束优化实现配电网络的增量电压/电流控制

arXiv - CS - Systems and Control Pub Date : 2024-05-03 DOI: arxiv-2405.02511

Antonin Colot, Elisabetta Perotti, Mevludin Glavic, Emiliano Dall'Anese

{"title":"Incremental Volt/Var Control for Distribution Networks via Chance-Constrained Optimization","authors":"Antonin Colot, Elisabetta Perotti, Mevludin Glavic, Emiliano Dall'Anese","doi":"arxiv-2405.02511","DOIUrl":"https://doi.org/arxiv-2405.02511","url":null,"abstract":"This paper considers an incremental Volt/Var control scheme for distribution\u0000systems with high integration of inverter-interfaced distributed generation\u0000(such as photovoltaic systems). The incremental Volt/Var controller is\u0000implemented with the objective of minimizing reactive power usage while\u0000maintaining voltages within safe limits sufficiently often. To this end, the\u0000parameters of the incremental Volt/Var controller are obtained by solving a\u0000chance-constrained optimization problem, where constraints are designed to\u0000ensure that voltage violations do not occur more often than a pre-specified\u0000probability. This approach leads to cost savings in a controlled, predictable\u0000way, while still avoiding significant over- or under-voltage issues. The\u0000proposed chance-constrained problem is solved using a successive convex\u0000approximation method. Once the gains are broadcast to the inverters, no\u0000additional communication is required since the controller is implemented\u0000locally at the inverters. The proposed method is successfully tested on a\u0000low-voltage 42-nodes network.","PeriodicalId":501062,"journal":{"name":"arXiv - CS - Systems and Control","volume":"26 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-05-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"140882206","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}

引用次数: 0

The Cambridge RoboMaster: An Agile Multi-Robot Research Platform 剑桥 RoboMaster：敏捷的多机器人研究平台

arXiv - CS - Systems and Control Pub Date : 2024-05-03 DOI: arxiv-2405.02198

Jan Blumenkamp, Ajay Shankar, Matteo Bettini, Joshua Bird, Amanda Prorok

{"title":"The Cambridge RoboMaster: An Agile Multi-Robot Research Platform","authors":"Jan Blumenkamp, Ajay Shankar, Matteo Bettini, Joshua Bird, Amanda Prorok","doi":"arxiv-2405.02198","DOIUrl":"https://doi.org/arxiv-2405.02198","url":null,"abstract":"Compact robotic platforms with powerful compute and actuation capabilities\u0000are key enablers for practical, real-world deployments of multi-agent research.\u0000This article introduces a tightly integrated hardware, control, and simulation\u0000software stack on a fleet of holonomic ground robot platforms designed with\u0000this motivation. Our robots, a fleet of customised DJI Robomaster S1 vehicles,\u0000offer a balance between small robots that do not possess sufficient compute or\u0000actuation capabilities and larger robots that are unsuitable for indoor\u0000multi-robot tests. They run a modular ROS2-based optimal estimation and control\u0000stack for full onboard autonomy, contain ad-hoc peer-to-peer communication\u0000infrastructure, and can zero-shot run multi-agent reinforcement learning (MARL)\u0000policies trained in our vectorized multi-agent simulation framework. We present\u0000an in-depth review of other platforms currently available, showcase new\u0000experimental validation of our system's capabilities, and introduce case\u0000studies that highlight the versatility and reliabilty of our system as a\u0000testbed for a wide range of research demonstrations. Our system as well as\u0000supplementary material is available online:\u0000https://proroklab.github.io/cambridge-robomaster","PeriodicalId":501062,"journal":{"name":"arXiv - CS - Systems and Control","volume":"27 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-05-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"140882203","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}

引用次数: 0