arXiv - CS - Systems and Control最新文献_第7页

Distributed Robust Learning based Formation Control of Mobile Robots based on Bioinspired Neural Dynamics 基于生物启发神经动力学的移动机器人分布式鲁棒学习编队控制

arXiv - CS - Systems and Control Pub Date : 2024-03-23 DOI: arxiv-2403.15716

Zhe Xu, Tao Yan, Simon X. Yang, S. Andrew Gadsden, Mohammad Biglarbegian

引用次数: 0

Risk-Calibrated Human-Robot Interaction via Set-Valued Intent Prediction 通过集值意图预测进行风险校准的人机交互

arXiv - CS - Systems and Control Pub Date : 2024-03-23 DOI: arxiv-2403.15959

Justin Lidard, Hang Pham, Ariel Bachman, Bryan Boateng, Anirudha Majumdar

{"title":"Risk-Calibrated Human-Robot Interaction via Set-Valued Intent Prediction","authors":"Justin Lidard, Hang Pham, Ariel Bachman, Bryan Boateng, Anirudha Majumdar","doi":"arxiv-2403.15959","DOIUrl":"https://doi.org/arxiv-2403.15959","url":null,"abstract":"Tasks where robots must cooperate with humans, such as navigating around a\u0000cluttered home or sorting everyday items, are challenging because they exhibit\u0000a wide range of valid actions that lead to similar outcomes. Moreover,\u0000zero-shot cooperation between human-robot partners is an especially challenging\u0000problem because it requires the robot to infer and adapt on the fly to a latent\u0000human intent, which could vary significantly from human to human. Recently,\u0000deep learned motion prediction models have shown promising results in\u0000predicting human intent but are prone to being confidently incorrect. In this\u0000work, we present Risk-Calibrated Interactive Planning (RCIP), which is a\u0000framework for measuring and calibrating risk associated with uncertain action\u0000selection in human-robot cooperation, with the fundamental idea that the robot\u0000should ask for human clarification when the risk associated with the\u0000uncertainty in the human's intent cannot be controlled. RCIP builds on the\u0000theory of set-valued risk calibration to provide a finite-sample statistical\u0000guarantee on the cumulative loss incurred by the robot while minimizing the\u0000cost of human clarification in complex multi-step settings. Our main insight is\u0000to frame the risk control problem as a sequence-level multi-hypothesis testing\u0000problem, allowing efficient calibration using a low-dimensional parameter that\u0000controls a pre-trained risk-aware policy. Experiments across a variety of\u0000simulated and real-world environments demonstrate RCIP's ability to predict and\u0000adapt to a diverse set of dynamic human intents.","PeriodicalId":501062,"journal":{"name":"arXiv - CS - Systems and Control","volume":"36 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-03-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"140297691","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

On the role of network structure in learning to coordinate with bounded rationality 论网络结构在有界理性协调学习中的作用

arXiv - CS - Systems and Control Pub Date : 2024-03-23 DOI: arxiv-2403.15683

Yifei Zhang, Marcos M. Vasconcelos

{"title":"On the role of network structure in learning to coordinate with bounded rationality","authors":"Yifei Zhang, Marcos M. Vasconcelos","doi":"arxiv-2403.15683","DOIUrl":"https://doi.org/arxiv-2403.15683","url":null,"abstract":"Many socioeconomic phenomena, such as technology adoption, collaborative\u0000problem-solving, and content engagement, involve a collection of agents\u0000coordinating to take a common action, aligning their decisions to maximize\u0000their individual goals. We consider a model for networked interactions where\u0000agents learn to coordinate their binary actions under a strict bound on their\u0000rationality. We first prove that our model is a potential game and that the\u0000optimal action profile is always to achieve perfect alignment at one of the two\u0000possible actions, regardless of the network structure. Using a stochastic\u0000learning algorithm known as Log Linear Learning, where agents have the same\u0000finite rationality parameter, we show that the probability of agents\u0000successfully agreeing on the correct decision is monotonically increasing in\u0000the number of network links. Therefore, more connectivity improves the accuracy\u0000of collective decision-making, as predicted by the phenomenon known as Wisdom\u0000of Crowds. Finally, we show that for a fixed number of links, a regular network\u0000maximizes the probability of success. We conclude that when using a network of\u0000irrational agents, promoting more homogeneous connectivity improves the\u0000accuracy of collective decision-making.","PeriodicalId":501062,"journal":{"name":"arXiv - CS - Systems and Control","volume":"180 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-03-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"140297463","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

Balancing Fairness and Efficiency in Energy Resource Allocations 平衡能源资源分配的公平与效率

arXiv - CS - Systems and Control Pub Date : 2024-03-22 DOI: arxiv-2403.15616

Jiayi Li, Matthew Motoki, Baosen Zhang

引用次数: 0

Transactive Local Energy Markets Enable Community-Level Resource Coordination Using Individual Rewards 交易型地方能源市场利用个人奖励实现社区层面的资源协调

arXiv - CS - Systems and Control Pub Date : 2024-03-22 DOI: arxiv-2403.15617

Daniel C. May, Petr Musilek

{"title":"Transactive Local Energy Markets Enable Community-Level Resource Coordination Using Individual Rewards","authors":"Daniel C. May, Petr Musilek","doi":"arxiv-2403.15617","DOIUrl":"https://doi.org/arxiv-2403.15617","url":null,"abstract":"ALEX (Autonomous Local Energy eXchange) is an economy-driven, transactive\u0000local energy market where each participating building is represented by a\u0000rational agent. Relying solely on building-level information, this agent\u0000minimizes its electricity bill by automating distributed energy resource\u0000utilization and trading. This study examines ALEX's capabilities to align\u0000participant and grid-stakeholder interests and assesses ALEX's impact on short-\u0000and long-term intermittence using a set of community net-load metrics, such as\u0000ramping rate, load factor, and peak load. The policies for ALEX's rational\u0000agents are generated using dynamic programming through value iteration in\u0000conjunction with iterative best response. This facilitates comparing ALEX and a\u0000benchmark energy management system, which optimizes building-level\u0000self-consumption, ramping rate, and peak net load. Simulations are performed\u0000using the open-source CityLearn2022 dataset to provide a pathway for\u0000benchmarking by future studies. The experiments demonstrate that ALEX enables\u0000the coordination of distributed energy resources across the community.\u0000Remarkably, this community-level coordination occurs even though the system is\u0000populated by agents who only access building-level information and selfishly\u0000maximize their own relative profit. Compared to the benchmark energy management\u0000system, ALEX improves across all metrics.","PeriodicalId":501062,"journal":{"name":"arXiv - CS - Systems and Control","volume":"35 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-03-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"140301346","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

On the Variational Interpretation of Mirror Play in Monotone Games 论单调博弈中镜像博弈的变式解释

arXiv - CS - Systems and Control Pub Date : 2024-03-22 DOI: arxiv-2403.15636

Yunian Pan, Tao Li, Quanyan Zhu

{"title":"On the Variational Interpretation of Mirror Play in Monotone Games","authors":"Yunian Pan, Tao Li, Quanyan Zhu","doi":"arxiv-2403.15636","DOIUrl":"https://doi.org/arxiv-2403.15636","url":null,"abstract":"Mirror play (MP) is a well-accepted primal-dual multi-agent learning\u0000algorithm where all agents simultaneously implement mirror descent in a\u0000distributed fashion. The advantage of MP over vanilla gradient play lies in its\u0000usage of mirror maps that better exploit the geometry of decision domains.\u0000Despite extensive literature dedicated to the asymptotic convergence of MP to\u0000equilibrium, the understanding of the finite-time behavior of MP before\u0000reaching equilibrium is still rudimentary. To facilitate the study of MP's\u0000non-equilibrium performance, this work establishes an equivalence between MP's\u0000finite-time primal-dual path (mirror path) in monotone games and the\u0000closed-loop Nash equilibrium path of a finite-horizon differential game,\u0000referred to as mirror differential game (MDG). Our construction of MDG rests on\u0000the Brezis-Ekeland variational principle, and the stage cost functional for MDG\u0000is Fenchel coupling between MP's iterates and associated gradient updates. The\u0000variational interpretation of mirror path in static games as the equilibrium\u0000path in MDG holds in deterministic and stochastic cases. Such a variational\u0000interpretation translates the non-equilibrium studies of learning dynamics into\u0000a more tractable equilibrium analysis of dynamic games, as demonstrated in a\u0000case study on the Cournot game, where MP dynamics corresponds to a linear\u0000quadratic game.","PeriodicalId":501062,"journal":{"name":"arXiv - CS - Systems and Control","volume":"32 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-03-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"140297608","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

Adaptive Dual Covariance Steering with Active Parameter Estimation 具有主动参数估计功能的自适应双协方差转向系统

arXiv - CS - Systems and Control Pub Date : 2024-03-22 DOI: arxiv-2403.15590

Jacob W. Knaup, Panagiotis Tsiotras

引用次数: 0

Autonomous Driving With Perception Uncertainties: Deep-Ensemble Based Adaptive Cruise Control 感知不确定性下的自动驾驶：基于深度集合的自适应巡航控制

arXiv - CS - Systems and Control Pub Date : 2024-03-22 DOI: arxiv-2403.15577

Xiao Li, H. Eric Tseng, Anouck Girard, Ilya Kolmanovsky

{"title":"Autonomous Driving With Perception Uncertainties: Deep-Ensemble Based Adaptive Cruise Control","authors":"Xiao Li, H. Eric Tseng, Anouck Girard, Ilya Kolmanovsky","doi":"arxiv-2403.15577","DOIUrl":"https://doi.org/arxiv-2403.15577","url":null,"abstract":"Autonomous driving depends on perception systems to understand the\u0000environment and to inform downstream decision-making. While advanced perception\u0000systems utilizing black-box Deep Neural Networks (DNNs) demonstrate human-like\u0000comprehension, their unpredictable behavior and lack of interpretability may\u0000hinder their deployment in safety critical scenarios. In this paper, we develop\u0000an Ensemble of DNN regressors (Deep Ensemble) that generates predictions with\u0000quantification of prediction uncertainties. In the scenario of Adaptive Cruise\u0000Control (ACC), we employ the Deep Ensemble to estimate distance headway to the\u0000lead vehicle from RGB images and enable the downstream controller to account\u0000for the estimation uncertainty. We develop an adaptive cruise controller that\u0000utilizes Stochastic Model Predictive Control (MPC) with chance constraints to\u0000provide a probabilistic safety guarantee. We evaluate our ACC algorithm using a\u0000high-fidelity traffic simulator and a real-world traffic dataset and\u0000demonstrate the ability of the proposed approach to effect speed tracking and\u0000car following while maintaining a safe distance headway. The\u0000out-of-distribution scenarios are also examined.","PeriodicalId":501062,"journal":{"name":"arXiv - CS - Systems and Control","volume":"28 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-03-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"140297609","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

Quantifying the Aggregate Flexibility of EV Charging Stations for Dependable Congestion Management Products: A Dutch Case Study 为可靠的拥堵管理产品量化电动汽车充电站的总体灵活性：荷兰案例研究

arXiv - CS - Systems and Control Pub Date : 2024-03-20 DOI: arxiv-2403.13367

Nanda Kishor Panda, Simon H. Tindemans

{"title":"Quantifying the Aggregate Flexibility of EV Charging Stations for Dependable Congestion Management Products: A Dutch Case Study","authors":"Nanda Kishor Panda, Simon H. Tindemans","doi":"arxiv-2403.13367","DOIUrl":"https://doi.org/arxiv-2403.13367","url":null,"abstract":"Electric vehicles (EVs) play a crucial role in the transition towards\u0000sustainable modes of transportation and thus are critical to the energy\u0000transition. As their number grows, managing the aggregate power of EV charging\u0000is crucial to maintain grid stability and mitigate congestion. This study\u0000analyses more than 500 thousand real charging transactions in the Netherlands\u0000to explore the challenge and opportunity for the energy system presented by EV\u0000growth and smart charging flexibility. Specifically, it analyses the collective\u0000ability to provide congestion management services according to the\u0000specifications of those services in the Netherlands. In this study, a\u0000data-driven model of charging behaviour is created to explore the implications\u0000of delivering dependable congestion management services at various aggregation\u0000levels and types of service. The probability of offering specific grid services\u0000by different categories of charging stations (CS) is analysed. These\u0000probabilities can help EV aggregators, such as charging point operators, make\u0000informed decisions about offering congestion mitigation products per relevant\u0000regulations and distribution system operators to assess their potential. The\u0000ability to offer different flexibility products, namely re-dispatch and\u0000capacity limitation, for congestion management, is assessed using various\u0000dispatch strategies. Next, machine learning models are used to predict the\u0000probability of CSs being able to deliver these products, accounting for\u0000uncertainties. Results indicate that residential charging locations have\u0000significant potential to provide both products during evening peak hours. While\u0000shared EVs offer better certainty regarding arrival and departure times, their\u0000small fleet size currently restricts their ability to meet the minimum order\u0000size of flexible products.","PeriodicalId":501062,"journal":{"name":"arXiv - CS - Systems and Control","volume":"132 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-03-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"140196614","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

Observer-Based Environment Robust Control Barrier Functions for Safety-critical Control with Dynamic Obstacles 基于观测器的环境鲁棒控制障碍函数，用于具有动态障碍物的安全关键控制

arXiv - CS - Systems and Control Pub Date : 2024-03-20 DOI: arxiv-2403.13288

Ying Shuai Quan, Jian Zhou, Erik Frisk, Chung Choo Chung

引用次数: 0