Scholarlink

<p>Connected and automated vehicles (CAVs) have improved traffic efficiency in mixed traffic systems, facilitating the development of diverse optimal control strategies. However, despite promising results in extensive numerical simulations and small-scale tests, most of these strategies rely heavily on specific parameters within mixed traffic systems. This paper introduces a novel data-driven control approach for mixed traffic systems with completely unknown parameters. In contrast to model-based control, the proposed approach leverages real-time state measurements and collects data to solve the optimal control problem, eliminating the need for prior knowledge of the system dynamics. In this context, an <InlineEquation ID="IEq2"> <InlineMediaObject> <ImageObject Color="BlackWhite" FileRef="12555_2024_928_Article_IEq1.gif" Format="GIF" Height="17" Rendition="HTML" Resolution="72" Type="Linedraw" Width="26" /> </InlineMediaObject> <EquationSource Format="TEX">\(\cal{H}_{2}\)</EquationSource> <EquationSource Format="MATHML"><math display="block"> <msub> <mrow> <mi mathvariant="script">H</mi> </mrow> <mrow> <mn class="MJX-tex-caligraphic" mathvariant="script">2</mn> </mrow> </msub> </math></EquationSource> </InlineEquation> norm minimization objective is formulated. Furthermore, this formulation is converted into a convex optimization problem by leveraging the technique of linear matrix inequalities (LMIs). Finally, through simulation results, it is verified that the proposed methods achieve desirable performance in controlling a mixed-vehicular network.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Data-driven \(\cal{H}_{2}\) Optimal Control of Connected Vehicles in Mixed Traffic

  • 期刊论文

  • 发表日期：

    2025年8月13日

查看全文
  • Xiaodi Wang,
  • Xuan Cai,
  • Lei Zhang,
  • Shuxin Liu

摘要

Connected and automated vehicles (CAVs) have improved traffic efficiency in mixed traffic systems, facilitating the development of diverse optimal control strategies. However, despite promising results in extensive numerical simulations and small-scale tests, most of these strategies rely heavily on specific parameters within mixed traffic systems. This paper introduces a novel data-driven control approach for mixed traffic systems with completely unknown parameters. In contrast to model-based control, the proposed approach leverages real-time state measurements and collects data to solve the optimal control problem, eliminating the need for prior knowledge of the system dynamics. In this context, an \(\cal{H}_{2}\) H 2 norm minimization objective is formulated. Furthermore, this formulation is converted into a convex optimization problem by leveraging the technique of linear matrix inequalities (LMIs). Finally, through simulation results, it is verified that the proposed methods achieve desirable performance in controlling a mixed-vehicular network.