Development of AI intelligent traffic monitoring system using self-powered censors

使用自供电传感器开发AI智能交通监控系统

• 批准号: 22K04371
• 负责人: 安藤 良輔
• 金额: $ 2.66万
• 依托单位: Toyota Transportation Research Institute
• 依托单位国家: 日本
• 项目类别: Grant-in-Aid for Scientific Research (C)
• 财政年份: 2022
• 资助国家: 日本
• 起止时间: 2022-04-01 至 2025-03-31
• 项目状态: 未结题
• 来源: https://kaken.nii.ac.jp/grant/KAKENHI-PROJECT-22K04371/
• 关键词: traffic engineering; sensors layout methods; monitoring network; traffic flow sensing; freeway; optimization goal; 交通量調査; AI; セルフ電源供給センサ

Aiming to provide theoretical support and a decision-making for the engineering practice of freeway monitoring network deployment, the relevant literature was comprehensively reviewed and the common methods from three aspects: precision-oriented research, cost-oriented research, and reliability-oriented research, were summarized. The model framework, algorithm characteristics and applicable scenarios of common deployment methods were systematically combed. Furthermore, the future developing direction was discussed considering the current gap between theoretical research and engineering practice as well as the new demands in the context of intelligent freeway construction and management. Results show that by combining traditional methods, e. g. traffic wave theory and planning model, with some methods, e. g. neural network, genetic algorithm and multi-objective dynamic deployment model, the sensor accuracy of main traffic parameters can be effectively improved at both section levels and network-level. Focused on cost, most research used biological heuristic algorithms to introduce cost constraint parameters or considering reducing cost as the optimization goal, in which the process of cost control can be reflected in the layout scheme formulation. To improve the reliability of monitoring network, various methods were utilized that generally followed two concepts: introducing sensor failure probability or reliability oriented global optimization. After years of development, the research on freeway monitoring networks can support most scenarios in engineering practice.

为了为高速公路监控网络部署提供理论支持和决策实践，对相关文献进行了全面审查，并从三个方面进行了共同的方法：面向精确的研究，面向成本的研究和面向可靠性的研究。系统地梳理了模型框架，算法特征和常见部署方法的适用场景。此外，考虑了理论研究与工程实践之间的当前差距以及在智能高速公路建设和管理的背景下的新需求，讨论了未来的发展方向。结果表明，通过结合传统方法，e。 g。交通波理论和计划模型，以及一些方法，e。 g。神经网络，遗传算法和多目标动态部署模型，主要交通参数的传感器精度可以在截面级别和网络级别上有效提高。专注于成本，大多数研究都使用生物启发式算法来引入成本约束参数或将降低成本作为优化目标，其中可以在布局方案制定中反映成本控制过程。为了提高监视网络的可靠性，利用了各种方法，通常遵循两个概念：引入传感器故障概率或面向可靠性的全局优化。经过多年的发展，高速公路监控网络的研究可以支持工程实践中的大多数情况。