| 引用本文: | 赵靖华,王浩男,汪介瑜,宫洵,解方喜,高炳钊.网联车辆速度规划及气路控制[J].控制理论与应用,2025,42(8):1523~1533.[点击复制] |
| ZHAO Jing-hua,WANG Hao-nan,WANG Jie-yu,GONG Xun,XIE Fang-xi,GAO Bing-zhao.Research on eco-velocity planning and air-path control of connected vehicles[J].Control Theory & Applications,2025,42(8):1523~1533.[点击复制] |
|
| 网联车辆速度规划及气路控制 |
| Research on eco-velocity planning and air-path control of connected vehicles |
| 摘要点击 3951 全文点击 210 投稿时间：2024-01-03 修订日期：2025-06-29 |
| 查看全文 查看/发表评论 下载PDF阅读器 HTML |
| DOI编号 10.7641/CTA.2024.40005 |
| 2025,42(8):1523-1533 |
| 中文关键词 发动机气路控制 车辆速度规划 神经网络 模型预测控制 |
| 英文关键词 air-path control of engine eco-velocity planning neural network model predictive control |
| 基金项目 国家自然科学基金面上项目(52472407),吉林省科技厅重点研发项目(20250201090GX)资助. |
|
| 中文摘要 |
| 新鲜空气和废气混合过程具有强非线性以及多尺度时滞问题,这对带VGT和EGR的汽油发动机节能减排
的气路控制提出了很高的要求.随着法规越来越严格,未来的车辆燃油排放测试要求在实时道路环境下进行,交通
灯、坡度等路况会改变车辆的运行工况,对燃油及排放控制具有很大干扰.智能网联技术的发展与普及,一方面使
得提前获取路况信息成为可能,另一方面也促进了面向节能减排的车辆速度规划技术的发展.针对本次“内燃动力
智能算法挑战赛”的发动机气路和车辆速度规划的控制要求,本文基于神经网络建模和模型预测控制技术,提出了
一种双闭环实时优化控制策略.基于比赛提供的高精度车辆模型,本文测试了实时神经网络模型作为前馈map的气
路控制的性能;在交通灯和限速等网联信息获取条件下,本文验证了速度规划模型预测控制方案的效果,讨论了采
样周期、预测时域、优化目标权重以及车辆质量变化对控制效果的敏感性. |
| 英文摘要 |
| The mixing process of fresh air and exhaust gas has strong nonlinearity and multi-scale time delay, which
poses high challenges for fuel economy improvement and emissions reduction of the air-path control of gasoline engines
with VGT and EGRsystems. With increasingly stringent regulations, future vehicle fuel consumption and emission testing
are required to be conducted in real-world driving conditions. Traffic lights, slopes, and other road conditions can alter
the operating conditions of vehicles, causing significant interference with fuel and emission controls. The development
and popularization of vehicle connected technology have made it possible to obtain road condition information in advance,
and has also promoted the development of eco-velocity planning for fuel economy improvement and emissions reduction.
Targeted at the competition requirements for engine air-path and eco-velocity planning of “internal combustion power intel
ligent algorithm challenge”, a dual closed-loop real-time optimization control strategy based on neural network modeling
and model predictive control technology is proposed in this paper. Based on the high-precision vehicle model provided
by the competition, the performance of real-time neural network models as feedforward controllers for air-path control is
tested. Under the conditions of obtaining networked information such as traffic lights and speed limits, the effectiveness of
the eco-velocity planning within the model predictive control framework is verified, and the sensitivity of sampling period,
prediction time domain, optimization objective weight, and vehicle mass changes to control effectiveness is analyzed. |
|
|
|
|
|