协同决策和协同控制算法测试场景数据集

为支撑智能网联车辆协同决策与协同控制算法的可复现评测，构建并发布面向协
同换道与匝道协同汇入的测试数据集。数据由 Matlab/Simulink 与 SUMO 联合仿真生成，并
在“虚拟交通环境+硬件在环+实体执行机构”平台上获取部分硬件在环数据。数据记录多
车在离散时间步下的纵向/横向位置、速度与加速度等运动学信息，覆盖城市道路高密度协
同换道与高速公路匝道协同汇入两类典型场景。为验证数据可靠性，构建了参数校核、异常
值监测、时间对齐以及插值重采样等质量控制流程，并采用仿真—HIL 轨迹一致性评估给出
RMSE、 MAE 及最大绝对误差等精度指标。该数据集可用于协同决策/控制算法测试、可行
域与安全性分析、学习型策略训练等。

To support the reproducible evaluation of collaborative decision-making and cooperative control algorithms for intelligent connected vehicles (ICVs), this work constructs and releases a test dataset focused on cooperative lane-changing and ramp merging scenarios. The dataset is generated via co-simulation between Matlab/Simulink and SUMO, and partial Hardware-in-the-Loop (HIL) data is collected on a platform integrating virtual traffic environment, HIL and physical actuators. The dataset records kinematic information such as longitudinal/lateral position, velocity and acceleration of multiple vehicles at discrete time steps, covering two typical scenarios: high-density cooperative lane-changing on urban roads and ramp merging on expressways. To verify the reliability of the dataset, quality control workflows including parameter verification, outlier detection, time alignment, interpolation and resampling are established. Additionally, precision metrics such as Root Mean Squared Error (RMSE), Mean Absolute Error (MAE) and maximum absolute error are provided via simulation-HIL trajectory consistency evaluation. This dataset can be applied to tests of collaborative decision-making/control algorithms, feasibility domain and safety analysis, training of learning-based strategies, and other related research works.

搜集汇总

背景与挑战

背景概述

该数据集是为支撑智能网联车辆协同决策与协同控制算法的可复现评测而构建的测试场景数据集，由Matlab/Simulink与SUMO联合仿真生成并包含部分硬件在环数据，覆盖城市道路高密度协同换道和高速公路匝道协同汇入两类典型场景，记录了多车的运动学信息，并经过严格的质量控制流程验证可靠性。