边缘云实时多任务计算结果数据集

本数据集旨在提供一套源自真实交通场景、由边缘云实时计算生成的结构化的多任务事件结果数据。数据采集依托国家级智能网联汽车示范区的路侧边缘计算节点，通过融合摄像头、毫米波雷达、激光雷达等多源传感器数据，由部署的实时多任务推理框架自动识别并输出路面异常事件。数据集覆盖时间为2025年12月，共包含约34万条事件记录，每条记录均包含事件类型、发生位置（经纬度）、时间戳、关联目标物ID、事件影响范围半径以及该识别任务在边缘侧的完成时间等关键字段。数据格式主要为JSON和CSV，已进行时间戳对齐、传感器噪声滤除及数据补全等预处理。本数据集为研究边缘计算环境下的任务调度优化、智能算法性能评估、交通流仿真与数字孪生系统构建提供了高质量、可直接利用的计算结果级数据资源，具有重要的科研与工程应用价值。

This dataset is designed to provide structured multi-task event outcome data generated by real-time edge cloud computing in real traffic scenarios. Data collection relies on roadside edge computing nodes within the national-level intelligent connected vehicle demonstration zone. By fusing multi-source sensor data including cameras, millimeter-wave radars, LiDARs and other devices, the deployed real-time multi-task inference framework automatically identifies and outputs road surface abnormal events. The dataset covers the period of December 2025, containing approximately 340,000 event records in total. Each record includes key fields such as event type, occurrence location (latitude and longitude), timestamp, associated target ID, radius of the event's impact range, and the completion time of the recognition task at the edge side. The main data formats are JSON and CSV, and preprocessing operations such as timestamp alignment, sensor noise filtering and data completion have been conducted on the dataset. This dataset offers high-quality, directly utilizable computational result-level data resources for research on task scheduling optimization in edge computing environments, intelligent algorithm performance evaluation, traffic flow simulation and digital twin system construction, and holds significant scientific research and engineering application value.

搜集汇总

背景与挑战

背景概述

该数据集源自国家级智能网联汽车示范区的路侧边缘计算节点，通过融合多源传感器数据，提供了2025年12月期间约34万条实时多任务事件记录，包含事件类型、位置、时间戳等关键字段，并已进行预处理。它为边缘计算任务调度优化、智能算法评估及交通数字孪生系统构建提供了高质量的计算结果级数据资源。