深圳市绿波运行监测场景停车率数据
收藏资源简介：
该数据主要应用于深圳市智能交通管理与信号灯优化场景。针对深圳市内特定交通路段（通过路段ID标识），监测其在特定时间段内的通行状况。通过展示路段的车辆总数、停车数以及计算得出的停车率指标，量化反映该路段的交通拥堵程度与通行效率。停车率越高，该路段的拥堵程度越严重。 这一数据成果为交通管理部门提供了客观的决策依据，能够辅助评估路段的信号控制方案效果，及时发现异常拥堵节点，从而服务于路口信号配时的优化调整，提升城市道路整体通行能力。数据采集： 依托回传的交通流数据，采集原始的车辆通行记录与位置信息。采集范围覆盖深圳市主要交通路段。 数据处理： 对原始数据进行清洗，去除重复、异常及缺失数据。对涉及车辆隐私的信息进行去标识化与脱敏处理，确保数据安全。 将处理后的数据进行结构化映射，明确每条数据的地理属性，包括城市名称、路段ID以及界定路段范围的起点经纬度和终点经纬度。 算法加工： 基于清洗后的数据进行时空聚合与统计计算，系统首先根据城市名称与路段ID，利用起点经纬度和终点经纬度划定具体的路段空间范围，并将数据按照精确的时间段（精确到小时的起止时间）进行切分聚合；接着，统计该时空范围内通行的车辆总数，并识别其中处于滞留或拥堵状态的停车数； 判断车辆停车状态算法的规则如下： 1.位置-距离转换：车辆定期上报的经纬度是地理坐标。我们采用 Haversine公式 进行球面三角学计算，将相邻两点的经纬度差值转换为地球表面的大圆直线距离（米）。该公式能有效修正地球曲率影响，比简单的平面欧氏距离更准确。 2.速度计算：获取上述两个位置点对应的时间戳，计算时间间隔（秒）。随后，用计算得到的位移距离除以时间间隔，即可得到这两个采样点之间的平均速度（米/秒）。 3.状态判定逻辑：设定一组基于实际场景和GPS定位误差（通常5-10米）的阈值进行判断： 4.停车状态：当计算出的连续速度值低于一个极小阈值（例如，0.1米/秒，约0.36公里/小时），并持续超过预设的时间窗口（如15秒），则判定为停车。单一低速点可能由GPS信号漂移导致，因此需要持续性条件来防误判。 （由于判断车辆停车状态算法涉及到每辆车的行驶数据，属于高敏感数据，所以并未在数据结构中展示。） 最后，采用比率计算模型，通过将停车数除以车辆总数计算得出停车率具体数值（即：停车率 = 停车数 / 车辆总数），从而形成量化反映路段通行效率与拥堵情况的结构化数据。
This dataset is primarily applied to the scenarios of intelligent traffic management and signal light optimization in Shenzhen City. For specific traffic road sections in Shenzhen (identified by road section IDs), the traffic conditions within specific time periods are monitored. By presenting the total number of vehicles, the number of stopped vehicles, and the calculated parking rate index, the traffic congestion level and traffic efficiency of the road section are quantitatively reflected. Higher parking rates indicate more severe congestion on the corresponding road section. This dataset provides objective decision-making basis for traffic management departments, assists in evaluating the effectiveness of signal control schemes for road sections, timely detects abnormal congestion nodes, and thus supports the optimization and adjustment of intersection signal timing to improve the overall traffic capacity of urban roads. Data Collection: Relying on the returned traffic flow data, original vehicle travel records and location information are collected. The collection scope covers major traffic road sections in Shenzhen. Data Processing: Clean the original data to remove duplicate, abnormal and missing values. Perform de-identification and desensitization processing on information involving vehicle privacy to ensure data security. Perform structured mapping on the processed data to clarify the geographic attributes of each data entry, including the city name, road section ID, and the start and end coordinates (latitude and longitude) that define the scope of the road section. Algorithm Processing: Perform spatiotemporal aggregation and statistical calculations based on the cleaned data. The system first delineates the specific spatial scope of the road section based on the city name and road section ID, using the start and end coordinates (latitude and longitude), and splits and aggregates the data according to precise time periods (with start and end times accurate to the hour); Next, count the total number of vehicles passing through this spatiotemporal range, and identify the number of stopped vehicles that are in a stationary or congested state; Rules of the vehicle parking status judgment algorithm are as follows: 1. Position-to-distance conversion: The latitude and longitude regularly reported by vehicles are geographic coordinates. We adopt the Haversine formula for spherical trigonometric calculations to convert the latitude and longitude difference between two adjacent points into the great-circle straight-line distance (in meters) on the Earth's surface. This formula can effectively correct the influence of Earth's curvature and is more accurate than simple planar Euclidean distance. 2. Speed calculation: Obtain the timestamps corresponding to the two aforementioned location points, and calculate the time interval (in seconds). Subsequently, divide the calculated displacement distance by the time interval to obtain the average speed (in meters per second) between these two sampling points. 3. Status judgment logic: Set a set of thresholds based on actual scenarios and GPS positioning errors (typically 5-10 meters) for judgment: 4. Parking status: When the calculated continuous speed value is lower than a very small threshold (for example, 0.1 m/s, approximately 0.36 km/h) and lasts for more than a preset time window (such as 15 seconds), it is judged as a parking state. A single low-speed point may be caused by GPS signal drift, so a continuous condition is required to avoid misjudgment. (Since the vehicle parking status judgment algorithm involves the driving data of each vehicle, which belongs to highly sensitive data, it is not displayed in the data structure.) Finally, adopt a ratio calculation model to calculate the specific value of the parking rate by dividing the number of stopped vehicles by the total number of vehicles (i.e., Parking Rate = Number of Stopped Vehicles / Total Number of Vehicles), thereby forming structured data that quantitatively reflects the traffic efficiency and congestion status of the road section.
