WO2017107790A1 - 一种基于大数据预测路段状况的方法及装置 - Google Patents

一种基于大数据预测路段状况的方法及装置 Download PDF

Info

Publication number
WO2017107790A1
WO2017107790A1 PCT/CN2016/109387 CN2016109387W WO2017107790A1 WO 2017107790 A1 WO2017107790 A1 WO 2017107790A1 CN 2016109387 W CN2016109387 W CN 2016109387W WO 2017107790 A1 WO2017107790 A1 WO 2017107790A1
Authority
WO
WIPO (PCT)
Prior art keywords
road
abnormal
data
road segment
driving data
Prior art date
Application number
PCT/CN2016/109387
Other languages
English (en)
French (fr)
Inventor
赵丹
刘智嘉
武凯
付登坡
Original Assignee
阿里巴巴集团控股有限公司
赵丹
刘智嘉
武凯
付登坡
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 阿里巴巴集团控股有限公司, 赵丹, 刘智嘉, 武凯, 付登坡 filed Critical 阿里巴巴集团控股有限公司
Priority to JP2018531408A priority Critical patent/JP2019505892A/ja
Publication of WO2017107790A1 publication Critical patent/WO2017107790A1/zh
Priority to US16/016,502 priority patent/US10977933B2/en

Links

Images

Classifications

    • GPHYSICS
    • G08SIGNALLING
    • G08GTRAFFIC CONTROL SYSTEMS
    • G08G1/00Traffic control systems for road vehicles
    • G08G1/01Detecting movement of traffic to be counted or controlled
    • G08G1/0104Measuring and analyzing of parameters relative to traffic conditions
    • G08G1/0108Measuring and analyzing of parameters relative to traffic conditions based on the source of data
    • G08G1/0112Measuring and analyzing of parameters relative to traffic conditions based on the source of data from the vehicle, e.g. floating car data [FCD]
    • GPHYSICS
    • G08SIGNALLING
    • G08GTRAFFIC CONTROL SYSTEMS
    • G08G1/00Traffic control systems for road vehicles
    • G08G1/01Detecting movement of traffic to be counted or controlled
    • G08G1/0104Measuring and analyzing of parameters relative to traffic conditions
    • G08G1/0125Traffic data processing
    • G08G1/0129Traffic data processing for creating historical data or processing based on historical data

Definitions

  • the invention belongs to the technical field of data processing, and in particular relates to a method and a device for predicting the condition of a road segment based on big data.
  • Pavement damage image collection including damage image acquisition and acquisition, digitization, compression coding, etc.
  • Pavement damage image analysis including road segmentation image segmentation, description, and classification.
  • image segmentation There are two main methods of image segmentation: boundary image segmentation based on region image segmentation.
  • the prior art has the problems of large time consumption, complicated image processing, and low accuracy. There is an urgent need for a more economical and practical method to locate damaged road sections and determine the specific damage type. The corresponding maintenance personnel went to the repair.
  • the object of the present invention is to provide a method and a device for predicting the condition of a road segment based on big data, so as to avoid the technical problem of low judgment efficiency and inaccurate judgment inherent in manual inspection of the road segment or image acquisition analysis.
  • a method for predicting a condition of a road segment based on big data comprising:
  • the road segment in the abnormal database it is determined whether the road segment is an abnormal road segment according to the number of abnormal data occurrences of the road segment;
  • the reason for predicting the road segment abnormality according to the preset model is provided to the user.
  • comparing the collected driving data with the normal observation sample to determine whether it is abnormal data including:
  • the determining whether the road segment is an abnormal road segment according to the number of occurrences of the abnormal data of the road segment includes:
  • the road condition evaluation value corresponding to the driving data is given a weight
  • the road condition evaluation value corresponding to the driving data is weighted, including:
  • the weight of the current driving data is increased according to the number of times the abnormal data is accumulated;
  • the weight of the current driving data is reduced according to the number of times the normal data is accumulated.
  • the determining, according to the product of the road condition evaluation value and the weight thereof, whether the road segment is an abnormal road segment includes:
  • the road segment When the product of the current road condition evaluation value and its weight is less than the set first threshold, the road segment is considered to be a normal road segment; and when the product of the front road condition evaluation value and its weight is greater than the set second threshold, the road segment is determined to be Abnormal road segment.
  • the method further includes: when determining whether the road segment is an abnormal road segment, the method further includes:
  • the road condition evaluation value of the corresponding road segment is calculated to be greater than the set third threshold value according to the collected driving data and the corresponding weight, the road segment is directly determined to be an abnormal road segment.
  • the invention also proposes a device for predicting the condition of a road segment based on big data, the device comprising:
  • a data acquisition module configured to collect driving data recorded by a driving vehicle on a road section
  • the abnormal data judging module is configured to compare the collected driving data with the normal observation sample to determine whether it is abnormal data, and if it is abnormal data, put the abnormal data and its corresponding road segment into the abnormal database, and continuously record the road segment.
  • Driving data ;
  • An abnormal road segment judging module is configured to generate a road segment in the abnormal database according to the road segment The number of abnormal data to determine whether the road segment is an abnormal road segment;
  • the abnormal cause analysis module is configured to provide a road segment that is determined to be an abnormal road segment, and predict the cause of the road segment abnormality according to a preset model, and provide the same to the user.
  • abnormal data judging module compares the collected driving data with the normal observation sample to determine whether it is abnormal data, the following operations are performed:
  • the abnormal road segment determining module determines whether the road segment is an abnormal road segment according to the number of times the abnormal data occurs in the road segment, the following operations are performed:
  • the road condition evaluation value corresponding to the driving data is given a weight
  • the abnormal road segment determining module assigns a weight to the road condition evaluation value corresponding to the driving data, including:
  • the weight of the current driving data is increased according to the number of times the abnormal data is accumulated;
  • the weight of the current driving data is reduced according to the number of times the normal data is accumulated.
  • abnormal road segment determining module performs the following operations when determining whether the road segment is an abnormal road segment according to the product of the road condition evaluation value and the weight thereof:
  • the road segment When the product of the current road condition evaluation value and its weight is less than the set first threshold, the road segment is considered to be a normal road segment; and when the product of the front road condition evaluation value and its weight is greater than the set second threshold, the road segment is determined to be Abnormal road segment.
  • the abnormal road segment determining module of the present invention determines whether the road segment is an abnormal road segment, the following operations are also performed:
  • the road condition evaluation value of the corresponding road segment is calculated to be greater than the set third threshold value according to the collected driving data and the corresponding weight, the road segment is directly determined to be an abnormal road segment.
  • the invention provides a method and a device for predicting the condition of a road segment based on big data, and collects abnormal driving data of a traveling vehicle on a road, determines whether it is abnormal data according to comparison with a normal observation sample, and analyzes the abnormal data to determine the road segment. situation.
  • the condition of the road section can be accurately predicted, the manpower and material resources are saved, and the damaged road section and the specific damage type can be located, which provides convenience for maintenance.
  • FIG. 1 is a flow chart of a method for predicting a condition of a road segment according to the present invention
  • FIG. 2 is a schematic structural diagram of an apparatus for predicting the condition of a road section according to the present invention.
  • a method for predicting a road segment condition based on big data includes:
  • Step S1 Collect driving data recorded by the driving vehicle on the road section.
  • This embodiment records the driving of the vehicle by a road surface detecting instrument distributed on the traveling vehicle.
  • the data for example, is issued to the passing vehicle at the entrance of the highway, and the pass card can also be used as a road surface detecting instrument for recording the driving data of the vehicle.
  • the specific driving data may include corresponding driving data caused by bumpy data, braking data, turning data, slip data, and the like which may occur on various road surfaces.
  • the driving data recorded by the vehicle is retrieved into the computer as the basic data for subsequent analysis after the high-speed exit recovers the pass card, and the more the collected driving data, the more accurate the subsequent analysis.
  • the embodiment can also collect the driving data through the in-vehicle navigation device or other devices having the data collection function, and details are not described herein again.
  • the collection of driving data on the same road section can be collected periodically, for example once a week.
  • the driving data is normal, it is not necessary to continue collecting during this week.
  • the driving data is abnormal, it can be judged whether the road section is abnormal by recording once every day or continuously recording multiple times in one day.
  • Step S2 Comparing the collected driving data with the normal observation sample to determine whether it is abnormal data. If it is abnormal data, the abnormal data and its corresponding road segment are placed in the abnormal database, and the driving data of the road segment is continuously recorded.
  • the normal observation sample is pre-stored, and the driving data that needs to be predicted is filtered to obtain abnormal data that deviates from normal, so as to analyze the abnormal data and determine the road section. situation.
  • the road condition evaluation value of the corresponding road section can be calculated according to the collected driving data, and the calculation formula of the road condition evaluation value S is as follows:
  • Different types of driving data such as s 1 is bump data, s 2 brake data, s 3 is brake curve data, and the like.
  • the driving data recorded during the running of the vehicle under the normal condition of the road section is a normal observation sample, and the road condition evaluation value S normal under the normal condition of the road section can be calculated, and the road condition evaluation value S under normal conditions is assumed.
  • the range of normal is:
  • the road condition evaluation value of the road section can be calculated, and compared with the road condition evaluation value of the normal observation sample. If the road condition evaluation value corresponding to the road section driving data is within the road condition evaluation value range of the normal observation sample of the road section, it is determined that the road section is a normal road section, and the driving data is normal data; otherwise, the road section is determined to be an abnormal road section, and the driving data is abnormal data.
  • the driving data is used as abnormal data, which needs to be saved for subsequent continuous analysis.
  • the saved abnormal data includes the road segment identification, the driving data, and the corresponding road condition evaluation value, so as to count the number of times the abnormal data appears in the road segment in the subsequent steps.
  • the present embodiment is for the road segment determined to be abnormal. It is necessary to retain the driving data for a period of time, whether it is abnormal data or not, it needs to be saved for subsequent judgment.
  • the history of one week is saved, and the abnormal data of the day is stored cyclically for subsequent analysis, and the expired data is naturally deleted.
  • Step S3 For the road segment in the abnormal database, determine whether the road segment is an abnormal road segment according to the number of times the abnormal data appears in the road segment.
  • the driving data for a period of time is continuously recorded.
  • one traffic card is arbitrarily issued every day to record the driving data, and the recording is recorded for a total of seven times. Then get the daily driving data of the road section within one week. Or on the same day, 7 passes will be issued to different cars, and each car will be recorded once in a number of cars, and a total of seven driving data will be obtained.
  • the present invention does not limit the number of specific records, and the more the number of records, the more accurate the results obtained.
  • the road segment is an abnormal road segment, and may include multiple methods. The following is described by way of an embodiment:
  • Embodiment 1 The number of consecutive occurrences of abnormal data is greater than a set threshold.
  • the road section If the number of consecutive abnormal data occurrences is greater than the set threshold, it is determined that the road section has an abnormal condition, and if the abnormal data appears discontinuous, the road section is considered to be normal.
  • Embodiment 2 Judging according to the ratio of the number of occurrences of abnormal data to the total number of driving data.
  • the abnormal data and the corresponding road segment are recorded in the abnormal database, and the driving data of the road segment is continuously recorded, and it is assumed that M times are recorded, wherein the abnormal data is N times, if N/M is greater than the setting If the threshold is reached, it is judged that an abnormal condition has occurred in the road section, otherwise it is determined to be normal.
  • Embodiment 3 The road segment that is not continuously abnormal data is placed in the observation database to continue observation.
  • the abnormal condition of the road section is an abnormal road section.
  • the road segments that are not continuously abnormal data they are placed in the observation database, and the driving data is continuously recorded for subsequent analysis.
  • the road condition evaluation value of the corresponding road section calculated according to the collected driving data is far beyond the range of the road condition evaluation value S normal under the normal condition, for example, exceeding the setting
  • a threshold it can also be directly judged that the road segment is an abnormal road segment. For example, a road surface suddenly breaks, and the risk of breaking occurs is very large.
  • the road condition evaluation value has exceeded the set threshold value, it is considered that the road section is problematic and needs to be processed immediately, otherwise If it is delayed for a few days, the road surface breakage has already formed, which may cause danger.
  • the road segment is placed in an observation database for continuous observation.
  • the method of this embodiment further includes the steps of:
  • the road condition evaluation value corresponding to the driving data is given a weight
  • this embodiment sets a weight W, and when the driving data of the road section is determined as abnormal data, the driving is increased.
  • the weight of the data when the driving data of the road section is determined to be normal data, the weight of the driving data is reduced.
  • the weight of the current driving data is calculated according to the following formula:
  • T dif is the number of times the abnormal data is accumulated from the time when the observation segment is added to the observation database to the current time
  • T nor is the number of times the normal data is accumulated from the time when the observation segment is added to the observation database to the current time.
  • the weight value can be determined according to the weight value, that is, when the weight value is less than a certain threshold, the road segment is considered to be a normal road segment, and is deleted from the observation database; and when the weight value is greater than a certain threshold, the road segment is determined. It is an abnormal road segment.
  • the road segment is considered to be a normal road segment; and when their product is greater than a certain threshold, the road segment is determined to be an abnormal road segment.
  • step S1 whether it is judged as an abnormal road segment or judged as a normal road segment, it will be The corresponding road segment and its driving data are deleted from the abnormal database and the observation database, and no continuous tracking is performed, but the conventional judgment is performed according to the flow of step S1.
  • Step S4 For the road segment that is determined to be an abnormal road segment, the reason for predicting the road segment abnormality according to the preset model is provided to the user.
  • the road section judged to be an abnormal road section it can be considered that the road surface of this section has been damaged. It is necessary to combine the performance data of different types of damaged pavements stored in the experience database to further determine the type and cause of the pavement damage. After analyzing the type of damage, send the corresponding maintenance personnel to repair, which greatly improves the efficiency of road detection. Of course, it can also be combined with other auxiliary methods, such as image analysis technology, to carry out in-depth detection and analysis of this road surface.
  • the preset model in this embodiment refers to the performance data of different types of damaged pavements stored in the experience database, the real-time maintenance experience database, the performance of the road surface is diverse, and the rich and diverse experience database updated in real time is more accurate.
  • the judgment of road surface damage provides a more reliable guarantee.
  • the device for predicting the condition of a road segment based on big data includes:
  • a data acquisition module configured to collect driving data recorded by a driving vehicle on a road section
  • the abnormal data judging module is configured to compare the collected driving data with the normal observation sample to determine whether it is abnormal data, and if it is abnormal data, put the abnormal data and its corresponding road segment into the abnormal database, and continuously record the road segment.
  • Driving data ;
  • the abnormal road segment determining module is configured to determine, according to the number of times the abnormal data is generated in the road segment, whether the road segment is an abnormal road segment;
  • the abnormal cause analysis module is configured to provide a road segment that is determined to be an abnormal road segment, and predict the cause of the road segment abnormality according to a preset model, and provide the same to the user.
  • the abnormal data judging module compares the collected driving data with the normal observation sample to determine whether it is abnormal data, and performs the following operations:
  • the abnormal road segment determining module performs the following operations when determining whether the road segment is an abnormal road segment according to the number of times the abnormal data occurs in the road segment:
  • the road condition evaluation value corresponding to the driving data is given a weight
  • the abnormal road segment determining module in this embodiment assigns a weight to the road condition evaluation value corresponding to the driving data, including:
  • the weight of the current driving data is increased according to the number of times the abnormal data is accumulated;
  • the weight of the current driving data is reduced according to the number of times the normal data is accumulated.
  • the abnormal road segment determining module performs the following operations when determining whether the road segment is an abnormal road segment according to the product of the road condition evaluation value and its weight:
  • the road segment When the product of the current road condition evaluation value and its weight is less than the set first threshold, the road segment is considered to be a normal road segment; and when the product of the front road condition evaluation value and its weight is greater than the set second threshold, the road segment is determined to be Abnormal road segment.
  • the abnormal road segment determining module performs the following operations when determining whether the road segment is an abnormal road segment:
  • the road data of the corresponding road segment is calculated according to the collected driving data and its corresponding weight
  • the evaluation value is greater than the set third threshold, it is directly determined that the road segment is an abnormal road segment.

Landscapes

  • Chemical & Material Sciences (AREA)
  • Analytical Chemistry (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Traffic Control Systems (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

一种基于大数据预测路段状况的方法及装置,所述方法采集道路路段上行驶车辆记录的行车数据(S1),将采集的行车数据与正常观测样本进行对比,判断是否为异常数据,如果是异常数据,则将该异常数据及其对应的路段放入异常数据库,并持续记录该路段的行车数据(S2);对于异常数据库中的路段,根据该路段出现异常数据的次数来判断该路段是否是异常路段(S3);对于判断为异常路段的路段,根据预设的模型预测路段异常的原因,提供给用户(S4)。所述装置包括数据采集模块、异常数据判断模块、异常路段判断模块和异常原因分析模块。该方法及装置通过大数据的分析可以准确预测路段的状况,节省了人力物力。

Description

一种基于大数据预测路段状况的方法及装置 技术领域
本发明属于数据处理技术领域,尤其涉及一种基于大数据预测路段状况的方法及装置。
背景技术
随着国民经济的高速发展,我国汽车产业发展也进入了一个新的时期,汽车已经进入家庭。时代的发展对道路交通的要求越来越高,公路的利用率相较于十几年前大大的增加。然而由于道路受汽车碾压、雨水侵蚀等因素的影响,经常会有某路段出现坑洼、路面断层等情况,这让公路的维修工作面临了严峻的考验。
传统的公路的维修工作依靠人工或图像采集去考察路段,公路维护人员需要经常沿路开车视察哪里有问题,这样既耗体力又浪费时间,同时也会有疏忽问题路段或者不能及时查看到问题的情况出现。其中基于图像采集去考察路段时,路面破损图像识别过程包括两个步骤:
路面破损图像采集,主要包括破损图像的采集和获取、数字化、压缩编码等;
路面破损图像分析,包括路面破损图像分割、描述、和分类等。图像分割的方法主要有二大类:基于边界图像分割、基于区域图像分割技术。
但是由于路面破损的种类繁多,加上破损程度很难用统一的解析式来描述,近些年,基于模糊逻辑、人工神经网络、专家系统等人工智能分类判断算法研究成为路面破损自动识别研究热点。
无论从哪种方式来看,现有技术都存在时间消耗大、图像处理复杂,准确性低的问题,急需更为经济和实际有效的方法来定位破损路段,并确定具体的损坏类型,以便派相应的维修人员前去维修。
发明内容
本发明的目的是提供一种基于大数据预测路段状况的方法及装置,以避免人工考察路段或图像采集分析所固有的判断效率低、判断不准确的技术问题。
为了实现上述目的,本发明技术方案如下:
一种基于大数据预测路段状况的方法,所述方法包括:
采集道路路段上行驶车辆记录的行车数据;
将采集的行车数据与正常观测样本进行对比,判断是否为异常数据,如果是异常数据,则将该异常数据及其对应的路段放入异常数据库,并持续记录该路段的行车数据;
对于异常数据库中的路段,根据该路段出现异常数据的次数来判断该路段是否是异常路段;
对于判断为异常路段的路段,根据预设的模型预测路段异常的原因,提供给用户。
进一步地,所述将采集的行车数据与正常观测样本进行对比,判断是否为异常数据,包括:
根据采集的行车数据及其对应的权重计算得到该路段对应的路况评价值;
将计算得到的路况评价值与正常观测样本对应的路况评价值范围进行对比,如果该路段行车数据对应的路况评价值在该路段正常观测样本的路况评价值范围内,则判定该路段为正常路段,其行车数据为正常数据,否则判定该路段为异常路段,其行车数据为异常数据。
进一步地,所述根据该路段出现异常数据的次数来判断该路段是否是异常路段,包括:
如果连续出现异常数据的次数大于设定的阈值,则判断该路段为异常路段,否则将该路段及其行车数据放入观察数据库;
对于放入观察数据库的路段,继续跟踪该路段的行车数据;
根据跟踪得到的行车数据中异常数据的次数及正常数据的次数,为行车数据对应的路况评价值赋予权重;
根据路况评价值及其权重的乘积,判断该路段是否为异常路段。
进一步地,所述为行车数据对应的路况评价值赋予权重,包括:
在当前行车数据被认定为异常数据时,根据累计出现异常数据的次数增加当前行车数据的权重;
在当前行车数据被认定为正常数据时,根据累计出现正常数据的次数减少当前行车数据的权重。
进一步地,所述根据路况评价值及其权重的乘积,判断该路段是否为异常路段,包括:
在当前路况评价值及其权重的乘积小于设定的第一阈值时,认为该路段为正常路段;而在前路况评价值及其权重的乘积大于设定的第二阈值时,判断该路段为异常路段。
本发明在判断路段是否是异常路段时,所述方法还包括:
如果根据采集的行车数据及其对应的权重计算得到其对应路段的路况评价值大于设定的第三阈值时,直接判断该路段是异常路段。
本发明还提出了一种基于大数据预测路段状况的装置,所述装置包括:
数据采集模块,用于采集道路路段上行驶车辆记录的行车数据;
异常数据判断模块,用于将采集的行车数据与正常观测样本进行对比,判断是否为异常数据,如果是异常数据,则将该异常数据及其对应的路段放入异常数据库,并持续记录该路段的行车数据;
异常路段判断模块,用于对于异常数据库中的路段,根据该路段出现 异常数据的次数来判断该路段是否是异常路段;
异常原因分析模块,用于对于判断为异常路段的路段,根据预设的模型预测路段异常的原因,提供给用户。
进一步地,所述异常数据判断模块在将采集的行车数据与正常观测样本进行对比,判断是否为异常数据时,执行如下操作:
根据采集的行车数据及其对应的权重计算得到该路段对应的路况评价值;
将计算得到的路况评价值与正常观测样本对应的路况评价值范围进行对比,如果该路段行车数据对应的路况评价值在该路段正常观测样本的路况评价值范围内,则判定该路段为正常路段,其行车数据为正常数据,否则判定该路段为异常路段,其行车数据为异常数据。
进一步地,所述异常路段判断模块在根据该路段出现异常数据的次数来判断该路段是否是异常路段时,执行如下操作:
如果连续出现异常数据的次数大于设定的阈值,则判断该路段为异常路段,否则将该路段及其行车数据放入观察数据库;
对于放入观察数据库的路段,继续跟踪该路段的行车数据;
根据跟踪得到的行车数据中异常数据的次数及正常数据的次数,为行车数据对应的路况评价值赋予权重;
根据路况评价值及其权重的乘积,判断该路段是否为异常路段。
进一步地,所述异常路段判断模块为行车数据对应的路况评价值赋予权重,包括:
在当前行车数据被认定为异常数据时,根据累计出现异常数据的次数增加当前行车数据的权重;
在当前行车数据被认定为正常数据时,根据累计出现正常数据的次数减少当前行车数据的权重。
进一步地,所述异常路段判断模块在根据路况评价值及其权重的乘积,判断该路段是否为异常路段时,执行如下操作:
在当前路况评价值及其权重的乘积小于设定的第一阈值时,认为该路段为正常路段;而在前路况评价值及其权重的乘积大于设定的第二阈值时,判断该路段为异常路段。
本发明所述异常路段判断模块在判断该路段是否是异常路段时,还执行如下操作:
如果根据采集的行车数据及其对应的权重计算得到其对应路段的路况评价值大于设定的第三阈值时,直接判断该路段是异常路段。
本发明提出的一种基于大数据预测路段状况的方法及装置,通过采集道路上行驶车辆的异常行驶数据,根据与正常观测样本的对比来判断是否是异常数据,并对异常数据进行分析确定路段状况。通过大数据的分析可以准确预测路段的状况,节省了人力物力,并能够定位破损路段和具体的损坏类型,为维修提供了方便。
附图说明
图1为本发明预测路段状况的方法流程图;
图2为本发明预测路段状况的装置结构示意图。
具体实施方式
下面结合附图和实施例对本发明技术方案做进一步详细说明,以下实施例不构成对本发明的限定。
如图1所示,本实施例一种基于大数据预测路段状况的方法,包括:
步骤S1、采集道路路段上行驶车辆记录的行车数据。
本实施例通过分布在行驶车辆上的路面检测仪器来记录车辆的行车 数据,例如在高速公路的入口向过往车辆发放通行卡,该通行卡作为路面检测仪器还可以用于记录车辆的行车数据。具体的行车数据可以包括颠簸数据、刹车数据、转弯数据、打滑数据等各种路面可能出现的状况导致的对应的行车数据。车辆记录的行车数据在高速出口收回通行卡后,导入到计算机作为后续分析的基础数据,采集的行车数据越多,后续的分析越准确。
本实施例还可通过车载导航设备或其他具有数据采集功能的设备实现对行车数据的采集,此处不再赘述。
容易理解的是,对于同一路段行车数据的采集,可以定时进行采集,例如一周一次。在行车数据正常时,在这一周内不需要继续采集。而在行车数据异常时,可以通过每天记录一次或一天内连续记录多次的方法来判断该路段是否异常。
步骤S2、将采集的行车数据与正常观测样本进行对比,判断是否为异常数据,如果是异常数据,则将该异常数据及其对应的路段放入异常数据库,并持续记录该路段的行车数据。
本实施例根据车辆在正常路段行驶过程中记录的行车数据,预先保存有正常观测样本,用此来过滤需要预测的行车数据,得到偏离正常的异常数据,以便后续对异常数据进行分析,判断路段状况。
在本实施例中,根据采集的行车数据可以计算得到其对应路段的路况评价值,路况评价值S的计算公式如下:
S=α1S12S2+…+αnSn
其中,s1~sn为不同类型的行车数据,α1~αn为不同类型行车数据对应的权重,满足1=α12+…+αn。不同类型的行车数据例如s1为颠簸数据,s2刹车数据,s3为刹转弯数据等等。
同样,本实施例以车辆在该路段正常状况下行驶过程中记录的行车数据为正常观测样本,可以计算得到该路段正常状况下的路况评价值Snormal,并假设正常状况下的路况评价值Snormal的范围为:
Snormal=[Snormal_low,Snormal_ig]。
这样在采集到行车数据后,就可以计算出该路段的路况评价值,与正常观测样本的路况评价值进行对比。如果该路段行车数据对应的路况评价值在该路段正常观测样本的路况评价值范围内,则判定该路段为正常路段,其行车数据为正常数据,否则判定该路段为异常路段,其行车数据为异常数据。
本实施例对于正常路段,不需要保存其行车数据。对于异常路段,将其行车数据作为异常数据,需要保存起来以便后续进行持续分析。保存的异常数据包括路段标识、行车数据、以及对应的路况评价值,以便与后续步骤中统计该路段出现异常数据的次数。
容易理解的是,对一个路段的预测不能仅单单依靠一次的异常数据,若路段存在异常,它的表现应该是连续或者间接连续的,因此为了提高准确性,本实施例对于判定为异常的路段,需要保留一段时间内的行车数据,无论是否是异常数据都需要保存,以便进行后续的判断。
例如对某一段道路,保存一周的历史记录,循环存储每天的异常数据,用于后续的分析,过期的数据自然删除。
步骤S3、对于异常数据库中的路段,根据该路段出现异常数据的次数来判断该路段是否是异常路段。
本实施例在根据行车数据判断某一路段是异常路段后,持续记录一段时间内的行车数据,例如对于同一路段,每天任意发放一张通行卡记录一次行车数据,记录一周,总共记录七次,则得到该路段一周内每天的行车数据。或者在同一天,发放7张通行卡给不同的车,通过多部车每车记录一次,共得到七次行车数据。本发明对具体记录的次数不做限制,记录的次数越多,得到的结果越准确。
容易理解的是,对于有异常数据的路段,统计其在一段时间内出现异常数据的次数,可以判断该路段是否破损。例如如果出现一次异常数据后,未记录到后续的异常数据,则可能是路面出现丢弃物或驾驶人员的操作造 成,或者出现误识别。而当出现一次异常数据后,连续几天记录到异常数据,则判断该路段出现破损等异常,需要派人去现场维修。
本实施例根据该路段连续出现异常数据的次数来进行判断该路段是否是异常路段,可以包括多种方法,以下通过实施例来进行说明:
实施例一、连续出现异常数据的次数大于设定的阈值。
如果连续出现异常数据的次数大于设定的阈值,则判断该路段出现异常状况,如果异常数据出现的不连续,则认为该路段正常。
实施例二、根据出现异常数据的次数占总的行车数据次数的比例来判断。
当行车数据出现异常时,将该异常数据及对应的路段记录在异常数据库中,并持续记录该路段的行车数据,假设记录了M次,其中异常数据为N次,如果N/M大于设定的阈值,则判断该路段出现了异常状况,否则判定为正常。
实施例三、对于不是连续出现异常数据的路段放入观察数据库继续观察。
首先如果连续出现异常数据的次数大于设定的阈值,则判断该路段出现异常状况,为异常路段。与实施例一不同的是,对于不是连续出现异常数据的路段,将其放入观察数据库,继续记录行车数据进行后续分析。
需要注意的是,在判断路段的行车数据是异常路段后,如果根据采集的行车数据计算得到其对应路段的路况评价值远远超出了正常状况下的路况评价值Snormal的范围,例如超过设定的一个阈值,则也可以直接判断该路段是异常路段。例如一个路面突然断裂,而发生断裂的危险权重是很大的,通过计算其路况评价值已经超出了设定的阀值,这时候就认为该路段就是有问题的,需要马上进行处理,否则还要拖延几天的话,路面断裂已经形成,可能造成危险。
容易理解的是,如果该路段有时出现异常数据,有时又不出现,则可能是破损情况不严重,或是采集的数据有误,需要进行持续的观察,以便 进一步确定是否出现破损等异常状况。
本实施例对于不连续出现异常数据的路段,将该路段放入观察数据库中进行持续观察。对于需要持续观察的路段,本实施例的方法还包括步骤:
对于放入观察数据库的路段,继续跟踪该路段的行车数据;
根据跟踪得到的行车数据中异常数据的次数及正常数据的次数,为行车数据对应的路况评价值赋予权重;
根据路况评价值及其权重的乘积,判断该路段是否为异常路段。
即对于偶然性的异常情况,行车数据表现为间接出现异常数据,不能很把握的确定其为破损路面,因此本实施例设置一个权重W,当该路段的行车数据被认定为异常数据时,增加行车数据的权重,当该路段的行车数据被认定为正常数据时,就减少行车数据的权重。
当前行车数据的权重按照如下公式计算:
Figure PCTCN2016109387-appb-000001
其中σ为常数,Tdif为从此路段加入观察数据库时间开始到当前时间累计出现异常数据的次数,Tnor为从此路段加入观察数据库时间开始到当前时间累计出现正常数据的次数。可见当前行车数据的权重W实时地发生变化,即累计的异常数据越多,权重值越大,累计的正常数据越多,权重值越小。
从而可以根据该权重值来进行判断,即当该权重值小于一定阈值时,认为该路段为正常路段,将其从观察数据库中删除;而当权重值大于设定的一定阈值时,判断该路段为异常路段。
或者根据路况评价值及其权重的乘积来判断,即当他们的乘积小于一定阈值时,认为该路段为正常路段;而当他们的乘积大于设定的一定阈值时,判断该路段为异常路段。
如果还是无法判断,则继续跟踪该路段的行车数据,持续进行判断。
需要说明的是,无论是判断为异常路段或是判断为正常路段后,都将 相应的路段及其行车数据从异常数据库和观察数据库中删除,不再进行持续跟踪,而是根据步骤S1的流程来进行常规的判断。
步骤S4、对于判断为异常路段的路段,根据预设的模型预测路段异常的原因,提供给用户。
对于判断为异常路段的路段,可以认为此段路面已经破损。需要结合经验数据库中保存的对不同类型的破损路面的表现数据,进一步的判断此路面破损的类型、原因等。当分析出来破损类型后,派相应的维修人员前去维修,这样大大提高了道路侦测的效率。当然这里也可以结合其它辅助的方式,如图像分析技术,针对性的对此路面进行深入的侦测、分析。
本实施例预设的模型是指经验数据库中保存的对不同类型的破损路面的表现数据,实时的维护经验数据库,路面的表现的情况是多样的,实时更新的丰富多样经验数据库,为更加精准的判断路面破损情况提供更可靠的保障。
如图2所示,本实施例一种基于大数据预测路段状况的装置,包括:
数据采集模块,用于采集道路路段上行驶车辆记录的行车数据;
异常数据判断模块,用于将采集的行车数据与正常观测样本进行对比,判断是否为异常数据,如果是异常数据,则将该异常数据及其对应的路段放入异常数据库,并持续记录该路段的行车数据;
异常路段判断模块,用于对于异常数据库中的路段,根据该路段出现异常数据的次数来判断该路段是否是异常路段;
异常原因分析模块,用于对于判断为异常路段的路段,根据预设的模型预测路段异常的原因,提供给用户。
本实施例异常数据判断模块在将采集的行车数据与正常观测样本进行对比,判断是否为异常数据时,执行如下操作:
根据采集的行车数据及其对应的权重计算得到该路段对应的路况评价值;
将计算得到的路况评价值与正常观测样本对应的路况评价值范围进行对比,如果该路段行车数据对应的路况评价值在该路段正常观测样本的路况评价值范围内,则判定该路段为正常路段,其行车数据为正常数据,否则判定该路段为异常路段,其行车数据为异常数据。
本实施例异常路段判断模块在根据该路段出现异常数据的次数来判断该路段是否是异常路段时,执行如下操作:
如果连续出现异常数据的次数大于设定的阈值,则判断该路段为异常路段,否则将该路段及其行车数据放入观察数据库;
对于放入观察数据库的路段,继续跟踪该路段的行车数据;
根据跟踪得到的行车数据中异常数据的次数及正常数据的次数,为行车数据对应的路况评价值赋予权重;
根据路况评价值及其权重的乘积,判断该路段是否为异常路段。
其中,本实施例异常路段判断模块为行车数据对应的路况评价值赋予权重,包括:
在当前行车数据被认定为异常数据时,根据累计出现异常数据的次数增加当前行车数据的权重;
在当前行车数据被认定为正常数据时,根据累计出现正常数据的次数减少当前行车数据的权重。
本实施例异常路段判断模块在根据路况评价值及其权重的乘积,判断该路段是否为异常路段时,执行如下操作:
在当前路况评价值及其权重的乘积小于设定的第一阈值时,认为该路段为正常路段;而在前路况评价值及其权重的乘积大于设定的第二阈值时,判断该路段为异常路段。
本实施例异常路段判断模块在判断该路段是否是异常路段时,还执行如下操作:
如果根据采集的行车数据及其对应的权重计算得到其对应路段的路 况评价值大于设定的第三阈值时,直接判断该路段是异常路段。
以上实施例仅用以说明本发明的技术方案而非对其进行限制,在不背离本发明精神及其实质的情况下,熟悉本领域的技术人员当可根据本发明作出各种相应的改变和变形,但这些相应的改变和变形都应属于本发明所附的权利要求的保护范围。

Claims (12)

  1. 一种基于大数据预测路段状况的方法,其特征在于,所述方法包括:
    采集道路路段上行驶车辆记录的行车数据;
    将采集的行车数据与正常观测样本进行对比,判断是否为异常数据,如果是异常数据,则将该异常数据及其对应的路段放入异常数据库,并持续记录该路段的行车数据;
    对于异常数据库中的路段,根据该路段出现异常数据的次数来判断该路段是否是异常路段;
    对于判断为异常路段的路段,根据预设的模型预测路段异常的原因,提供给用户。
  2. 根据权利要求1所述的预测路段状况的方法,其特征在于,所述将采集的行车数据与正常观测样本进行对比,判断是否为异常数据,包括:
    根据采集的行车数据及其对应的权重计算得到该路段对应的路况评价值;
    将计算得到的路况评价值与正常观测样本对应的路况评价值范围进行对比,如果该路段行车数据对应的路况评价值在该路段正常观测样本的路况评价值范围内,则判定该路段为正常路段,其行车数据为正常数据,否则判定该路段为异常路段,其行车数据为异常数据。
  3. 根据权利要求2所述的预测路段状况的方法,其特征在于,所述根据该路段出现异常数据的次数来判断该路段是否是异常路段,包括:
    如果连续出现异常数据的次数大于设定的阈值,则判断该路段为异常路段,否则将该路段及其行车数据放入观察数据库;
    对于放入观察数据库的路段,继续跟踪该路段的行车数据;
    根据跟踪得到的行车数据中异常数据的次数及正常数据的次数,为行车数据对应的路况评价值赋予权重;
    根据路况评价值及其权重的乘积,判断该路段是否为异常路段。
  4. 根据权利要求3所述的预测路段状况的方法,其特征在于,所述为行车数据对应的路况评价值赋予权重,包括:
    在当前行车数据被认定为异常数据时,根据累计出现异常数据的次数增加当前行车数据的权重;
    在当前行车数据被认定为正常数据时,根据累计出现正常数据的次数减少当前行车数据的权重。
  5. 根据权利要求4所述的预测路段状况的方法,其特征在于,所述根据路况评价值及其权重的乘积,判断该路段是否为异常路段,包括:
    在当前路况评价值及其权重的乘积小于设定的第一阈值时,认为该路段为正常路段;而在前路况评价值及其权重的乘积大于设定的第二阈值时,判断该路段为异常路段。
  6. 根据权利要求2所述的预测路段状况的方法,其特征在于,所述方法还包括:
    如果根据采集的行车数据及其对应的权重计算得到其对应路段的路况评价值大于设定的第三阈值时,直接判断该路段是异常路段。
  7. 一种基于大数据预测路段状况的装置,其特征在于,所述装置包括:
    数据采集模块,用于采集道路路段上行驶车辆记录的行车数据;
    异常数据判断模块,用于将采集的行车数据与正常观测样本进行对比,判断是否为异常数据,如果是异常数据,则将该异常数据及其对应的路段放入异常数据库,并持续记录该路段的行车数据;
    异常路段判断模块,用于对于异常数据库中的路段,根据该路段出现异常数据的次数来判断该路段是否是异常路段;
    异常原因分析模块,用于对于判断为异常路段的路段,根据预设的模型预测路段异常的原因,提供给用户。
  8. 根据权利要求7所述的预测路段状况的装置,其特征在于,所述异常数据判断模块在将采集的行车数据与正常观测样本进行对比,判断是否为异常数据时,执行如下操作:
    根据采集的行车数据及其对应的权重计算得到该路段对应的路况评价值;
    将计算得到的路况评价值与正常观测样本对应的路况评价值范围进行对比,如果该路段行车数据对应的路况评价值在该路段正常观测样本的路况评价值范围内,则判定该路段为正常路段,其行车数据为正常数据,否则判定该路段为异常路段,其行车数据为异常数据。
  9. 根据权利要求8所述的预测路段状况的装置,其特征在于,所述异常路段判断模块在根据该路段出现异常数据的次数来判断该路段是否是异常路段时,执行如下操作:
    如果连续出现异常数据的次数大于设定的阈值,则判断该路段为异常路段,否则将该路段及其行车数据放入观察数据库;
    对于放入观察数据库的路段,继续跟踪该路段的行车数据;
    根据跟踪得到的行车数据中异常数据的次数及正常数据的次数,为行车数据对应的路况评价值赋予权重;
    根据路况评价值及其权重的乘积,判断该路段是否为异常路段。
  10. 根据权利要求9所述的预测路段状况的装置,其特征在于,所述异常路段判断模块为行车数据对应的路况评价值赋予权重,包括:
    在当前行车数据被认定为异常数据时,根据累计出现异常数据的次数增加当前行车数据的权重;
    在当前行车数据被认定为正常数据时,根据累计出现正常数据的次数减少当前行车数据的权重。
  11. 根据权利要求10所述的预测路段状况的装置,其特征在于,所述异常路段判断模块在根据路况评价值及其权重的乘积,判断该路段是否 为异常路段时,执行如下操作:
    在当前路况评价值及其权重的乘积小于设定的第一阈值时,认为该路段为正常路段;而在前路况评价值及其权重的乘积大于设定的第二阈值时,判断该路段为异常路段。
  12. 根据权利要求8所述的预测路段状况的装置,其特征在于,所述异常路段判断模块在判断该路段是否是异常路段时,还执行如下操作:
    如果根据采集的行车数据及其对应的权重计算得到其对应路段的路况评价值大于设定的第三阈值时,直接判断该路段是异常路段。
PCT/CN2016/109387 2015-12-22 2016-12-12 一种基于大数据预测路段状况的方法及装置 WO2017107790A1 (zh)

Priority Applications (2)

Application Number Priority Date Filing Date Title
JP2018531408A JP2019505892A (ja) 2015-12-22 2016-12-12 ビッグデータに基づいて道路状態を予測する方法及び装置
US16/016,502 US10977933B2 (en) 2015-12-22 2018-06-22 Method and apparatus for predicting road conditions based on big data

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201510976430.1 2015-12-22
CN201510976430.1A CN106910334B (zh) 2015-12-22 2015-12-22 一种基于大数据预测路段状况的方法及装置

Related Child Applications (1)

Application Number Title Priority Date Filing Date
US16/016,502 Continuation US10977933B2 (en) 2015-12-22 2018-06-22 Method and apparatus for predicting road conditions based on big data

Publications (1)

Publication Number Publication Date
WO2017107790A1 true WO2017107790A1 (zh) 2017-06-29

Family

ID=59089014

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2016/109387 WO2017107790A1 (zh) 2015-12-22 2016-12-12 一种基于大数据预测路段状况的方法及装置

Country Status (4)

Country Link
US (1) US10977933B2 (zh)
JP (1) JP2019505892A (zh)
CN (1) CN106910334B (zh)
WO (1) WO2017107790A1 (zh)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112070239A (zh) * 2020-11-11 2020-12-11 上海森亿医疗科技有限公司 基于用户数据建模的分析方法、系统、介质及设备
US10977933B2 (en) 2015-12-22 2021-04-13 Alibaba Group Holding Limited Method and apparatus for predicting road conditions based on big data

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112815956B (zh) * 2019-11-18 2022-06-14 百度在线网络技术(北京)有限公司 道路情况确定的方法及装置
CN112614342B (zh) * 2020-12-10 2022-08-30 大唐高鸿智联科技(重庆)有限公司 一种道路异常事件的预警方法、车载设备及路侧设备

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2006048501A (ja) * 2004-08-06 2006-02-16 Denso Corp 路面情報収集システム及びそれに用いる車載装置とサーバ
CN101246645A (zh) * 2008-04-01 2008-08-20 东南大学 一种识别离群交通数据的方法
CN201927175U (zh) * 2011-01-05 2011-08-10 中国科学院深圳先进技术研究院 智能交通系统信息采集器
CN103185724A (zh) * 2011-12-28 2013-07-03 富士通株式会社 路面检查装置
CN103975372A (zh) * 2011-12-06 2014-08-06 三菱电机株式会社 中心侧系统及车辆侧系统
CN104504903A (zh) * 2014-12-31 2015-04-08 北京赛维安讯科技发展有限公司 交通事件采集装置及方法、交通事件监测系统及方法
CN104933863A (zh) * 2015-06-02 2015-09-23 福建工程学院 一种交通道路中异常路段识别的方法及系统

Family Cites Families (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR100625096B1 (ko) * 2006-03-27 2006-09-15 주식회사 윈스테크넷 트래픽 변화량과 해킹 위협률의 상호 연관성 분석에 기초한예경보 방법 및 그 시스템
US20080103835A1 (en) * 2006-10-31 2008-05-01 Caterpillar Inc. Systems and methods for providing road insurance
SE0602606L (sv) * 2006-12-05 2008-06-06 Volvo Lastvagnar Ab Ett förfarande för att bestämma skicket hos en vägbana och ett förfarande för att generera en logg över ett fordons användning
CN102409599B (zh) * 2011-09-22 2013-09-04 中国科学院深圳先进技术研究院 道路路面检测方法及系统
CN103745595B (zh) * 2012-10-17 2016-08-03 中国电信股份有限公司 分析路况信息的方法和系统以及路况分析服务器
CN104751629B (zh) * 2013-12-31 2017-09-15 中国移动通信集团公司 一种交通事件的检测方法和系统
CN204311328U (zh) * 2014-12-04 2015-05-06 陕西中大机械集团有限责任公司 一种路面平整度实时监控系统
CN104929024B (zh) * 2015-06-15 2017-02-01 广西大学 路面平整度检测仪及路面平整度测量方法
CN106910334B (zh) 2015-12-22 2019-12-24 阿里巴巴集团控股有限公司 一种基于大数据预测路段状况的方法及装置

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2006048501A (ja) * 2004-08-06 2006-02-16 Denso Corp 路面情報収集システム及びそれに用いる車載装置とサーバ
CN101246645A (zh) * 2008-04-01 2008-08-20 东南大学 一种识别离群交通数据的方法
CN201927175U (zh) * 2011-01-05 2011-08-10 中国科学院深圳先进技术研究院 智能交通系统信息采集器
CN103975372A (zh) * 2011-12-06 2014-08-06 三菱电机株式会社 中心侧系统及车辆侧系统
CN103185724A (zh) * 2011-12-28 2013-07-03 富士通株式会社 路面检查装置
CN104504903A (zh) * 2014-12-31 2015-04-08 北京赛维安讯科技发展有限公司 交通事件采集装置及方法、交通事件监测系统及方法
CN104933863A (zh) * 2015-06-02 2015-09-23 福建工程学院 一种交通道路中异常路段识别的方法及系统

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US10977933B2 (en) 2015-12-22 2021-04-13 Alibaba Group Holding Limited Method and apparatus for predicting road conditions based on big data
CN112070239A (zh) * 2020-11-11 2020-12-11 上海森亿医疗科技有限公司 基于用户数据建模的分析方法、系统、介质及设备

Also Published As

Publication number Publication date
US20180301025A1 (en) 2018-10-18
JP2019505892A (ja) 2019-02-28
US10977933B2 (en) 2021-04-13
CN106910334B (zh) 2019-12-24
CN106910334A (zh) 2017-06-30

Similar Documents

Publication Publication Date Title
CN110197588B (zh) 一种基于gps轨迹数据的大货车驾驶行为评估方法及装置
WO2017107790A1 (zh) 一种基于大数据预测路段状况的方法及装置
CN111553902B (zh) 一种基于大数据的公路路面安全监测系统
CN104408925B (zh) 基于陈列雷达的交叉口运行状态评价方法
CN110164132B (zh) 一种道路交通异常的检测方法及系统
CN116644373B (zh) 基于人工智能的汽车流量数据分析管理系统
CN110176139A (zh) 一种基于dbscan+的道路拥堵识别可视化方法
CN104809878A (zh) 利用公交车gps数据检测城市道路交通异常状态的方法
CN111640304B (zh) 面向连续流交通设施的交通拥堵传播特征自动化量化提取方法
CN107590999B (zh) 一种基于卡口数据的交通状态判别方法
CN113870570B (zh) 一种基于etc的路网运行状态方法、系统和存储介质
WO2012024976A1 (zh) 一种交通信息处理方法及装置
CN103065469A (zh) 行程时间的确定方法和装置
CN112767684A (zh) 一种基于收费数据的高速公路交通拥堵检测方法
CN116168356B (zh) 一种基于计算机视觉的车辆损伤判别方法
CN101075377A (zh) 基于偏最小二乘原理的高速公路交通事件自动检测方法
CN109344903B (zh) 基于车载感知数据的城市道路路面故障实时检测方法
CN113158141B (zh) 基于大数据的新能源汽车超载检测方法
CN104809869A (zh) 基于升降式限高架的高架桥入口匝道交通状态判别方法
CN115497306A (zh) 一种基于gis数据的速度区间权重计算方法
CN112927497B (zh) 一种浮动车识别方法、相关方法和装置
Jang Data-Cleaning Technique for Reliable Real-Life Travel Time Estimation: Use of Dedicated Short-Range Communications Probes on Rural Highways
Hang et al. Research on expressway state recognition based on Beidou positioning and navigation data
Abou Chacra et al. Road Defect Detection in Street View Images using Texture Descriptors and Contour Maps
CN111369062B (zh) 车辆动力学指标提取方法和事故风险值预测方法

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 16877601

Country of ref document: EP

Kind code of ref document: A1

ENP Entry into the national phase

Ref document number: 2018531408

Country of ref document: JP

Kind code of ref document: A

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 16877601

Country of ref document: EP

Kind code of ref document: A1