WO2018072362A1 - 车辆实时轨迹预测方法及预测系统 - Google Patents

车辆实时轨迹预测方法及预测系统 Download PDF

Info

Publication number
WO2018072362A1
WO2018072362A1 PCT/CN2017/073558 CN2017073558W WO2018072362A1 WO 2018072362 A1 WO2018072362 A1 WO 2018072362A1 CN 2017073558 W CN2017073558 W CN 2017073558W WO 2018072362 A1 WO2018072362 A1 WO 2018072362A1
Authority
WO
WIPO (PCT)
Prior art keywords
time
vehicle
real
heading angle
angle change
Prior art date
Application number
PCT/CN2017/073558
Other languages
English (en)
French (fr)
Inventor
刘均
李磊
Original Assignee
深圳市元征科技股份有限公司
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 深圳市元征科技股份有限公司 filed Critical 深圳市元征科技股份有限公司
Publication of WO2018072362A1 publication Critical patent/WO2018072362A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/16—Anti-collision systems

Definitions

  • the invention relates to a vehicle network communication technology, in particular to a vehicle real-time trajectory prediction method and a prediction system.
  • ITS Intelligent Transportation System
  • the primary goal of driving is to achieve a high degree of safety in driving a car, to identify potential safety incidents, to promptly and accurately provide early warning and driving assistance, to effectively circumvent accidents and to plan driving routes.
  • the vehicle position information is provided by the Global Navigation Satellite System (GNSS).
  • GNSS Global Navigation Satellite System
  • the basic principle is to fit the driving state of the vehicle based on historical data. Furthermore, the trajectory of the vehicle is predicted.
  • GNSS data records can be used as a source to fit the vehicle's driving trajectory, but from the perspective of artificial intelligence, a large number of vehicle GNSS data can be used as the basis for excavating the driving state of the vehicle.
  • trajectory prediction techniques such as Markov chain method, inertial navigation method, Kalman filter method, particle filter method, least square method, Gaussian mixture model, neural network, machine learning method, topology theory and Other more complicated trajectory prediction methods.
  • the above various trajectory prediction methods exist at the theoretical research level, and there are few algorithms that can be used for vehicle trajectory prediction. Because there are computational resources, GNSS drift, storage resources, effectiveness, dependence on geographic information, and other vehicle location information in the embedded system, the above algorithms need to undergo a lot of mathematical operations, resulting in waste of resources.
  • the main object of the present invention is to provide a vehicle real-time trajectory prediction method and prediction system, aiming at Simplify the calculation process, reduce resource waste, and improve the accuracy of prediction.
  • the method for predicting real-time trajectory of a vehicle proposed by the present invention comprises the following steps:
  • the real-time vehicle parameters include a real-time heading angle of the vehicle, a time, a real-time speed, a real-time longitude of the vehicle, a real-time latitude, and a curvature of the vehicle travel trajectory, according to the heading angle and time, Obtaining a heading angle change rate, obtaining a vehicle travel path radius from a rate of change of the heading angle in a continuous state and the real-time speed;
  • the next time vehicle travel trajectory is obtained according to the real-time longitude of the vehicle, the real-time latitude, the arc of the vehicle travel trajectory, and the obtained travel radius of the vehicle.
  • the step S1 comprises the following steps:
  • S1a collecting real-time vehicle parameters from the GNSS dynamic data according to a preset frequency, wherein the real-time vehicle parameters include a real-time heading angle of the vehicle, a time, a real-time speed, a real-time longitude of the vehicle, a real-time latitude, and a curvature of the vehicle travel trajectory;
  • S1d obtains the radius of the traveling path of the vehicle according to the heading angle change rate and the real-time speed in the continuous state described above.
  • the step S2 comprises the following steps:
  • step S2c comparing the difference value of the heading angle change rate with the size of the preset threshold, if the difference value of the heading angle change rate is less than the preset threshold, step S3 is performed; otherwise, it ends.
  • the selected confidence interval is 50% to 100%.
  • the step S2 comprises the following steps:
  • the confidence rate interval is selected, and based on the difference between the head rate change rate and the confidence rate interval
  • the mapping relationship determines a preset threshold of the difference value of the heading angle change rate
  • step S2c comparing the difference value of the heading angle change rate with the size of the preset threshold, if the difference value of the heading angle change rate is less than the preset threshold, step S3 is performed; otherwise, it ends.
  • the step S2 comprises the following steps:
  • step S2c comparing the difference value of the heading angle change rate with the size of the preset threshold, if the difference value of the heading angle change rate is less than the preset threshold, step S3 is performed; otherwise, it ends.
  • the selected confidence interval is 50% to 100%.
  • the present invention also provides a vehicle real-time trajectory prediction system, wherein the system includes:
  • a radius acquisition module which collects real-time vehicle parameters from GNSS dynamic data, wherein the vehicle real-time parameters include a real-time heading angle of the vehicle, time, real-time speed, real-time longitude of the vehicle, real-time latitude, and arc of the vehicle travel trajectory, according to the heading angle and Time, obtaining a heading angle change rate, obtaining a vehicle travel path radius from the heading angle change rate in a continuous state and a real-time speed in the parameter;
  • a confidence rate screening module calculates a confidence rate reflecting a change state of the travel path of the vehicle according to the real-time heading angle and time of the vehicle, and compares the confidence rate with a preset confidence rate interval;
  • the trajectory prediction module obtains the vehicle travel trajectory of the next time according to the vehicle real-time longitude, the real-time latitude, the arc of the vehicle travel trajectory, and the obtained vehicle travel radius, if the confidence rate falls within the confidence rate interval.
  • the radius acquiring module comprises:
  • the collecting unit collects real-time vehicle parameters from the GNSS dynamic data according to a preset frequency, wherein the real-time vehicle parameters include a real-time heading angle of the vehicle, a time, a real-time speed, a real-time longitude of the vehicle, a real-time latitude, and a curvature of the vehicle driving track;
  • a conversion unit that obtains a real-time heading angle change rate of the vehicle by calculating the heading angle and time;
  • Filtering unit filtering the above-mentioned real-time heading angle change rate to filter out the heading angle change rate in a discontinuous state
  • the calculating unit obtains the radius of the traveling path of the vehicle according to the heading angle change rate in the continuous state and the real-time speed.
  • the confidence rate screening module comprises:
  • the confidence rate obtaining unit calculates a difference value of the heading angle change rate filtered by the filtering unit, and determines a difference value of the current heading angle change rate according to a mapping relationship between the difference value of the heading angle change rate and the confidence rate interval. Confidence rate interval;
  • a threshold module for determining a difference value of the heading angle change rate according to a mapping relationship between the selected confidence rate interval and the difference between the difference value of the heading angle change rate and the confidence rate interval;
  • a screening unit that compares the difference value of the heading angle change rate with the preset threshold value. If the difference value of the heading angle change rate is less than the preset threshold value, if the confidence rate falls within a preset confidence rate interval, The vehicle real-time longitude, the real-time latitude, the arc of the vehicle travel trajectory, and the obtained vehicle travel radius are obtained to obtain the next-time vehicle travel trajectory; otherwise, the end.
  • the selected confidence interval interval in the threshold module is 50% to 100%.
  • the confidence rate screening module comprises:
  • the confidence rate obtaining unit calculates a difference value of the heading angle change rate filtered by the filtering unit, and determines a difference value of the current heading angle change rate according to a mapping relationship between the difference value of the heading angle change rate and the confidence rate interval. Confidence rate interval;
  • a threshold module for determining a difference value of the heading angle change rate according to a mapping relationship between the selected confidence rate interval and the difference between the difference value of the heading angle change rate and the confidence rate interval;
  • a screening unit that compares the difference value of the heading angle change rate with the preset threshold value. If the difference value of the heading angle change rate is less than the preset threshold value, if the confidence rate falls within a preset confidence rate interval, The vehicle real-time longitude, the real-time latitude, the arc of the vehicle travel trajectory, and the obtained vehicle travel radius are obtained to obtain the next-time vehicle travel trajectory; otherwise, the end.
  • the confidence rate screening module comprises:
  • the confidence rate obtaining unit calculates a difference value of the heading angle change rate filtered by the filtering unit, and determines a difference value of the current heading angle change rate according to a mapping relationship between the difference value of the heading angle change rate and the confidence rate interval. Confidence rate interval;
  • a threshold module for determining a difference value of the heading angle change rate according to a mapping relationship between the selected confidence rate interval and the difference between the difference value of the heading angle change rate and the confidence rate interval;
  • a screening unit that compares the difference value of the heading angle change rate with the preset threshold value. If the difference value of the heading angle change rate is less than the preset threshold value, if the confidence rate falls within a preset confidence rate interval, The vehicle real-time longitude, the real-time latitude, the arc of the vehicle travel trajectory, and the obtained vehicle travel radius are obtained to obtain the next-time vehicle travel trajectory; otherwise, the end.
  • the selected confidence interval interval in the threshold module is 50% to 100%.
  • the filtering unit performs filtering processing using a second-order low-pass filter.
  • the vehicle travels on the earth, the earth approximates the sphere, and the trajectory of the vehicle can be approximated by a circular arc.
  • the trajectory of the vehicle is a circular arc, and the vehicle is first collected from the GNSS dynamic data.
  • Parameters including real-time heading angle of vehicle, time and real-time speed, real-time latitude of vehicle, real-time latitude, and arc of vehicle travel trajectory; firstly, according to heading angle and time, the heading angle change rate is obtained, from the heading angle in continuous state The rate of change and the real-time speed can obtain the radius of the travel path of the vehicle; and the time rate and the rate of change of the heading angle in the continuous state can calculate a confidence rate reflecting the change state of the travel path of the vehicle, and compare the confidence rate with a preset confidence rate interval; if the confidence rate falls within a preset confidence rate interval, the next moment is obtained according to the vehicle real-time longitude, the real-time latitude, the arc of the vehicle travel trajectory, and the obtained vehicle travel radius in the above parameters.
  • the scheme relies on the vehicle's own GNSS dynamic data information to predict the vehicle trajectory at the next moment. Compared with the prior art, it does not need to rely on other vehicle geographical location information, and does not need to consider factors such as GNSS drift and effectiveness, so there is no need to go through a lot of mathematical operations. Saving computer and storage resources; and in the step of radius prediction in the present application, since the rate of change of the heading angle in a continuous state is adopted, that is, the heading angle data in which the heading angle change rate is in a discontinuous state is considered to be a device Factors such as road conditions or operations that affect large data, if used, will result in distortion of the prediction radius. In order to improve the accuracy of the prediction, these data are not used as radius prediction data.
  • the concept of confidence rate is introduced to change the driving path of the vehicle.
  • the state is considered. If the confidence rate is within the predetermined interval, the vehicle is determined to be in a stable driving state. To improve the accuracy of the prediction, the data in the unsteady driving state is filtered out; and the prediction accuracy is improved by filtering the prediction data twice. .
  • FIG. 1 is a flow chart showing a method for predicting a real-time trajectory of a vehicle according to an embodiment of the present invention
  • FIG. 2 is a schematic diagram of functional modules of a real-time trajectory prediction system for a vehicle according to an embodiment of the present invention
  • FIG. 3 is a schematic diagram of a travel path of a vehicle in a real-time trajectory prediction method for a vehicle according to the present invention.
  • first, second, and the like in the present invention are used for the purpose of description only, and are not to be construed as indicating or implying their relative importance or implicitly indicating the number of technical features indicated.
  • features defining “first” or “second” may include at least one of the features, either explicitly or implicitly.
  • the meaning of "a plurality” is at least two, such as two, three, etc., unless specifically defined otherwise.
  • the terms "connected”, “fixed” and the like should be understood broadly, unless otherwise clearly defined and limited.
  • “fixed” may be a fixed connection, or may be a detachable connection, or may be integrated; It may be a mechanical connection or an electrical connection; it may be directly connected or indirectly connected through an intermediate medium, and may be an internal connection of two elements or an interaction relationship of two elements unless explicitly defined otherwise.
  • the specific meanings of the above terms in the present invention can be understood on a case-by-case basis.
  • the invention provides a real-time trajectory prediction method for a vehicle.
  • the vehicle trajectory change can be represented by a length of arc.
  • the radius of the instantaneous traveling trajectory of the vehicle can be calculated. Referring to FIG. 1 and FIG. 3, in an embodiment of the present invention, The method includes the following steps:
  • the real-time vehicle parameters include the vehicle real-time heading angle, time, real-time speed, real-time vehicle longitude, real-time latitude, and arc of the vehicle travel trajectory, according to Obtaining a heading angle and time in the parameter, obtaining a heading angle change rate, obtaining a radius of the vehicle travel path by a rate of change of the heading angle in a continuous state and the real-time speed;
  • the next-time vehicle travel trajectory is obtained by calculation according to the real-time longitude, the real-time latitude, the arc of the vehicle traveling, and the above-mentioned vehicle travel radius.
  • the trajectory confidence interval of the rapid change track is used to establish a model to determine the vehicle trajectory of the vehicle at the next moment, and the model only depends on its own Current and historical GNSS data information to predict the vehicle trajectory at the next moment, compared to the prior art (the embedded system relies on other vehicle location information, there are computing resources, GNSS drift, storage resources, effectiveness, etc.),
  • the algorithm does not need to go through a lot of mathematical operations, saving computer data, and the calculation accuracy is higher.
  • step S1 includes the following steps:
  • the real-time parameters of the vehicle include a real-time heading angle of the vehicle, time, real-time speed, real-time longitude of the vehicle, real-time latitude, and arc of the vehicle travel trajectory;
  • the acquisition frequency may be according to The setting needs to be set.
  • the acquisition frequency is 10 Hz;
  • the collected parameters are set according to specific needs.
  • the parameters in this embodiment include the real-time heading angle ⁇ , speed ⁇ , longitude and latitude of the vehicle;
  • the vehicle real-time heading angle change rate ⁇ is obtained by calculating the heading angle ⁇ and the time t;
  • ⁇ 1 is a heading angle at time t1
  • ⁇ 2 is a heading angle at time t2
  • the sample points of the real-time heading angles of the plurality of vehicles are obtained from the GNSS dynamic data, and the functions of taking the time as the abscissa and the heading angle as the ordinates can be simulated by using the sample points, and the function is obtained by deriving the time of the function.
  • the abscissa and the change rate of the vehicle heading angle are functions of the ordinate.
  • the instantaneous heading angle change rate of the vehicle at each moment is obtained by substituting into the function.
  • the heading angle change rate is filtered to prevent the discontinuous drastically changing heading angle change rate as the input signal to obtain the radius values from positive, negative and infinity; this will directly affect the radius calculation result. If the radius calculation is wrong, Directly causing the trajectory prediction of the vehicle to deviate from the actual trajectory of the next moment of the vehicle. In order to avoid the trajectory prediction as the input data in this special case, it is necessary to filter out the drastically changing heading angle change rate of the discontinuous state through this step;
  • a second-order low-pass filter is used for filtering processing, and the second-order low-pass filter expression conforms to:
  • u is the heading angle, where n ⁇ 3; at 0.32 Hz ⁇ f 0 ⁇ 0.34 Hz, 0 ⁇ ⁇ ⁇ 2, 100 ms ⁇ Ts ⁇ 400 ms.
  • f 0 is preferably 0.33 Hz
  • ⁇ is preferably 1
  • Ts is preferably 100 ms.
  • R is the radius of travel
  • v is the vehicle speed acquired from GNSS
  • ⁇ is the heading angle change rate
  • ⁇ 1 is the heading angle at time t1
  • ⁇ 2 is the heading angle at time t2.
  • the travel radius can be obtained from the angular velocity of the heading angle (ie, the unit of ⁇ , ⁇ is the angle/second) and the speed of the vehicle (ie, the unit of v, v is m/s):
  • l is the angular velocity of the heading angle
  • the unit is radians/second
  • the minimum value of velocity v is not zero.
  • the filtered vehicle travel radius R is obtained, and if the vehicle radius is greater than the critical value, the vehicle path is considered to be a straight line.
  • step S1 is effective for the steady state of the vehicle, but when the driving state changes drastically, the calculated predicted radius is discounted. Therefore, a confidence rate reflecting the trueness of the predicted radius when the drastic driving state changes is required. A high confidence rate means that the vehicle is in a stable driving state, and you can safely use the method of calculating the predicted radius in the previous step. Otherwise, the calculated prediction radius is discarded.
  • the angular velocity of the heading angle (ie, the unit of ⁇ , ⁇ is angle/second) is passed through a second-order low-pass filter transfer function:
  • step S2 specifically includes:
  • the difference value of the heading angle change rate refers to the secondary derivation of the heading angle with respect to the time function, reflecting the state of the heading angle change rate. If the difference value of the heading angle change rate is constant, the heading is indicated. The angular change is stable and the vehicle is in a stable driving state; the magnitude of the change can be judged by a preset threshold value which is obtained based on a plurality of experimental data. If the difference value of the heading angle change rate is less than the preset threshold, the driving radius of the vehicle is considered to be high and can be adopted; if the difference value of the heading angle change rate is greater than or equal to the preset threshold, the driving radius confidence rate of the vehicle is considered Low; give up, refuse to accept, end this data forecast;
  • the difference value of the heading angle change rate (ie, the heading angle second difference value) in the step S2a is in accordance with the following expression:
  • the confidence rate interval A is selected to be greater than 50%, and the selected confidence interval interval can be obtained through experiments.
  • the mapping relationship between the difference value of the heading angle change rate and the confidence rate interval is also obtained through experiments.
  • the predetermined confidence rate interval and Table 1, the preset threshold corresponding to the difference value of the heading angle change rate is 2.5, that is, when the second difference value of the heading angle is less than 2.5, the corresponding confidence rate interval satisfies 50% ⁇ A ⁇ 100%. ;
  • step S2c comparing the difference value of the heading angle change rate with the size of the preset threshold value, if the difference value of the heading angle change rate is when the second difference value of the heading angle is less than a preset threshold value of 2.5, then the confidence rate is selected Within the range of 50% ⁇ A ⁇ 100%, then proceed to step S3;
  • the difference value of the heading angle change rate is greater than or equal to the preset threshold of 2.5, the confidence rate is considered not If the 50% ⁇ A ⁇ 100% is not satisfied within the confidence rate interval, the data prediction is ended.
  • the threshold value in step S2c is 2.5, that is, when the secondary difference value of the heading angle is less than 2.5 and the corresponding confidence rate is greater than 50%, the heading angle change is considered to be stable, and the vehicle is in a stable driving state, when the second difference value of the heading angle is greater than 2.5.
  • the corresponding confidence rate is less than 50%, it is considered that the heading angle change is unstable, the vehicle is in an unstable driving state, and the vehicle trajectory prediction corresponding to the vehicle travel radius obtained by the previous calculation is abandoned, or the data considered invalid is not As the current trajectory prediction input.
  • vehicle safety events such as vehicle rear-end collision, vehicle collision in bad weather environment, and the like.
  • An embodiment of the present invention provides a real-time trajectory prediction system for a vehicle, the system comprising: a radius acquisition module, a confidence rate screening module, and a trajectory prediction module 3;
  • the radius acquisition module collects real-time vehicle parameters from the GNSS dynamic data, wherein the real-time vehicle parameters include the vehicle real-time heading angle, time, real-time speed, vehicle real-time longitude, real-time latitude, and arc of the vehicle travel trajectory, according to the heading angle in the parameter And time, obtaining a heading angle change rate, obtaining a vehicle travel path radius from the heading angle change rate in a continuous state and a real-time speed in the above parameters;
  • the confidence rate screening module calculates a confidence rate reflecting the change state of the travel path of the vehicle according to the time and the real-time heading angle in the above parameters, and compares the confidence rate with a preset confidence rate interval;
  • the trajectory prediction module 3 if the confidence rate falls within the confidence rate interval, inputting the radius of the vehicle travel path as a radius signal to the trajectory acquisition module 3; obtaining the vehicle at the next moment according to the parameter input by the radius acquisition module and the radius signal Track; if the confidence rate is not within the confidence rate interval, the radius signal is not input to the trajectory acquisition module 3, and the current data prediction is ended.
  • the invention is a sub-patent implementation method in the patent of the vehicle safety event precise warning based on the DSRC V2X communication and the multi-sensor system, that is, the vehicle trajectory prediction algorithm of the vehicle GNSS data, It is proposed to predict the path trajectory of the vehicle at the next moment based on current and historical GNSS data.
  • GNSS data records can be used as a source to fit the vehicle's driving trajectory.
  • a large number of vehicle GNSS data can be used as the basis for excavating the driving state of the vehicle; the present invention proposes a speed, acceleration, and heading based on GNSS data.
  • the angle, curvature, and rapid change trajectory confidence interval are used to establish the model to determine the vehicle's trajectory at the next moment of the vehicle, which is simpler than the prior art algorithm and saves a lot of computing resources.
  • the radius acquisition module includes an acquisition unit 11, a conversion unit 12, a filtering unit 13, and a calculation unit 14; wherein:
  • the collecting unit 11 collects real-time vehicle parameters from the GNSS dynamic data according to a preset frequency; specifically, the data collector collects the parameter data to be used from the GNSS dynamic data according to the set frequency; and the real-time heading angle in the above parameters
  • the data is sent to the converting unit 12, and the real-time speed data is sent to the calculating unit 14, and the longitude, latitude and traveling arc data are sent to the trajectory prediction module;
  • the converting unit 12 obtains the real-time heading angle and time of the vehicle from the collecting unit 11, and obtains the real-time heading angle change rate of the vehicle by the heading angle; obtains the real-time heading angle and time of the vehicle from the collecting unit 11, and sets the calculation in the program.
  • the formula obtains the heading angle change rate; and inputs it to the filtering unit;
  • the filtering unit 13 performs filtering processing on the real-time heading angle change rate to filter out the heading angle change rate in the discontinuous state; and uses a second-order low-pass filter to perform filtering processing; specifically, a filtering circuit;
  • the calculating unit 14 obtains the real-time speed of the vehicle from the collecting unit 11, and obtains the radius of the traveling path of the vehicle according to the heading angle change rate and the real-time speed in the continuous state;
  • the confidence rate screening module includes a confidence rate acquisition unit 21, a threshold value unit 22, and a screening unit 23, wherein: a second-order low-pass filter is used for filtering processing; specifically, a filter circuit;
  • the confidence rate acquisition unit 21 calculates a difference value of the heading angle change rate filtered by the filtering unit, and determines a current heading angle change rate according to a pre-stored mapping relationship between the difference value of the heading angle change rate and the confidence rate interval. a confidence interval corresponding to the difference value;
  • the threshold unit 22 selects a confidence rate interval, and determines a preset threshold value of the difference value of the heading angle change rate according to a mapping relationship between the selected confidence rate interval and the difference between the difference value of the heading angle change rate and the confidence rate interval;
  • the screening unit 23 compares the difference value of the heading angle change rate with the preset threshold value, and if the difference value of the heading angle change rate is smaller than the preset threshold, if the confidence rate falls within a preset confidence rate area
  • the vehicle travel trajectory of the next moment is obtained according to the real-time longitude of the vehicle, the real-time latitude, the curvature of the vehicle travel trajectory, and the obtained travel radius of the vehicle; otherwise, it ends.
  • the vehicle travel path radius is input as a radius signal to the trajectory prediction module; if the heading angle change rate If the difference value is greater than or equal to the preset threshold, it is considered not to be within the confidence rate interval; then the radius signal is not input to the prediction module to end the current data prediction;
  • the trajectory prediction module 3 obtains the vehicle trajectory of the next time according to the real-time longitude of the vehicle, the real-time latitude, the arc of the vehicle travel trajectory, and the radius information of the driving radius of the vehicle in the screening unit. Receiving the real-time longitude of the vehicle from the collection unit 11, the real-time latitude, the arc of the vehicle travel trajectory, and the radius information from the screening unit 23, by calculating the above data according to a preset formula, the vehicle travel trajectory can be obtained at the next moment.
  • Each vehicle itself periodically transmits local vehicle GNSS travel trajectory prediction information based on short-range communication technology and simultaneously receives surrounding vehicle GNSS travel trajectory prediction information; local vehicles and remote vehicles are in different positions, and safety occurs at this time.
  • the GNSS predicted trajectory of the local vehicle and the predicted trajectory of the remote GNSS will coincide at some point in the future, and a warning will be generated on the local vehicle or the remote vehicle when a security event is about to occur.
  • vehicle safety events such as vehicle rear-end collisions, vehicle collisions in bad weather conditions, and the like.

Landscapes

  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Traffic Control Systems (AREA)
  • Navigation (AREA)

Abstract

一种基于车辆实时轨迹预测方法及预测系统，该方法包括：从GNSS动态数据中采集车辆实时参数，其中参数中包括车辆实时航向角、时间及实时速度，由处于连续状态的所述航向角的变化率和上述实时速度，获得车辆行驶路径半径；根据上述参数计算反映车辆行驶路径变化状态的置信率，并比较置信率与预设的置信率区间，如果置信率落在置信率区间内，则进行轨迹预测，如果置信率不在置信率区间内，则结束本次数据预测；根据上述参数以及获得的车辆行驶半径获得下一时刻车辆行驶轨迹。该方法依赖车辆自身GNSS动态数据信息来预测下一时刻车辆行驶轨迹，无须经过大量数学运算，且计算精度更高。

Description

车辆实时轨迹预测方法及预测系统 技术领域
本发明涉及车联网通信技术，尤其是一种车辆实时轨迹预测方法及预测系统。
背景技术
目前，在智能交通系统(Intelligent Transport System或者Intelligent Transportation System，简称ITS)发展过程中，实现精确高效安全的驾驶是智能交通终极目标。驾驶首要目标就是实现汽车驾驶的高度安全性，发现潜在的安全事故，及时精确的提出预警和驾驶辅助，有效的规避事故和规划行驶路径。
目前车辆位置信息都是由全球导航卫星系统(Global Navigation Satellite System，以下简称GNSS)提供的，基于GNSS的车辆轨迹预测技术目前有较多的版本，基本原理就是根据历史数据拟合汽车行驶状态，进而对车辆的行驶轨迹进行预测。
车辆在行驶过程中，大量的GNSS数据记录可以作为拟合车辆行驶轨迹的来源，但是从人工智能角度来看，大量车辆GNSS数据可以作为挖掘车辆行驶状态的基础。目前现有的轨迹预测技术较多，如基于马尔可夫链法、惯性导航法、卡尔曼滤波法、粒子滤波法、最小二乘法、高斯混合模型、神经网络、机器学习法、拓扑学理论以及其他较为复杂的轨迹预测方法。以上各种轨迹预测方法存在于理论研究水平，真正能用于车辆轨迹预测的算法较少。因为在嵌入式系统中存在着计算资源、GNSS飘移、存储资源、实效性、依赖于地理信息、依赖于其他车辆位置信息等要求，上述算法需要经过大量的数学运算，造成资源浪费。
发明内容
本发明的主要目的是提供一种车辆实时轨迹预测方法及预测系统，旨在 简化运算过程，减少资源浪费，并提高预测的精确度。
为实现上述目的，本发明提出的车辆实时轨迹预测方法，包括以下步骤：
S1，从GNSS动态数据中采集车辆实时参数，其中所述车辆实时参数包括车辆实时航向角、时间、实时速度、车辆实时经度、实时纬度及车辆行驶轨迹的弧度，根据所述航向角及时间，获得航向角变化率，由处于连续状态的所述航向角的变化率和所述实时速度，获得车辆行驶路径半径；
S2，根据所述时间及实时航向角计算反映车辆行驶路径变化状态的置信率，并比较所述置信率与预设的置信率区间；
S3，如果置信率落在预设的置信率区间内，则根据所述车辆实时经度、实时纬度、车辆行驶轨迹的弧度以及获得的所述车辆行驶半径获得下一时刻车辆行驶轨迹。
优选地，所述步骤S1包括以下步骤：
S1a，按照预设频率从GNSS动态数据中采集车辆实时参数，其中所述车辆实时参数包括车辆实时航向角、时间、实时速度、车辆实时经度、实时纬度及车辆行驶轨迹的弧度；
S1b，由航向角及时间通过计算获得车辆实时航向角变化率；
S1c，对上述实时航向角变化率进行滤波处理，以过滤掉处于非连续状态的航向角变化率；
S1d，根据上述处于连续状态的航向角变化率以及实时速度，获得车辆行驶路径半径。
优选地，所述步骤S2包括以下步骤：
S2a，计算经过上述步骤S1c后的航向角变化率的差分值；
S2b，选定置信率区间，并根据航向角变化率的差分值与置信率区间之间的映射关系确定航向角变化率的差分值的预设阈值；
S2c，比较航向角变化率的差分值与上述预设阀值的大小，如果航向角变化率的差分值小于上述预设阈值，则执行步骤S3；否则，结束。
优选地，所述步骤S2b中，选定的置信率区间为50％～100％。
优选地，所述步骤S2包括以下步骤：
S2a，计算经过上述步骤S1c后的航向角变化率的差分值；
S2b，选定置信率区间，并根据航向角变化率的差分值与置信率区间之间 的映射关系确定航向角变化率的差分值的预设阈值；
S2c，比较航向角变化率的差分值与上述预设阀值的大小，如果航向角变化率的差分值小于上述预设阈值，则执行步骤S3；否则，结束。
优选地，所述步骤S2包括以下步骤：
S2a，计算经过上述步骤S1c后的航向角变化率的差分值；
S2b，选定置信率区间，并根据航向角变化率的差分值与置信率区间之间的映射关系确定航向角变化率的差分值的预设阈值；
S2c，比较航向角变化率的差分值与上述预设阀值的大小，如果航向角变化率的差分值小于上述预设阈值，则执行步骤S3；否则，结束。
优选地，所述步骤S2b中，选定的置信率区间为50％～100％。
另外，本发明还提供一种车辆实时轨迹预测系统，其中，该系统包括：
半径获取模块，从GNSS动态数据中采集车辆实时参数，其中所述车辆实时参数包括车辆实时航向角、时间、实时速度、车辆实时经度、实时纬度及车辆行驶轨迹的弧度，根据所述航向角及时间，获得航向角变化率，由处于连续状态的所述航向角变化率和所述参数中的实时速度，获得车辆行驶路径半径；
置信率筛选模块，根据所述车辆实时航向角及时间计算反映车辆行驶路径变化状态的置信率，并比较所述置信率与预设的置信率区间；
轨迹预测模块，如果置信率落在置信率区间内，根据半径获取模块输入的所述车辆实时经度、实时纬度、车辆行驶轨迹的弧度以及获得的所述车辆行驶半径获得下一时刻车辆行驶轨迹。
优选地，所述半径获取模块包括：
采集单元，按照预设频率从GNSS动态数据中采集车辆实时参数，其中所述车辆实时参数包括车辆实时航向角、时间、实时速度、车辆实时经度、实时纬度及车辆行驶轨迹的弧度；
转换单元，由所述航向角及时间通过计算获得车辆实时航向角变化率；
过滤单元，对上述实时航向角变化率进行滤波处理，以过滤掉处于非连续状态的航向角变化率；
计算单元，根据上述处于连续状态的航向角变化率以及所述实时速度，获得车辆行驶路径半径。
优选地，所述置信率筛选模块包括：
置信率获取单元，计算经过上述过滤单元过滤后的航向角变化率的差分值；并根据航向角变化率的差分值与置信率区间之间的映射关系确定当前航向角变化率的差分值对应的置信率区间；
阈值模块，选定置信率区间，根据选定置信率区间以及航向角变化率的差分值与置信率区间之间的映射关系的确定航向角变化率的差分值的预设阈值；
筛选单元，比较航向角变化率的差分值与上述预设阀值的大小，如果航向角变化率的差分值小于上述预设阈值，如果置信率落在预设的置信率区间内，则根据所述车辆实时经度、实时纬度、车辆行驶轨迹的弧度以及获得的所述车辆行驶半径获得下一时刻车辆行驶轨迹；否则，结束。
优选地，所述阈值模块中选定的置信率区间为50％～100％。
优选地，所述置信率筛选模块包括：
置信率获取单元，计算经过上述过滤单元过滤后的航向角变化率的差分值；并根据航向角变化率的差分值与置信率区间之间的映射关系确定当前航向角变化率的差分值对应的置信率区间；
阈值模块，选定置信率区间，根据选定置信率区间以及航向角变化率的差分值与置信率区间之间的映射关系的确定航向角变化率的差分值的预设阈值；
筛选单元，比较航向角变化率的差分值与上述预设阀值的大小，如果航向角变化率的差分值小于上述预设阈值，如果置信率落在预设的置信率区间内，则根据所述车辆实时经度、实时纬度、车辆行驶轨迹的弧度以及获得的所述车辆行驶半径获得下一时刻车辆行驶轨迹；否则，结束。
优选地，所述置信率筛选模块包括：
置信率获取单元，计算经过上述过滤单元过滤后的航向角变化率的差分值；并根据航向角变化率的差分值与置信率区间之间的映射关系确定当前航向角变化率的差分值对应的置信率区间；
阈值模块，选定置信率区间，根据选定置信率区间以及航向角变化率的差分值与置信率区间之间的映射关系的确定航向角变化率的差分值的预设阈值；
筛选单元，比较航向角变化率的差分值与上述预设阀值的大小，如果航向角变化率的差分值小于上述预设阈值，如果置信率落在预设的置信率区间内，则根据所述车辆实时经度、实时纬度、车辆行驶轨迹的弧度以及获得的所述车辆行驶半径获得下一时刻车辆行驶轨迹；否则，结束。
优选地，所述阈值模块中选定的置信率区间为50％～100％。
优选地，所述过滤单元采用二阶低通滤波器进行滤波处理。
本发明技术方案中，车辆在地球上行驶，地球近似球体，车辆的行驶轨迹可近似用一段圆弧表示，本方案中假定车辆的行驶轨迹是一段圆弧，首先从GNSS动态数据中采集车辆实时参数，这些参数中包括车辆实时航向角、时间及实时速度车辆实时经度、实时纬度、车辆行驶轨迹的弧度；首先根据航向角及时间，获得航向角变化率，由处于连续状态的所述航向角的变化率和实时速度，能够获得车辆行驶路径半径；再由时间参数及处于连续状态的所述航向角的变化率能够计算出反映车辆行驶路径变化状态的置信率，并比较所述置信率与预设的置信率区间；如果置信率落在预设的置信率区间内，则根据上述参数中的车辆实时经度、实时纬度、车辆行驶轨迹的弧度以及获得的所述车辆行驶半径获得下一时刻车辆行驶轨迹。本方案依赖车辆自身GNSS动态数据信息来预测下一时刻的车辆行驶轨迹，相对现有技术，无须依赖其他车辆地理位置信息，不用考虑GNSS漂移、实效性等因素，因此无须经过大量的数学运算，节约了计算机和存储资源；且本申请中半径预测的步骤中，由于采用的是处于连续状态的所述航向角的变化率，即航向角变化率处于非连续状态的航向角数据认为是受设备、路况或操作等因素影响较大的数据，如果采用则会导致预测半径失真，为提高预测的准确性，因此这些数据不作为半径预测数据；同时，引入置信率概念，对于车辆的行驶路径变化状态予于考虑，如果置信率在预定区间内，则认定车辆处于稳定行驶状态，为提高预测的准确性，过滤掉处于非稳定行驶状态的数据；通过两次对预测数据的过滤，提高预测精度。
附图说明
为了更清楚地说明本发明实施例或现有技术中的技术方案，下面将对实施例或现有技术描述中所需要使用的附图作简单地介绍，显而易见地，下面 描述中的附图仅仅是本发明的一些实施例，对于本领域普通技术人员来讲，在不付出创造性劳动的前提下，还可以根据这些附图示出的结构获得其他的附图。
图1为本发明一实施例提供的车辆实时轨迹预测方法的流程图示意图；
图2为本发明一实施例提供的车辆实时轨迹预测系统的功能模块示意图；
图3为本发明提供的车辆实时轨迹预测方法中车辆的行驶轨迹示意图。
本发明目的的实现、功能特点及优点将结合实施例，参照附图做进一步说明。
具体实施方式
下面将结合本发明实施例中的附图，对本发明实施例中的技术方案进行清楚、完整地描述，显然，所描述的实施例仅仅是本发明的一部分实施例，而不是全部的实施例。基于本发明中的实施例，本领域普通技术人员在没有作出创造性劳动前提下所获得的所有其他实施例，都属于本发明保护的范围。
需要说明，本发明实施例中所有方向性指示(诸如上、下、左、右、前、后……)仅用于解释在某一特定姿态(如附图所示)下各部件之间的相对位置关系、运动情况等，如果该特定姿态发生改变时，则该方向性指示也相应地随之改变。
另外，在本发明中如涉及“第一”、“第二”等的描述仅用于描述目的，而不能理解为指示或暗示其相对重要性或者隐含指明所指示的技术特征的数量。由此，限定有“第一”、“第二”的特征可以明示或者隐含地包括至少一个该特征。在本发明的描述中，“多个”的含义是至少两个，例如两个，三个等，除非另有明确具体的限定。
在本发明中，除非另有明确的规定和限定，术语“连接”、“固定”等应做广义理解，例如，“固定”可以是固定连接，也可以是可拆卸连接，或成一体；可以是机械连接，也可以是电连接；可以是直接相连，也可以通过中间媒介间接相连，可以是两个元件内部的连通或两个元件的相互作用关系，除非另有明确的限定。对于本领域的普通技术人员而言，可以根据具体情况理解上述术语在本发明中的具体含义。
另外，本发明各个实施例之间的技术方案可以相互结合，但是必须是以 本领域普通技术人员能够实现为基础，当技术方案的结合出现相互矛盾或无法实现时应当认为这种技术方案的结合不存在，也不在本发明要求的保护范围之内。
本发明提出一种车辆实时轨迹预测方法。
在瞬时状态，车辆轨迹变化可以用一段弧长来表示，根据在瞬时时间航向角变化以及速度变化，可以计算出车辆瞬时行驶轨迹的半径，请参照图1、图3，在本发明一实施例中，该方法包括以下步骤：
S1，预测车辆行驶路径半径；具体从GNSS动态数据中采集车辆实时参数，其中所述车辆实时参数包括车辆实时航向角、时间、实时速度、车辆实时经度、实时纬度及车辆行驶轨迹的弧度，根据参数中的航向角及时间，获得航向角变化率，由处于连续状态的所述航向角的变化率和上述实时速度，获得车辆行驶路径半径；
S2，根据上述参数中的时间及实时航向角计算反映车辆行驶路径变化状态的置信率，并比较所述置信率与预设的置信率区间；
S3，如果置信率落在预设的置信率区间内，则根据所述实时经度、实时纬度、车辆绕行驶的弧度及上述车辆行驶半径通过计算获得下一时刻车辆行驶轨迹。
本发明的技术方案，基于GNSS数据的速度、加速度、航向角、曲率以及车辆行驶轨迹的弧度、急速变化轨迹置信区间来建立模型判断车辆下一个时刻的车辆行驶轨迹，该模型仅仅依赖于自身的当前和历史GNSS数据信息来预测下一个时刻的车辆行驶轨迹，相对于现有技术(嵌入式系统依赖于其他车辆位置信息，存在着计算资源、GNSS飘移、存储资源、实效性等要求)，上述算法无须经过大量的数学运算，节约计算机资料，且计算精度更高。
进一步地，所述步骤S1包括以下步骤：
S1a，按照预设频率从GNSS动态数据中采集车辆实时参数，其中所述车辆实时参数包括车辆实时航向角、时间、实时速度、车辆实时经度、实时纬度及车辆行驶轨迹的弧度；采集频率可根据需要进行设定，本实施例中采集频率为10Hz；采集的参数根据具体需要进行设置，本实施例中的参数包括车辆实时航向角λ、速度ν、经度和纬度；
S1b，由航向角λ及时间t通过计算获得车辆实时航向角变化率α；
由：
Figure PCTCN2017073558-appb-000001
获得航向角变化率α；其中λ1为t1时刻的航向角；λ2为t2时刻的航向角；
从GNSS动态数据中获得多个车辆实时航向角的样本点，通过这些样本点可模拟获得以时间为横坐标、以航向角为纵坐标的函数，对所述函数对时间进行求导获得时间为横坐标、以车辆航向角变化率为纵坐标的函数，每个时刻车辆瞬时航向角变化率都通过代入该函数中获得。
S1c，对上述实时航向角变化率进行滤波处理，以过滤掉处于非连续状态的航向角变化率；
因为存在道路噪声(复杂道路状况影响GNSS传感器数据输出)、传感器噪声(自身的飘移)以及驾驶员开车行为噪声(复杂驾驶行为)，均会对预测半径产生影响，此时需要对实时计算出来的航向角变化率进行滤波处理以阻止非连续的剧烈变动的航向角变化率作为输入信号得到从正、负值、无穷大的半径值；这样就会直接影响半径的计算结果，如果半径计算错误，会直接导致车辆的轨迹预测偏离车辆下一刻实际轨迹，为了避免这一特殊情况下的数据作为输入数据进行轨迹预测，因此需要通过此步骤将非连续状态的剧烈变动的航向角变化率过滤掉；
具体采用二阶低通滤波器进行滤波处理，所述二阶低通滤波器表达式符合：
Figure PCTCN2017073558-appb-000002
对上述表达式(1)进行离散后得到以下关系式：
Figure PCTCN2017073558-appb-000003
其中初始化条件选取为：y1＝u1，y2＝u2；ω0＝2πf0，f0为截止频率，ζ为阻尼系数，Ts为采样时间，y为航向角变化率的差分值，u为航向角，其中n≥3；在0.32Hz≤f0≤0.34Hz，0≤ζ≤2，100ms≤Ts≤400ms。本实施例中f0优选为0.33Hz，ζ优选为1，Ts优选为100ms。上述公式及公式中参数选取能有效过滤掉非连续状态的航向角变化率，过滤效果更佳。
S1d，参照图3，根据上述处于连续状态的航向角变化率α以及实时速度ν，获得车辆行驶路径半径R:
根据公式：R＝v/α和
Figure PCTCN2017073558-appb-000004
可得：
Figure PCTCN2017073558-appb-000005
R为行驶半径，v为从GNSS获取的车辆速度，α为航向角变化率；λ1为t1时刻的航向角；λ2为t2时刻的航向角。
对预测的车辆行驶半径进行滤除噪声：
由航向角的角速度(即α，α的单位为角度/秒)和车辆的速度(即v，v的单位为米/秒)能够获得行驶半径：
R＝v/α；
又由
Figure PCTCN2017073558-appb-000006
(即将角速度的单位由角度/秒转换为弧度/秒)获得曲率：
Figure PCTCN2017073558-appb-000007
其中l为航向角的角速度，单位为弧度/秒，速度v最小值不为零。
上述曲率经二阶低通滤波器传递函数：
Figure PCTCN2017073558-appb-000008
过滤后获得车辆原始曲率ρ；
由
Figure PCTCN2017073558-appb-000009
获得过滤后的车辆行驶半径R，如果车辆半径大于临界值，则认为车辆路径为直线。
步骤S1计算是针对车辆稳定状态行驶有效，但是在驾驶状态剧烈变化时，计算出的预测半径会打折扣。因而需要一个在剧烈驾驶状态变化时的反映预测半径真实性的置信率。置信率高则表示车辆处于稳定行驶状态，可以放心的使用上一步的计算预测半径的方法。反之，则放弃计算出来的预测半径。
计算预测的车辆行驶半径的置信率：
航向角的角速度(即α，α的单位为角度/秒)经二阶低通滤波器传递函数：
Figure PCTCN2017073558-appb-000010
过滤后取绝对值，查看航向角的二次差分值与置信率参数对比表获得置信率。
进一步地，所述步骤S2具体包括：
S2a，计算经过上述步骤S1c后的航向角变化率的差分值；
这里航向角变化率的差分值指的是对航向角相对时间函数的二次求导，反映航向角变化率的状态，如果航向角变化率的差分值一直不变，表示航向 角变化是稳定的，车辆处于稳定行驶状态；变化的大小可以通过预先设定的阈值进行判断，这个阈值是根据由多次实验数据获得的。如果航向角变化率的差分值小于预设阈值，则认为车辆的行驶半径置信率高，可以采纳；反之如果航向角变化率的差分值大于或等于预设阈值，则认为车辆的行驶半径置信率低；放弃，不予采纳，结束本次数据预测；
优选地，所述步骤S2a中航向角变化率的差分值(即航向角二次差分值)符合以下表达式：
Figure PCTCN2017073558-appb-000011
其中初始化条件选取为：y1＝0，y2＝0；ω’0＝2πf0’，f0’为置信率的截止频率，ζ’为置信率阻尼系数，T’s为计算置信率的离散化采样时间，y为航向角二次差分值，u为航向角。其中0.33Hz≤f0’≤1Hz，0≤ζ’≤2，100ms≤T’s≤400ms；进一步地，f0’优选为1Hz，ζ’优选为1，T’s优选为100ms。上述公式及公式中的参数选取能够使得航向角二次差分值与置信率的映射关系更为明显，通过航向角二次差分值能够更为准确的反映置信率，即反映了车辆行驶状态。
S2b，选定置信率区间，并根据航向角变化率的差分值与置信率区间之间的映射关系(参见下面的表一)确定航向角变化率的差分值的预设阈值；
本实施例中，置信率区间A选定大于50％，选定的置信率区间可以通过实验获得，航向角变化率的差分值与置信率区间之间的映射关系也是通过实验获得的，通过选定的置信率区间以及表一，能够获得航向角变化率的差分值对应的预设阈值是2.5，即当航向角二次差分值小于2.5时，对应置信率区间满足50％<A≤100％；
其中航向角的二次差分值与置信率参数对比参见表一：
Figure PCTCN2017073558-appb-000012
S2c，比较航向角变化率的差分值与上述预设阀值的大小，如果航向角变化率的差分值即当航向角二次差分值小于预设阈值2.5时，则认为在置信率在选定的范围内即满足50％<A≤100％，那么进行步骤S3；
如果航向角变化率的差分值大于或等于预设阈值2.5时，则认为置信率不 在置信率区间范围内即不满足50％<A≤100％；则结束本次数据预测。
步骤S2c中的阈值为2.5，即当航向角二次差分值小于2.5，对应置信率大于50％时，认为航向角变化是稳定的，车辆处于稳定行驶状态，当航向角二次差分值大于2.5时，对应置信率小于50％时，认为航向角变化是不稳定的，车辆处于不稳定行驶状态，放弃上一部计算获得的车辆行驶半径对应的车辆轨迹预测，或者说认为无效的数据，不作为本次轨迹预测输入。以实现对车辆轨迹准确预测，从而减少车辆安全事件如车辆追尾、恶劣天气环境下车辆碰撞等的发生。
参见图3，表明车辆将会以这个半径一直行驶下去，通过车辆实时经度、纬度能够判断车辆所处的位置，θ为车辆行驶轨迹的弧度，可通过GNSS动态数据获得，根据几何关系：S＝θ*R，R为车辆行驶半径，可以近似得到车辆预测轨迹S。
此时要区分道路是笔直状况的场景，由实验数据得出，如果车辆的行驶半径大于2000米时，则认为道路笔直的，如果车辆的行驶半径小于或等于2000米，则认为车辆行驶轨迹是圆弧。
本发明一实施例提供一种车辆实时轨迹预测系统，该系统包括：半径获取模块、置信率筛选模块和轨迹预测模块3；
半径获取模块，从GNSS动态数据中采集车辆实时参数，其中所述车辆实时参数包括车辆实时航向角、时间、实时速度、车辆实时经度、实时纬度及车辆行驶轨迹的弧度,根据参数中的航向角及时间，获得航向角变化率，由处于连续状态的所述航向角变化率和上述参数中的实时速度，获得车辆行驶路径半径；
置信率筛选模块，根据上述参数中的时间及实时航向角计算反映车辆行驶路径变化状态的置信率，并比较所述置信率与预设的置信率区间；
轨迹预测模块3，如果置信率落在置信率区间内，则将上述车辆行驶路径半径作为半径信号向轨迹获取模块3输入；根据半径获取模块输入的上述参数以及上述半径信号获得下一时刻车辆行驶轨迹；如果置信率不在置信率区间内，则不向所述轨迹获取模块3输入半径信号，结束本次数据预测。
本发明是在DSRC V2X通信以及多传感器系统基础上的车辆安全事件精确预警专利中的子专利实现方式，即车载GNSS数据的车辆轨迹预测算法， 提出基于当前和历史的GNSS数据预测车辆下一个时刻的路径轨迹。
车辆在行驶过程中，大量的GNSS数据记录可以作为拟合车辆行驶轨迹的来源，大量车辆GNSS数据可以作为挖掘车辆行驶状态的基础；本发明通过提出了一种基于GNSS数据的速度、加速度、航向角、曲率以及急速变化轨迹置信区间来建立模型，判断车辆下一个时刻的车辆行驶轨迹，相对于现有技术算法更为简单，节约了大量运算资源。
进一步地，参见图2，所述半径获取模块包括采集单元11、转换单元12、过滤单元13和计算单元14；其中：
采集单元11，按照预设频率从GNSS动态数据中采集车辆实时参数；具体为数据采集器，按照设定的频率从GNSS动态数据中采集需要使用的参数数据；并将上述参数中的实时航向角数据发送给转换单元12，将实时速度数据发送给计算单元14，将经度、纬度及行驶的弧度数据发送给轨迹预测模块；
转换单元12，从采集单元中11获取车辆实时航向角、时间，由航向角通过计算获得车辆实时航向角变化率；从采集单元11中获取车辆实时航向角及时间，通过在程序中设定计算公式获得航向角变化率；并输入给过滤单元；
过滤单元13，对上述实时航向角变化率进行滤波处理，以过滤掉处于非连续状态的航向角变化率；采用二阶低通滤波器进行滤波处理；具体可以是一个滤波电路；
计算单元14，从采集单元11中获取车辆实时速度，根据上述处于连续状态的航向角变化率以及实时速度，获得车辆行驶路径半径；
置信率筛选模块包括置信率获取单元21、阈值单元22、筛选单元23，其中：采用二阶低通滤波器进行滤波处理；具体可以是一个滤波电路；
置信率获取单元21，对经过上述过滤单元过滤后的航向角变化率的差分值进行计算；并预存的根据航向角变化率的差分值与置信率区间之间的映射关系确定当前航向角变化率的差分值对应的置信率区间；
阈值单元22，选定置信率区间，根据选定置信率区间以及航向角变化率的差分值与置信率区间之间的映射关系的确定航向角变化率的差分值的预设阈值；
筛选单元23，比较航向角变化率的差分值与上述预设阀值的大小，如果航向角变化率的差分值小于上述预设阈值，如果置信率落在预设的置信率区 间内，则根据所述车辆实时经度、实时纬度、车辆行驶轨迹的弧度以及获得的所述车辆行驶半径获得下一时刻车辆行驶轨迹；否则，结束。具体地，如果航向角变化率的差分值小于上述预设阈值，则认为在选定的置信率区间范围内，则将上述车辆行驶路径半径作为半径信号向轨迹预测模块输入；如果航向角变化率的差分值大于或等于预设阈值，则认为不在置信率区间范围内；则不向所述轨迹获预测模块输入半径信号，结束本次数据预测；
轨迹预测模块3具体根据上述车辆实时经度、实时纬度、车辆行驶轨迹的弧度以及所述筛选单元中表征车辆行驶半径的半径信息获得下一时刻车辆行驶轨迹。接收来自于采集单元11的车辆实时经度、实时纬度、车辆行驶轨迹的弧度以及来自于筛选单元23的半径信息，通过对上述数据按照预设的公式进行计算，就能下一时刻车辆行驶轨迹。
每个车辆本身会在短距离通信技术的基础上周期性的发送本地车辆GNSS行驶轨迹预测信息并同时接收周围车辆GNSS行驶轨迹预测信息；本地车辆和远程车辆处于不同的位置，此时若发生安全事故，则需要本地车辆的GNSS预测轨迹和远程的GNSS的预测轨迹会在将来某一时刻重合，当将要发生安全事件时，在本地车辆或者远程车辆上产生警告。从而减少车辆安全事件如车辆追尾、恶劣天气环境下车辆碰撞等的发生。
以上所述仅为本发明的优选实施例，并非因此限制本发明的专利范围，凡是在本发明的发明构思下，利用本发明说明书及附图内容所作的等效结构变换，或直接/间接运用在其他相关的技术领域均包括在本发明的专利保护范围内。

Claims (15)

  1. 一种车辆实时轨迹预测方法，其特征在于，该方法包括以下步骤：
    S1，从GNSS动态数据中采集车辆实时参数，其中所述车辆实时参数包括车辆实时航向角、时间、实时速度、车辆实时经度、实时纬度及车辆行驶轨迹的弧度，根据所述航向角及时间，获得航向角变化率，由处于连续状态的所述航向角的变化率和所述实时速度，获得车辆行驶路径半径；
    S2，根据所述时间及实时航向角计算反映车辆行驶路径变化状态的置信率，并比较所述置信率与预设的置信率区间；
    S3，如果置信率落在预设的置信率区间内，则根据所述车辆实时经度、实时纬度、车辆行驶轨迹的弧度以及获得的所述车辆行驶半径获得下一时刻车辆行驶轨迹。
  2. 根据权利要求1所述的车辆实时轨迹预测方法，其特征在于，所述步骤S1包括以下步骤：
    S1a，按照预设频率从GNSS动态数据中采集车辆实时参数，其中所述车辆实时参数包括车辆实时航向角、时间、实时速度、车辆实时经度、实时纬度及车辆行驶轨迹的弧度；
    S1b，由航向角及时间通过计算获得车辆实时航向角变化率；
    S1c，对上述实时航向角变化率进行滤波处理，以过滤掉处于非连续状态的航向角变化率；
    S1d，根据上述处于连续状态的航向角变化率以及实时速度，获得车辆行驶路径半径。
  3. 根据权利要求2所述的车辆实时轨迹预测方法，其特征在于，所述步骤S2包括以下步骤：
    S2a，计算经过上述步骤S1c后的航向角变化率的差分值；
    S2b，选定置信率区间，并根据航向角变化率的差分值与置信率区间之间的映射关系确定航向角变化率的差分值的预设阈值；
    S2c，比较航向角变化率的差分值与上述预设阀值的大小，如果航向角变 化率的差分值小于上述预设阈值，则执行步骤S3；否则，结束。
  4. 根据权利要求2述的车辆实时轨迹预测方法，其特征在于，所述步骤S2b中，选定的置信率区间为50％～100％。
  5. 根据权利要求4所述的车辆实时轨迹预测方法，其特征在于，所述步骤S2包括以下步骤：
    S2a，计算经过上述步骤S1c后的航向角变化率的差分值；
    S2b，选定置信率区间，并根据航向角变化率的差分值与置信率区间之间的映射关系确定航向角变化率的差分值的预设阈值；
    S2c，比较航向角变化率的差分值与上述预设阀值的大小，如果航向角变化率的差分值小于上述预设阈值，则执行步骤S3；否则，结束。
  6. 根据权利要求1所述的车辆实时轨迹预测方法，其特征在于，所述步骤S2包括以下步骤：
    S2a，计算经过上述步骤S1c后的航向角变化率的差分值；
    S2b，选定置信率区间，并根据航向角变化率的差分值与置信率区间之间的映射关系确定航向角变化率的差分值的预设阈值；
    S2c，比较航向角变化率的差分值与上述预设阀值的大小，如果航向角变化率的差分值小于上述预设阈值，则执行步骤S3；否则，结束。
  7. 根据权利要求6述的车辆实时轨迹预测方法，其特征在于，所述步骤S2b中，选定的置信率区间为50％～100％。
  8. 一种车辆实时轨迹预测系统，其特征在于，该系统包括：
    半径获取模块，从GNSS动态数据中采集车辆实时参数，其中所述车辆实时参数包括车辆实时航向角、时间、实时速度、车辆实时经度、实时纬度及车辆行驶轨迹的弧度，根据所述航向角及时间，获得航向角变化率，由处于连续状态的所述航向角变化率和所述参数中的实时速度，获得车辆行驶路径半径；
    置信率筛选模块，根据所述车辆实时航向角及时间计算反映车辆行驶路径变化状态的置信率，并比较所述置信率与预设的置信率区间；
    轨迹预测模块，如果置信率落在置信率区间内，根据半径获取模块输入的所述车辆实时经度、实时纬度、车辆行驶轨迹的弧度以及获得的所述车辆行驶半径获得下一时刻车辆行驶轨迹。
  9. 根据权利要求8所述的车辆实时轨迹预测系统，其特征在于，所述半径获取模块包括：
    采集单元，按照预设频率从GNSS动态数据中采集车辆实时参数，其中所述车辆实时参数包括车辆实时航向角、时间、实时速度、车辆实时经度、实时纬度及车辆行驶轨迹的弧度；
    转换单元，由所述航向角及时间通过计算获得车辆实时航向角变化率；
    过滤单元，对上述实时航向角变化率进行滤波处理，以过滤掉处于非连续状态的航向角变化率；
    计算单元，根据上述处于连续状态的航向角变化率以及所述实时速度，获得车辆行驶路径半径。
  10. 根据权利要求9所述的车辆实时轨迹预测系统，其特征在于，所述置信率筛选模块包括：
    置信率获取单元，计算经过上述过滤单元过滤后的航向角变化率的差分值；并根据航向角变化率的差分值与置信率区间之间的映射关系确定当前航向角变化率的差分值对应的置信率区间；
    阈值模块，选定置信率区间，根据选定置信率区间以及航向角变化率的差分值与置信率区间之间的映射关系的确定航向角变化率的差分值的预设阈值；
    筛选单元，比较航向角变化率的差分值与上述预设阀值的大小，如果航向角变化率的差分值小于上述预设阈值，如果置信率落在预设的置信率区间内，则根据所述车辆实时经度、实时纬度、车辆行驶轨迹的弧度以及获得的所述车辆行驶半径获得下一时刻车辆行驶轨迹；否则，结束。
  11. 根据权利要求9述车的车辆实时轨迹预测系统，其特征在于，所述阈值模块中选定的置信率区间为50％～100％。
  12. 根据权利要求11所述的车辆实时轨迹预测系统，其特征在于，所述置信率筛选模块包括：
    置信率获取单元，计算经过上述过滤单元过滤后的航向角变化率的差分值；并根据航向角变化率的差分值与置信率区间之间的映射关系确定当前航向角变化率的差分值对应的置信率区间；
    阈值模块，选定置信率区间，根据选定置信率区间以及航向角变化率的差分值与置信率区间之间的映射关系的确定航向角变化率的差分值的预设阈值；
    筛选单元，比较航向角变化率的差分值与上述预设阀值的大小，如果航向角变化率的差分值小于上述预设阈值，如果置信率落在预设的置信率区间内，则根据所述车辆实时经度、实时纬度、车辆行驶轨迹的弧度以及获得的所述车辆行驶半径获得下一时刻车辆行驶轨迹；否则，结束。
  13. 根据权利要求8所述的车辆实时轨迹预测系统，其特征在于，所述置信率筛选模块包括：
    置信率获取单元，计算经过上述过滤单元过滤后的航向角变化率的差分值；并根据航向角变化率的差分值与置信率区间之间的映射关系确定当前航向角变化率的差分值对应的置信率区间；
    阈值模块，选定置信率区间，根据选定置信率区间以及航向角变化率的差分值与置信率区间之间的映射关系的确定航向角变化率的差分值的预设阈值；
    筛选单元，比较航向角变化率的差分值与上述预设阀值的大小，如果航向角变化率的差分值小于上述预设阈值，如果置信率落在预设的置信率区间内，则根据所述车辆实时经度、实时纬度、车辆行驶轨迹的弧度以及获得的所述车辆行驶半径获得下一时刻车辆行驶轨迹；否则，结束。
  14. 根据权利要求13述车的车辆实时轨迹预测系统，其特征在于，所述 阈值模块中选定的置信率区间为50％～100％。
  15. 根据权利要求9所述的车辆实时轨迹预测系统，其特征在于，所述过滤单元采用二阶低通滤波器进行滤波处理。
PCT/CN2017/073558 2016-10-20 2017-02-15 车辆实时轨迹预测方法及预测系统 WO2018072362A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201610920986.3A CN106384540B (zh) 2016-10-20 2016-10-20 车辆实时轨迹预测方法及预测系统
CN201610920986.3 2016-10-20

Publications (1)

Publication Number Publication Date
WO2018072362A1 true WO2018072362A1 (zh) 2018-04-26

Family

ID=57958124

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2017/073558 WO2018072362A1 (zh) 2016-10-20 2017-02-15 车辆实时轨迹预测方法及预测系统

Country Status (2)

Country Link
CN (1) CN106384540B (zh)
WO (1) WO2018072362A1 (zh)

Cited By (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111103566A (zh) * 2019-11-28 2020-05-05 中国科学院上海微系统与信息技术研究所 一种航向角确定方法、系统、电子设备及存储介质
CN112714020A (zh) * 2020-12-28 2021-04-27 青岛海尔科技有限公司 数据有效性的确定方法和装置、存储介质及电子装置
CN113723316A (zh) * 2021-09-01 2021-11-30 杭州智诚惠通科技有限公司 一种车辆识别方法、装置、设备及存储介质
CN113997941A (zh) * 2021-09-14 2022-02-01 武汉光庭信息技术股份有限公司 一种自车行驶状态分析方法、系统、电子设备及存储介质
CN114248761A (zh) * 2021-12-10 2022-03-29 重庆长安汽车股份有限公司 目标轨迹离散化方法、泊车路径跟踪误差优化方法及系统
CN115733867A (zh) * 2022-12-29 2023-03-03 阿维塔科技(重庆)有限公司 一种车载gps数据上传方法、系统及设备
CN116645828A (zh) * 2023-05-26 2023-08-25 赛马物联科技(宁夏)有限公司 一种物流车辆轨迹偏离报警方法及系统

Families Citing this family (16)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106384540B (zh) * 2016-10-20 2019-04-19 深圳市元征科技股份有限公司 车辆实时轨迹预测方法及预测系统
CN106875746B (zh) * 2017-02-27 2020-09-29 山东师范大学 一种汽车安全驾驶预警系统与方法
DE102017109731A1 (de) * 2017-05-05 2018-11-08 Konecranes Global Corporation Verfahren und System zum Betrieb eines automatisch geführten Transportfahrzeugs für Container
US10649458B2 (en) 2017-09-07 2020-05-12 Tusimple, Inc. Data-driven prediction-based system and method for trajectory planning of autonomous vehicles
CN108111173B (zh) * 2017-12-27 2021-03-26 东软集团股份有限公司 轨迹压缩方法、装置、存储介质及电子设备
CN109765886B (zh) * 2018-12-21 2022-05-24 北京智行者科技有限公司 车辆跟随的目标轨迹识别方法
CN109782754B (zh) * 2018-12-25 2022-05-06 东软睿驰汽车技术（沈阳）有限公司 一种车辆控制方法及装置
CN109922439B (zh) * 2019-03-29 2020-10-16 重庆长安汽车股份有限公司 多传感器数据的融合方法、基于v2x的目标检测方法及系统
CN112537296A (zh) * 2019-09-23 2021-03-23 北京新能源汽车股份有限公司 一种紧急制动装置、汽车和制动控制方法
CN111063210A (zh) * 2019-12-26 2020-04-24 广东星舆科技有限公司 一种基于高精度定位的车辆监控方法及设备
CN111693723B (zh) * 2020-05-29 2022-04-22 阿波罗智联(北京)科技有限公司 速度预测方法、装置和电子设备
CN111661061B (zh) * 2020-06-18 2022-02-11 东软睿驰汽车技术（沈阳）有限公司 一种车辆行驶意图识别方法及装置
CN111754605A (zh) * 2020-06-28 2020-10-09 深圳壹账通智能科技有限公司 用户轨迹绘制方法、装置、计算机设备及存储介质
CN112839855B (zh) * 2020-12-31 2022-07-12 华为技术有限公司 一种轨迹预测方法与装置
CN113191672B (zh) * 2021-05-20 2022-01-04 哈尔滨工业大学 基于空间数据和vrptw的村镇垃圾分类转运规划方法及系统
CN113247000B (zh) * 2021-07-16 2021-09-17 浙江所托瑞安科技集团有限公司 车辆转向识别方法、装置、设备和存储介质

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US7610146B2 (en) * 1997-10-22 2009-10-27 Intelligent Technologies International, Inc. Vehicle position determining system and method
CN102542634A (zh) * 2012-01-13 2012-07-04 北京理工大学 一种目标车辆行驶状态测量系统
CN102608641A (zh) * 2012-03-30 2012-07-25 江苏物联网研究发展中心 单轴陀螺仪和单轴加速度计的车载组合导航系统及方法
CN103339472A (zh) * 2010-12-03 2013-10-02 高通股份有限公司 用于gnss车辆导航的受惯性传感器辅助的航向和定位
CN104678415A (zh) * 2013-11-29 2015-06-03 现代摩比斯株式会社 车辆位置推测系统及其方法
CN105718750A (zh) * 2016-01-29 2016-06-29 长沙理工大学 一种车辆行驶轨迹的预测方法及系统
CN106384540A (zh) * 2016-10-20 2017-02-08 深圳市元征科技股份有限公司 车辆实时轨迹预测方法及预测系统

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US7610146B2 (en) * 1997-10-22 2009-10-27 Intelligent Technologies International, Inc. Vehicle position determining system and method
CN103339472A (zh) * 2010-12-03 2013-10-02 高通股份有限公司 用于gnss车辆导航的受惯性传感器辅助的航向和定位
CN102542634A (zh) * 2012-01-13 2012-07-04 北京理工大学 一种目标车辆行驶状态测量系统
CN102608641A (zh) * 2012-03-30 2012-07-25 江苏物联网研究发展中心 单轴陀螺仪和单轴加速度计的车载组合导航系统及方法
CN104678415A (zh) * 2013-11-29 2015-06-03 现代摩比斯株式会社 车辆位置推测系统及其方法
CN105718750A (zh) * 2016-01-29 2016-06-29 长沙理工大学 一种车辆行驶轨迹的预测方法及系统
CN106384540A (zh) * 2016-10-20 2017-02-08 深圳市元征科技股份有限公司 车辆实时轨迹预测方法及预测系统

Cited By (13)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111103566A (zh) * 2019-11-28 2020-05-05 中国科学院上海微系统与信息技术研究所 一种航向角确定方法、系统、电子设备及存储介质
CN111103566B (zh) * 2019-11-28 2023-09-05 中国科学院上海微系统与信息技术研究所 一种航向角确定方法、系统、电子设备及存储介质
CN112714020A (zh) * 2020-12-28 2021-04-27 青岛海尔科技有限公司 数据有效性的确定方法和装置、存储介质及电子装置
CN113723316A (zh) * 2021-09-01 2021-11-30 杭州智诚惠通科技有限公司 一种车辆识别方法、装置、设备及存储介质
CN113723316B (zh) * 2021-09-01 2024-04-16 杭州智诚惠通科技有限公司 一种车辆识别方法、装置、设备及存储介质
CN113997941A (zh) * 2021-09-14 2022-02-01 武汉光庭信息技术股份有限公司 一种自车行驶状态分析方法、系统、电子设备及存储介质
CN113997941B (zh) * 2021-09-14 2023-10-13 武汉光庭信息技术股份有限公司 一种自车行驶状态分析方法、系统、电子设备及存储介质
CN114248761A (zh) * 2021-12-10 2022-03-29 重庆长安汽车股份有限公司 目标轨迹离散化方法、泊车路径跟踪误差优化方法及系统
CN114248761B (zh) * 2021-12-10 2024-06-04 重庆长安汽车股份有限公司 目标轨迹离散化方法、泊车路径跟踪误差优化方法及系统
CN115733867B (zh) * 2022-12-29 2024-04-16 阿维塔科技(重庆)有限公司 一种车载gps数据上传方法、系统及设备
CN115733867A (zh) * 2022-12-29 2023-03-03 阿维塔科技(重庆)有限公司 一种车载gps数据上传方法、系统及设备
CN116645828B (zh) * 2023-05-26 2024-04-16 赛马物联科技(宁夏)有限公司 一种物流车辆轨迹偏离报警方法及系统
CN116645828A (zh) * 2023-05-26 2023-08-25 赛马物联科技(宁夏)有限公司 一种物流车辆轨迹偏离报警方法及系统

Also Published As

Publication number Publication date
CN106384540A (zh) 2017-02-08
CN106384540B (zh) 2019-04-19

Similar Documents

Publication Publication Date Title
WO2018072362A1 (zh) 车辆实时轨迹预测方法及预测系统
CN108919802B (zh) 无人驾驶的车辆行驶方法和装置
CN113879295B (zh) 轨迹预测方法及装置
Tao et al. Lane marking aided vehicle localization
KR102196827B1 (ko) 휠 회전의 광학 추적을 이용한 관측된 차량들의 상태 및 위치 예측
Song et al. Enhancing GPS with lane-level navigation to facilitate highway driving
WO2018072350A1 (zh) 车辆轨迹预测方法及装置
US10089867B2 (en) Anomalous travel location detection device and anomalous travel location detection method
EP3418997A1 (en) Detection and estimation of variable speed signs
CN103348395B (zh) 交通拥堵检测设备和车辆控制设备
EP3717996B1 (en) System and method for vehicle sensing and analysis
CN105675006B (zh) 一种道路偏离检测方法
Park et al. Vehicle positioning based on velocity and heading angle observer using low-cost sensor fusion
JPWO2018143237A1 (ja) 情報処理装置、サーバ装置、情報処理システム、情報処理方法、およびプログラム
US10532750B2 (en) Method, device and system for wrong-way driver detection
US10769942B2 (en) Method, device and system for wrong-way driver detection
CN113227831B (zh) 基于多传感器数据融合的护栏估计方法和车载设备
US11328596B2 (en) Parking prediction
JP2012108084A (ja) 交差点における進行方向予測装置、進行方向予測方法およびプログラム
CN114199267B (zh) 车辆的车道偏离预警测评方法、装置及系统
Ahmed et al. Minimizing GPS dependency for a vehicle’s trajectory identification by using data from smartphone inertial sensors and onboard diagnostics device
Yang et al. Optimal multi-sensor multi-vehicle (MSMV) localization and mobility tracking
Worrall et al. A probabilistic method for detecting impending vehicle interactions
US20180038696A1 (en) A system for use in a vehicle
CN111326002A (zh) 一种自动驾驶汽车环境感知的预测方法、装置及系统

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: 17861778

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 17861778

Country of ref document: EP

Kind code of ref document: A1