CN116198520B - Short-time prediction method, system and storable medium for driving behavior at ramp - Google Patents

Short-time prediction method, system and storable medium for driving behavior at ramp Download PDF

Info

Publication number
CN116198520B
CN116198520B CN202310163003.6A CN202310163003A CN116198520B CN 116198520 B CN116198520 B CN 116198520B CN 202310163003 A CN202310163003 A CN 202310163003A CN 116198520 B CN116198520 B CN 116198520B
Authority
CN
China
Prior art keywords
driving
driving behavior
intention
hmm
prediction
Prior art date
Legal status (The legal status 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 status listed.)
Active
Application number
CN202310163003.6A
Other languages
Chinese (zh)
Other versions
CN116198520A (en
Inventor
糜江
朱艳燕
胡强
马俊
彭理群
李明
王代君
徐周
邹瑶
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Jiangxi Intelligent Transportation Affairs Center
Jiangxi Provincial Transportation Comprehensive Administrative Law Enforcement Supervision And Administration Bureau
Jiangxi Academy Of Transportation Sciences Co ltd
East China Jiaotong University
Original Assignee
Jiangxi Intelligent Transportation Affairs Center
Jiangxi Provincial Transportation Comprehensive Administrative Law Enforcement Supervision And Administration Bureau
Jiangxi Academy Of Transportation Sciences Co ltd
East China Jiaotong University
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 Jiangxi Intelligent Transportation Affairs Center, Jiangxi Provincial Transportation Comprehensive Administrative Law Enforcement Supervision And Administration Bureau, Jiangxi Academy Of Transportation Sciences Co ltd, East China Jiaotong University filed Critical Jiangxi Intelligent Transportation Affairs Center
Priority to CN202310163003.6A priority Critical patent/CN116198520B/en
Publication of CN116198520A publication Critical patent/CN116198520A/en
Application granted granted Critical
Publication of CN116198520B publication Critical patent/CN116198520B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B60—VEHICLES IN GENERAL
    • B60W—CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
    • B60W50/00—Details of control systems for road vehicle drive control not related to the control of a particular sub-unit, e.g. process diagnostic or vehicle driver interfaces
    • B60W50/0097—Predicting future conditions
    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B60—VEHICLES IN GENERAL
    • B60W—CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
    • B60W40/00—Estimation or calculation of non-directly measurable driving parameters for road vehicle drive control systems not related to the control of a particular sub unit, e.g. by using mathematical models
    • B60W40/08—Estimation or calculation of non-directly measurable driving parameters for road vehicle drive control systems not related to the control of a particular sub unit, e.g. by using mathematical models related to drivers or passengers
    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B60—VEHICLES IN GENERAL
    • B60W—CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
    • B60W40/00—Estimation or calculation of non-directly measurable driving parameters for road vehicle drive control systems not related to the control of a particular sub unit, e.g. by using mathematical models
    • B60W40/10—Estimation or calculation of non-directly measurable driving parameters for road vehicle drive control systems not related to the control of a particular sub unit, e.g. by using mathematical models related to vehicle motion
    • B60W40/105—Speed
    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B60—VEHICLES IN GENERAL
    • B60W—CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
    • B60W2520/00—Input parameters relating to overall vehicle dynamics
    • B60W2520/10—Longitudinal speed
    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B60—VEHICLES IN GENERAL
    • B60W—CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
    • B60W2554/00—Input parameters relating to objects
    • B60W2554/40—Dynamic objects, e.g. animals, windblown objects
    • B60W2554/404—Characteristics
    • B60W2554/4041—Position

Landscapes

  • Engineering & Computer Science (AREA)
  • Automation & Control Theory (AREA)
  • Transportation (AREA)
  • Mechanical Engineering (AREA)
  • Physics & Mathematics (AREA)
  • Mathematical Physics (AREA)
  • Human Computer Interaction (AREA)
  • Traffic Control Systems (AREA)

Abstract

The invention discloses a short-time prediction method, a short-time prediction system and a storage medium for driving behaviors at a ramp, which relate to the technical field of prediction of the driving behaviors at the ramp, and firstly determine a predicted point of the driving behaviors; obtaining an observation state sequence formed by the driving behavior of the vehicle and a corresponding steering lamp use state sequence, and calculating by using a driving intention HMM recognition model to obtain a maximum probability hidden state sequence; counting the proportion values of hidden state nodes of different types to determine the driving intention represented by the observation state sequence; calculating a driving intention transition probability to determine a driving intention at a driving behavior prediction point; and calculating the probability of executing each driving behavior at the driving behavior prediction point, and taking the driving behavior corresponding to the maximum probability value as a prediction result. The invention can scientifically and accurately predict the driving behavior of the driver in the ramp scene.

Description

Short-time prediction method, system and storable medium for driving behavior at ramp
Technical Field
The invention relates to the technical field of driving behavior prediction at a ramp, in particular to a short-time driving behavior prediction method, a short-time driving behavior prediction system and a storage medium at the ramp.
Background
Research shows that improper behavior of a driver is a main cause of traffic accidents, and the driving behavior directly affects road traffic capacity and traffic safety. Therefore, the research is of great practical significance in identifying and predicting the driving behavior intention.
The mapping relationship between the driving intention of the driver in the ramp and the implementation of the driving behavior has uncertainty, for example, the action of 'stepping on the brake pedal' implemented by the driver at a certain moment can be implemented under the driving intention of straight driving or under the driving intention of turning. The existing driving behavior prediction method has low prediction precision, and particularly has complex prediction of the driving behavior in the ramp environment, the prediction precision is low, and the requirement cannot be met, so how to scientifically and accurately predict the driving behavior of a driver in the ramp environment is a problem which needs to be solved by the technicians in the field.
Disclosure of Invention
In view of the above, the invention provides a method, a system and a storable medium for predicting driving behavior at a ramp in short time.
In order to achieve the above object, the present invention provides the following technical solutions:
a short-time prediction method for driving behavior at a ramp comprises the following steps:
step 1, obtaining the distance between the current position of the vehicle and a front intersections n Current speed of vehiclevDistance is tos n Comparing with preset geographic scale, and based on the current speed of the vehiclevDetermining the geographic scale with the smallest distance in front of the vehicle as a driving behavior prediction points n+1 ；
Step 2, obtaining the driving behavior currently executed by the vehicleThe driving behavior performed +.>Together constitute the observation state sequence->，/>The method comprises the steps of carrying out a first treatment on the surface of the Acquisition of the observation State sequence->The use states of the steering lamps corresponding to each observation state node form a sequence flag of the use states of the steering lamps;
step 3, based on the steering lamp use state sequence flag and the observation state sequenceCalculating by using driving intention HMM recognition model to obtain maximum probability hidden state sequence +.>；
Step 4, hiding state sequence based on maximum probabilityColumn ofCounting the proportion value of the total number of hidden state nodes occupied by different types of hidden state nodes to determine an observation state sequence +.>The represented driving intention;
step 5, according to the observation state sequenceCalculating the driving intention transition probability to determine the driving behavior prediction point s n+1 Driving intention at;
step 6, predicting the point based on driving behaviors n+1 The driving intention at the position, calculating the predicted point of driving behaviors n+1 Probability of executing each driving behavior is carried out, and driving behavior corresponding to the maximum probability value is taken as a prediction objectI.e. driving behavior prediction pointss n+1 Prediction results of driving behavior at the location.
Optionally, the driving behavior includes stepping on a brake pedal B, stepping on a clutch pedal C, stepping on an accelerator pedal D, and switching a gear E, and the driving intention includes straight runningAnd turn->Wherein turn->Including left turn->And right turn->。
Optionally, an optionalIn the step 3, the maximum probability hidden state sequence is calculatedThe method of (1) is as follows:
step 3.1, based on the steering lamp use state sequence flag, observing a state sequenceDivided into 2 observation state subsequences, i.e. +.>；
Step 3.2, for the observation state subsequenceUsing a first driving intention HMM recognition model and a Viterbi algorithm to calculate a maximum probability hidden state sequence to obtain an observation state subsequence +.>Corresponding maximum probability hidden state subsequence +.>The method comprises the steps of carrying out a first treatment on the surface of the For observation state subsequence->Using a second driving intention HMM recognition model and a Viterbi algorithm to calculate the maximum probability hidden state sequence to obtain an observation state subsequence +.>Corresponding maximum probability hidden state subsequence +.>The method comprises the steps of carrying out a first treatment on the surface of the The first driving intention HMM recognition model represents a driving intention HMM recognition model before turning on a turn signal lamp, and the second driving intention HMM recognition model represents a driving intention HMM recognition model after turning on the turn signal lamp;
step 3.3, sub-sequence of observed statesCorresponding maximum probability hidden state subsequence +.>And observe state subsequencesCorresponding maximum probability hidden state subsequence +.>Combining to obtain the observation state sequence +.>Corresponding maximum probability hidden state sequence +.>。
Optionally, the method for constructing the driving intention HMM recognition model includes:
based on the turning-on state of the turning lamp, two Hidden Markov Model (HMM) network models are established, wherein the HMM network models are a first HMM network model corresponding to the first HMM network model before turning on the turning lamp and a second HMM network model corresponding to the second HMM network model after turning on the turning lamp; the hidden state layer of the HMM network model of the first hidden Markov model is the driving intention state held by the driver before turning on the steering lampAnd->The hidden state layer of the HMM network model of the second hidden Markov model is the driving intention state +.>And->The method comprises the steps of carrying out a first treatment on the surface of the The observation state layer nodes of the HMM network model of the two hidden Markov models are implemented for the driverB, C, D, E of the driving behavior of (a);
expanding the first hidden Markov model HMM network model and the second hidden Markov model HMM network model according to the geographic position of the vehicle, and taking a preset geographic scale mark as an expansion point during expansion;
calculating a driving intention transition probability function of each unfolded point after unfolding, and counting driving behavior intention mapping probabilities of each unfolded point after unfolding to obtain a final first driving intention HMM recognition model and a second driving intention HMM recognition model, wherein the first driving intention HMM recognition model and the second driving intention HMM recognition model jointly form a driving intention HMM recognition model.
Optionally, the time required for the vehicle to travel from the current position to the position on the front geographic scale is taken as the predicted time lengthWith the predicted duration +.>And predicting the driving behavior of the ramp vehicle for the time interval.
Optionally, the current position of the vehicle is obtained through a vehicle-mounted GPS or a geographic information electronic tag buried on the ground, and the current speed of the vehicle is obtained through a vehicle-mounted information acquisition systemvDriving behavior and turn signal usage status.
A short-term prediction system of driving behavior at a ramp, comprising:
the driving behavior prediction point determining module is used for obtaining the distance between the current position of the vehicle and the front intersections n Current speed of vehiclevDistance is tos n Comparing with preset geographic scale, and based on the current speed of the vehiclevDetermining the geographic scale with the smallest distance in front of the vehicle as a driving behavior prediction points n+1 ；
The driving information acquisition module is used for acquiring the driving behavior currently executed by the vehicleExecuted driving behaviorTogether constitute the observation state sequence->，/>The method comprises the steps of carrying out a first treatment on the surface of the Acquisition of the observation State sequence->The use states of the steering lamps corresponding to each observation state node form a sequence flag of the use states of the steering lamps;
the maximum probability hidden state sequence calculation module is used for calculating a state sequence flag and an observation state sequence based on the steering lamp use state sequenceCalculating by using driving intention HMM recognition model to obtain maximum probability hidden state sequence +.>；
An observation state sequence driving intention judging module for hiding the state sequence based on the maximum probabilityCounting the proportion value of the total number of hidden state nodes occupied by different types of hidden state nodes to determine an observation state sequence +.>The represented driving intention;
the prediction point driving intention judging module is used for judging the driving intention according to the observation state sequenceCalculating the driving intention transition probability to determine the predicted point of driving behaviors n+1 Driving intention at;
predictive point drivingA driving behavior prediction module for predicting a point based on driving behaviors n+1 The driving intention at the position, calculating the predicted point of driving behaviors n+1 Probability of executing each driving behavior is carried out, and driving behavior corresponding to the maximum probability value is taken as a prediction objectI.e. driving behavior prediction pointss n+1 Prediction results of driving behavior at the location.
A computer readable storage medium having stored thereon a computer program which when executed by a processor implements a method of short-term prediction of driving behaviour at a ramp as claimed in any one of the preceding claims.
According to the technical scheme, the invention discloses a short-time prediction method, a short-time prediction system and a storage medium for driving behaviors at a ramp, and has the following beneficial effects compared with the prior art:
the method comprises the steps of respectively constructing a first Hidden Markov Model (HMM) network model and a second Hidden Markov Model (HMM) network model by taking the state of a steering lamp as a demarcation point, and expanding according to a preset geographic scale on the basis to obtain a first driving intention HMM recognition model and a second driving intention HMM recognition model which are used for calculating different observation state sequence subsequences and comprehensively obtaining a maximum probability hidden state sequence; determining the driving intention represented by the observation state sequence based on the maximum probability hidden state sequence, and calculating the driving intention transition probability to determine the driving intention at the driving behavior prediction point; and further calculating the probability of executing each driving behavior at the driving behavior prediction point, and finally predicting the driving behavior according to the driving behavior corresponding to the maximum probability value. According to the technical scheme, factors such as different lengths of the driving roads and mutual influences among vehicles are considered, and the driving behavior of a driver in a ramp scene can be predicted scientifically and accurately.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings that are required to be used in the embodiments or the description of the prior art will be briefly described below, and it is obvious that the drawings in the following description are only embodiments of the present invention, and that other drawings can be obtained according to the provided drawings without inventive effort for a person skilled in the art.
FIG. 1 is a schematic diagram of a graphical model of intersection driving behavior and intent;
FIG. 2 is a schematic illustration of a driving intent HMM recognition model deployed by vehicle geographic location;
FIG. 3 is a schematic diagram of a predictive process of driving intent before use of turn signals;
FIG. 4 is a schematic diagram of a driving intent prediction process after use of turn signals;
FIG. 5 is a step diagram of a method for predicting driving behavior at a ramp in short time;
fig. 6 is a schematic diagram of a system for predicting driving behavior at a ramp in a short time according to the present invention.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
The embodiment of the invention discloses a short-time prediction method for driving behavior at a ramp, which is shown in fig. 5 and comprises the following steps:
step 1, obtaining the distance between the current position of the vehicle and a front intersections n Current speed of vehiclevDistance is tos n Comparing with preset geographic scale, and based on the current speed of the vehiclevDetermining the geographic scale with the smallest distance in front of the vehicle as a driving behavior prediction points n+1 。
In the specific implementation process, the system can be electrically powered by a vehicle-mounted GPS or geographic information buried on the groundThe sub-tag acquires the current position of the vehicle, and acquires the current speed of the vehicle through the vehicle-mounted information acquisition systemvAnd subsequent driving behavior and turn signal usage status, etc.
Step 2, obtaining the driving behavior currently executed by the vehicleThe driving behavior performed +.>Together constitute the observation state sequence->，/>The method comprises the steps of carrying out a first treatment on the surface of the Acquisition of the observation State sequence->The value of the steering lamp use state corresponding to each observation state node is 0 when a vehicle enters a geographic position scale coverage area in an intersection, and the value of the steering lamp use state is 1 when an information acquisition system captures that a driver turns on the steering lamp, so as to form a steering lamp use state sequence flag;
in the present embodiment, the driving behavior includes stepping on the foot brake pedal B, stepping on the clutch pedal C, stepping on the accelerator pedal D, and shifting the gear E, and the driving intention includes straight runningAnd turn->Wherein turn->Including left turn->And right turn->。
Step 3, based on the steering lamp use state sequence flag and the observation state sequenceCalculating by using driving intention HMM recognition model to obtain maximum probability hidden state sequence +.>。
The construction method of the driving intention HMM recognition model comprises the following steps:
1. construction of graph model
The relationship between the driving behavior and the intention of the driver implemented in the ramp can be directly described by using the graph model, and in order to facilitate the graph model expression, the observation probabilities between the 3 driving intention types and the transition probabilities thereof in the ramp and the characteristic driving behavior and the intention (neglecting the proportional influence of the "using the steering lamp" on the other driving behaviors) are defined as shown in table 1.
Table 1 graph model node and relationship probability definition table
Wherein, the liquid crystal display device comprises a liquid crystal display device,(K epsilon {3,4} or K epsilon { L, R }) is the driving intention state (type) held by the driver in the whole intersection, and before turning on the turn signal, the state can only be "straight driving intention->"AND" turning Driving intention>"1 of the following; after turning on the steering lamp, the driving intention state of the driving behavior map implemented by the driver is changed from the option of turning or not to the option of turning in what direction, and at this time, the driving intention state can only be left turning driving intention->"and" right turn driving intention>"1 of the above. 4 driving actions of "pedal brake pedal B", "pedal clutch pedal C", "pedal accelerator pedal D", "shift gear E(KE {3,4}, orK∈{L,R' s) to>(KE {3,4}, orK∈{L,R};Y∈{B,C,D,E-or) indicate->The specific value can be obtained by counting data collected by a real vehicle driving test.
Based on the turning-on state of the turn signal, two Hidden Markov Model (HMM) network models are established, see FIG. 1, respectively a first HMM network model (left side) corresponding to the first HMM before turning on the turn signal and a second HMM network model (right side) corresponding to the second HMM after turning on the turn signal; the hidden state layer of the HMM network model of the first hidden Markov model is the driving intention state held by the driver before turning on the steering lampAnd->The hidden state layer of the HMM network model of the second hidden Markov model is the driving intention state +.>And->The method comprises the steps of carrying out a first treatment on the surface of the HMM network model of two hidden Markov modelsThe observation state layer nodes are all driving behaviors B, C, D, E executed by the driver.
2. Model improvement
Because the road length of the vehicle is different and the vehicle speed is easy to be influenced by the peripheral traffic flow when the actual vehicle driving test is developed, the first hidden Markov model HMM network model and the second hidden Markov model HMM network model are developed according to the geographic position of the vehicle, the preset geographic scale is used as the development point when the vehicle is developed, meanwhile, the separation effect of using the steering lamp in the road section (or the moment of using the steering lamp is considered to be positioned at the two ends of the road section) is ignored, at the moment, the 2 hidden Markov model HMM network models are required to be developed according to the whole intersection length, and the driving behavior is realizedY(Y∈{B,C,D,E}) and driving intents(KE {3,4}, orK∈{L,RAnd }) carrying out statistics according to the mapping relation between the real vehicle driving test data. The expanded driving intention HMM recognition model is shown in FIG. 2, including the first driving intention HMM recognition model +.>And a second driving intention HMM recognition model +.>。
In this embodiment, geographical position scale marks "80, 60, 40, 20, 0" are set as expansion points during actual driving test, and statistics are performed for driving behavior types implemented at each expansion point for 3 driving intentions of straight, left-turn and right-turn, so as to obtain driving behavior when the steering lamp is not used (or the position of the steering lamp is considered to be infinitely close to the intersection) in all the sectionsY(Y∈{B,C,D,E}) and driving intents(KE {3,4 }) mapping probability value ∈>、/>、/>、/>(N=1, 2,3,4, 5), driving behavior when all segments are "using turn lights (or considered" using turn lights "located infinitely close to segment entrance)"YIs +.>(K∈{L,RProbability value of mapping relation ∈0->、/>、/>、/>The variables mentioned above will be substituted->(K∈{3,4,L,R};Y∈{B,C,D,E}). Driving behavior acquired at geographic position scaleY(Y∈{B,C,D,EAccording to })、/>、/>、/>Or->、/>、/>、/>To calculate->(K∈{3,4,L,R}) is performed depending on whether the driving behavior "use turn signal" has been performed.
Types of driving intents at each scale point(N=1,2,3,4,5;K∈{3,4,L,RProbability values of mutual transitions +.>(N=1, 2,3, 4), wherein +.>Indicating driving intention transition probability when driver does not turn on steering lamp, < >>The driving intention transition probability after the driver turns on the turn signal is indicated.
3. Model parameters
(1) Calculating driving intention transition probability function of each unfolding point after unfolding
Driving intent at different geographic location scale scales(N=1,2,3,4,5;K∈{3,4,L,R}) to self or other driving intents ∈>(/>Time->,/>Time->；/>Time->,/>Time->) The probability value of the occurrence of the transition isq N Andq' N what transition probability value is taken at the scale depends on whether the driving behaviour "using turn signal" has been performed. Combining the geographical position information of the vehicle when the related steering lamp is started in the driving test data of the real vehicle, the method can knowq N The variation over the full segment has the following law: the position where the driver turns on the turn signal is concentrated at 40m from the intersection, and is set to be corresponding to the driving intention +.>To->Transition probability->Maximum point (++)>Minimum point), set ∈>The value of this scale is 0.66, the other +.>、/>(N=1, 2,4, 5) the content-acceptable value is analyzed according to the driving intention transition probability described above. When a driver enters the intersection, the driver has driving intention +.>(turning on no turn signal), driving intention ++as the vehicle is located gradually closer to the intersection>To->Transition probability approach->Maximum point (++)>Minimum point), set ∈>The value of this scale is 0.9, the other +.>、/>(N=1, 2,3, 4) the analysis content acceptable values according to the aforementioned driving intention transition probability are shown in table 2.
TABLE 2q N Value table at geographical position scale
Using Gauss Amp function pairsNonlinear fitting is carried out along with the change rule of the scale marks of the geographic position to obtain +.>The expression is
(1)
Corresponding toTo->Transition probability of->Then use the formula->Can be obtained by the expression of
(2)
In the formulas (1) and (2)sFor the distance between the position of the vehicle and the intersection, the vehicle issSubstituting into (1) and (2) can determine all positions in the road section including scale marks 80, 60, 40, 20 and 0、/>Values.
Using Exp Dec1 function pairsScale change gauge along with geographical position scaleThe law is fitted non-linearly to obtain +.>The expression is
(3)
Corresponding toTo->Transition probability of->Then use the formula->Can be obtained by the expression of
(4)
In the same way, willsSubstituting into (3) and (4) can determine the positions of intersections including scale marks 80, 60, 40, 20 and 0、/>Values.
(2) Statistics of driving behavior intention mapping probability of each unfolding point after unfolding
Driving behavior for full-segment "no turn light" timeY(Y∈{B,C,D,E}) and driving intents(KE {3,4 }) mapping probability value ∈>、/>、/>、/>(N=1, 2,3,4, 5), the selected statistical samples should include an equivalent number of straight driving data and turning driving data. In the actual driving test process, the tested driver must turn on the steering lamp before turning; therefore, the test data of the steering lamp position closest to the intersection is selected from all the turning behavior data as a statistical sample, the driver uses the steering lamp position closest to the intersection (the scale is 20 m) to turn 27 times (the left and right turns are regarded as effective) in total, the 27 times of test data are used as turning behavior data parts in the statistical sample, 27 times of test data are randomly extracted from all the straight behavior data to be used as straight behavior data parts in the statistical sample, and 2 parts of test data are synthesized to count driving behaviorsY(Y∈{B,C,D,E}) and driving intention->(KE {3,4 })) mapping probability values、/>、/>、(N=1, 2,3, 4) is shown in table 3. In practical application, the probability value of the mapping relation between adjacent scale marks can be obtained through linear function fitting.
TABLE 3 Table 3、/>、/>、/>Value table at geographical position scale
Driving behavior after "use of turn signal" for full segmentY(Y∈{B,C,D,E}) and driving intents(K∈{L,RProbability value of mapping relation ∈0->、/>、/>、/>(N=1, 2,3,4, 5), the chosen statistical samples should also include an equivalent number of straight driving data and turning driving data. And selecting test data of the 'using the steering lamp position closest to the entrance of the section' from all turning behavior data as statistical samples, wherein the total number of turning behaviors of a driver using the steering lamp position closest to the entrance of the section (scale mark is 80 m) is 8, and the number of samples is small. Therefore, the turning behavior data at 60m of the ruler scale is added and received to 40 times of test data in total to be used as a turning behavior data part in a statistical sample, 40 times of test data are randomly extracted from all the straight behavior data to be used as a straight behavior data part in the statistical sample, and the driving behavior is counted by combining 2 parts of test dataY(Y∈{B,C,D,E}) and driving intention->(K∈{L,RProbability value of mapping relation ∈0->、/>、/>、/>(N=2, 3,4, 5) is shown in table 4.
TABLE 4 Table 4、/>、/>、/>Value table at geographical position scale
(3) Hidden state transition matrix
Describing driving intent using 2 different hidden state transition matrices by driving behavior and intent HMM deployed in vehicle geographic locationIs a transfer relationship of (a). Before the driver uses the turn signal, use the matrix +.>Description of driving intention->、/>Transfer relationship between->The element of->. Intersection driving intention HMM hidden state transition matrix +.>Expressed as follows
(5)
After the driver uses the turn signal, the matrix is usedDescription of driving intention->、/>Transfer relation between->The element of->. For "turn signal on" vehicles, pedestrians identify the driver as having +.>Has a high possibility of having +.>To->Metastasis, or->To->The transfer judgment mode is simpler, so in the embodimentThe value is 0.5. Intersection driving intention HMM hidden state transition matrix +.>Expressed as follows
(6)
Corresponding initial hidden state distribution matrixCan be defined as +.>。
(4) Observation matrix
And hidden state transition matrixAnd->Correspondingly, driving behaviors developed according to geographic positions of vehicles and driving behaviors corresponding to intended HMMs by using 2 different observation state matrixesY(Y∈{B,C,D,E}) and driving intention->Is a mapping relation of (a) to (b). Before the driver uses the turn signal, use the matrix +.>Description of the inventionYAnd driving intentionFigure->、/>Mapping probability between>The elements of the list can be obtained by looking up a table. Intersection driving intention HMM hidden state transition matrix +.>The expression is as follows:
(7)
after the driver uses the turn signal, the matrix is usedDescription of the inventionYIs +.>、/>The probability of the mapping between them,the medium elements may be obtained by querying. Intersection driving intention HMM hidden state transition matrix +.>The expression is as follows:
(8)
in practical application, the probability value of the mapping relation between adjacent scale marksOr->The method can be obtained by fitting a linear function and substituting the geographical position of the vehicle (represented by the distance between the vehicle and the intersection) into the linear function for solving.
After obtaining the driving intention HMM recognition model, calculating by using the driving intention HMM recognition model to obtain the maximum probability hidden state sequenceThe specific steps of (a) are as follows:
step 3.1, based on the steering lamp use state sequence flag, observing a state sequenceDivided into 2 observation state subsequences, i.e. +.>；
Step 3.2, for the observation state subsequenceUsing a first driving intention HMM recognition model and a Viterbi algorithm to calculate a maximum probability hidden state sequence to obtain an observation state subsequence +.>Corresponding maximum probability hidden state subsequence +.>，，/>Last 1 driving behaviour before turning on the turn signal for the driver +.>Corresponding maximum probability hidden state,/>The method comprises the steps of carrying out a first treatment on the surface of the For observation state subsequence->Using a second driving intention HMM recognition model and a Viterbi algorithm to calculate the maximum probability hidden state sequence to obtain an observation state subsequence +.>Corresponding maximum probability hidden state subsequence +.>，/>，/>First driving behavior after turning on the turn signal for the driver>Corresponding maximum probability hidden state,/>The method comprises the steps of carrying out a first treatment on the surface of the The first driving intention HMM recognition model represents a driving intention HMM recognition model before turning on a turn signal lamp, and the second driving intention HMM recognition model represents a driving intention HMM recognition model after turning on the turn signal lamp;
step 3.3, sub-sequence of observed statesCorresponding maximum probability hidden state subsequence +.>And observe state subsequencesCorresponding maximum probability hidden state subsequence +.>Sequentially and smoothlySequence combination, obtaining the observation state sequence +.>Corresponding maximum probability hidden state sequence +.>，/>。
Step 4, if the sequence flag is all 0, then from(equivalent +.>) To implement for driverThe driving intention type held at the time is determined, the intention type is +.>Or->The method comprises the steps of carrying out a first treatment on the surface of the If the sequence flag contains a value of 1, then the sub-sequence +.>Carry out +.>The driving intention type held at the time is determined, the intention type is +.>Or->. Counting the number of driving intention nodes (the number of hidden state nodes) of different types; calculating the total number of nodes of various driving intentionsnProportional value of +.>(iFor the number of driving intention types, at the intersection, no matter what the sequence flag is,i2) to determine the entire driving behavior sequence +.>Represented driving intention.
Step 5, according to the observation state sequenceCalculating the driving intention transition probability to determine the driving behavior prediction point s n+1 Driving intention at.
Step 6, predicting the point based on driving behaviors n+1 The driving intention at the position, calculating the predicted point of driving behaviors n+1 Probability of executing each driving behavior is carried out, and driving behavior corresponding to the maximum probability value is taken as a prediction objectI.e. driving behavior prediction pointss n+1 Prediction results of driving behavior at the location.
Specifically, it is provided withFor the predicted driving behavior object, its execution is the predicted point +.>. If the driving behavior sequence->The type of driving intention represented is +.>Will->Substitution to solve for +.>And->And comparing if->It is considered that the driving intention is not shifted, i.e. +.>. According to->Deriving that the driver is holding->On the premise of (1) the most likely driving behavior to be performed, i.e. +.>、/>、/>、/>The driving behavior type corresponding to the maximum value in the range is a prediction objectIn particular forms of (2). If the driving behavior sequence->The type of driving intention represented is +.>Follow-up prediction +.>The method for which driving behavior type is the same as described above. If the driving behavior sequence->The type of driving intention represented is +.>And->In 1, it is considered that the driving intention of the driver to turn in a certain direction is not shifted, and +.>、、/>、/>The driving behavior type corresponding to the maximum value in the range is a predicted object +.>In particular forms of (2).
Regarding the prediction duration, when the driving behavior action of the driver is predicted in the intersection, the information acquisition system compares the geographic position scales, and extracts the driving behavior action information corresponding to the geographic position scales of the driver. In predicting the durationIn terms of arrangement, since the driving behavior information implemented at the scale mark positions "80, 60, 40, 20, 0" is valid, the time required for the vehicle to travel from the current geographical position to the front scale mark is the predicted time length +.>This value is related to the current speed of the vehicle +.>And (5) correlation. During the vehicle advancing process, the system needs to judge whether a driver uses a steering lamp or not; if the driverExecuting the driving behavior, the driving behavior and the intention HMM model parameters used by the driving behavior prediction model are selected from +.>Turn to->In the prediction duration +.>On the premise of no change, recalculating the time from the moment when the driver turns on the turn signal to the position of the next scale mark +.>The method comprises the steps of carrying out a first treatment on the surface of the After this driving behavior prediction is completed, the system can resume the use +.>As a set period of time for the next driving behavior prediction.
The embodiment of the invention also provides a short-time prediction system for driving behavior at a ramp, referring to fig. 6, which comprises:
the driving behavior prediction point determining module is used for obtaining the distance between the current position of the vehicle and the front intersections n Current speed of vehiclevDistance is tos n Comparing with preset geographic scale, and based on the current speed of the vehiclevDetermining the geographic scale with the smallest distance in front of the vehicle as a driving behavior prediction points n+1 ；
The driving information acquisition module is used for acquiring the driving behavior currently executed by the vehicleExecuted driving behaviorTogether constitute the observation state sequence->，/>The method comprises the steps of carrying out a first treatment on the surface of the Acquisition of the observation State sequence->The use states of the steering lamps corresponding to each observation state node form a sequence flag of the use states of the steering lamps;
the maximum probability hidden state sequence calculation module is used for calculating a state sequence flag and an observation state sequence based on the steering lamp use state sequenceCalculating by using driving intention HMM recognition model to obtain maximum probability hidden state sequence +.>；
An observation state sequence driving intention judging module for hiding the state sequence based on the maximum probabilityCounting the proportion value of the total number of hidden state nodes occupied by different types of hidden state nodes to determine an observation state sequence +.>The represented driving intention;
the prediction point driving intention judging module is used for judging the driving intention according to the observation state sequenceCalculating the driving intention transition probability to determine the predicted point of driving behaviors n+1 Driving intention at;
the prediction point driving behavior prediction module is used for predicting a point based on driving behaviors n+1 The driving intention at the position, calculating the predicted point of driving behaviors n+1 Probability of executing each driving behavior is carried out, and driving behavior corresponding to the maximum probability value is taken as a prediction objectI.e. driving behavior prediction pointss n+1 Prediction results of driving behavior at the location.
For the system module disclosed in the embodiment, since the system module corresponds to the method disclosed in the embodiment, the description is simpler, and the relevant points refer to the description of the method section.
Further, the present invention also provides a computer readable storage medium, on which a computer program is stored, which when executed by a processor implements a method for short-term prediction of driving behavior at a ramp as described in any one of the above.
For the technical solution of the present invention, specific examples are listed below for illustration:
and randomly extracting certain driving behavior data in the experiment to be used as a test sample for predicting the driving behavior of the intersection. Information data such as various driving behaviors and vehicle speed of the tested person are automatically acquired by a vehicle-mounted information acquisition system, and the geographic position information of the tested person 'stepping on a foot brake pedal (observing a tail brake lamp)' and 'using a steering lamp' is recorded by an outside-vehicle experimenter in combination with a geographic position scale. Comparing and analyzing 2 groups of driving behavior data, screening 9 driving behaviors related to driving intention types in intersections, and sequentially stepping on an accelerator pedal according to the execution sequence of actions, wherein the 9 driving behaviors are sequentiallyD(80 m) ", step on foot brake pedalB(75 m) ", step on foot brake pedalB(70 m) ", step on foot brake pedalB(63 m) "," pedal accelerator pedalD(60 m) "," use turn signal lamp (45 m) "," tread on clutch pedal "C(40 m) "," shift gearE(35 m) ", step on foot brake pedalB(27 m) ", the distance between the vehicle and the intersection ahead when the driving behavior is to be performed is shown in brackets, and 8 driving behaviors usable for the prediction of the driving behavior of the intersection are taken out of them. Setting a driving behavior sequence composed of the 8 driving behaviors executed by the driverLet's consider according to the known conditions>Middle->(/>) In the "use turn signal" front and rear situation (according to +.>Value judgment) is known->Comprising 2 subsequences->Wherein->A Viterbi algorithm and a first driving intention HMM recognition model pair are used for +.>And carrying out driving intention recognition. The first driving intention HMM recognition model parameter cases are shown in table 5; and is opposite toThe model parameters to be used are +.>Is a second driving intention HMM recognition model. />And->Driving behavior +.>The corresponding model parameters are shown in table 5.
TABLE 5 HMM model parameter Table for Driving behaviors
Will be shown in Table 5Substituting corresponding model parameters into Viterbi algorithm to solve +.>Most likely corresponding driving intention subsequence->，/>The solution process is shown in fig. 3. At the position of the scale mark 80m, the driver steps on the accelerator pedal) When the driver is estimated to have the straight driving intention according to the prediction model>Is greater than the turning driving intentionBut as the vehicle position gets closer to the intersection the driver performs +.>、/>、/>、/>From the prediction model, it is estimated that the driver has turning driving intention +.>The probability of (a) is greater than straight driving intention +.>。
When the driver does not use the steering lamp, the predictive model is based on the driving behavior sequence executed by the driverThe variation of the type of driving intention held by the driver is estimated as indicated by the direction of the thick solid arrow in FIG. 3 to obtain +.>Corresponding most likely corresponding driving intention subsequence +.>Is->The method comprises the steps of carrying out a first treatment on the surface of the Wherein turning driving intention->Is greater than the number of straight driving intentions>Number, according to statistics, ->The type of driving intention representing the driver at this stage is turning driving intention +.>And then the driver turns on the steering lamp at the position of the scale mark 45m, which indicates that the established driving behavior prediction model is effective for the driving intention recognition and behavior and action prediction when the driver does not turn on the steering lamp in the intersection.
Since the driving behavior to be executed includes "using a turn signal", the sequence is based on the executed driving behaviorFor the driverIs predicted from the driving behavior of (a) only according to the subsequence thereof +.>And (3) obtaining the product. Let in Table 5->Substituting corresponding model parameters into Viterbi algorithm to solve +.>Most likely corresponding driving intention subsequence->，/>The solution process is shown in fig. 4.
After the driver uses the steering lamp, the predictive model is based on the sequence of driving actions performed by the driverThe variation of the type of driving intention held by the driver is estimated as indicated by the direction of the thick solid arrow in FIG. 4 to obtain +.>Corresponding most likely corresponding driving intention subsequence +.>Is->The method comprises the steps of carrying out a first treatment on the surface of the Wherein right turn driving intention->Is greater than the number of straight driving intentions>Number, and finally driving behaviour +.>The corresponding driving intention type is +.>According to statistics, ->The type of driving intention representing the driver at this stage is turning driving intention +.>. After recognizing that the type of driving intention already held by the driver is "right turn intention", it is known from the look-up table 5 that the driving behavior most likely to be performed by the driver at the distance intersection 20m is still "step on the brake pedalB”(/>As the highest value of the same ratio).
In the present specification, each embodiment is described in a progressive manner, and each embodiment is mainly described in a different point from other embodiments, and identical and similar parts between the embodiments are all enough to refer to each other.
The previous description of the disclosed embodiments is provided to enable any person skilled in the art to make or use the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the invention. Thus, the present invention is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims (8)

1. The short-time prediction method for the driving behavior at the ramp is characterized by comprising the following steps of:
step 1, obtaining the distance between the current position of the vehicle and a front intersections n Current speed of vehiclevDistance is tos n Comparing with preset geographic scale, and based on the current speed of the vehiclevDetermining a geographic region where a distance ahead of the vehicle is minimalThe scale mark is used as a driving behavior prediction points n+1 ；
Step 2, obtaining the driving behavior currently executed by the vehicleThe driving behavior performed +.>Together constitute the observation state sequence->，/>The method comprises the steps of carrying out a first treatment on the surface of the Acquisition of the observation State sequence->The use states of the steering lamps corresponding to each observation state node form a sequence flag of the use states of the steering lamps;
step 3, based on the steering lamp use state sequence flag and the observation state sequenceCalculating by using driving intention HMM recognition model to obtain maximum probability hidden state sequence +.>；
Step 4, based on the maximum probability hidden state sequenceCounting the proportion value of the total number of hidden state nodes occupied by different types of hidden state nodes to determine an observation state sequence +.>The represented driving intention;
step 5, according to the observation state sequenceCalculating the driving intention transition probability to determine the driving behavior prediction point s n+1 Driving intention at;
step 6, predicting the point based on driving behaviors n+1 The driving intention at the position, calculating the predicted point of driving behaviors n+1 Probability of executing each driving behavior is carried out, and driving behavior corresponding to the maximum probability value is taken as a prediction objectI.e. driving behavior prediction pointss n+1 Prediction results of driving behavior at the location.
2. The short-term prediction method of driving behavior at a ramp according to claim 1, wherein the driving behavior includes a step on a foot brake pedal B, a step on a clutch pedal C, a step on an accelerator pedal D, a shift range E, and the driving intention includes straight runningAnd turn->Wherein turn->Including left turn->And right turn->。
3. The method for predicting the driving behavior in a short time at a ramp according to claim 1, wherein in the step 3, a maximum probability hidden state sequence is calculatedThe method of (1) is as follows:
step 3.1, based on the steering lamp use state sequence flag, observing a state sequenceDivided into 2 observation state subsequences, i.e. +.>；
Step 3.2, for the observation state subsequenceUsing a first driving intention HMM recognition model and a Viterbi algorithm to calculate a maximum probability hidden state sequence to obtain an observation state subsequence +.>Corresponding maximum probability hidden state subsequence +.>The method comprises the steps of carrying out a first treatment on the surface of the For observation state subsequence->Using a second driving intention HMM recognition model and a Viterbi algorithm to calculate the maximum probability hidden state sequence to obtain an observation state subsequence +.>Corresponding maximum probability hidden state subsequence +.>The method comprises the steps of carrying out a first treatment on the surface of the The first driving intention HMM recognition model represents a driving intention HMM recognition model before turning on a turn signal lamp, and the second driving intention HMM recognition model represents a driving intention HMM recognition model after turning on the turn signal lamp;
step 3.3, sub-sequence of observed statesCorresponding maximum probability hidden state subsequence +.>And observation state subsequence->Corresponding maximum probability hidden state subsequence +.>Combining to obtain the observation state sequence +.>Corresponding maximum probability hidden state sequence +.>。
4. The short-term prediction method of driving behavior at a ramp according to claim 2, wherein the construction method of the driving intention HMM recognition model is as follows:
based on the turning-on state of the turning lamp, two Hidden Markov Model (HMM) network models are established, wherein the HMM network models are a first HMM network model corresponding to the first HMM network model before turning on the turning lamp and a second HMM network model corresponding to the second HMM network model after turning on the turning lamp; the hidden state layer of the HMM network model of the first hidden Markov model is the driving intention state held by the driver before turning on the steering lampAnd->The hidden state layer of the HMM network model of the second hidden Markov model is the driving intention state +.>And->The method comprises the steps of carrying out a first treatment on the surface of the The observation state layer nodes of the two Hidden Markov Model (HMM) network models are the driving behaviors B, C, D, E executed by the driver;
expanding the first hidden Markov model HMM network model and the second hidden Markov model HMM network model according to the geographic position of the vehicle, and taking a preset geographic scale mark as an expansion point during expansion;
calculating a driving intention transition probability function of each unfolded point after unfolding, and counting driving behavior intention mapping probabilities of each unfolded point after unfolding to obtain a final first driving intention HMM recognition model and a second driving intention HMM recognition model, wherein the first driving intention HMM recognition model and the second driving intention HMM recognition model jointly form a driving intention HMM recognition model.
5. The short-term prediction method for driving behavior at ramp according to claim 1, wherein the time required for the vehicle to travel from the current location to the front geographic scale is taken as the prediction durationWith the predicted duration +.>And predicting the driving behavior of the ramp vehicle for the time interval.
6. The short-term prediction method for driving behavior at ramp according to claim 1, wherein the current position of the vehicle is obtained by a vehicle-mounted GPS or a geographic information electronic tag buried on the ground, and the current speed of the vehicle is obtained by a vehicle-mounted information acquisition systemvDriving behavior and turn signal usage status.
7. A short-term prediction system of driving behavior at a ramp, comprising:
the driving behavior prediction point determining module is used for obtaining the distance between the current position of the vehicle and the front intersections n Current speed of vehiclevDistance is tos n Comparing with preset geographic scale, and based on the current speed of the vehiclevDetermining the geographic scale with the smallest distance in front of the vehicle as a driving behavior prediction points n+1 ；
The driving information acquisition module is used for acquiring the driving behavior currently executed by the vehicleExecuted driving behaviorTogether constitute the observation state sequence->，/>The method comprises the steps of carrying out a first treatment on the surface of the Acquisition of the observation State sequence->The use states of the steering lamps corresponding to each observation state node form a sequence flag of the use states of the steering lamps;
the maximum probability hidden state sequence calculation module is used for calculating a state sequence flag and an observation state sequence based on the steering lamp use state sequenceCalculating by using driving intention HMM recognition model to obtain maximum probability hidden state sequence +.>；
An observation state sequence driving intention judging module for hiding the state sequence based on the maximum probabilityCounting the proportion value of the total number of hidden state nodes occupied by different types of hidden state nodes to determine an observation state sequence +.>The represented driving intention;
the prediction point driving intention judging module is used for judging the driving intention according to the observation state sequenceCalculating the driving intention transition probability to determine the predicted point of driving behaviors n+1 Driving intention at;
the prediction point driving behavior prediction module is used for predicting a point based on driving behaviors n+1 The driving intention at the position, calculating the predicted point of driving behaviors n+1 Probability of executing each driving behavior is carried out, and driving behavior corresponding to the maximum probability value is taken as a prediction objectI.e. driving behavior prediction pointss n+1 Prediction results of driving behavior at the location.
8. A computer-readable storage medium, on which a computer program is stored, which computer program, when being executed by a processor, implements a method for short-term prediction of driving behavior at a ramp as claimed in any one of claims 1 to 6.
CN202310163003.6A 2023-02-24 2023-02-24 Short-time prediction method, system and storable medium for driving behavior at ramp Active CN116198520B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202310163003.6A CN116198520B (en) 2023-02-24 2023-02-24 Short-time prediction method, system and storable medium for driving behavior at ramp

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202310163003.6A CN116198520B (en) 2023-02-24 2023-02-24 Short-time prediction method, system and storable medium for driving behavior at ramp

Publications (2)

Publication Number Publication Date
CN116198520A CN116198520A (en) 2023-06-02
CN116198520B true CN116198520B (en) 2023-09-22

Family

ID=86507369

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202310163003.6A Active CN116198520B (en) 2023-02-24 2023-02-24 Short-time prediction method, system and storable medium for driving behavior at ramp

Country Status (1)

Country Link
CN (1) CN116198520B (en)

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
DE102015222964A1 (en) * 2015-11-20 2017-05-24 Honda Motor Co., Ltd. A driver intention prediction device and method and vehicle comprising such a device
CN106740864A (en) * 2017-01-12 2017-05-31 北京交通大学 A kind of driving behavior is intended to judge and Forecasting Methodology
CN109460023A (en) * 2018-11-09 2019-03-12 上海理工大学 Driver's lane-changing intention recognition methods based on Hidden Markov Model
CN114701509A (en) * 2022-03-11 2022-07-05 北京科技大学 Hybrid traffic flow-oriented driving intention identification method and system

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP4797948B2 (en) * 2006-11-22 2011-10-19 株式会社デンソー Driving behavior estimation method and apparatus
US10486707B2 (en) * 2016-01-06 2019-11-26 GM Global Technology Operations LLC Prediction of driver intent at intersection

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
DE102015222964A1 (en) * 2015-11-20 2017-05-24 Honda Motor Co., Ltd. A driver intention prediction device and method and vehicle comprising such a device
CN106740864A (en) * 2017-01-12 2017-05-31 北京交通大学 A kind of driving behavior is intended to judge and Forecasting Methodology
CN109460023A (en) * 2018-11-09 2019-03-12 上海理工大学 Driver's lane-changing intention recognition methods based on Hidden Markov Model
CN114701509A (en) * 2022-03-11 2022-07-05 北京科技大学 Hybrid traffic flow-oriented driving intention identification method and system

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
面向人机协同共驾的驾驶行为短时预测方法研究;喻恺, 彭理群, 丁雪, 贺宜;交通信息与安全;第37卷(第216期);第42-48、57页 *

Also Published As

Publication number Publication date
CN116198520A (en) 2023-06-02

Similar Documents

Publication Publication Date Title
CN109840660B (en) Vehicle characteristic data processing method and vehicle risk prediction model training method
Jing et al. Building Tianjin driving cycle based on linear discriminant analysis
US20190004517A1 (en) Method and apparatus for generating automatic driving strategy
EP2710572B1 (en) Vehicle data analysis method and vehicle data analysis system
Chen et al. Analysis of factors affecting the severity of automated vehicle crashes using XGBoost model combining POI data
CN107590999B (en) Traffic state discrimination method based on checkpoint data
CN114435138A (en) Vehicle energy consumption prediction method and device, vehicle and storage medium
Tian et al. Evaluating the effectiveness of V2V-based lane speed monitoring application: a simulation study
Wheeler et al. Factor graph scene distributions for automotive safety analysis
CN114077785A (en) Method and device for constructing simulation test scene of vehicle
CN116046008A (en) Situation awareness-based route planning method, system and efficiency evaluation device
Montanari et al. Pattern recognition for driving scenario detection in real driving data
CN110909907A (en) Method and device for predicting fuel consumption of truck and storage medium
CN111824169A (en) Method for reducing exhaust gas emissions of a drive system of a vehicle having an internal combustion engine
CN116198520B (en) Short-time prediction method, system and storable medium for driving behavior at ramp
CN116597690B (en) Highway test scene generation method, equipment and medium for intelligent network-connected automobile
CN114692713A (en) Driving behavior evaluation method and device for automatic driving vehicle
CN116753938A (en) Vehicle test scene generation method, device, storage medium and equipment
CN115520216A (en) Driving state judging method and device, computer equipment and storage medium
CN113570870A (en) Distributed intersection average delay estimation method, device, equipment and storage medium
CN115959135A (en) Enhanced vehicle operation
CN114722931A (en) Vehicle-mounted data processing method and device, data acquisition equipment and storage medium
CN114116854A (en) Track data processing method, device, equipment and storage medium
Jomrich et al. Lane Accurate Detection of Map Changes based on Low Cost Smartphone Data.
CN108197150B (en) Analysis method and system for urban parking difficult area

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant