CN114613127A - Driving risk prediction method based on multi-layer multi-dimensional index system - Google Patents

Driving risk prediction method based on multi-layer multi-dimensional index system Download PDF

Info

Publication number
CN114613127A
CN114613127A CN202210124973.0A CN202210124973A CN114613127A CN 114613127 A CN114613127 A CN 114613127A CN 202210124973 A CN202210124973 A CN 202210124973A CN 114613127 A CN114613127 A CN 114613127A
Authority
CN
China
Prior art keywords
vehicle
risk
index
node
lane
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.)
Granted
Application number
CN202210124973.0A
Other languages
Chinese (zh)
Other versions
CN114613127B (en
Inventor
熊晓夏
何禹
蔡英凤
刘擎超
王海
沈钰杰
陈龙
景鹏
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.)
Jiangsu University
Original Assignee
Jiangsu 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 Jiangsu University filed Critical Jiangsu University
Priority to CN202210124973.0A priority Critical patent/CN114613127B/en
Publication of CN114613127A publication Critical patent/CN114613127A/en
Application granted granted Critical
Publication of CN114613127B publication Critical patent/CN114613127B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/01—Detecting movement of traffic to be counted or controlled
    • G08G1/0104—Measuring and analyzing of parameters relative to traffic conditions
    • G08G1/0125—Traffic data processing
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G06Q10/063—Operations research, analysis or management
    • G06Q10/0635—Risk analysis of enterprise or organisation activities
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G06Q10/063—Operations research, analysis or management
    • G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
    • G06Q10/06393—Score-carding, benchmarking or key performance indicator [KPI] analysis
    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/01—Detecting movement of traffic to be counted or controlled
    • G08G1/0104—Measuring and analyzing of parameters relative to traffic conditions
    • G08G1/0137—Measuring and analyzing of parameters relative to traffic conditions for specific applications
    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/16—Anti-collision systems
    • G08G1/166—Anti-collision systems for active traffic, e.g. moving vehicles, pedestrians, bikes
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
    • Y02T10/00—Road transport of goods or passengers
    • Y02T10/10—Internal combustion engine [ICE] based vehicles
    • Y02T10/40—Engine management systems

Landscapes

  • Business, Economics & Management (AREA)
  • Human Resources & Organizations (AREA)
  • Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Strategic Management (AREA)
  • Physics & Mathematics (AREA)
  • Economics (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Development Economics (AREA)
  • Educational Administration (AREA)
  • Game Theory and Decision Science (AREA)
  • Marketing (AREA)
  • Operations Research (AREA)
  • Quality & Reliability (AREA)
  • Tourism & Hospitality (AREA)
  • General Business, Economics & Management (AREA)
  • Theoretical Computer Science (AREA)
  • Analytical Chemistry (AREA)
  • Chemical & Material Sciences (AREA)
  • Traffic Control Systems (AREA)

Abstract

The invention provides a driving risk prediction method based on a multi-level multi-dimensional index system, which comprises the steps of firstly, constructing a driving risk index system comprising an interval traffic flow risk index, a peripheral vehicle group risk index and an adjacent vehicle risk index from three levels of a macroscopic traffic flow, a middle vehicle-watching group and a microscopic vehicle by combining four dimensions of a road environment, a vehicle space-time distribution, a vehicle motion state and vehicle performance; constructing a driving risk prediction matter element extension model based on an index system; and finally, substituting each index in the driving risk index system into the matter element extension model, determining the comprehensive association degree of different driving risk grades of the target vehicle, and obtaining the driving risk prediction grade. The invention can more comprehensively, systematically and accurately depict the driving risk level of the vehicles on the highway.

Description

Driving risk prediction method based on multi-layer multi-dimensional index system
Technical Field
The invention relates to the technical field of traffic safety evaluation and intelligent traffic, in particular to a driving risk prediction method based on a multi-layer multi-dimensional index system.
Background
With the development of the current intelligent networking automobile technology, how to further improve the driving safety of the vehicle by utilizing the networking technology becomes one of the key points and difficulties in the development of the intelligent automobile industry. The driving risk evaluation method widely used at present is mainly based on a traffic conflict technology, and compared with the traditional method for researching traffic safety problems by historical accident data, the traffic conflict technology has the advantages of large data sample size, short data acquisition period, high data precision and the like. The conflict measurement index can identify the traffic conflict through a preset threshold value, and quantify the occurrence probability or severity of the traffic conflict, and the main types of the conflict measurement index comprise a risk avoidance behavior measurement index, a time/space proximity measurement index, a vehicle motion characteristic index and a conflict energy index. Most of the conflict measurement indexes are microscopic conflict indexes oriented to a single scene, and have specific use principles and limitations, such as the possibility that the Time To Collision (TTC) can only reflect longitudinal conflict accidents, and the Lane Change Risk Index (LCRI) can only reflect the possibility and the severity of transverse conflicts.
In recent years, some researches are focused on exploring the relationship between microscopic conflict indexes by methods such as social network analysis, fuzzy logic, TOPSIS (approximate ideal solution), Logit model discrete selection and the like so as to achieve the purpose of predicting the driving risk more accurately by using better index combination; however, the relevance and difference of vehicle operation characteristics between the own vehicle and surrounding vehicles, local vehicle groups and road section traffic flow are not systematically considered in the current research. Under the environment of the Internet of vehicles, the characteristic data can be acquired in real time, and a multi-layer multi-dimensional driving risk index system is established.
Disclosure of Invention
In view of this, the invention provides a driving risk prediction method based on a multi-level multi-dimensional index system, which is beneficial to improving the accuracy of risk prediction, so that the purpose of preventing traffic accidents is achieved.
The present invention achieves the above-described object by the following technical means.
A driving risk prediction method based on a multi-layer multi-dimensional index system specifically comprises the following steps:
constructing a driving risk index system comprising interval traffic flow risk indexes, surrounding vehicle group risk indexes and adjacent vehicle risk indexes by combining the road environment, the vehicle space-time distribution, the vehicle motion state and the vehicle performance from three aspects of macroscopic traffic flow, a central vehicle group and microscopic vehicles;
taking the running risk of vehicles on the highway as an object to be evaluated, taking each index in a running risk index system as an object characteristic, determining a classical domain and a segment domain based on historical track data of the vehicles on the highway section, determining a weight based on a game theory combined weighting method, and constructing a running risk prediction object element extension model based on the index system;
and acquiring the road environment of the running road section of the target vehicle, the spatial distribution and motion state of the vehicle and the vehicle performance in real time in the Internet of vehicles environment, calculating each index in the running risk index system in real time, substituting the index into the matter element extension model, and determining the comprehensive association degree of different running risk levels of the target vehicle to obtain the running risk prediction level.
According to the further technical scheme, an interval traffic flow risk index is constructed by combining a macroscopic traffic flow level with a road environment and vehicle space-time distribution, and the interval traffic flow risk index comprises interval lane complexity, interval traffic flow instability and interval traffic comprehensive risk.
According to a further technical scheme, the complexity of the inter-zone lane is calculated according to a network diagram for constructing the inter-zone lane;
the construction process of the inter-zone lane network graph comprises the following steps: taking each lane in each road section of the research interval as a node of the interval lane network graph, determining directed edges in the interval lane network graph according to the driving direction, and determining the edge weight according to the complexity of the interaction process among the lane nodes to obtain the interval lane network graph;
the complexity of the interval lane is defined as the ratio of the sum of point weights of all nodes in the interval lane network graph to 2 times of the number of the nodes;
the point weight of the node is defined as the sum of the edge weights associated with it:
Figure BDA0003500040510000021
wherein W'iIs node c'iDot weight of, w'ijIs node c'iIs a starting point, c'jWeight of the directed edge as end point, w'jiIs node c'jIs a starting point, c'iWeight of the directed edge as endpoint, N'iIs and node c'iThere is a connected set of neighbors.
According to a further technical scheme, the interval traffic flow instability is calculated according to a constructed interval traffic flow network diagram;
the construction process of the inter-area traffic flow network diagram comprises the following steps: taking a running vehicle in a certain range in front of a target vehicle as a node, and constructing an interval traffic flow network diagram according to an edge connection rule and an edge weight determination rule among the nodes;
the side connecting rule is as follows: when any one of the following conditions (1) and (2) is satisfied, the vehicle node ciAnd cjHas the following connecting edges:
(1)ciand cjFor adjacent vehicles traveling in the same lane, namely, the following conditions are satisfied simultaneously:
①LLi＝LLj
②{ck∈C|(LLk＝LLi) And is
Figure BDA0003500040510000022
(2)ciAnd cjAdjacent vehicles traveling in two adjacent lanes, having a longitudinal distance of no more than 150m, with opposite lateral velocities or lateral accelerations, i.e. satisfying both:
①|LLi-LLj|＝1
②|xi-xj|<150
③(LLi-LLj)vyi<0 or (LL)i-LLj)ayi<0 or (LL)i-LLj)vyj>0 or (LL)i-LLj)ayj>0；
Wherein: LL (LL)i、LLj、LLkAll represent lane numbers, ckIs any one vehicle node, C is the set of all vehicle nodes in a certain range in front of the target vehicle, xkIs node ckCentroid position abscissa, xi、xjAre respectively node ciAnd cjCenter of mass position abscissa of, ayi、ayjAre respectively node ciAnd cjTransverse acceleration of vyi、vyjAre respectively node ciAnd cjThe lateral velocity of (d);
the edge weight determination rule is specifically as follows: if vehicle node ciAnd cjIf the two nodes have connecting edges, the undirected edge weight for connecting the two nodes is defined as the speed difference of the two vehicles within a unit distance;
the unstable degree of the inter-zone traffic flow is the average unit weight of the inter-zone traffic flow network diagram, namely:
Figure BDA0003500040510000031
wherein: i is2The interval traffic flow instability index is obtained;
Figure BDA0003500040510000032
is node ciAnd is defined as node ciPoint weight W ofiDegree of node d theretoiThe ratio of (a) to (b).
A further technical scheme is that the comprehensive risk of the interval traffic I3＝I1×I2In which I1Is an interval lane complexity index.
According to the further technical scheme, peripheral vehicle cluster risk indexes are constructed by combining vehicle space-time distribution and vehicle motion states from the view vehicle cluster level, and the peripheral vehicle cluster risk indexes comprise peripheral vehicle cluster density, peripheral vehicle cluster operation entropy and peripheral vehicle cluster comprehensive risks.
According to a further technical scheme, the process for acquiring the density of the peripheral vehicle cluster comprises the following steps:
based on relative longitudinal and transverse time distances X between the target vehicle and the surrounding vehiclesjCalculating the distribution density LTD of the peripheral vehicle cluster for the vehicle cluster distribution characteristic variable, and using the LTD as the peripheral vehicle cluster density index I4Namely:
Figure BDA0003500040510000033
Figure BDA0003500040510000034
wherein: j e {1,2, …, Nj}，NjThe number of surrounding vehicles; mu and Λ ═ diag (σ)x 2,σy 2) Respectively is the mean value and the variance of the two-dimensional Gaussian distribution; (x)0,y0)TAnd (x)j,yj)TPosition coordinates, L, of the target vehicle and the jth vehicle in the periphery thereof, respectively0And LjRespectively the length, U, of the target vehicle and the jth vehicle in its periphery0And UjThe width of the target vehicle and the jth vehicle around the target vehicle respectively, (v)x0,vy0) And (v)xj,vyj) The longitudinal and transverse speeds of the target vehicle and the jth vehicle around the target vehicle are respectively.
According to a further technical scheme, the process for acquiring the running entropy of the peripheral vehicle group comprises the following steps:
establishing a peripheral vehicle group running entropy index I by taking the speed difference and the acceleration difference between a target vehicle and peripheral vehicles as characteristic variables of the running state of the vehicles and combining the distribution density of the peripheral vehicle group5：
Figure BDA0003500040510000041
Wherein v is0、a0Speed, acceleration, v, of the target vehicle, respectivelyj、ajRespectively the speed and acceleration, k, of the jth vehicle around it1And k2Are weight coefficients.
Further technical scheme, comprehensive risk index I of peripheral vehicle groups6＝I4×I5。
According to the further technical scheme, adjacent vehicle risk indexes based on longitudinal and transverse collision avoidance deceleration and vehicle maximum braking deceleration are constructed on the microscopic vehicle level by combining the vehicle motion state and the vehicle performance;
adjacent vehicle risk indicator I7The following formula is satisfied:
Figure BDA0003500040510000042
wherein: RI (Ri)xjIs the risk probability, RI, of the target vehicle colliding with the jth vehicle in the periphery thereof in the longitudinal directionyjIs the risk probability, N, of the target vehicle colliding with the jth vehicle in the periphery thereof in the transverse directionjThe number of adjacent vehicles; the collision risk probability is determined by the ratio of the vehicle deceleration to the vehicle maximum braking deceleration.
The invention has the beneficial effects that:
(1) the invention comprehensively considers the influence of the operation characteristics of three levels of macroscopic traffic flow, central vehicle group and microscopic vehicles on the vehicle running risk, combines four dimensions of road environment, vehicle space-time distribution, vehicle motion state and vehicle performance, constructs a highway running risk index system comprising an interval traffic flow risk index, a peripheral vehicle group risk index and an adjacent vehicle risk index, and can more comprehensively, systematically and accurately depict the running risk level of the highway vehicles.
(2) The invention is based on a graph theory method, creatively constructs an interval lane network diagram and an interval traffic flow network diagram, provides an interval lane complexity and interval traffic flow instability index based on the network diagram, and provides an effective means for better evaluating the influence of a road traffic environment on driving risks.
(3) The invention provides a peripheral vehicle group running entropy index representing the uncertainty degree of the running state of a vehicle group based on an information entropy theory and by combining the distribution density of the peripheral vehicle group from the difference of running characteristics of a vehicle and peripheral vehicles, and provides a method for better predicting the running risk level from the time-space relationship between the vehicle and the peripheral vehicles.
Drawings
FIG. 1 is a block diagram of a driving risk prediction process based on a multi-layer multi-dimensional index system according to the present invention;
FIG. 2(a) is a side entry lane node to main line lane node diagram according to the present invention;
fig. 2(b) is a diagram of nodes from a main lane node to a lateral front main lane node according to the present invention;
FIG. 2(c) is a diagram of a main line lane node to side entry lane node according to the present invention;
FIG. 3 is a conceptual diagram of an inter-zone traffic network diagram according to the present invention;
fig. 4 is a schematic view of a collision risk fault tree of a vehicle around a microscopic level according to the present invention.
Detailed Description
The invention will be further described with reference to the following figures and specific examples, but the scope of the invention is not limited thereto.
As shown in fig. 1, the driving risk prediction method based on the multi-layer multi-dimensional index system of the present invention specifically includes the following steps:
step (1), constructing a multi-level and multi-dimensional driving risk index system: constructing a highway driving risk index system comprising an interval traffic flow risk index, a peripheral vehicle group risk index and an adjacent vehicle risk index from three levels of a macroscopic traffic flow, a central vehicle group and a microscopic vehicle respectively by combining four dimensions of road environment, vehicle space-time distribution, vehicle motion state and vehicle performance
Step (1.1), from a macroscopic traffic flow level, combining two dimensions of interval lane layout (belonging to road environment) and interval traffic flow distribution (belonging to vehicle space-time distribution), constructing interval traffic flow risk indexes, specifically comprising three indexes of interval lane complexity, interval traffic flow instability and interval traffic comprehensive risk
Step (1.1.1), constructing an interval lane network graph and calculating the complexity of the interval lane so as to represent the overall complexity of the running of the interval lane of the road where the target vehicle is located
Step (1.1.1.1), taking a 900m range in front of a target vehicle as a research interval, dividing the research interval into 6 road sections by taking 150m as a unit, taking each lane in each road section as a node of an interval lane network graph, determining a directed edge (arc) in the interval lane network graph according to a driving direction, determining an edge weight according to the complexity of an interaction process among the lane nodes, and obtaining a weighted directed network graph of the interval lane; determination of the side rights refer to table 1:
TABLE 1 edge weights for interval lane network graph
Figure BDA0003500040510000051
W 'of'ijIs a lane node c'iIs a starting point, c'jA weight of a directed edge being an endpoint; delta is a road alignment correction coefficient, and when the road curvature is larger than a preset threshold qthIf so, δ is taken to be 1.5, otherwise, 1 is taken.
Step (1.1.1.2), calculating the point weight, node c 'of each lane node in the inter-zone lane network graph of step (1.1.1.1)'iW 'of'iDefined as the sum of the edge weights associated with it, i.e.:
Figure BDA0003500040510000061
wherein, N'iIs and node c'iThere is a connected set of neighbors.
Step (1.1.1.3), calculating the average point weight of all lane nodes in the inter-zone lane network graph
Figure BDA0003500040510000062
The method is defined as the ratio of the sum of point weights of all nodes in the network graph of the section lane to 2 times of the number of the nodes, and the sum is used as a complexity index I of the section lane1Namely:
Figure BDA0003500040510000063
and N' is the number of the lane nodes in the interval lane network graph.
Step (1.1.2), constructing an interval traffic flow network diagram and calculating the instability of the interval traffic flow to represent the differentiation degree of the overall motion state of the vehicle in the road interval of the target vehicle
Step (1.1.2.1), taking a running vehicle in the range of 900m in front of a target vehicle as a node, and constructing an undirected graph of the inter-regional traffic flow network according to an edge connection rule and an edge weight determination rule among the nodes (as shown in figure 3)
Step (1.1.2.1.1), establishing a dynamic interval coordinate system by taking the intersection point of the road cross section at the position of the centroid of the target vehicle and the central dividing strip as an origin, taking the advancing direction along the lane as a positive x-axis and taking the clockwise rotation direction of the x-axis as a positive y-axis; simultaneously, numbering and marking lanes, wherein the number of an inner lane is 1, and the numbers of other lanes are sequentially increased by 1 from inside to outside; the nodes of all vehicles within 900m in front of the target vehicle are collected into a group
Figure BDA0003500040510000064
(K is the total number of vehicles), any one of which is the vehicle node ckHas a centroid position coordinate of (x)k,yk) The number of the lane is LLkLongitudinal and transverse velocities, respectively, vxk、vykLongitudinal acceleration and lateral acceleration are respectively axk、ayk。
Step (1.1.2.1.2) of, when either of the following conditions (1) and (2) is satisfied, the vehicle node ciAnd cjHas the following connecting edges:
(1)ciand cjFor adjacent vehicles traveling in the same lane, namely, the following conditions are satisfied simultaneously:
①LLi＝LLj
②{ck∈C|(LLk＝LLi) And is
Figure BDA0003500040510000065
(2)ciAnd cjAdjacent vehicles having a longitudinal distance of not more than 150m, with opposite lateral velocities or lateral accelerations, for driving in two adjacent lanes, i.e. satisfying simultaneously:
①|LLi-LLj|＝1
②|xi-xj|<150
③(LLi-LLj)vyi<0 or (LL)i-LLj)ayi<0 or (LL)i-LLj)vyj>0 or (LL)i-LLj)ayj>0。
Step (1.1.2.1.3), if the vehicle node ciAnd cjWith a connecting edge, the undirected edge right connecting two nodes is defined as the speed difference of two vehicles within a unit distance:
Figure BDA0003500040510000071
step (1.1.2.2), calculating the point weight of each vehicle node in the inter-zone traffic flow network diagram of step (1.1.2.1); node ciPoint weight W ofiDefined as the sum of the edge weights associated with it, i.e.:
Figure BDA0003500040510000072
wherein N isiIs and node ciThere is a connected set of neighbors.
Step (1.1.2.3), calculating the unit weight of each vehicle node in the inter-zone traffic flow network diagram of step (1.1.2.1); node (C)Point ciUnit weight of
Figure BDA0003500040510000073
Is defined as node ciPoint weight W ofiDegree of node d theretoiThe ratio of (a) to (b), i.e.:
Figure BDA0003500040510000074
wherein node ciDegree d ofiDefined as the number of edges connected to it.
And (1.1.2.4) calculating the average unit weight of the regional traffic flow network diagram in the step (1.1.2.1) by taking the degree of each vehicle node as the weight, and taking the average unit weight as a regional traffic flow instability index I2Namely:
Figure BDA0003500040510000075
where K is the total number of vehicle nodes in the graph.
Step (1.1.3), based on the complexity index I of the section lane1And interval traffic flow instability index I2And constructing an interval traffic comprehensive risk index I3Namely:
I3＝I1×I2
step (1.2), constructing a peripheral vehicle group risk index from the aspect of observing the vehicle group by combining two dimensions of the spatial distribution (belonging to the vehicle space-time distribution) of the peripheral vehicle group and the motion characteristics (belonging to the vehicle motion state) of the peripheral vehicle group, wherein the peripheral vehicle group risk index specifically comprises three indexes of the concentration of the peripheral vehicle group, the operation entropy of the peripheral vehicle group and the comprehensive risk of the peripheral vehicle group
And (1.2.1) calculating the density index of the surrounding vehicle group by taking the dynamic interval coordinate system of the step (1.1.2.1.1) as the basis and the target vehicle as the center and considering the interaction influence of the spatial distribution form of the surrounding vehicle group (including the adjacent front vehicle, left front vehicle, right front vehicle, left vehicle, right vehicle, rear vehicle, left rear vehicle and right rear vehicle) on the target vehicle
Step (1.2.1.1), defining the relative longitudinal and transverse time distance between the target vehicle and the peripheral jth vehicle as a variable Xj：
Figure BDA0003500040510000081
Wherein (x)0,y0)TAnd (x)j,yj)TPosition coordinates, L, of the target vehicle and the jth vehicle in the periphery thereof, respectively0And LjRespectively the length, U, of the target vehicle and the jth vehicle in its periphery0And UjThe widths of the target vehicle and the jth vehicle around the target vehicle, respectively, (v)x0,vy0) And (v)xj,vyj) The longitudinal and transverse speeds of the target vehicle and the jth vehicle around the target vehicle are respectively.
Step (1.2.1.2), based on two-dimensional Gaussian distribution probability density function, relative longitudinal and transverse time distances X between the target vehicle and the surrounding vehiclesj(j∈{1,2,…,Nj}，NjNumber of surrounding vehicles) as a vehicle cluster distribution characteristic variable, and calculating the distribution density LTD of the surrounding vehicle cluster and using the distribution density LTD as a surrounding vehicle cluster density index I4Namely:
Figure BDA0003500040510000082
where μ and Λ ═ diag (σ)x 2,σy 2) The mean and variance of the two-dimensional Gaussian distribution are respectively, and for the convenience of calculation, the mean value mu is (0,0)TAnd variance Λ ═ diag (1,1) (two-dimensional standard normal distribution).
Step (1.2.2), based on the information entropy theory, taking the speed difference and the acceleration difference between the target vehicle and the surrounding vehicles as the characteristic variables of the vehicle running state, and combining the distribution density of the surrounding vehicle group to establish the running entropy index I of the surrounding vehicle group5To represent the uncertainty of the running state of the vehicle group:
Figure BDA0003500040510000083
wherein v is0、a0Speed, acceleration, v, of the target vehicle, respectivelyj、ajRespectively the speed and acceleration, k, of the jth vehicle around it1And k2As a weight coefficient, satisfy k1+k21 (may be k)1＝k2＝0.5)。
Step (1.2.3), based on the density index I of the surrounding vehicle cluster4And peripheral vehicle group running entropy index I5Building comprehensive risk index I of surrounding vehicle groups6Namely:
I6＝I4×I5
step (1.3), constructing an adjacent vehicle risk index based on longitudinal and transverse collision avoidance deceleration and vehicle maximum braking deceleration from a microscopic vehicle level by combining two dimensions of adjacent vehicle motion characteristics (belonging to vehicle motion state) and vehicle braking performance (belonging to vehicle performance)
Step (1.3.1), taking the target vehicle as the center, calculating the collision avoidance deceleration DRAC of each vehicle adjacent to the periphery in the driving process so as to represent the interactive collision risk between the adjacent vehicles
Step (1.3.1.1), the j (j epsilon {1,2, …, N) th adjacent to the target vehiclej}，NjNumber of adjacent vehicles) longitudinal collision avoidance deceleration of the vehicle
Figure BDA0003500040510000091
The calculation method of (2) is as follows:
Figure BDA0003500040510000092
Figure BDA0003500040510000093
Figure BDA0003500040510000094
wherein l represents the front vehicle number of the two vehicles (0 represents the target vehicle, j represents the jth vehicle around the target vehicle), f represents the rear vehicle number of the two vehicles, DRACxIndicating the minimum longitudinal deceleration required by the rear vehicle to avoid a longitudinal collision with the front vehicle.
Step (1.3.1.2), the transverse collision avoidance deceleration of the target vehicle and the adjacent jth vehicle
Figure BDA0003500040510000095
The calculation method of (2) is as follows:
Figure BDA0003500040510000096
Figure BDA0003500040510000097
Figure BDA0003500040510000098
wherein i represents the number of the left side vehicle of the two vehicles, r represents the number of the right side vehicle of the two vehicles, DRACyIndicating the minimum lateral deceleration of a vehicle on one side required to avoid a lateral collision with another vehicle on the other side.
Step (1.3.2), considering vehicle braking performance, defining the risk probability that the target vehicle collides with the jth vehicle around the target vehicle in the longitudinal direction and the transverse direction respectively as follows:
Figure BDA0003500040510000099
Figure BDA00035000405100000910
wherein the MADRxAnd MADRyMaximum braking longitudinal deceleration and maximum braking of the vehicle, respectivelyLateral deceleration.
Step (1.3.3), as shown in FIG. 4, based on the fault tree principle, to get the own vehicle (target vehicle) and each vehicle c aroundj(j＝1,…,Nj) Longitudinal collision and lateral collision of the vehicle as basic events to the vehicle and the peripheral vehicles cjThe collision is an intermediate event, the total collision risk of the own vehicle and the surrounding vehicles is taken as an overhead event, and the adjacent vehicle risk index I is calculated according to the following formula7：
Figure BDA0003500040510000101
Wherein RIxjRIyjThe "and relation" indicating a basic event that two vehicles collide in reality only when the target vehicle satisfies the collision conditions with the jth vehicle in the periphery thereof in both the longitudinal and lateral directions.
Step (2), constructing an extension model of the driving risk prediction matter element based on an index system: taking the running risk of the vehicles on the expressway as an object to be evaluated, taking each index in the running risk index system in the step (1) as an object characteristic, determining a classical domain and a segment domain based on historical track data of the vehicles on the expressway section, determining a weight based on a game theory combined weighting method, and obtaining a running risk prediction matter element extension model
And (2.1) taking the running risk of the vehicles on the highway as an object E to be evaluated, and taking n-7 indexes { I) in the running risk index system in the step (1)1,I2,…,InLet the n features correspond to a quantity of t1,t2,…,tnAnd (E, I, t) constructing an n-dimensional matter element R:
Figure BDA0003500040510000102
step (2.2), dividing the driving risk of the highway vehicles into 3 grades of low, medium and high risks, and respectively using EjJ is 1, …, m represents; note tji＝(aji,bji) I 1, …, n, j 1, …, m, etcObtaining the value range of the ith characteristic of the object under the level j to obtain the classical domain matter element Rj：
Figure BDA0003500040510000103
Step (2.2.1), selecting highway section samples with different lane numbers, road line shapes and access distribution based on a highway geographic information system, and calculating interval lane complexity indexes I of all samples according to the method in the step (1)1。
Step (2.2.2), historical vehicle track data samples of the road sections in different months, days of the week and time periods in the step (2.2.1) are collected, and interval traffic flow instability indexes I of all samples are calculated according to the method in the step (1)2And regional traffic comprehensive risk index I3。
Step (2.2.3), on the basis of the historical vehicle track data of the road section acquired in the step (2.2.2), screening target vehicle samples with adjacent vehicles in the longitudinal distance range of 150m before and after the driving process, and calculating the peripheral vehicle cluster density index I of all the samples according to the method in the step (1)4Peripheral vehicle group running entropy index I5And comprehensive risk index I of peripheral vehicle group6And an adjacent vehicle risk indicator I7。
Step (2.2.4), respectively calculating the obtained characteristics Ii(I-1, 2, …,7) numerical value samples are sorted from small to large, and 15%, 50% and 85% quantiles of sample values are respectively selected as characteristics IiDividing threshold values of low, middle and high three grades, namely determining object characteristics I under different grades jiValue range (a)ji,bji)，i＝1,2,…,7，j＝1,2,3。
Step (2.3) recording (a)pi,bpi) Is (a)ji,bji) J is 1, …, m, which is the union of all the grade value ranges of the ith characteristic of the object, and the area-saving object element R is obtainedp：
Figure BDA0003500040510000111
Step (2.4), the object characteristics { I ] are calculated by using a combined weighting method based on the game theory1,I2,…,InThe combination of } is weighted
Step (2.4.1), calculating each characteristic of things { I) by a Delphi method1,I2,…,InSubjective weighting of }
Step (2.4.1.1), invite experts in the field to the respective characteristics I1,I2,…,InAnd (4) scoring when n is 7, wherein the score value ranges from {1,2,3,4 and 5}, and taking the arithmetic average value of the scores of the features as the final score of each feature:
Figure BDA0003500040510000112
wherein z isijThe score given to the ith feature for the jth expert, ne is the number of experts, zi' is the final score of the ith feature, and the subjective weight of the feature is the proportion of the final score to the sum of all index scores:
Figure BDA0003500040510000113
step (2.4.1.2), the consistency of the expert evaluation opinions is checked by using a Kendall method so as to fully consider the consistency of all the expert opinions; if the test fails, the expert scoring is carried out again, otherwise, the subjective weight Z of each feature is obtainedS＝{zs1,zs2,…,zsn}。
Step (2.4.2), calculating each characteristic of the object by an entropy weight method { I }1,I2,…,InObjective weight of }
Step (2.4.2.1), for the characteristic I of step (2.2.4)iThe numerical sample of i-1, 2, …, n is normalized:
Figure BDA0003500040510000114
wherein u isgiIs characterized byiOriginal value of g-th sample, uimax、uiminRespectively represent characteristics IiMaximum and minimum values, x, in all sample datagiIs characterized byiNormalized values for the g-th sample.
Step (2.4.2.2), finding out each characteristic Ii(i ═ 1,2, …, n) entropy values for samples:
Figure BDA0003500040510000115
wherein eiIs characterized byiInformation entropy of (1), kiRepresents the characteristic IiNumber of samples of (1), pgiIs characterized byiThe g sample value accounts for the proportion of the sum of all sample values of the feature:
Figure BDA0003500040510000116
step (2.4.2.3) of calculating objective weights Z for the features by entropyO＝{zo1,zo2,…,zon}：
Figure BDA0003500040510000121
Wherein z isoiIs the objective weight of the ith feature.
Step (2.4.3), calculating each characteristic { I ] of the object by a game theory combination weighting method1,I2,…,InThe combining weights of
Step (2.4.3.1), setting the subjective weight Z in the step (2.4.1)S＝{zs1,zs2,…,zsnIs a vector Z1Step (2.4.2) said objective weight ZO＝{zo1,zo2,…,zonIs a vector Z2The linear combination coefficient of the two weight vectors is alpha ═{α1,α2Where (the parameters to be determined), the combined weight ω of each feature is { ω ═ ω1,ω2,…,ωnThe method is as follows:
Figure BDA0003500040510000122
step (2.4.3.2), according to the idea of game theory, using omega and Z1、Z2The minimum dispersion of (c) is the optimization objective, namely:
Figure BDA0003500040510000123
obtaining undetermined parameter alpha according to the optimized first derivative conditionkOptimum value of (k 1,2)
Figure BDA0003500040510000124
Figure BDA0003500040510000125
Step (2.4.3.3), for
Figure BDA0003500040510000126
Carrying out normalization processing, namely:
Figure BDA0003500040510000127
finally, the characteristics of the object { I }1,I2,…,InThe optimal combining weight of } is determined
Figure BDA0003500040510000128
Figure BDA0003500040510000129
Step (2.5), OverallThe step (2.1) to the step (2.4) are carried out to obtain a driving risk prediction object element extension model { R, Rj,Rp,ω*}。
And (3) predicting the online driving risk in real time: acquiring information such as road environment of a running road section of a target vehicle, spatial distribution and motion state of the vehicle, vehicle performance and the like in real time in an internet of vehicles environment, calculating each index in the driving risk index system in the step (1) in real time, substituting the indexes into the matter element extension model in the step (2), determining the comprehensive relevance of different driving risk levels of the target vehicle, and obtaining the driving risk prediction level
Step (3.1), acquiring information such as road environment of a running road section of the target vehicle S, spatial distribution and motion state of the vehicle, vehicle performance and the like in real time in the Internet of vehicles environment, and calculating in real time to obtain each index { I in the step one1,I2,…,InCorresponding magnitude tS1,tS2,…,tSnAnd obtaining an object element to be evaluated:
Figure BDA0003500040510000131
step (3.2), calculating each characteristic value t obtained in the step (3.1)Si(I ═ 1,2, …, n) and object characteristic I in step (2)iRange of values for different grades (a)ji,bji) Degree of association k of (j ═ 1,2,3)Sj(tSi)：
Figure BDA0003500040510000132
Where ρ isji(tSi,tji) The ith characteristic quantity value t of the object S to be evaluatedSiMagnitude range t of (point) and classical domain level jjiDistance (finite interval):
Figure BDA0003500040510000133
in the same way, rhoji(tSi,tpi) The ith characteristic quantity value t of the object S to be evaluatedSiRange of magnitudes t of (point) and level j of sectionpiDistance (finite interval):
Figure BDA0003500040510000134
|tjil is the magnitude range t of the classical domain level jjiMode (finite interval):
|tji|＝|bji-aji|
and (3.3) calculating the comprehensive association degree of the object to be evaluated S and the driving risk grade j equal to 1,2 and 3 on the basis of the association degree of each characteristic grade obtained in the step (3.2):
Figure BDA0003500040510000135
and (3.4) the driving risk prediction value of the target vehicle S at the current moment is the grade with the highest comprehensive association degree:
J＝argmaxj∈{1,2,3}(Kj(ES))。
the present invention is not limited to the above-described embodiments, and any obvious improvements, substitutions or modifications can be made by those skilled in the art without departing from the spirit of the present invention.

Claims (10)

1. A driving risk prediction method based on a multi-layer multi-dimensional index system is characterized by comprising the following steps:
constructing a driving risk index system comprising an interval traffic flow risk index, a peripheral vehicle group risk index and an adjacent vehicle risk index from three levels of a macroscopic traffic flow, a central vehicle group and a microscopic vehicle by combining a road environment, vehicle space-time distribution, a vehicle motion state and vehicle performance;
taking the running risk of vehicles on the highway as an object to be evaluated, taking each index in a running risk index system as an object characteristic, determining a classical domain and a segment domain based on historical track data of the vehicles on the highway section, determining a weight based on a game theory combined weighting method, and constructing a running risk prediction object element extension model based on the index system;
and acquiring the road environment of the running road section of the target vehicle, the spatial distribution and motion state of the vehicle and the vehicle performance in real time in the Internet of vehicles environment, calculating each index in the running risk index system in real time, substituting the index into the matter element extension model, and determining the comprehensive association degree of different running risk levels of the target vehicle to obtain the running risk prediction level.
2. The driving risk prediction method according to claim 1, wherein an inter-zone traffic flow risk index is constructed from a macroscopic traffic flow level by combining a road environment and a vehicle space-time distribution, and the inter-zone traffic flow risk index includes inter-zone lane complexity, inter-zone traffic flow instability and an inter-zone traffic comprehensive risk.
3. The driving risk prediction method according to claim 2, wherein the inter-zone lane complexity is calculated according to a constructed inter-zone lane network map;
the construction process of the inter-zone lane network graph comprises the following steps: taking each lane in each road section of the research interval as a node of the interval lane network graph, determining directed edges in the interval lane network graph according to the driving direction, and determining the edge weight according to the complexity of the interaction process among the lane nodes to obtain the interval lane network graph;
the complexity of the interval lane is defined as the ratio of the sum of point weights of all nodes in the interval lane network graph to 2 times of the number of the nodes;
the point weight of the node is defined as the sum of the edge weights associated with it:
Figure FDA0003500040500000011
wherein W'iIs node c'iDot weight of, w'ijIs node c'iIs a starting point, c'jWeight of the directed edge as end point, w'jiIs node c'jIs a starting point, c'iWeight of oriented edge, N ', as end point'iIs and node c'iThere is a connected set of neighbors.
4. The driving risk prediction method according to claim 3, wherein the interval traffic flow instability is calculated according to a constructed interval traffic flow network diagram;
the construction process of the inter-area traffic flow network diagram comprises the following steps: taking a running vehicle in a certain range in front of a target vehicle as a node, and constructing an interval traffic flow network diagram according to an edge connection rule and an edge weight determination rule among the nodes;
the side connecting rule is as follows: when any one of the following conditions (1) and (2) is satisfied, the vehicle node ciAnd cjHas the following connecting edges:
(1)ciand cjFor adjacent vehicles traveling in the same lane, namely, the following conditions are satisfied simultaneously:
①LLi＝LLj
②
Figure FDA0003500040500000021
(2)ciand cjAdjacent vehicles traveling in two adjacent lanes, having a longitudinal distance of no more than 150m, with opposite lateral velocities or lateral accelerations, i.e. satisfying both:
①|LLi-LLj|＝1
②|xi-xj|<150
③(LLi-LLj)vyi<0 or (LL)i-LLj)ayi<0 or (LL)i-LLj)vyj>0 or (LL)i-LLj)ayj>0；
Wherein: LL (LL)i、LLj、LLkAll represent lane numbers, ckIs any one vehicle node, and C is all vehicle nodes in a certain range in front of the target vehicleSet of (2), xkIs node ckCenter of mass position abscissa, xi、xjAre respectively node ciAnd cjCenter of mass position abscissa of, ayi、ayjAre respectively node ciAnd cjTransverse acceleration of vyi、vyjAre respectively node ciAnd cjThe lateral velocity of (d);
the edge weight determination rule is specifically as follows: if vehicle node ciAnd cjIf the two nodes have connecting edges, the undirected edge weight for connecting the two nodes is defined as the speed difference of the two vehicles within a unit distance;
the unstable degree of the inter-zone traffic flow is the average unit weight of the inter-zone traffic flow network diagram, namely:
Figure FDA0003500040500000022
wherein: i is2The interval traffic flow instability index is obtained;
Figure FDA0003500040500000023
is node ciAnd is defined as node ciPoint weight W ofiDegree of node d theretoiThe ratio of (a) to (b).
5. The driving risk prediction method according to claim 4, wherein the comprehensive risk of inter-zone traffic I3＝I1×I2In which I1Is an interval lane complexity index.
6. The driving risk prediction method according to claim 1, wherein from the perspective vehicle group level, the peripheral vehicle group risk indicators are constructed by combining the vehicle space-time distribution and the vehicle motion state, and include the peripheral vehicle group density, the peripheral vehicle group operation entropy and the peripheral vehicle group comprehensive risk.
7. The driving risk prediction method according to claim 6, wherein the acquisition process of the surrounding vehicle group density is:
based on relative longitudinal and transverse time distances X between the target vehicle and the surrounding vehiclesjCalculating the distribution density LTD of the peripheral vehicle cluster for the vehicle cluster distribution characteristic variable, and using the LTD as the peripheral vehicle cluster density index I4Namely:
Figure FDA0003500040500000031
Figure FDA0003500040500000032
wherein: j e {1,2, …, Nj}，NjThe number of surrounding vehicles; mu and Λ ═ diag (σ)x 2,σy 2) Respectively is the mean value and the variance of the two-dimensional Gaussian distribution; (x)0,y0)TAnd (x)j,yj)TPosition coordinates, L, of the target vehicle and its peripheral jth vehicle, respectively0And LjRespectively the length, U, of the target vehicle and the jth vehicle in its periphery0And UjThe widths of the target vehicle and the jth vehicle around the target vehicle, respectively, (v)x0,vy0) And (v)xj,vyj) The longitudinal and transverse speeds of the target vehicle and the jth vehicle around the target vehicle are respectively.
8. The driving risk prediction method according to claim 7, wherein the process of obtaining the running entropy of the surrounding vehicle group is as follows:
establishing a peripheral vehicle group running entropy index I by taking the speed difference and the acceleration difference between a target vehicle and peripheral vehicles as characteristic variables of the running state of the vehicles and combining the distribution density of the peripheral vehicle group5：
Figure FDA0003500040500000033
Wherein v is0、a0Speed, acceleration, v, of the target vehicle, respectivelyj、ajRespectively the speed and acceleration, k, of the jth vehicle around it1And k2Are weight coefficients.
9. The driving risk prediction method according to claim 7, wherein the peripheral vehicle group integrated risk index I6＝I4×I5。
10. The driving risk prediction method according to claim 1, characterized in that an adjacent vehicle risk index based on the longitudinal and lateral collision avoidance deceleration and the vehicle maximum braking deceleration is constructed from a microscopic vehicle level in combination with the vehicle motion state and the vehicle performance;
adjacent vehicle risk indicator I7The following formula is satisfied:
Figure FDA0003500040500000034
wherein: RI (Ri)xjIs the risk probability, RI, of the target vehicle colliding with the jth vehicle in the periphery thereof in the longitudinal directionyjIs the risk probability, N, of the target vehicle colliding with the jth vehicle in the periphery thereof in the transverse directionjThe number of adjacent vehicles; the collision risk probability is determined by the ratio of the vehicle deceleration to the vehicle maximum braking deceleration.
CN202210124973.0A 2022-02-10 2022-02-10 Driving risk prediction method based on multi-layer multi-dimensional index system Active CN114613127B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202210124973.0A CN114613127B (en) 2022-02-10 2022-02-10 Driving risk prediction method based on multi-layer multi-dimensional index system

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202210124973.0A CN114613127B (en) 2022-02-10 2022-02-10 Driving risk prediction method based on multi-layer multi-dimensional index system

Publications (2)

Publication Number Publication Date
CN114613127A true CN114613127A (en) 2022-06-10
CN114613127B CN114613127B (en) 2023-04-07

Family

ID=81859643

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202210124973.0A Active CN114613127B (en) 2022-02-10 2022-02-10 Driving risk prediction method based on multi-layer multi-dimensional index system

Country Status (1)

Country Link
CN (1) CN114613127B (en)

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP2950294A1 (en) * 2014-05-30 2015-12-02 Honda Research Institute Europe GmbH Method and vehicle with an advanced driver assistance system for risk-based traffic scene analysis
CN109448369A (en) * 2018-10-26 2019-03-08 中交第公路勘察设计研究院有限公司 Highway real time execution Risk Calculation method
CN111402586A (en) * 2020-03-26 2020-07-10 华东交通大学 Road meteorological environment forecasting and early warning control system and method based on Internet of vehicles
CN112508392A (en) * 2020-12-02 2021-03-16 云南省交通规划设计研究院有限公司 Dynamic evaluation method for traffic conflict risk of hidden danger road section of mountain area double-lane highway
CN113436432A (en) * 2021-05-24 2021-09-24 武汉理工大学 Method for predicting short-term traffic risk of road section by using road side observation data
CN113762473A (en) * 2021-08-25 2021-12-07 江苏大学 Complex scene driving risk prediction method based on multi-space-time diagram
US20220036735A1 (en) * 2018-05-08 2022-02-03 Tsinghua University Real-time driving risk assessment method employing equivalent force and device thereof

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP2950294A1 (en) * 2014-05-30 2015-12-02 Honda Research Institute Europe GmbH Method and vehicle with an advanced driver assistance system for risk-based traffic scene analysis
US20220036735A1 (en) * 2018-05-08 2022-02-03 Tsinghua University Real-time driving risk assessment method employing equivalent force and device thereof
CN109448369A (en) * 2018-10-26 2019-03-08 中交第公路勘察设计研究院有限公司 Highway real time execution Risk Calculation method
CN111402586A (en) * 2020-03-26 2020-07-10 华东交通大学 Road meteorological environment forecasting and early warning control system and method based on Internet of vehicles
CN112508392A (en) * 2020-12-02 2021-03-16 云南省交通规划设计研究院有限公司 Dynamic evaluation method for traffic conflict risk of hidden danger road section of mountain area double-lane highway
CN113436432A (en) * 2021-05-24 2021-09-24 武汉理工大学 Method for predicting short-term traffic risk of road section by using road side observation data
CN113762473A (en) * 2021-08-25 2021-12-07 江苏大学 Complex scene driving risk prediction method based on multi-space-time diagram

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
GIOVANNI DE NUNZIO 等: "Urban traffic Eco-Driving: A macroscopic steady-state analysis", 《 2014 EUROPEAN CONTROL CONFERENCE (ECC)》 *
张英杰: "基于驾驶工况预测的安全辅助策略研究", 《中国优秀博硕士学位论文全文数据库(硕士) 工程科技Ⅱ辑》 *
李倩: "基于实测微观驾驶状态的交通安全风险分析及模型校正", 《中国优秀博硕士学位论文全文数据库(硕士) 工程科技Ⅱ辑》 *

Also Published As

Publication number Publication date
CN114613127B (en) 2023-04-07

Similar Documents

Publication Publication Date Title
CN107229973B (en) Method and device for generating strategy network model for automatic vehicle driving
CN107742193B (en) A kind of driving Risk Forecast Method based on time-varying state transition probability Markov chain
Khayyam et al. Adaptive cruise control look-ahead system for energy management of vehicles
CN105006147B (en) A kind of Link Travel Time estimating method based on road spatial and temporal association
CN106095963A (en) Vehicle drive behavior analysis big data public service platform under the Internet+epoch
CN114312830B (en) Intelligent vehicle coupling decision model and method considering dangerous driving conditions
CN111292534A (en) Traffic state estimation method based on clustering and deep sequence learning
CN111063204B (en) Expressway vehicle speed prediction model training method based on toll station flow
Wang et al. A learning-based personalized driver model using bounded generalized Gaussian mixture models
Khodayari et al. Design of an improved fuzzy logic based model for prediction of car following behavior
CN110182217A (en) A kind of traveling task complexity quantitative estimation method towards complicated scene of overtaking other vehicles
CN111907523B (en) Vehicle following optimizing control method based on fuzzy reasoning
CN110979342A (en) Working condition information acquisition method for vehicle global energy management control
Mardiati et al. Review of microscopic model for traffic flow
CN114384916A (en) Adaptive decision-making method and system for off-road vehicle path planning
Liu et al. Vehicle state estimation based on unscented Kalman filtering and a genetic-particle swarm algorithm
Wang et al. ARIMA Model and Few‐Shot Learning for Vehicle Speed Time Series Analysis and Prediction
Khodayari et al. Modify car following model by human effects based on locally linear neuro fuzzy
CN115062202A (en) Method, device, equipment and storage medium for predicting driving behavior intention and track
CN114613127B (en) Driving risk prediction method based on multi-layer multi-dimensional index system
Feng et al. A Fuzzy Logic‐Based Approach for Humanized Driver Modelling
Wang et al. Dynamic traffic prediction based on traffic flow mining
CN114701517A (en) Multi-target complex traffic scene automatic driving solution based on reinforcement learning
CN110610611B (en) Driving safety evaluation method for intelligent network-connected vehicle in mixed-driving traffic flow
CN106529118A (en) Two-tuple linguistic AHP based intelligent automobile human-simulated steering control performance evaluation method

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