CN114662967B - Unmanned driving collision risk assessment method and system based on dynamic Bayesian network - Google Patents

Unmanned driving collision risk assessment method and system based on dynamic Bayesian network Download PDF

Info

Publication number
CN114662967B
CN114662967B CN202210363894.5A CN202210363894A CN114662967B CN 114662967 B CN114662967 B CN 114662967B CN 202210363894 A CN202210363894 A CN 202210363894A CN 114662967 B CN114662967 B CN 114662967B
Authority
CN
China
Prior art keywords
collision risk
vehicle
network
road traffic
risk
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
CN202210363894.5A
Other languages
Chinese (zh)
Other versions
CN114662967A (en
Inventor
黄文成
张寅�
于耀程
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.)
Southwest Jiaotong University
Original Assignee
Southwest 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 Southwest Jiaotong University filed Critical Southwest Jiaotong University
Priority to CN202210363894.5A priority Critical patent/CN114662967B/en
Publication of CN114662967A publication Critical patent/CN114662967A/en
Application granted granted Critical
Publication of CN114662967B publication Critical patent/CN114662967B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • 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
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N7/00—Computing arrangements based on specific mathematical models
    • G06N7/01—Probabilistic graphical models, e.g. probabilistic networks
    • 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
    • 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
    • G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
    • G06Q50/40—Business processes related to the transportation industry
    • 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

  • Engineering & Computer Science (AREA)
  • Business, Economics & Management (AREA)
  • Human Resources & Organizations (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Strategic Management (AREA)
  • Economics (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Development Economics (AREA)
  • General Business, Economics & Management (AREA)
  • Tourism & Hospitality (AREA)
  • Marketing (AREA)
  • Data Mining & Analysis (AREA)
  • Educational Administration (AREA)
  • General Engineering & Computer Science (AREA)
  • Quality & Reliability (AREA)
  • Operations Research (AREA)
  • Evolutionary Computation (AREA)
  • Game Theory and Decision Science (AREA)
  • Artificial Intelligence (AREA)
  • Mathematical Analysis (AREA)
  • Computational Mathematics (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Software Systems (AREA)
  • Mathematical Physics (AREA)
  • Computing Systems (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Pure & Applied Mathematics (AREA)
  • Mathematical Optimization (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Algebra (AREA)
  • Probability & Statistics with Applications (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Evolutionary Biology (AREA)
  • Health & Medical Sciences (AREA)
  • General Health & Medical Sciences (AREA)
  • Primary Health Care (AREA)
  • Traffic Control Systems (AREA)

Abstract

The invention discloses an unmanned driving collision risk assessment method and system based on a dynamic Bayesian network, which are characterized in that a network collision risk assessment model is constructed to obtain a network collision risk by acquiring the current road traffic condition and preprocessing the current road traffic condition to obtain a division performance index and the divided road traffic condition; constructing a vehicle collision risk estimation model, and evaluating vehicle collision risk probability according to the network collision risk; constructing a dynamic Bayesian network model, and carrying out unmanned driving collision risk evaluation on the vehicle collision risk probability in multiple periods; according to the method, the collision risk of the automatic driving vehicle is evaluated by constructing the interactive perception-dynamic Bayesian model and combining the influence of the network-level risk, early warning can be carried out according to the evaluation risk, certain reference is provided for the development of the automatic driving in the aspect of safety, and the problem that the influence of the road traffic environment is not considered in the estimation of the collision risk of the vehicle in the driving process of the vehicle, namely the safety is not high enough is solved.

Description

Unmanned driving collision risk assessment method and system based on dynamic Bayesian network
Technical Field
The invention relates to the technical field of driving collision risk assessment, in particular to an unmanned driving collision risk assessment method and system based on a dynamic Bayesian network.
Background
The automatic driving technology is continuously developed, the safety of the automatic driving technology is concerned by the public, the requirement on the risk assessment technology under the automatic driving environment is higher and higher, the considered aspect of the automatic driving technology is more complex and comprehensive, the collision accident which is one of the road driving accident forms is frequently generated, and the automatic driving technology is also one of the risk conditions which are firstly considered as the road driving safety. Therefore, in order to ensure that the evaluation of the collision risk of the automatic driving is safer, a driver or a vehicle is prompted to take measures, the possibility of collision threat is reduced, and the factors influencing the probability of the collision accident of the vehicle are various, at present, a system influencing the probability of the collision accident of the vehicle is mainly considered to be established from a vehicle layer, a kinematics layer and a sensor measurement layer of the automatic driving, however, the different road traffic environments can also influence the occurrence of the driving collision accident, and the probability of the collision accident of the road environments with different traffic flow attributes is different.
Disclosure of Invention
Aiming at the defects in the prior art, the invention provides an unmanned driving collision risk assessment method based on interactive perception-dynamic Bayes, which is used for assessing the collision risk of an automatic driving vehicle by constructing an interactive perception-dynamic Bayes model and combining the influence of network-level risks, can perform early warning according to the assessed risk and provides a certain reference for promoting the development of automatic driving in the aspect of safety.
In order to achieve the purpose of the invention, the invention adopts the technical scheme that:
in one aspect, a method for assessing risk of unmanned collision based on a dynamic Bayesian network comprises the following steps:
s1, collecting current road traffic conditions, and preprocessing the current road traffic conditions to obtain dividing performance indexes and divided road traffic conditions;
s2, constructing a network collision risk evaluation model, and obtaining a network collision risk according to the division performance indexes;
s3, constructing a vehicle collision risk estimation model, and evaluating vehicle collision risk probability according to the network collision risk;
and S4, evaluating through a dynamic Bayesian network model based on the vehicle collision risk probability of not less than one time period.
Preferably, step S1 specifically includes:
collecting current road traffic conditions, classifying the current road traffic conditions to obtain road traffic conditions in a dangerous state, road traffic conditions in a safe state and division performance indexes; wherein, the dividing performance indexes are respectively expressed as:
Figure BDA0003585065770000021
Figure BDA0003585065770000022
/>
Figure BDA0003585065770000023
wherein W is the overall accuracy of classification, D is the accuracy of judgment as a dangerous state, S is the accuracy of judgment as a safe state, and T is pz To correctly judge the number of dangerous states, T aq To correctly judge the number of safe states, F pz For the number of misjudged dangerous states, F aq The number of false positives as a safety state.
Preferably, step S2 is specifically:
constructing a network collision risk evaluation model, evaluating the probability of the road section having the traffic condition which is easy to collide in the current road traffic condition according to the division performance indexes, and obtaining the network collision risk, wherein a collision risk calculation formula in the network collision risk evaluation model is represented as follows:
Figure BDA0003585065770000031
Figure BDA0003585065770000032
the TC is a road traffic condition, values of 0 and 1 respectively correspond to two conditions of a safe state and a dangerous state, the NIR is a network-level risk condition, the P (NLR = "danger") is a probability of the network-level risk condition in the dangerous state, and the P (NLR = "safe") is a probability of the network-level risk condition in the safe state.
Preferably, step S3 is specifically:
constructing a vehicle collision risk estimation model, wherein the model is expressed as:
Figure BDA0003585065770000033
wherein,
Figure BDA0003585065770000034
the collision risk probability of the vehicle under the dangerous condition of the current vehicle at the moment t, N is the total number of the vehicles around the current vehicle, VLR t For the current vehicle condition at time t,
Figure BDA0003585065770000035
is the total status of the surrounding vehicle at time (t-1), based on>
Figure BDA0003585065770000036
The total situation of the risk of movement of the vehicle around the time (t-1),
Figure BDA0003585065770000037
for the network collision risk situation at time t, device for combining or screening>
Figure BDA0003585065770000038
And &>
Figure BDA0003585065770000039
Parameters of vehicle motion risk, surrounding vehicle risk and network collision risk respectively;
and calculating the probability of the vehicle collision risk by using the network collision risk and combining the vehicle collision risk estimation model.
Preferably, the network collision risk parameter calculation formula in step S3 is represented as:
Figure BDA0003585065770000041
wherein W is the overall accuracy of classification, D is the accuracy of judgment as a dangerous state, S is the accuracy of judgment as a safe state,
Figure BDA0003585065770000042
a vehicle whose surroundings pose no threat to the current vehicle, is considered as 0>
Figure BDA0003585065770000043
Greater than 0, a surrounding vehicle is considered, which represents a threat to the current vehicle>
Figure BDA0003585065770000044
The network collision risk under the safety condition at the moment t;
Figure BDA0003585065770000045
and the network collision risk of the current vehicle in the dangerous condition at the moment t.
In another aspect, a system for assessing risk of unmanned collision based on a dynamic bayesian network comprises:
the road traffic preprocessing module is used for acquiring the current road traffic condition and preprocessing the current road traffic condition to obtain a dividing performance index and the divided road traffic condition;
the network collision risk evaluation module is used for constructing a network collision risk evaluation model and obtaining a network collision risk according to the division performance indexes;
the vehicle collision risk estimation module is used for constructing a vehicle collision risk estimation model and evaluating the vehicle collision risk probability according to the network collision risk;
and the dynamic Bayesian network module is used for constructing a dynamic Bayesian network model and evaluating the vehicle collision risk probability in multiple periods.
The invention has the following beneficial effects:
collecting the current road traffic condition, preprocessing the current road traffic condition to obtain a division performance index and the divided road traffic condition, constructing a network collision risk evaluation model, and obtaining a network collision risk by using the division performance index; constructing a vehicle collision risk estimation model, and evaluating vehicle collision risk probability according to the network collision risk; constructing a dynamic Bayesian network model, and carrying out unmanned collision risk assessment on the vehicle collision risk probability in multiple periods; the collision risk of the automatic driving vehicle can be evaluated by constructing an interactive perception-dynamic Bayesian model and combining the influence of network-level risk, early warning can be performed according to the evaluation risk, certain reference is provided for promoting the development of the automatic driving in the aspect of safety, and the problem that the influence of the vehicle collision risk in the estimation process is not considered, namely the safety is not high enough, is solved.
Drawings
FIG. 1 is a flowchart illustrating steps of a method for evaluating a collision risk of an unmanned vehicle based on a dynamic Bayesian network according to the present invention;
fig. 2 is a schematic structural diagram of a dynamic bayesian network model according to an embodiment of the present invention.
Detailed Description
The following description of the embodiments of the present invention is provided to facilitate the understanding of the present invention by those skilled in the art, but it should be understood that the present invention is not limited to the scope of the embodiments, and it will be apparent to those skilled in the art that various changes may be made without departing from the spirit and scope of the invention as defined and defined by the appended claims, and all changes that can be made by the invention using the inventive concept are intended to be protected.
In one aspect, as shown in fig. 1, an embodiment of the present invention provides a method for assessing risk of unmanned collision based on a dynamic bayesian network, including the following steps:
s1, collecting current road traffic conditions, and preprocessing the current road traffic conditions to obtain dividing performance indexes and divided road traffic conditions;
in the embodiment of the invention, the road traffic condition can be divided into two conditions of easy collision and safety, which are obtained by a machine learning classifier, the common method generally comprises k neighbor, neural network, support vector machine and the like, a proper road traffic condition evaluation index is selected, certain traffic data is used to divide the traffic condition into two types through a machine learning classification process, namely: road traffic conditions in a dangerous state and road traffic conditions in a safe state.
Preferably, step S1 is specifically:
collecting current road traffic conditions, classifying the current road traffic conditions to obtain road traffic conditions in a dangerous state, road traffic conditions in a safe state and division performance indexes; wherein, the dividing performance indexes are respectively expressed as:
Figure BDA0003585065770000061
Figure BDA0003585065770000062
Figure BDA0003585065770000063
wherein W is the overall accuracy of classification, D is the accuracy of judgment as a dangerous state, S is the accuracy of judgment as a safe state, and T pz To correctly determine the number of dangerous states, T aq To correctly judge the number of safe states, F pz For the number of erroneously determined dangerous states, F aq The number of false positives as a safety state.
S2, constructing a network collision risk evaluation model, and obtaining a network collision risk according to the division performance indexes;
preferably, step S2 is specifically:
constructing a network collision risk evaluation model, and evaluating the probability of the traffic condition which is easy to collide in the road section in the current road traffic condition according to the division performance indexes to obtain the network collision risk, wherein a collision risk calculation formula in the network collision risk evaluation model is represented as follows:
Figure BDA0003585065770000071
Figure BDA0003585065770000072
the TC is a road traffic condition, the values of 0 and 1 respectively correspond to two situations, namely a safe state and a dangerous state, the NIR is a network-level risk condition, the P (NLR = "danger") is the probability of the network-level risk condition in the dangerous state, and the P (NLR = "safe") is the probability of the network-level risk condition in the safe state.
In the embodiment of the invention, when TC =1, the classification result of the road traffic condition is easy to collide, the probability of collision at the network level is the median of the overall accuracy and the accuracy of judging danger, and when TC =0, the classification result of the current road traffic condition is safe;
the network-level collision risk can be evaluated and the collision risk condition of the road traffic can be mastered by classifying the road traffic network into two categories of safety and danger and representing the collision risk of the road network by using the division performance indexes, namely the collision risk is known from the road network layer in the driving process and is used as a part of collision risk estimation, so that the comprehensive and comprehensive driving collision risk evaluation is facilitated.
S3, constructing a vehicle collision risk estimation model, and evaluating vehicle collision risk probability according to the network collision risk;
preferably, step S3 is specifically:
constructing a vehicle collision risk estimation model, wherein the model is expressed as:
Figure BDA0003585065770000073
wherein,
Figure BDA0003585065770000074
the probability of the collision risk of the vehicle under the dangerous condition of the current vehicle at the moment t, namely the probability of the current vehicle under the dangerous condition at the moment t under the condition of grasping the total condition of the vehicles around the moment (t-1), the total condition of the motion risks of the vehicles around the moment (t-1) and the network collision risk condition at the moment t, wherein N is the total number of the vehicles around the current vehicle, VLR t For the current vehicle situation at instant t,>
Figure BDA0003585065770000081
is the total status of the surrounding vehicle at time (t-1), based on>
Figure BDA0003585065770000082
Is the total condition of the risk of vehicle movement around the moment (t-1)>
Figure BDA0003585065770000083
In case of a network collision risk situation at time t->
Figure BDA0003585065770000084
And &>
Figure BDA0003585065770000085
Parameters of vehicle motion risk, surrounding vehicle risk and network collision risk are respectively set;
and calculating the probability of the vehicle collision risk by using the network collision risk and combining the vehicle collision risk estimation model.
Preferably, the network collision risk parameter in step S3
Figure BDA0003585065770000086
The calculation is expressed as:
Figure BDA0003585065770000087
wherein W is the overall accuracy of classification, D is the accuracy of judgment as a dangerous state, S is the accuracy of judgment as a safe state,
Figure BDA0003585065770000088
a vehicle whose surroundings pose no threat to the current vehicle, is considered as 0>
Figure BDA0003585065770000089
Greater than 0, a surrounding vehicle is considered, which represents a threat to the current vehicle>
Figure BDA00035850657700000810
Network collision risk under a safe condition at the moment t;
Figure BDA00035850657700000811
network collision risk of the current vehicle under a dangerous condition at the moment t;
in the embodiments of the present invention, wherein
Figure BDA00035850657700000812
There are four calculation formulas, which respectively correspond to four situations that the network collision risk condition is dangerous and safe, and whether the surrounding vehicles constitute a threat combination, and the judgment result is based on the judgment result>
Figure BDA00035850657700000813
A value of 0 indicates that there is no surrounding vehicle threatening the current vehicle,,, or>
Figure BDA00035850657700000814
When the current vehicle is more than 0, vehicles which form threats to the current vehicle exist around the vehicle; risk of vehicle movement
Figure BDA00035850657700000815
The current vehicle senses the speed of the nth vehicle around and the distance between the nth vehicle and the current vehicle by using a sensor to obtain the collision time of the nth vehicle, and judges the collision timeWhether the collision time is less than the critical collision time or not is judged to judge whether the vehicle threatens the current vehicle or not and correspondingly gets the judgment result>
Figure BDA0003585065770000091
Taking the value of (A); surrounding vehicle risk->
Figure BDA0003585065770000092
The method mainly comprises the steps that according to data interaction between a current vehicle and an nth vehicle around, whether dangerous driving conditions exist in the vehicle around is mastered, and then the value of the vehicle around is obtained; parameters of network collision risk
Figure BDA0003585065770000093
The method is mainly obtained by calculating the division performance index of the road traffic network condition.
In the embodiment of the invention, the vehicle motion risk parameter
Figure BDA0003585065770000094
Is characterized by:
Figure BDA0003585065770000095
wherein TTC is the time to collision, TTC C Critical collision time, when the collision time is less than the critical collision time, a dangerous condition, when f K Taking 1; when the collision time is not less than the critical collision time, a safe condition is established, in which f K Take 0.
In the embodiment of the invention, the surrounding vehicle risk parameter
Figure BDA0003585065770000096
Expressed as:
Figure BDA0003585065770000097
wherein, when dangerous driving action appears in vehicle around, be dangerous state, promptly:
Figure BDA0003585065770000098
when the dangerous driving behavior of the surrounding vehicle does not appear, the safe state is that: />
Figure BDA0003585065770000099
In the embodiment of the invention, the collision risk condition of the current vehicle is comprehensively estimated from three aspects of the surrounding vehicle level collision risk, the vehicle motion risk and the network level collision risk by means of data interaction and sensor perception in combination with the evaluation indexes of the network level risk in the step S2, the method for estimating the vehicle collision risk comprehensively considers the road traffic network condition, the surrounding vehicle driving state and the vehicle motion condition, and compared with the method for estimating the collision risk which only considers a single factor and the like, the method for estimating the vehicle collision risk is more comprehensive, when the current vehicle perceives the surrounding vehicle level collision risk and the vehicle motion risk to be only at a lower level, but the road traffic condition is poorer, the probability result obtained by estimating the collision risk of the current vehicle is at a relatively higher level due to the consideration of the network level collision risk, so as to remind a driver to keep alert, and ensure that the driving process is safer.
And S4, evaluating through a dynamic Bayesian network model based on the vehicle collision risk probability of not less than one time period.
In the embodiment of the invention, traffic condition data of 30 seconds, 1 minute, 3 minutes and 5 minutes are acquired at intervals, and the dynamic Bayesian network model is utilized in the whole driving process, so that the vehicle collision risk is dynamically estimated to master the collision risk condition in the current driving process of the vehicle, and the collision risk condition is correspondingly pre-warned;
the collision risk of the vehicle at the next moment can be estimated by using the current vehicle-level collision risk condition, the motion risk condition and the network-level collision risk condition estimated at the next moment. As shown in fig. 2, the collision risk of the current vehicle at the time t is calculated by using the collision risk of the vehicles around the time t-1, the motion risk of the vehicles around the time t-1 and the network-level risk at the time t as input through the vehicle collision risk estimation model constructed in the step S3; the collision risk of the current vehicle at the time t +1 is obtained by taking the collision risk of the vehicle around the time t, the vehicle motion risk around the time t and the network-level risk at the time t +1 as input and calculating through the vehicle collision risk estimation model constructed in the step S3, wherein the vehicle kinematic risks around the time t-1, the time t and the time t +1 are respectively obtained by sensor measurement and model calculation at the time t-1, the time t and the time t + 1. After a certain time interval is set, data interaction in the aspects of networks, vehicles, sensors and the like is carried out once every fixed time interval, and therefore the dynamic Bayesian network model can estimate the collision risk of the current vehicle.
An unmanned collision risk assessment system based on a dynamic Bayesian network, comprising:
the road traffic preprocessing module is used for acquiring the current road traffic condition and preprocessing the current road traffic condition to obtain a dividing performance index and the divided road traffic condition;
the network collision risk evaluation module is used for constructing a network collision risk evaluation model and obtaining a network collision risk according to the division performance indexes;
the vehicle collision risk estimation module is used for constructing a vehicle collision risk estimation model and evaluating the vehicle collision risk probability according to the network collision risk;
and the dynamic Bayesian network module is used for constructing a dynamic Bayesian network model and evaluating the vehicle collision risk probability in multiple periods.
The unmanned collision risk assessment system based on the dynamic Bayesian network provided by the embodiment of the invention has all the beneficial effects of the unmanned collision risk assessment method based on the dynamic Bayesian network.
The present invention is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
The principle and the implementation mode of the invention are explained by applying specific embodiments in the invention, and the description of the embodiments is only used for helping to understand the method and the core idea of the invention; meanwhile, for a person skilled in the art, according to the idea of the present invention, there may be variations in the specific embodiments and the application scope, and in summary, the content of the present specification should not be construed as a limitation to the present invention.
It will be appreciated by those of ordinary skill in the art that the embodiments described herein are intended to assist the reader in understanding the principles of the invention and are to be construed as being without limitation to such specifically recited embodiments and examples. Those skilled in the art can make various other specific changes and combinations based on the teachings of the present invention without departing from the spirit of the invention, and these changes and combinations are within the scope of the invention.

Claims (3)

1. The unmanned driving collision risk assessment method based on the dynamic Bayesian network is characterized by comprising the following steps of:
s1, collecting current road traffic conditions, and preprocessing the current road traffic conditions to obtain dividing performance indexes and divided road traffic conditions; the method specifically comprises the following steps:
collecting current road traffic conditions, classifying the current road traffic conditions to obtain road traffic conditions in a dangerous state, road traffic conditions in a safe state and division performance indexes; wherein, the dividing performance indexes are respectively expressed as:
Figure FDA0004056225480000011
Figure FDA0004056225480000012
Figure FDA0004056225480000013
wherein W is the overall accuracy of classification, D is the accuracy of judgment as a dangerous state, S is the accuracy of judgment as a safe state, and T pz To correctly determine the number of dangerous states, T aq To correctly judge the number of safe states, F pz Number of misjudged dangerous states，F aq The number of false positives as a safety state;
s2, constructing a network collision risk evaluation model, and obtaining a network collision risk according to the division performance indexes; the method specifically comprises the following steps:
constructing a network collision risk evaluation model, and evaluating the probability of the traffic condition which is easy to collide in the road section in the current road traffic condition according to the division performance indexes to obtain the network collision risk, wherein a collision risk calculation formula in the network collision risk evaluation model is represented as follows:
Figure FDA0004056225480000021
Figure FDA0004056225480000022
the TC is a road traffic condition, values of 0 and 1 respectively correspond to two conditions of a safe state and a dangerous state, the NLR is a network level risk condition, the P (NLR = "danger") is the probability of the network level risk condition in the dangerous state, and the P (NLR = "safe") is the probability of the network level risk condition in the safe state;
s3, constructing a vehicle collision risk estimation model, and evaluating vehicle collision risk probability according to the network collision risk; the method specifically comprises the following steps:
constructing a vehicle collision risk estimation model, wherein the model is expressed as follows:
Figure FDA0004056225480000023
/>
wherein,
Figure FDA0004056225480000024
the collision risk probability of the vehicle under the dangerous condition of the current vehicle at the moment t, N is the total number of the vehicles around the current vehicle, VLR t Is the current vehicle condition at time t,
Figure FDA0004056225480000025
is the total status of the surrounding vehicle at time (t-1), based on>
Figure FDA0004056225480000026
The total situation of the risk of movement of the vehicle around the time (t-1),
Figure FDA0004056225480000027
for the network collision risk situation at time t, device for selecting or keeping>
Figure FDA0004056225480000028
And &>
Figure FDA0004056225480000029
Parameters of vehicle motion risk, surrounding vehicle risk and network collision risk are respectively set;
calculating the probability of the vehicle collision risk by combining the network collision risk and a vehicle collision risk estimation model;
and S4, evaluating through a dynamic Bayesian network model based on the vehicle collision risk probability of not less than one time period.
2. The unmanned aerial vehicle collision risk assessment method based on dynamic Bayesian network as claimed in claim 1, wherein the network collision risk parameter calculation formula in step S3 is expressed as:
Figure FDA0004056225480000031
wherein W is the overall accuracy of classification, D is the accuracy of judgment as a dangerous state, S is the accuracy of judgment as a safe state,
Figure FDA0004056225480000032
a vehicle in the surroundings which does not pose a threat to the current vehicle, when 0>
Figure FDA0004056225480000033
Greater than 0 for a vehicle in the surroundings which poses a threat to the current vehicle, based on the vehicle status>
Figure FDA0004056225480000034
The network collision risk under the safety condition at the moment t;
Figure FDA0004056225480000035
and the network collision risk of the current vehicle in the dangerous condition at the moment t.
3. A system for unmanned collision risk assessment based on dynamic bayesian network using the method of claim 1, comprising:
the road traffic preprocessing module is used for acquiring the current road traffic condition and preprocessing the current road traffic condition to obtain a division performance index and divided road traffic conditions;
the network collision risk evaluation module is used for constructing a network collision risk evaluation model and obtaining a network collision risk according to the division performance indexes;
the vehicle collision risk estimation module is used for constructing a vehicle collision risk estimation model and evaluating the vehicle collision risk probability according to the network collision risk;
and the dynamic Bayesian network module is used for constructing a dynamic Bayesian network model and evaluating the vehicle collision risk probability in multiple periods.
CN202210363894.5A 2022-04-07 2022-04-07 Unmanned driving collision risk assessment method and system based on dynamic Bayesian network Active CN114662967B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202210363894.5A CN114662967B (en) 2022-04-07 2022-04-07 Unmanned driving collision risk assessment method and system based on dynamic Bayesian network

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202210363894.5A CN114662967B (en) 2022-04-07 2022-04-07 Unmanned driving collision risk assessment method and system based on dynamic Bayesian network

Publications (2)

Publication Number Publication Date
CN114662967A CN114662967A (en) 2022-06-24
CN114662967B true CN114662967B (en) 2023-04-07

Family

ID=82034413

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202210363894.5A Active CN114662967B (en) 2022-04-07 2022-04-07 Unmanned driving collision risk assessment method and system based on dynamic Bayesian network

Country Status (1)

Country Link
CN (1) CN114662967B (en)

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP2869283A1 (en) * 2013-10-31 2015-05-06 Inria Institut National de Recherche en Informatique et en Automatique Method and system of driving assistance for collision avoidance
CN105761548A (en) * 2016-04-14 2016-07-13 西安电子科技大学 Intersection collision-avoiding method based on dynamic Bayes network
WO2018220418A1 (en) * 2017-06-02 2018-12-06 Toyota Motor Europe Driving assistance method and system
CN110834644A (en) * 2019-10-30 2020-02-25 中国第一汽车股份有限公司 Vehicle control method and device, vehicle to be controlled and storage medium
EP3706034A1 (en) * 2019-03-06 2020-09-09 Robert Bosch GmbH Movement prediction of pedestrians useful for autonomous driving
CN112015843A (en) * 2020-09-02 2020-12-01 中国科学技术大学 Driving risk situation assessment method and system based on multi-vehicle intention interaction result
CN113792655A (en) * 2021-09-14 2021-12-14 京东鲲鹏(江苏)科技有限公司 Intention identification method and device, electronic equipment and computer readable medium

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US11577750B2 (en) * 2018-11-08 2023-02-14 Bayerische Motoren Werke Aktiengesellschaft Method and apparatus for determining a vehicle comfort metric for a prediction of a driving maneuver of a target vehicle

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP2869283A1 (en) * 2013-10-31 2015-05-06 Inria Institut National de Recherche en Informatique et en Automatique Method and system of driving assistance for collision avoidance
CN105761548A (en) * 2016-04-14 2016-07-13 西安电子科技大学 Intersection collision-avoiding method based on dynamic Bayes network
WO2018220418A1 (en) * 2017-06-02 2018-12-06 Toyota Motor Europe Driving assistance method and system
EP3706034A1 (en) * 2019-03-06 2020-09-09 Robert Bosch GmbH Movement prediction of pedestrians useful for autonomous driving
CN110834644A (en) * 2019-10-30 2020-02-25 中国第一汽车股份有限公司 Vehicle control method and device, vehicle to be controlled and storage medium
CN112015843A (en) * 2020-09-02 2020-12-01 中国科学技术大学 Driving risk situation assessment method and system based on multi-vehicle intention interaction result
CN113792655A (en) * 2021-09-14 2021-12-14 京东鲲鹏(江苏)科技有限公司 Intention identification method and device, electronic equipment and computer readable medium

Non-Patent Citations (4)

* Cited by examiner, † Cited by third party
Title
Dongye Sun 等.A highway crash risk assessment method based on traffic safety state division.《plos one》.2020,第1-14页. *
Mohammad Bahram 等.A Combined Model- and Learning-Based Framework for Interaction-Aware Maneuver Prediction.《IEEE TRANSACTIONS ON INTELLIGENT TRANSPORTATION SYSTEMS》.2016,第17卷(第6期),第1538-1550页. *
Wencheng Huang 等.Railway dangerous goods transportation system risk analysis: An Interpretive Structural Modeling and Bayesian Network combining approach.《Reliability Engineering and System Safety》.2020,第204卷第1-10页. *
仝兆景 等.基于 DBN 的特种车辆前向防撞推理模型.《测控技术》.2019,第38卷(第10期),第56-60页. *

Also Published As

Publication number Publication date
CN114662967A (en) 2022-06-24

Similar Documents

Publication Publication Date Title
CN106314428B (en) Collision avoidance system, collision avoidance method and motor vehicle
Åsljung et al. Using extreme value theory for vehicle level safety validation and implications for autonomous vehicles
US7579942B2 (en) Extra-vehicular threat predictor
Peng et al. Uncertainty evaluation of object detection algorithms for autonomous vehicles
CN104182618B (en) A kind of method for early warning that knocks into the back based on Bayesian network
EP1716540B1 (en) System and method for detecting a passing vehicle from dynamic background using robust information fusion
CN111469802A (en) Seat belt state determination system and method
US20150200957A1 (en) Detection of false vehicle-to-vehicle emergency brake light messages
US9950700B2 (en) Road surface condition detection with multi-scale fusion
US9064419B2 (en) Driver condition assessment device
Peng et al. Intelligent method for identifying driving risk based on V2V multisource big data
EP3940487B1 (en) Estimation of probability of collision with increasing severity level for autonomous vehicles
Muehlfeld et al. Statistical behavior modeling for driver-adaptive precrash systems
CN116901975B (en) Vehicle-mounted AI security monitoring system and method thereof
CN111860210A (en) Method and device for detecting separation of hands from steering wheel, electronic equipment and storage medium
CN110901385A (en) Active speed limiting method based on fatigue state of driver
Kuhn et al. Introspective failure prediction for autonomous driving using late fusion of state and camera information
CN110097571B (en) Quick high-precision vehicle collision prediction method
CN111563468A (en) Driver abnormal behavior detection method based on attention of neural network
CN118014349A (en) Tunnel accident risk assessment method, system and equipment
CN114662967B (en) Unmanned driving collision risk assessment method and system based on dynamic Bayesian network
CN115865988B (en) Passenger ship passenger tread event monitoring system and method by utilizing mobile phone sensor network
CN116597516A (en) Training method, classification method, detection method, device, system and equipment
CN113283286B (en) Driver abnormal behavior detection method and device
Popescu et al. Cut-in maneuver recognition and behavior generation using bayesian networks and fuzzy logic

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