CN115273456A - Method and system for judging illegal driving of two-wheeled electric vehicle and storage medium - Google Patents

Method and system for judging illegal driving of two-wheeled electric vehicle and storage medium Download PDF

Info

Publication number
CN115273456A
CN115273456A CN202210685350.0A CN202210685350A CN115273456A CN 115273456 A CN115273456 A CN 115273456A CN 202210685350 A CN202210685350 A CN 202210685350A CN 115273456 A CN115273456 A CN 115273456A
Authority
CN
China
Prior art keywords
electric vehicle
wheeled electric
network model
target
driving
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
CN202210685350.0A
Other languages
Chinese (zh)
Other versions
CN115273456B (en
Inventor
孙宁
宋娟
姜川
孙晓琳
马梦丽
孟维宇
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.)
Beijing Chewang Technology Development Co ltd
Original Assignee
Beijing Chewang Technology Development Co ltd
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 Beijing Chewang Technology Development Co ltd filed Critical Beijing Chewang Technology Development Co ltd
Priority to CN202210685350.0A priority Critical patent/CN115273456B/en
Publication of CN115273456A publication Critical patent/CN115273456A/en
Application granted granted Critical
Publication of CN115273456B publication Critical patent/CN115273456B/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
    • G08G1/0129—Traffic data processing for creating historical data or processing based on historical data
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G06N3/084—Backpropagation, e.g. using gradient descent
    • 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
    • G08G1/01—Detecting movement of traffic to be counted or controlled
    • G08G1/017—Detecting movement of traffic to be counted or controlled identifying vehicles
    • 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

  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Engineering & Computer Science (AREA)
  • Analytical Chemistry (AREA)
  • Theoretical Computer Science (AREA)
  • Chemical & Material Sciences (AREA)
  • Computational Linguistics (AREA)
  • General Health & Medical Sciences (AREA)
  • Biomedical Technology (AREA)
  • Biophysics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Data Mining & Analysis (AREA)
  • Evolutionary Computation (AREA)
  • Artificial Intelligence (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Health & Medical Sciences (AREA)
  • Traffic Control Systems (AREA)
  • Electric Propulsion And Braking For Vehicles (AREA)

Abstract

The invention relates to a method, a system and a storage medium for judging illegal driving of a two-wheeled electric vehicle, wherein the method comprises the following steps: training an originally generated confrontation network model based on collected historical driving data of the two-wheeled electric vehicle to obtain a target generated confrontation network model; substituting the current driving data of the two-wheeled electric vehicle to be tested into a target to generate a confrontation network model, and obtaining a current violation index of the two-wheeled electric vehicle to be tested; and judging whether the current violation index is larger than a preset violation index, and if so, judging that the two-wheeled electric vehicle to be detected has violation driving. According to the method, the characteristics that the confrontation network model is generated, a labeling training set does not need to be constructed, modeling is carried out on complex high-dimensional data, and the proportion of illegal driving data of the two-wheeled electric vehicle in an original training data set is small are utilized, so that the process of manually screening combination characteristics is omitted, and the accuracy of the recognition result of the illegal driving behavior of the two-wheeled electric vehicle is improved.

Description

Method and system for judging illegal driving of two-wheeled electric vehicle and storage medium
Technical Field
The invention relates to the technical field of vehicle supervision, in particular to a method and a system for judging illegal driving of a two-wheeled electric vehicle and a storage medium.
Background
At present, in order to solve various potential safety hazards existing in the practical use of the two-wheeled electric vehicle, related departments have developed various specifications to constrain the behavior of the two-wheeled electric vehicle, and in order to guarantee the effective execution of the specifications, a supervision method needs to be introduced, so that when the two-wheeled electric vehicle enters an illegal state, the illegal behavior is timely recognized and corresponding prompt and record are made. The traditional method arranges manpower to identify violations, but the method has high manpower cost. The other method is based on the traditional supervised or unsupervised machine learning method, but the supervised machine learning method needs to manually label a large amount of two-wheel vehicle operation historical data according to driving specifications, so that the cost is high, the coverage rate is limited, and the training requirements of a supervised machine learning model are often not met; although the unsupervised machine learning method can avoid the construction of a label set, the existing unsupervised method still has the problems of poor modeling performance on complex high-dimensional data and the like; meanwhile, the proportion of illegal behaviors in a sample set is small, and the traditional machine learning method usually requires that the quantity of positive and negative samples is equal to achieve a relatively ideal effect.
Therefore, it is necessary to provide a technical solution to solve the existing problems.
Disclosure of Invention
In order to solve the technical problem, the invention provides a method and a system for judging illegal driving of a two-wheeled electric vehicle and a storage medium.
The invention discloses a method for judging illegal driving of a two-wheeled electric vehicle, which adopts the technical scheme as follows:
training an originally generated confrontation network model based on collected historical driving data of a plurality of two-wheeled electric vehicles to obtain a target generated confrontation network model;
acquiring current driving data of a two-wheeled electric vehicle to be tested, substituting the current driving data into the target to generate an confrontation network model, and obtaining a current violation index of the two-wheeled electric vehicle to be tested;
and judging whether the current violation index is larger than a preset violation index to obtain a judgment result, and if so, judging that the two-wheeled electric vehicle to be detected runs in violation.
The method for judging illegal driving of the two-wheeled electric vehicle has the following beneficial effects:
according to the method, by utilizing the characteristics that the generated countermeasure network model does not need to construct a labeling training set and is modeled on complex high-dimensional data and according to the characteristic that the proportion of the illegal driving data of the two-wheeled electric vehicle in the original training data set is small, the method provided by the invention saves the process of manually screening the combination characteristics and improves the accuracy of the recognition result of the illegal driving behavior of the two-wheeled electric vehicle.
On the basis of the scheme, the method for judging illegal driving of the two-wheeled electric vehicle can be further improved as follows.
Further, the original generator of the originally generated confrontation network model adopts a laminated stride convolution layer, and the original discriminator of the originally generated confrontation network model adopts a standard CNN network structure.
Further, the training of the countermeasure network model generated according to all the historical driving data and the original generation countermeasure network model to obtain the target generation countermeasure network model comprises the following steps:
training the original generator and the original discriminator based on a training optimization function and all historical driving data to obtain a target generator and a target discriminator;
obtaining the target generation confrontation network model according to the target generator and the target discriminator; wherein the training optimization function is:
Figure BDA0003695566670000021
Figure BDA0003695566670000022
Figure BDA0003695566670000023
for the training optimization function, x is a feature vector formed by any historical driving data, D1As the primitive discriminator, G1For the primitive generator, z is the input variable of the primitive generation countermeasure network, pdata(x) Distribution obeyed by x, D1(x) Is the output result of x after passing through the original discriminator,
Figure BDA0003695566670000024
denotes when x obeys pdata(x) When being distributed, D1(x) Expected value of the logarithm of (1), pz(z) is the uniform distribution obeyed by z,
Figure BDA0003695566670000031
denotes when z obeys the distribution pzWhen (z) is 1-D1(G1(z)) a desired value of the logarithm of the value.
Further, the substituting the current driving data into the target to generate a confrontation network model to obtain a current violation index of the two-wheeled electric vehicle to be tested includes:
substituting the current driving data into an objective function corresponding to the objective generation countermeasure network model for iterative calculation until the objective function value is minimum, and obtaining a current violation index of the two-wheeled electric vehicle to be tested; wherein the objective function is: l (z) = (1- λ) ∑ x1- G2(z)|+λ∑|f(x1)-f(G2(z))|；x1For the current driving data, G2For the object generator, G2(z) is the data generated after the input variable z is acted on by the target generator, f isThe object discriminator D2L (z) is the objective function, for measuring G2(z) and x1Similarity of (2); λ is a predetermined weight, L (z)γ) Is the current violation index, zγThe z value for the time when L (z) converged to a minimum.
Further, the historical driving data of any two-wheeled electric vehicle comprises: driving speed, driving longitude, driving latitude, heading angle, driving gear, vehicle lamp state, longitudinal acceleration, transverse acceleration, motor output speed, motor torque, yaw angular velocity, front wheel speed, rear wheel speed, battery remaining capacity and road direction.
Further, still include: and when the judgment result is yes, sending violation prompt information to the two-wheeled electric vehicle to be detected.
The invention discloses a system for judging illegal driving of a two-wheeled electric vehicle, which adopts the technical scheme as follows:
the method comprises the following steps: the device comprises a construction module, an operation module and a judgment module;
the building module is used for: training an originally generated confrontation network model based on collected historical driving data of the two-wheeled electric vehicle to obtain a target generated confrontation network model;
the operation module is used for: acquiring current driving data of a two-wheeled electric vehicle to be detected, substituting the current driving data into the target to generate a confrontation network model, and obtaining a current violation index of the two-wheeled electric vehicle to be detected;
the determination module is configured to: and judging whether the current violation index is greater than a preset violation index to obtain a judgment result, and if so, judging that the two-wheeled electric vehicle to be detected has violation driving.
The system for judging illegal driving of the two-wheeled electric vehicle has the following beneficial effects:
the system provided by the invention utilizes the characteristics that the generated countermeasure network model does not need to construct a labeling training set and is modeled on complex high-dimensional data, and according to the characteristic that the proportion of the illegal driving data of the two-wheeled electric vehicle in the original training data set is very small, the system provided by the invention saves the process of manually screening the combination characteristics, and simultaneously improves the accuracy of the recognition result of the illegal driving behavior of the two-wheeled electric vehicle.
On the basis of the scheme, the system for judging illegal driving of the two-wheeled electric vehicle can be further improved as follows.
Further, the original generator of the originally generated confrontation network model adopts a laminated stride convolution layer, and the original discriminator of the originally generated confrontation network model adopts a standard CNN network structure.
Further, the building module is specifically configured to:
training the original generator and the original arbiter based on a training optimization function and all historical driving data to obtain a target generator and a target arbiter;
obtaining the target generation confrontation network model according to the target generator and the target discriminator; wherein the training optimization function is:
Figure BDA0003695566670000041
Figure BDA0003695566670000042
Figure BDA0003695566670000043
for the training optimization function, x is a feature vector formed by any historical driving data, D1As the original discriminator, G1For the primitive generator, z is the input variable of the primitive generation countermeasure network, pdata(x) Distribution obeyed by x, D1(x) Is the output result of x after passing through the original discriminator,
Figure BDA0003695566670000044
denotes when x obeys pdata(x) When being distributed, D1(x) Expected value of the logarithm of (1), pz(z) is the uniform distribution obeyed by z,
Figure BDA0003695566670000045
denotes when z obeys distribution pzWhen (z) is 1-D1(G1(z)) a desired value of the logarithm of the value.
The technical scheme of the storage medium of the invention is as follows:
the storage medium stores instructions that, when read by the computer, cause the computer to execute the steps of a method for determining illegal driving of a two-wheeled electric vehicle according to the present invention.
Drawings
Fig. 1 is a schematic flow chart of a method for determining illegal driving of a two-wheeled electric vehicle according to an embodiment of the present invention;
fig. 2 is a schematic structural diagram of a system for determining illegal driving of a two-wheeled electric vehicle according to an embodiment of the present invention.
Detailed Description
As shown in fig. 1, a method for determining illegal driving of a two-wheeled electric vehicle according to an embodiment of the present invention includes the following steps:
s1, training an originally generated confrontation network model based on collected historical driving data of a plurality of two-wheeled electric vehicles to obtain a target generated confrontation network model.
Wherein, the historical driving data of arbitrary two-wheeled electric motor car all includes: speed of travel X0Longitude X1Latitude X2Course angle X3Speed X4Lamp state X of vehicle5Longitudinal acceleration X6Transverse acceleration X7Output speed X of the motor8Motor torque X9Yaw angular velocity X10Front wheel speed X11Speed of rear wheel X12And remaining battery capacity X13In the direction of the road X14。
Wherein the originally generated antagonistic network model is Generated Antagonistic Network (GAN).
The method comprises the following steps of constructing and generating an original input feature vector X of a countermeasure network through any historical driving data; x = [ X =0，X1，X2，X3，X4，X5，X6，X7，X8，X9，X10，X11，X12，X13， X14]And combining each X of the constructs into a training data set that originally generated the confrontation network model.
Wherein the target generation countermeasure network model is: and training the originally generated confrontation network model through a training data set to obtain the generated confrontation network model.
S2, obtaining current driving data of the two-wheeled electric vehicle to be tested, substituting the current driving data into the target to generate a confrontation network model, and obtaining a current violation index of the two-wheeled electric vehicle to be tested.
The two-wheeled electric vehicle to be tested is the two-wheeled electric vehicle needing to be judged, the data types selected by the current driving data and the historical driving data are the same, and the current driving data and the historical driving data all comprise the 14 index characteristics.
Wherein, the current violation index is: and generating a value corresponding to the closest point in the distribution of the characteristic vector X and the G of the two-wheeled electric vehicle to be detected by carrying out reverse iteration on the target function corresponding to the confrontation network model according to the target.
And S3, judging whether the current violation index is larger than a preset violation index to obtain a judgment result, and if so, judging that the two-wheeled electric vehicle to be detected has violation driving.
The preset violation index is a value preset by a user, and can be adjusted according to the severity of violation judgment in actual use.
Specifically, if the current violation index is larger than the preset violation index, the two-wheeled electric vehicle is judged to be in violation; otherwise, judging that the two-wheeled electric vehicle runs normally.
Preferably, the original generator of the originally generated confrontation network model adopts a laminated stride convolution layer, and the original discriminator of the originally generated confrontation network model adopts a standard CNN network structure.
The primitive generation confrontation network model includes a generator (corresponding to the primitive generator in the present embodiment) and a discriminator (corresponding to the primitive discriminator in the present embodiment).
In particular, the primitive generator G1For learning the distribution of the potentially uniformly distributed random variables z to x in the training dataset, a convolutional decoder network constructed from a stride convolutional layer stack is selected as the original generator in this embodiment; wherein the stack represents the way the decoder network is built as a raw generator; the step-by-step convolutional layer represents that when the convolutional layer performs mapping operation on input data, the step length of the convolutional kernel moved each time is larger than 1 (namely, the convolutional kernel moved has the characteristic of step-by-step movement). Primitive discriminator D1For outputting a probability estimation that a feature vector X representing a given input (e.g., a feature vector X corresponding to current driving data of the two-wheeled electric vehicle to be tested in this embodiment) is derived from training data in a training data set or from G (z) generation, a standard CNN network structure is selected as an original discriminator in this embodiment. The standard CNN network belongs to the category of neural networks, and specifically, a network structure is formed by stacking a group of convolutional layers, so as to implement mapping from an input space to an output space.
Preferably, the training to obtain the target generation countermeasure network model according to all the historical driving data and the originally generated countermeasure network model includes:
and training the original generator and the original arbiter based on a training optimization function and all historical driving data to obtain a target generator and a target arbiter.
And obtaining the target generation confrontation network model according to the target generator and the target discriminator. Wherein the training optimization function is:
Figure BDA0003695566670000071
Figure BDA0003695566670000072
Figure BDA0003695566670000073
for the training optimization function, xFeature vectors formed for any historical driving data, D1As the primitive discriminator, G1For the primitive generator, z is the input variable of the primitive generation countermeasure network, pdata(x) Distribution obeyed by x, D1(x) Is the output result of x after passing through the original discriminator,
Figure BDA0003695566670000074
denotes when x obeys pdata(x) When being distributed, D1(x) Expected value of the logarithm of (1), pz(z) is the uniform distribution obeyed by z,
Figure BDA0003695566670000075
denotes when z obeys the distribution pzWhen (z) is 1-D1(G1(z)) a desired value of the logarithm of the value.
Wherein, a training optimization function pair G is adopted1And D1Training the parameters to obtain a generator G meeting the user requirements2And discriminator D2。
It should be noted that the training process for the raw generator and the raw discriminator in the present embodiment is the prior art. Specifically, as can be seen from the above training optimization function, the training process includes the following three steps:
(1) the original discriminator training process is as follows: fixed primitive generator G1Iteratively adjusts the original discriminator D1Up to V (D)1，G1) The value converges to a maximum value;
(2) the training process of the raw generator is as follows: fixed primitive discriminator D1Parametric, iterative adjustment of the original Generator G1Up to V (D)1，G1) The value converges to a minimum value;
(3) the process of steps 1 and 2 is cyclically executed until G1And V (D)1，G1) Until the parameters of (2) converge steadily.
Preferably, the substituting the current driving data into the target generation countermeasure network model to obtain the current violation index of the two-wheeled electric vehicle to be tested includes:
substituting the current driving data into a target function corresponding to the target generation countermeasure network model for iterative computation until the target function value is minimum, and obtaining a current violation index of the two-wheeled electric vehicle to be tested; wherein the objective function is: l (z) = (1- λ) ∑ x1- G2(z)|+λ∑|f(x1)-f(G2(z))|；x1For the current driving data, G2For the object generator, G2(z) is data generated after the input variable z is acted by the target generator, f is the target discriminator D2L (z) is the objective function, for measuring G2(z) and x1The similarity of (2); λ is a predetermined weight, L (z)γ) Is the current violation index, zγThe z value corresponding to the minimum convergence of L (z).
Wherein L (z) is used for measuring data G generated after any z is acted by a target generator2(z) and x1To a similar extent, indirectly react x1Belonging to the degree of normal driving data, so that the larger L (z), the larger x1The greater the likelihood of a drive violation.
It should be noted that the objective function in this embodiment is obtained by performing weighted average on two indexes. The first part is: sigma | x1-G2(z) |, used for two-wheeled electric vehicle x to be measured1And z via target generator G2The result of the action is expressed in x1The distance in the feature space is the similarity of scale measurement, and the similarity describes the similarity under the view angle of the target generator; the second part is Σ | f (x)1)- f(G2(z)) |, for measuring the x of two-wheeled electric vehicle to be measured1And target generator G2The similarity of the generated result in the space after the intermediate layer f of the target discriminator is mapped, wherein the similarity describes the similarity under the visual angle of the target discriminator; and obtaining the fused objective function L (z) by carrying out weighted average on the two indexes according to the parameter lambda.
In this embodiment, the trained target generation countermeasure network model is integrated into the roadside device, and when the roadside device receives the current driving data broadcasted by the peripheral two-wheeled electric vehicles through the V2X BSM message in real time through the PC5 interface, the road testing device can monitor the illegal driving behaviors of the two-wheeled electric vehicles on both sides of the road.
Preferably, the historical driving data of any two-wheeled electric vehicle comprises: driving speed, driving longitude, driving latitude, heading angle, driving gear, lamp state, longitudinal acceleration, lateral acceleration, motor output speed, motor torque, yaw rate, front wheel speed, rear wheel speed, battery remaining capacity and road direction.
Preferably, the method further comprises the following steps: and when the judgment result is yes, sending violation prompt information to the two-wheeled electric vehicle to be detected.
Specifically, when judging that the two-wheeled electric vehicle to be detected runs in an illegal manner, sending illegal prompt information to the two-wheeled electric vehicle to be detected, wherein the prompt information can be: informing the driver of safety, etc. For example, when the road test equipment determines that a certain two-wheeled electric vehicle runs illegally, the road test equipment sends prompt information back to the illegal two-wheeled electric vehicle through the PC5 interface and informs a driver of safety through the human-machine interaction interface HMI.
Besides sending violation prompt information to the two-wheeled electric vehicle to be tested, the two-wheeled electric vehicle to be tested can be registered or manually checked to further confirm whether the two-wheeled electric vehicle runs in violation or not.
According to the technical scheme, the characteristics that the anti-network model is generated, a labeling training set does not need to be constructed, modeling is carried out on complex high-dimensional data, and the characteristic that the proportion of the illegal driving data of the two-wheeled electric vehicle in an original training data set is small is utilized.
Fig. 2 is a system 200 for determining illegal driving of a two-wheeled electric vehicle according to an embodiment of the present invention, including: a construction module 210, an operation module 220, and a determination module 230;
the building module 210 is configured to: training an originally generated confrontation network model based on collected historical driving data of a plurality of two-wheeled electric vehicles to obtain a target generated confrontation network model;
the operation module 220 is configured to: acquiring current driving data of a two-wheeled electric vehicle to be tested, substituting the current driving data into the target to generate an confrontation network model, and obtaining a current violation index of the two-wheeled electric vehicle to be tested;
the determination module 230 is configured to: and judging whether the current violation index is larger than a preset violation index to obtain a judgment result, and if so, judging that the two-wheeled electric vehicle to be detected runs in violation.
Preferably, the original generator of the originally generated confrontation network model adopts a laminated stride convolution layer, and the original discriminator of the originally generated confrontation network model adopts a standard CNN network structure.
Preferably, the building module is specifically configured to:
training the original generator and the original arbiter based on a training optimization function and all historical driving data to obtain a target generator and a target arbiter;
obtaining the target generation confrontation network model according to the target generator and the target discriminator; wherein the training optimization function is:
Figure BDA0003695566670000101
Figure BDA0003695566670000102
Figure BDA0003695566670000103
for the training optimization function, x is a feature vector formed by any historical driving data, D1As the original discriminator, G1For the primitive generator, z is the input variable of the primitive generation countermeasure network, pdata(x) Distribution obeyed by x, D1(x) Is the output result of x after passing through the original discriminator,
Figure BDA0003695566670000104
denotes when x obeys pdata(x) When being distributed, D1(x) Expected value of the logarithm of (1), pz(z) is the uniform distribution obeyed by z,
Figure BDA0003695566670000105
denotes when z obeys the distribution pzWhen (z) is 1-D1(G1(z)) a desired value of the logarithm of the value.
According to the technical scheme, the characteristics that the anti-network model is generated, a labeling training set does not need to be constructed, modeling is carried out on complex high-dimensional data, and the characteristic that the proportion of the illegal driving data of the two-wheeled electric vehicle in an original training data set is small is utilized.
For the above steps for realizing the corresponding functions of each parameter and each module in the system 200 for determining illegal driving of a two-wheeled electric vehicle according to the present invention, reference may be made to each parameter and step in the above embodiment of a method for determining illegal driving of a two-wheeled electric vehicle, which are not described herein again.
An embodiment of the present invention provides a storage medium, including: the storage medium stores instructions, and when the computer reads the instructions, the computer is caused to execute the steps of the method for determining illegal driving of a two-wheeled electric vehicle as described above, which may specifically refer to the parameters and steps in the embodiment of the method for determining illegal driving of a two-wheeled electric vehicle, and are not described herein again.
Computer storage media such as: flash disks, portable hard disks, and the like.
As will be appreciated by one skilled in the art, the present invention may be embodied as methods, systems, and storage media.
Thus, the present invention may be embodied in the form of: the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment (including firmware, resident software, micro-code, etc.) or an embodiment combining software and hardware aspects that may all generally be referred to herein as a "circuit," module "or" system. Furthermore, in some embodiments, the invention may also be embodied in the form of a computer program product in one or more computer-readable media having computer-readable program code embodied in the medium. Any combination of one or more computer-readable media may be employed. The computer readable medium may be a computer readable signal medium or a computer readable storage medium. A computer readable storage medium may be, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the foregoing. More specific examples (a non-exhaustive list) of the computer readable storage medium would include the following: an electrical connection having one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing. In the context of this document, a computer readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device. Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention, and that variations, modifications, substitutions and alterations can be made to the above embodiments by those of ordinary skill in the art within the scope of the present invention.

Claims (10)

1. A two-wheeled electric vehicle illegal driving judgment method is characterized by comprising the following steps:
training an originally generated confrontation network model based on collected historical driving data of the two-wheeled electric vehicle to obtain a target generated confrontation network model;
acquiring current driving data of a two-wheeled electric vehicle to be detected, substituting the current driving data into the target to generate a confrontation network model, and obtaining a current violation index of the two-wheeled electric vehicle to be detected;
and judging whether the current violation index is greater than a preset violation index to obtain a judgment result, and if so, judging that the two-wheeled electric vehicle to be detected has violation driving.
2. The two-wheeled electric vehicle illegal driving judgment method according to claim 1, characterized in that an original generator of the original generation countermeasure network model adopts a laminated stride convolution layer, and an original discriminator of the original generation countermeasure network model adopts a standard CNN network structure.
3. The method for judging illegal driving of the two-wheeled electric vehicle according to claim 2, wherein the training of the target generation countermeasure network model according to all historical driving data and the original generation countermeasure network model comprises:
training the original generator and the original discriminator based on a training optimization function and all historical driving data to obtain a target generator and a target discriminator;
obtaining the target generation confrontation network model according to the target generator and the target discriminator; wherein the training optimization function is:
Figure FDA0003695566660000011
Figure FDA0003695566660000012
Figure FDA0003695566660000013
for the training optimization function, x is a feature vector formed by any historical driving data, D1As the primitive discriminator, G1For the primitive generator, z is the input variable of the primitive generation countermeasure network, pdata(x) Distribution obeyed by x, D1(x) X is passed through the original discriminatorThe result of the latter output is that,
Figure FDA0003695566660000014
denotes when x obeys pdata(x) When being distributed, D1(x) Expected value of the logarithm of (1), pz(z) is the uniform distribution obeyed by z,
Figure FDA0003695566660000021
denotes when z obeys distribution pzWhen (z) is 1-D1(G1(z)) expected value of the log value.
4. The method for determining illegal driving of the two-wheeled electric vehicle according to claim 3, wherein the step of substituting the current driving data into the target generation countermeasure network model to obtain a current violation index of the two-wheeled electric vehicle to be tested comprises the steps of:
substituting the current driving data into an objective function corresponding to the objective generation countermeasure network model for iterative calculation until the objective function value is minimum, and obtaining a current violation index of the two-wheeled electric vehicle to be tested; wherein the objective function is: l (z) = (1- λ) ∑ x1-G2(z)|+λ∑|f(x1)-f(G2(z))|；x1For the current driving data, G2For the object generator, G2(z) is data generated by the input variable z being acted on by the target generator, f is the target discriminator D2L (z) is the objective function, for measuring G2(z) and x1Similarity of (2); λ is a predetermined weight, L (z)γ) Is the current violation index, zγThe z value corresponding to the minimum convergence of L (z).
5. The method for determining illegal driving of the two-wheeled electric vehicle according to claim 1, wherein the historical driving data of any two-wheeled electric vehicle comprises: driving speed, driving longitude, driving latitude, heading angle, driving gear, vehicle lamp state, longitudinal acceleration, transverse acceleration, motor output speed, motor torque, yaw angular velocity, front wheel speed, rear wheel speed, battery remaining capacity and road direction.
6. The method for determining illegal driving of a two-wheeled electric vehicle according to any one of claims 1 to 5, further comprising: and when the judgment result is yes, sending violation prompt information to the two-wheeled electric vehicle to be detected.
7. A two-wheeled electric vehicle illegal driving judgment system is characterized by comprising: the device comprises a construction module, an operation module and a judgment module;
the building module is used for: training an originally generated confrontation network model based on collected historical driving data of the two-wheeled electric vehicle to obtain a target generated confrontation network model;
the operation module is used for: acquiring current driving data of a two-wheeled electric vehicle to be tested, substituting the current driving data into the target to generate an confrontation network model, and obtaining a current violation index of the two-wheeled electric vehicle to be tested;
the determination module is configured to: and judging whether the current violation index is greater than a preset violation index to obtain a judgment result, and if so, judging that the two-wheeled electric vehicle to be detected has violation driving.
8. The two-wheeled electric vehicle illegal driving judgment system according to claim 7, characterized in that the original generator of the original generation countermeasure network model adopts a laminated stride convolution layer, and the original discriminator of the original generation countermeasure network model adopts a standard CNN network structure.
9. The system for determining illegal driving of a two-wheeled electric vehicle according to claim 8, wherein the construction module is specifically configured to:
training the original generator and the original arbiter based on a training optimization function and all historical driving data to obtain a target generator and a target arbiter;
obtaining the target generation confrontation network model according to the target generator and the target discriminator; wherein the training optimization function is:
Figure FDA0003695566660000031
Figure FDA0003695566660000032
Figure FDA0003695566660000033
for the training optimization function, x is a feature vector formed by any historical driving data, D1As the primitive discriminator, G1For the primitive generator, z is the input variable of the primitive generation countermeasure network, pdata(x) Distribution obeyed by x, D1(x) Is the output result of x after passing through the original discriminator,
Figure FDA0003695566660000034
denotes when x obeys pdata(x) When being distributed, D1(x) Expected value of the logarithm of (1), pz(z) is the uniform distribution obeyed by z,
Figure FDA0003695566660000035
denotes when z obeys the distribution pzWhen (z) is 1-D1(G1(z)) expected value of the log value.
10. A storage medium having stored therein instructions that, when read by a computer, cause the computer to execute a two-wheeled electric vehicle illegal driving determination method according to any one of claims 1 to 6.
CN202210685350.0A 2022-06-14 2022-06-14 Method, system and storage medium for judging illegal running of two-wheeled electric vehicle Active CN115273456B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202210685350.0A CN115273456B (en) 2022-06-14 2022-06-14 Method, system and storage medium for judging illegal running of two-wheeled electric vehicle

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202210685350.0A CN115273456B (en) 2022-06-14 2022-06-14 Method, system and storage medium for judging illegal running of two-wheeled electric vehicle

Publications (2)

Publication Number Publication Date
CN115273456A true CN115273456A (en) 2022-11-01
CN115273456B CN115273456B (en) 2023-08-29

Family

ID=83762090

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202210685350.0A Active CN115273456B (en) 2022-06-14 2022-06-14 Method, system and storage medium for judging illegal running of two-wheeled electric vehicle

Country Status (1)

Country Link
CN (1) CN115273456B (en)

Citations (17)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103093624A (en) * 2013-01-09 2013-05-08 东南大学 Signalized intersection non-motor vehicle illegal cross-street behavior automatic judging method
AU2014218477A1 (en) * 2008-01-07 2014-09-18 Chargepoint, Inc. Network-controlled charging system for electric vehicles
CN108875766A (en) * 2017-11-29 2018-11-23 北京旷视科技有限公司 Method, apparatus, system and the computer storage medium of image procossing
US20190094867A1 (en) * 2017-09-28 2019-03-28 Nec Laboratories America, Inc. Generative adversarial inverse trajectory optimization for probabilistic vehicle forecasting
US20190107405A1 (en) * 2017-04-27 2019-04-11 Beijing Didi Infinity Technology And Development Co., Ltd. Systems and methods for route planning
CN110225446A (en) * 2018-03-01 2019-09-10 北京嘀嘀无限科技发展有限公司 A kind of system, method, apparatus and storage medium identifying driving behavior
CN110717433A (en) * 2019-09-30 2020-01-21 华中科技大学 A traffic violation analysis method and device based on deep learning
WO2020029356A1 (en) * 2018-08-08 2020-02-13 杰创智能科技股份有限公司 Method employing generative adversarial network for predicting face change
CN112017447A (en) * 2020-08-20 2020-12-01 北京赛博星通科技有限公司 Method and system for judging vehicle converse violation based on GPS position information
CN112149511A (en) * 2020-08-27 2020-12-29 深圳市点创科技有限公司 Method, terminal and device for detecting violation of driver based on neural network
CN112464749A (en) * 2020-11-11 2021-03-09 鹏城实验室 Traffic scene abnormal target detection method based on rules and learning
CN112466003A (en) * 2019-09-06 2021-03-09 顺丰科技有限公司 Vehicle state detection method, device, server and storage medium
CN113160575A (en) * 2021-03-15 2021-07-23 超级视线科技有限公司 Traffic violation detection method and system for non-motor vehicles and drivers
CN113380021A (en) * 2020-03-10 2021-09-10 深圳市丰驰顺行信息技术有限公司 Vehicle state detection method, device, server and computer-readable storage medium
CN113450571A (en) * 2021-09-01 2021-09-28 深圳市鼎粤科技有限公司 Traffic intersection-based driving direction reminding method and device and storage medium
CN113936465A (en) * 2021-10-26 2022-01-14 公安部道路交通安全研究中心 Traffic incident detection method and device
CN114548298A (en) * 2022-02-25 2022-05-27 阿波罗智联(北京)科技有限公司 Model training method, traffic information processing method, device, equipment and storage medium

Patent Citations (17)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
AU2014218477A1 (en) * 2008-01-07 2014-09-18 Chargepoint, Inc. Network-controlled charging system for electric vehicles
CN103093624A (en) * 2013-01-09 2013-05-08 东南大学 Signalized intersection non-motor vehicle illegal cross-street behavior automatic judging method
US20190107405A1 (en) * 2017-04-27 2019-04-11 Beijing Didi Infinity Technology And Development Co., Ltd. Systems and methods for route planning
US20190094867A1 (en) * 2017-09-28 2019-03-28 Nec Laboratories America, Inc. Generative adversarial inverse trajectory optimization for probabilistic vehicle forecasting
CN108875766A (en) * 2017-11-29 2018-11-23 北京旷视科技有限公司 Method, apparatus, system and the computer storage medium of image procossing
CN110225446A (en) * 2018-03-01 2019-09-10 北京嘀嘀无限科技发展有限公司 A kind of system, method, apparatus and storage medium identifying driving behavior
WO2020029356A1 (en) * 2018-08-08 2020-02-13 杰创智能科技股份有限公司 Method employing generative adversarial network for predicting face change
CN112466003A (en) * 2019-09-06 2021-03-09 顺丰科技有限公司 Vehicle state detection method, device, server and storage medium
CN110717433A (en) * 2019-09-30 2020-01-21 华中科技大学 A traffic violation analysis method and device based on deep learning
CN113380021A (en) * 2020-03-10 2021-09-10 深圳市丰驰顺行信息技术有限公司 Vehicle state detection method, device, server and computer-readable storage medium
CN112017447A (en) * 2020-08-20 2020-12-01 北京赛博星通科技有限公司 Method and system for judging vehicle converse violation based on GPS position information
CN112149511A (en) * 2020-08-27 2020-12-29 深圳市点创科技有限公司 Method, terminal and device for detecting violation of driver based on neural network
CN112464749A (en) * 2020-11-11 2021-03-09 鹏城实验室 Traffic scene abnormal target detection method based on rules and learning
CN113160575A (en) * 2021-03-15 2021-07-23 超级视线科技有限公司 Traffic violation detection method and system for non-motor vehicles and drivers
CN113450571A (en) * 2021-09-01 2021-09-28 深圳市鼎粤科技有限公司 Traffic intersection-based driving direction reminding method and device and storage medium
CN113936465A (en) * 2021-10-26 2022-01-14 公安部道路交通安全研究中心 Traffic incident detection method and device
CN114548298A (en) * 2022-02-25 2022-05-27 阿波罗智联(北京)科技有限公司 Model training method, traffic information processing method, device, equipment and storage medium

Non-Patent Citations (4)

* Cited by examiner, † Cited by third party
Title
杨懿男;齐林海;王红;苏林萍;: "基于生成对抗网络的小样本数据生成技术研究", 电力建设, no. 05 *
林万华: "一种基于RFID 的电动自行车违章管理系统", 《中国科技信息》, no. 1 *
温惠英;张伟罡;赵胜;: "基于生成对抗网络的车辆换道轨迹预测模型", 华南理工大学学报(自然科学版), no. 05 *
高忠文: "基于生成对抗网络改进的更快速区域卷积神经网络交通标志检测", 《汽车技术》, no. 2 *

Also Published As

Publication number Publication date
CN115273456B (en) 2023-08-29

Similar Documents

Publication Publication Date Title
CN110091751B (en) Electric automobile endurance mileage prediction method, device and medium based on deep learning
CN107169567B (en) Method and device for generating decision network model for automatic vehicle driving
CN112466119B (en) A method and system for predicting vehicle following speed using vehicle-road coordination data
CN114815605A (en) Automatic driving test case generation method and device, electronic equipment and storage medium
CN116490858A (en) Adaptive generation and evaluation of autonomous vehicle key scenarios
CN110196587A (en) Vehicular automatic driving control strategy model generating method, device, equipment and medium
CN113696890B (en) Lane keeping method, apparatus, device, medium, and system
CN114021840A (en) Channel switching strategy generation method and device, computer storage medium and electronic equipment
Jiang et al. Efficient and unbiased safety test for autonomous driving systems
CN114544191A (en) Automatic driving test scene quantitative evaluation method and related equipment
CN114021080A (en) Trajectory prediction model training and trajectory prediction method, apparatus, device and medium
CN116513236A (en) Adaptive vehicle omnidirectional driving assistance method and device based on driving risk field
CN112462759B (en) Evaluation method, system and computer storage medium of rule control algorithm
CN114248780A (en) IDM-LSTM combined following model establishing method considering driver style
CN115273456A (en) Method and system for judging illegal driving of two-wheeled electric vehicle and storage medium
CN115979679B (en) Actual road test method, device and storage medium for automatic driving system
JP7532681B2 (en) Generating unknown unsafe scenarios,improving autonomous vehicles,computer systems
CN117516956A (en) Methods, systems and computer program products for objectively evaluating the performance of ADAS/ADS systems
CN118254802A (en) Driving style identification method, device, computer readable medium and vehicle
CN115688254A (en) Automatic driving risk prediction method and device, storage medium and electronic equipment
Zhang et al. Model Validation of Adaptive Cruise Control in Vehicles Utilizing Real-World Driving Data
Li et al. Traffic Modeling Considering Motion Uncertainties
CN116945907B (en) New energy electric automobile mileage calculation method and system
Yamauchi et al. Adaptive identification method of vehicle modeling according to the fluctuation of road and running situation in autonomous driving
CN118494526A (en) Method and device for predicting lane change intention of peripheral vehicle, electronic equipment and storage medium

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