CN110705854A - Driving level evaluation method and system - Google Patents

Abstract

The invention discloses a driving level evaluation method and system, belonging to the technical field of intelligent driving. Cloud models of various driving states are used to construct an evaluation cloud map corresponding to the driving level. When acquiring real-time driving data of an unknown level driver, construct its corresponding level evaluation cloud map, and compare it with the cloud maps of various driving levels constructed to determine the real driving level of the unknown level driver. This scheme can deal with the uncertainty of the driving level and evaluate the driving level of the driver.

Description

A driving level assessment method and system

technical field

The invention relates to the technical field of intelligent driving, in particular to a driving level evaluation method and system.

Background technique

Intelligent driving technology involves information engineering, control science and engineering, computer science, mechanical engineering, mathematical science, life science and many other disciplines. It is an important symbol for measuring a country's scientific research strength and industrial level. The emergence of intelligent driving has fundamentally changed the traditional way of vehicle driving, freeing the driver from the "vehicle-road-human" closed-loop system. It uses advanced electronic and information technology to control the driving of vehicles, so that the routine, long-lasting and fatigued operations in driving activities can be completed automatically. People only do advanced purposeful operations, which can greatly improve the efficiency and safety of the transportation system. application prospects. At the same time, the research on intelligent driving technology will greatly enhance my country's core competitiveness in automotive active safety systems, which is of great strategic significance to enhancing the independent innovation capabilities of my country's automotive electronic products and automotive industry.

However, due to different driving levels, drivers with different driving levels have different occupation of road resources and their impact on congestion. Therefore, it is necessary to evaluate the driving level of drivers to provide a basis for studying the relationship between right of way and urban congestion. Solve the problem of urban traffic congestion to the greatest extent.

SUMMARY OF THE INVENTION

The purpose of the present invention is to overcome the deficiencies or defects of the prior art, so as to evaluate the driving level of the driver and solve the problem of urban traffic congestion to the greatest extent.

In order to achieve the above object, the present invention adopts a kind of driving level evaluation method, which comprises the following steps:

S100, classifying the driving state of the driver, and setting evaluation indicators for various driving states;

S200, based on the expert evaluation method, obtain the evaluation value of each driving state;

S300, input the evaluation value of each driving state under the same driving level into the reverse cloud generator, and calculate the number of representations of the cloud model in each driving state;

S400. Input the characterization number of the cloud model in each driving state under the same driving level into the forward cloud generator, obtain the quantitative position of the cloud droplet of the driving level in the number domain space and the certainty of the qualitative concept, and construct the Cloud map for assessment of driving level;

S500. Repeat steps S200 to S400 to obtain evaluation cloud maps corresponding to different driving levels;

S600, obtaining real-time driving data of the driver to be evaluated, and executing the steps S200 to S400 to obtain a cloud map of the level evaluation of the driver to be evaluated;

S700. Compare the level evaluation cloud image of the driver to be evaluated with the evaluation cloud images corresponding to the different driving levels, and determine the driving level of the driver to be evaluated.

Further, the evaluation value of each driving state under the same driving level is input into the reverse cloud generator, and the number of representations of the cloud model in each driving state is calculated, including:

Obtain the evaluation value of the same driving state of N drivers of the same driving level as the cloud drop corresponding to the driving state;

According to the cloud drop corresponding to the driving state, calculate the mean value of the cloud drop, the variance of the cloud drop and the entropy of the cloud drop respectively;

Calculate the superentropy of the cloud drop according to the variance of the cloud drop and the entropy of the cloud drop;

The mean value of the cloud droplets, the entropy of the cloud droplets, and the superentropy of the cloud droplets are used as the characterization numbers of the cloud model corresponding to the driving state.

Further, the characteristic number of the cloud model in each driving state under the same driving level is input into the forward cloud generator, and the quantitative position of the cloud droplet in the number domain space and the certainty of the qualitative concept of the driving level are obtained. , construct an evaluation cloud map for this driving level, including:

S401. According to the number of representations of the cloud model in each driving state under the same driving level, establish a comprehensive cloud model A=(E _x , _{E n} , He ) for evaluating the driving level, and the number of representations of the comprehensive cloud model are the expected value _Ex , the entropy _En and the super- _{entropy He} , respectively;

以及一个均值为E_n，标准差为

的正态随机数x；S402. According to the expected value Ex, the entropy _En , the _superentropy He, and the given cloud droplet number _N , obtain a normal random number with a mean value of _Ex and a standard deviation of _He.

and a mean E _n with a standard deviation of

The normal random number x of ;

令x是定性概念的一次具体量化值，令y是x的确定度；S403. Calculation

Let x be a specific quantitative value of a qualitative concept, and let y be the degree of certainty of x;

S404. Repeat steps S402 to S403 until N cloud droplets are generated;

S405, output the quantitative positions of the cloud droplets of N drivers of the same level in the number domain space and the certainty of the qualitative concept (x, y);

S406 , obtaining the quantitative position of the cloud droplet of the driving level in the number domain space and the certainty of the qualitative concept, and constructing an evaluation cloud map of the driving level.

Further, the level evaluation cloud map of the driver to be evaluated is compared with the evaluation cloud maps corresponding to the different driving levels, and the driving level of the driver to be evaluated is determined, including:

Calculate the similarity between the level evaluation cloud image of the driver to be evaluated and the evaluation cloud images corresponding to the different driving levels;

The driving level corresponding to the evaluation cloud image with the maximum similarity is taken as the driving level of the driver to be evaluated.

Further, the driving status of the driver includes violations, driving directions and long-term status;

The evaluation indicators of violations include the traffic conditions stipulated by traffic lights, the driving conditions at the specified speed, and the driving conditions in the specified lanes;

The evaluation index of driving orientation includes lateral deviation and azimuth deviation;

The evaluation indicators of long-term state include driving emotional stability, ability to judge running speed, ability to respond to emergencies, and ability to judge time and space in different driving environments.

Further, the cloud models corresponding to the violation situation, the driving direction and the long-term state are respectively A ₁ =(E _1x , E _1n , H _1e ), A ₂ =(E _2x , E _2n , H _2e ) and A ₃ = (E _3x , E _3n , H _3e ), establishing a comprehensive cloud model for evaluating the driving level according to the number of representations of the cloud model in each driving state under the same driving level, including:

则计算云模型A′＝A₁∪A₂＝A₂；Assuming E _1x ≤ E _2x , if

Then the computing cloud model A′=A ₁ ∪ A ₂ =A ₂ ;

则计算新的云模型A′＝A₁∪A₂＝A₁；if

Then calculate the new cloud model A′=A ₁ ∪ A ₂ =A ₁ ;

Through the _three characterization parameters in the new cloud models A' and A3, a comprehensive cloud model A=( _{Ex ,E n ,} H _e ) for the evaluation of the driver's driving behavior with respect to the current driving level is obtained.

On the other hand, a driving level evaluation system is adopted, including: a classification module, an evaluation value calculation module, a cloud model calculation module, a first evaluation cloud image construction module, a second evaluation cloud image construction module, and a comparison module;

The classification module is used to classify the driving states of drivers with different driving levels, and set the evaluation indicators of various driving states;

The evaluation value calculation module is used to obtain the weight coefficient of the influence degree of the evaluation index on the driving state based on the expert evaluation method, and obtain the evaluation value of each driving state;

The cloud model calculation module is used to input the evaluation value of each driving state under the same driving level into the reverse cloud generator, and calculate the number of representations of the cloud model in each driving state;

The first evaluation cloud image building module is used to input the characterization numbers of the cloud model in each driving state under the same driving level into the forward cloud generator, and obtain the quantitative position and qualitative concept of the cloud droplet in the number domain space for this driving level The certainty of the driving level is determined, and the evaluation cloud map of the driving level is constructed, so as to obtain the evaluation cloud map corresponding to different driving levels;

The second evaluation cloud image building module is used to obtain the real-time driving data of the driver to be evaluated, and obtain the level evaluation cloud image of the driver to be evaluated;

The comparison module is configured to compare the level evaluation cloud image of the driver to be evaluated with the evaluation cloud images corresponding to the different driving levels to determine the driving level of the driver to be evaluated.

Further, the cloud model calculation module includes a cloud drop acquisition unit, a parameter calculation unit, a super entropy calculation unit and a cloud model calculation unit;

The cloud drop acquisition unit is used to obtain the evaluation value of the same driving state of N drivers of the same driving level, as the cloud drop corresponding to the driving state;

The parameter calculation unit is used to calculate the mean value of the cloud drop, the variance of the cloud drop and the entropy of the cloud drop respectively according to the cloud drop corresponding to the driving state;

The super entropy calculation unit is used to calculate the super entropy of the cloud drop according to the variance of the cloud drop and the entropy of the cloud drop;

The cloud model computing unit is configured to construct a cloud model by using the mean value of the cloud droplets, the entropy of the cloud droplets, and the superentropy of the cloud droplets as the characterization numbers of the cloud model corresponding to the driving state.

Further, the first evaluation cloud image construction module includes a comprehensive cloud model construction unit, a normal random number calculation unit, a quantitative qualitative unit, a certainty calculation unit and an evaluation cloud image construction unit;

The comprehensive cloud model building unit is used for establishing a comprehensive cloud model A=(E _x , _{E n} , He ) for evaluating the driving level according to the number of representations of the cloud model in each driving state under the same driving level. The characterization numbers of the cloud model are the expected value Ex, the entropy _En and the _{hyperentropy He,} respectively;

正态随机数

的均值为E_x、标准差为H_e，以及一个正态随机数x，正态随机数x的均值为E_n，标准差为

The normal random number calculation unit _is used to obtain a normal random number according to the expected value Ex, entropy En and super _{entropy He} and the given number _N of cloud droplets

normal random number

The mean is _Ex , the standard deviation is He, and a normal random number _x , the mean of the normal random number _x is En, and the standard deviation is

Quantitative and qualitative unit for calculation Let x be a specific quantitative value of a qualitative concept, and let y be the degree of certainty of x;

The certainty calculation unit is used to output the quantitative position of the cloud droplets of the N same-level drivers in the number domain space and the certainty (x, y) of the qualitative concept when N cloud droplets are generated;

The evaluation cloud map construction unit is used for constructing the evaluation cloud map of the driving level according to the quantitative position of the cloud droplet in the number domain space and the certainty of the qualitative concept.

Further, the comparison module includes a similarity calculation unit and a comparison unit;

The similarity calculation unit is used to calculate the similarity between the level evaluation cloud image of the driver to be evaluated and the evaluation cloud images corresponding to the different driving levels;

The comparison unit is configured to compare the calculated similarity, and use the driving level corresponding to the evaluation cloud image with the maximum similarity as the driving level of the driver to be evaluated.

Compared with the prior art, the present invention has the following technical effects: the present invention uses the reverse cloud generator to calculate the cloud model of each driving state by classifying the driving states of drivers with different driving levels, and according to the same driving level The cloud model of various driving states is constructed, and the evaluation cloud map corresponding to the driving level is constructed. When acquiring real-time driving data of an unknown level driver, construct its corresponding level evaluation cloud map, and compare it with the cloud maps of various driving levels constructed to determine the real driving level of the unknown level driver. This scheme can deal with the uncertainty of the driving level and evaluate the driving level of the driver.

Description of drawings

Below in conjunction with the accompanying drawings, the specific embodiments of the present invention are described in detail:

FIG. 1 is a schematic flowchart of a driving level assessment method;

Figure 2 is a diagram of a smart car driving behavior evaluation system;

FIG. 3 is a schematic diagram of driving state characterization parameters;

FIG. 4 is a schematic flowchart of a driving level evaluation system.

Detailed ways

To further illustrate the features of the present invention, please refer to the following detailed description and accompanying drawings of the present invention. The attached drawings are for reference and description only, and are not intended to limit the protection scope of the present invention.

As shown in FIG. 1 , this embodiment discloses a driving level evaluation method, which includes the following steps S100-S700:

S100, classifying the driving state of the driver, and setting evaluation indicators for various driving states;

S200, based on the expert evaluation method, obtain the evaluation value of each driving state;

S300, input the evaluation value of each driving state under the same driving level into the reverse cloud generator, and calculate the number of representations of the cloud model in each driving state;

S400. Input the characterization number of the cloud model in each driving state under the same driving level into the forward cloud generator, obtain the quantitative position of the cloud droplet of the driving level in the number domain space and the certainty of the qualitative concept, and construct the Cloud map for assessment of driving level;

S500. Repeat steps S200 to S400 to obtain evaluation cloud maps corresponding to different driving levels;

S600, obtaining real-time driving data of the driver to be evaluated, and executing the steps S200 to S400 to obtain a cloud map of the level evaluation of the driver to be evaluated;

S700. Compare the level evaluation cloud image of the driver to be evaluated with the evaluation cloud images corresponding to the different driving levels, and determine the driving level of the driver to be evaluated.

Among them, the driving level of the driver includes normal driving level, low driving level (rookie) and high driving level (race driver). The research on the analysis of driving behavior and psychological characteristics of different drivers can classify the driving status of drivers into: violations, driving directions and long-term status. The evaluation indicators for each type of driving state are:

(1) Violation: Indicates the driver's compliance with traffic rules during driving, which is divided into no violation, general violation and serious violation. This indicator reflects the driver's ability to master rules and regulations and driving quality. Through the assistance of electronic eyes at intersections, on-board cameras and traffic police departments, the traffic violations of different drivers while driving can be obtained. The evaluation indicators include: (1-1) traffic conditions specified by traffic lights; (1-2) driving conditions at specified speeds; (1-3) driving conditions in specified lanes.

(2) Driving direction: It indicates the habitual relative position of the vehicle and the center of the road when the driver is driving. This indicator reflects the driver's driving habits. If the driving habits are good, the driver should drive in the middle of the road. The road information obtained by the actual vehicle information collection system, and the lane line detection method or the road detection method are used to detect the left and right lane lines of the current lane on the current road of the smart car.

As shown in Figure 3, the two characterization parameters of lateral deviation and directional deviation are used to evaluate the current driving direction and the degree of center deviation of the smart car. The specific calculation formula is: lateral deviation = right deviation - left deviation; Bias = Right Azimuth - Left Azimuth. Evaluation indicators include: (2-1) lateral deviation; (2-2) azimuth deviation.

(3) Long-term state: Indicates the degree to which the driver maintains the driving state during the continuous driving process. This indicator reflects changes in driving proficiency and mental state while driving. It can be analyzed through psychological tests, interviews and vehicle monitoring records to obtain the status of different drivers after long-term driving. The evaluation indicators include: (3-1) Emotional stability of driving; (3-2) Ability to judge running speed; (3-3) Ability to respond to emergencies; (3-4) Ability to judge time and space in different driving environments.

It should be noted that each evaluation index set in this embodiment is disconnected from each other. Each detailed evaluation index has a corresponding, direct qualitative agreement according to the specific evaluation standard.

However, it should be understood that the principle of this embodiment is not limited to the driving state index classification and its evaluation index proposed above, and may include all reasonable index classifications and evaluation indexes for analyzing the driving state.

Further, in the above-mentioned step S200: obtaining the evaluation value of each driving state based on the expert evaluation method, taking the violation situation in the driving state as an example, the weight coefficient w _i of each driving index is defined as:

Among them, the importance of each index is analyzed based on the comprehensive mechanism, and the "quality" of each evaluation index in the case of violation is determined as M _i . Set an ideal value for the index value of the evaluation object Calculate the Euclidean distance between the index value x=(x ₁ , x ₂ , x ₃ ) and the ideal value, and obtain the evaluation value of traffic violation in the driving state of the jth driver. The specific evaluation formula is:

可以基于专家评价方法来设定。Among them, j=1,2,...,N, represents the jth driver among N drivers of the same driving level, i=1, 2, 3 respectively represent the evaluation indicators of violations (1-1) by traffic The traffic light stipulates the traffic situation; (1-2) The driving situation according to the specified speed; (1-3) The driving situation according to the specified lane. Each evaluation index "quality" _Mi and ideal value

It can be set based on an expert evaluation method.

Similarly, taking the violation of regulations in the driving state as an example, the above step S300: input the evaluation value of each driving state under the same driving level into the reverse cloud generator, and calculate the number of representations of the cloud model in each driving state, Specifically include:

_Obtain the evaluation value Y ₁ =(Y ₁₁ , Y ₁₂ , .

Calculate the mean value of cloud droplets, the formula is as follows:

Calculate the variance of cloud droplets, the formula is as follows:

To calculate the entropy of cloud droplets, the formula is as follows:

To calculate the superentropy of cloud droplets, the formula is as follows:

The digital feature A ₁ =(E _1x , E _1n , H _1e ) of the output cloud droplet.

In the same way, the cloud model numerical features of the driving position and the long-term state in the driving state are obtained, respectively A ₂ =(E _2x , E _2n , H _2e ) and A ₃ =(E _3x , E _3n , H _3e ).

Further, the above-mentioned step S400: the characteristic number of the cloud model in each driving state under the same driving level is input into the forward cloud generator, and the quantitative position and qualitative concept of the cloud droplet of the driving level in the number domain space are obtained. Determine the degree of certainty, and construct the evaluation cloud map of the driving level, including the following steps S401-S406:

S401. According to the number of representations of the cloud model in each driving state under the same driving level, establish a comprehensive cloud model A=(E _x , _{E n} , He ) for evaluating the driving level, and the number of representations of the comprehensive cloud model are the expected value _Ex , the entropy _En and the super- _{entropy He} , respectively;

以及一个均值为E_n，标准差为

的正态随机数x；S402. According to the expected value Ex, the entropy _En , the _superentropy He, and the given cloud droplet number _N , obtain a normal random number with a mean value of _Ex and a standard deviation of _He.

and a mean E _n with a standard deviation of

The normal random number x of ;

令x是定性概念的一次具体量化值，令y是x的确定度；S403. Calculation

Let x be a specific quantitative value of a qualitative concept, and let y be the degree of certainty of x;

S404. Repeat steps S402 to S403 until N cloud droplets are generated;

S405, output the quantitative positions of the cloud droplets of N drivers of the same level in the number domain space and the certainty of the qualitative concept (x, y);

S406 , obtaining the quantitative position of the cloud droplet of the driving level in the number domain space and the certainty of the qualitative concept, and constructing an evaluation cloud map of the driving level.

Specifically, the above step S401: According to the number of representations of the cloud model in each driving state under the same driving level, a comprehensive cloud model for evaluating the driving level is established, specifically:

则计算云模型A′＝A₁∪A₂＝A₂；Assuming E _1x ≤ E _2x , if

Then the computing cloud model A′=A ₁ ∪ A ₂ =A ₂ ;

则计算新的云模型A′＝A₁∪A₂＝A₁；if

Then calculate the new cloud model A′=A ₁ ∪ A ₂ =A ₁ ;

Through the _three characterization parameters in the new cloud models A' and A3, a comprehensive cloud model A=( _{Ex ,E n ,} H _e ) for the evaluation of the driver's driving behavior with respect to the current driving level is obtained.

Assuming E _1x ≤ E _2x , if |E _1x -E _2x |<|3(E _1n +E _2n )|, then the new cloud model A′=(E′ _x ,E′ _n ,H′ _e ) The characterization parameters can be calculated according to the following formula:

则A′＝A₁∪A₂＝A₁。If |E _1x -E _2x |≥|3(E _1n +E _2n )|, then the new cloud model A' is represented by two cloud models A ₁ and A ₂ , that is, if Then A'=A ₁ ∪ A ₂ =A ₂ . Conversely, if

Then A'=A ₁ ∪ A ₂ =A ₁ .

Through the _three characterization parameters in the cloud models A' and A3, a comprehensive cloud model A ₌ (Ex, _En , He) for evaluating the driving behavior of _rookies is obtained.

Further, the above-mentioned step S700: compare the level evaluation cloud map of the driver to be evaluated with the evaluation cloud maps corresponding to the different driving levels, and determine the driving level of the driver to be evaluated, including:

Calculate the similarity between the level evaluation cloud image of the driver to be evaluated and the evaluation cloud images corresponding to the different driving levels;

The driving level corresponding to the evaluation cloud image with the maximum similarity is taken as the driving level of the driver to be evaluated.

Specifically, real-time driving data of a driver with an unknown level is recorded, the average value of different evaluation indicators in different driving states is obtained, and the evaluation cloud map X ^* of the driver with unknown level is depicted. The evaluation contours X ^* are compared with the evaluation contours A ^* , B ^* , and C ^* , respectively, wherein the evaluation contours A ^* , B ^* , and C ^* represent low, normal, and high driving levels, respectively. If the evaluation contour X ^* is more similar to the evaluation contour A ^* than the evaluation contour B ^* and C ^* , it can be judged that the driver's driving level is a low driving level; if the evaluation contour X ^* is similar to the evaluation contour B ^* If the degree of similarity is greater than the similarity with the evaluation cloud maps A ^* and C ^* , it can be judged that the driver's driving level is the normal driving level; similarly, if the evaluation cloud map X ^* and the evaluation cloud map C ^* are more similar than the evaluation cloud map A ^* and If the similarity of B ^* is high, it can be judged that the driver's driving level is a high driving level.

This embodiment analyzes a large amount of experimental data through the cloud model, and effectively realizes the conversion of uncertainty between qualitative and quantitative in the process of driving behavior judgment, so that the smart car has the same driving habits as the driver, and can deal with different driving levels. Deterministic, simulating the driving style of drivers with different driving levels, can more realistically study the problems and solutions encountered in real-time traffic. At the same time, through the occupancy of road resources by drivers with different driving levels and their impact on congestion, the relationship between the right of way and urban traffic congestion can be studied, so as to solve the problem of urban traffic congestion to the greatest extent.

As shown in FIG. 4 , this embodiment discloses a driving level evaluation system, including: a classification module 10 , an evaluation value calculation module 20 , a cloud model calculation module 30 , a first evaluation cloud map construction module 40 , and a second evaluation cloud map construction module 50 and a comparison module 60;

The classification module 10 is used for classifying the driving states of drivers with different driving levels, and setting evaluation indexes of various driving states;

The evaluation value calculation module 20 is configured to obtain the weight coefficient of the influence degree of the evaluation index on the driving state based on the expert evaluation method, and obtain the evaluation value of each driving state;

The cloud model calculation module 30 is configured to input the evaluation value of each driving state under the same driving level into the reverse cloud generator, and calculate the number of representations of the cloud model in each driving state;

The first evaluation cloud image building module 40 is used to input the characteristic number of the cloud model in each driving state under the same driving level into the forward cloud generator to obtain the quantitative position and qualitative position of the cloud droplet in the number domain space of the driving level. The certainty of the concept, construct the evaluation cloud map of the driving level, so as to obtain the evaluation cloud map corresponding to different driving levels;

The second evaluation cloud image building module 50 is used to obtain the real-time driving data of the driver to be evaluated, and obtain the level evaluation cloud image of the driver to be evaluated;

The comparison module 60 is configured to compare the level evaluation cloud image of the driver to be evaluated with the evaluation cloud images corresponding to the different driving levels to determine the driving level of the driver to be evaluated.

Specifically, the cloud model calculation module 30 includes a cloud droplet acquisition unit, a parameter calculation unit, a super entropy calculation unit and a cloud model calculation unit;

The cloud drop acquisition unit is used to obtain the evaluation value of the same driving state of N drivers of the same driving level, as the cloud drop corresponding to the driving state;

The parameter calculation unit is used to calculate the mean value of the cloud drop, the variance of the cloud drop and the entropy of the cloud drop respectively according to the cloud drop corresponding to the driving state;

The super entropy calculation unit is used to calculate the super entropy of the cloud drop according to the variance of the cloud drop and the entropy of the cloud drop;

The cloud model computing unit is configured to construct a cloud model by using the mean value of the cloud droplets, the entropy of the cloud droplets, and the superentropy of the cloud droplets as the characterization numbers of the cloud model corresponding to the driving state.

Specifically, the first evaluation cloud image construction module 40 includes a comprehensive cloud model construction unit, a normal random number calculation unit, a quantitative qualitative unit, a certainty calculation unit, and an evaluation cloud image construction unit;

The comprehensive cloud model building unit is used for establishing a comprehensive cloud model A=(E _x , _{E n} , He ) for evaluating the driving level according to the number of representations of the cloud model in each driving state under the same driving level. The characterization numbers of the cloud model are the expected value Ex, the entropy _En and the _{hyperentropy He,} respectively;

正态随机数

的均值为E_x、标准差为H_e，以及一个正态随机数x，正态随机数x的均值为E_n，标准差为

The normal random number calculation unit _is used to obtain a normal random number according to the expected value Ex, entropy En and super _{entropy He} and the given number _N of cloud droplets

normal random number

The mean is _Ex , the standard deviation is He, and a normal random number _x , the mean of the normal random number _x is En, and the standard deviation is

令x是定性概念的一次具体量化值，令y是x的确定度；Quantitative and qualitative unit for calculation

Let x be a specific quantitative value of a qualitative concept, and let y be the degree of certainty of x;

The certainty calculation unit is used to output the quantitative position of the cloud droplets of the N same-level drivers in the number domain space and the certainty (x, y) of the qualitative concept when N cloud droplets are generated;

The evaluation cloud map construction unit is used for constructing the evaluation cloud map of the driving level according to the quantitative position of the cloud droplet in the number domain space and the certainty of the qualitative concept.

Specifically, the comparison module 60 includes a similarity calculation unit and a comparison unit;

The similarity calculation unit is used to calculate the similarity between the level evaluation cloud image of the driver to be evaluated and the evaluation cloud images corresponding to the different driving levels;

The comparison unit is configured to compare the calculated similarity, and use the driving level corresponding to the evaluation cloud image with the maximum similarity as the driving level of the driver to be evaluated.

It should be noted that the similarity calculation process in this embodiment is:

Through the root mean square error formula, the N cloud droplets generated by the level evaluation cloud map of the driver to be evaluated are compared with the N cloud droplets generated by the time series corresponding to the evaluation cloud map corresponding to different driving levels. The calculation formula is as follows:

Among them, λ ₁ and λ ₂ represent the error coefficient of the quantitative position x in the number domain space and the certainty y of the qualitative concept; N represents the number of cloud droplets; r=1, 2, 3 represent the low driving level, normal driving, respectively level and advanced driving level; p indicates the driving level to be assessed.

If the RSME ₁ value is the smallest, the driving level to be assessed belongs to the low driving level; if the RSME ₂ value is the smallest, the driving level to be assessed belongs to the normal driving level; if the RSME ₃ value is the smallest, the driving level to be assessed belongs to the advanced driving level .

It should be noted that the conventional driving level evaluation method avoids the randomness, ambiguity and statistical qualitative indicators of evaluation, which makes the result of driving level evaluation inaccurate. However, this embodiment proposes to establish a qualitative-to-quantitative conversion relationship based on the cloud model theory by using the forward cloud and reverse cloud generators in the cloud model, which integrates subjective judgment and objective data, and avoids the need to manually determine the driving level evaluation. Subjective influence includes qualitative indicators that are difficult to be counted, fuzzy and random in the evaluation problem, which greatly improves the credibility and scientificity of driving level evaluation.

The above are only preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention shall be included in the protection of the present invention. within the range.

Claims (10)

1. A driving level evaluation method characterized by comprising:

s100, classifying the driving states of the driver, and setting evaluation indexes of various driving states;

s200, obtaining an evaluation value of each driving state based on an expert evaluation method;

s300, inputting the evaluation value of each driving state under the same driving level into a reverse cloud generator, and calculating the characterization number of the cloud model in each driving state;

s400, inputting the characterization number of the cloud model in each driving state under the same driving level into a forward cloud generator, obtaining the quantitative position of the cloud droplets of the driving level in a number domain space and the certainty degree of a qualitative concept, and constructing an evaluation cloud picture of the driving level;

s500, repeatedly executing the steps S200-S400 to obtain evaluation cloud pictures corresponding to different driving levels;

s600, acquiring real-time driving data of a driver to be evaluated, executing the steps S200-S400, and acquiring a horizontal evaluation cloud picture of the driver to be evaluated;

s700, comparing the level evaluation cloud picture of the driver to be evaluated with the evaluation cloud pictures corresponding to different driving levels, and determining the driving level of the driver to be evaluated.

2. The driving level assessment method according to claim 1, wherein the inputting of the assessment value for each driving state at the same driving level into the inverse cloud generator, and the calculating of the characterization number of the cloud model in each driving state comprises:

acquiring evaluation values of N drivers with the same driving state at the same driving level as cloud droplets corresponding to the driving state;

respectively calculating the mean value of the cloud droplets, the variance of the cloud droplets and the entropy of the cloud droplets according to the cloud droplets corresponding to the driving state;

calculating the super entropy of the cloud droplets according to the variance of the cloud droplets and the entropy of the cloud droplets;

and taking the mean value of the cloud droplets, the entropy of the cloud droplets and the super entropy of the cloud droplets as the characterization numbers of the cloud model corresponding to the driving state.

3. The driving level assessment method according to claim 2, wherein the inputting the characterization number of the cloud model in each driving state of the same driving level into the forward cloud generator to obtain the quantitative position and the certainty degree of the qualitative concept of the cloud droplet of the driving level in the number domain space, and constructing the assessment cloud map of the driving level comprises:

s401, establishing a comprehensive cloud model A (E) for evaluating the driving level according to the characterization number of the cloud model in each driving state under the same driving level_x,E_n,H_e) The characterization numbers of the comprehensive cloud model are respectively expected values E_xEntropy E_nAnd entropy H_e；

S402, according to the expected value E_xEntropy E_nAnd entropy H_eAnd a given cloud drop number N, to obtain an average value E_xStandard deviation of H_eNormal random number of

And a mean value of E_nStandard deviation of

The normal random number x of (a);

s403, calculatingLet x be a specific quantization value of the qualitative concept and let y be the certainty of x;

s404, repeatedly executing the steps S402-S403 until N cloud droplets are generated;

s405, outputting the quantitative positions of the cloud droplets of N drivers at the same level in a number domain space and the certainty degree (x, y) of a qualitative concept;

s406, obtaining the quantitative position of the cloud droplets of the driving level in the number domain space and the certainty degree of the qualitative concept, and constructing an evaluation cloud picture of the driving level.

4. The driving level assessment method according to claim 1, wherein the comparing the level assessment cloud of the driver to be assessed with the assessment clouds corresponding to the different driving levels to determine the driving level of the driver to be assessed comprises:

calculating the similarity between the level evaluation cloud picture of the driver to be evaluated and the evaluation cloud pictures corresponding to different driving levels;

and taking the driving level corresponding to the evaluation cloud picture with the maximum similarity as the driving level of the driver to be evaluated.

5. The driving level assessment method of claim 3, wherein the driving status of the driver includes a violation, a driving direction, and a long-term status;

the evaluation indexes of the violation conditions comprise the passing conditions specified by traffic lights, the running conditions at specified speed and the running conditions at specified lanes;

the evaluation indexes of the running direction comprise a lateral deviation condition and a direction deviation condition;

the evaluation indexes of the long-term state comprise driving emotion stability, running speed judgment capability, emergency handling response capability and time and space judgment capability of different driving environments.

6. The method of claim 5 wherein the cloud models for the violation, the driving direction, and the long-term condition are A₁＝(E_1x,E_1n,H_1e)，A₂＝(E_2x,E_2n,H_2e) And A₃＝(E_3x,E_3n,H_3e) The establishing of the comprehensive cloud model for evaluating the driving level according to the characterization number of the cloud model in each driving state under the same driving level comprises the following steps:

suppose E_1x≤E_2xIf, if

Then compute cloud model a' ═ a₁∪A₂＝A₂；

If it is not

Then a new cloud model a' ═ a is computed₁∪A₂＝A₁；

By new cloud models A' and A₃Of the current driving level, resulting in an integrated cloud model a ═ (E) for the driving behavior assessment of the driver at the current driving level_x,E_n,H_e)。

7. A driving level evaluation system characterized by comprising: the cloud model evaluation system comprises a classification module, an evaluation value calculation module, a cloud model calculation module, a first evaluation cloud picture construction module, a second evaluation cloud picture construction module and a comparison module;

the classification module is used for classifying the driving states of drivers with different driving levels and setting evaluation indexes of various driving states;

the evaluation value calculation module is used for obtaining a weight coefficient of the influence degree of the evaluation index on the driving state based on an expert evaluation method and obtaining an evaluation value of each driving state;

the cloud model calculation module is used for inputting the evaluation value of each driving state under the same driving level into the reverse cloud generator and calculating the characterization number of the cloud model in each driving state;

the first evaluation cloud picture construction module is used for inputting the characterization number of the cloud model in each driving state under the same driving level into the forward cloud generator, obtaining the quantitative position of cloud droplets of the driving level in a number domain space and the certainty degree of a qualitative concept, and constructing the evaluation cloud picture of the driving level so as to obtain the evaluation cloud pictures corresponding to different driving levels;

the second evaluation cloud picture construction module is used for acquiring real-time driving data of the driver to be evaluated and acquiring a level evaluation cloud picture of the driver to be evaluated;

the comparison module is used for comparing the level evaluation cloud picture of the driver to be evaluated with the evaluation cloud pictures corresponding to different driving levels to determine the driving level of the driver to be evaluated.

8. The driving level assessment system according to claim 7, wherein the cloud model calculation module comprises a cloud droplet acquisition unit, a parameter calculation unit, a hyper-entropy calculation unit, and a cloud model calculation unit;

the cloud droplet acquisition unit is used for acquiring the evaluation values of N drivers with the same driving level in the same driving state as the cloud droplets corresponding to the driving state;

the parameter calculation unit is used for respectively calculating the mean value of the cloud droplets, the variance of the cloud droplets and the entropy of the cloud droplets according to the cloud droplets corresponding to the driving state;

the super-entropy calculation unit is used for calculating the super-entropy of the cloud droplets according to the variance of the cloud droplets and the entropy of the cloud droplets;

and the cloud model computing unit is used for constructing a cloud model by taking the mean value of the cloud droplets, the entropy of the cloud droplets and the super entropy of the cloud droplets as the representation numbers of the cloud model corresponding to the driving state.

9. The driving level evaluation system of claim 8, wherein the first evaluation cloud picture construction module comprises a comprehensive cloud model construction unit, a normal random number calculation unit, a quantitative qualitative unit, a certainty factor calculation unit, and an evaluation cloud picture construction unit;

the comprehensive cloud model building unit is used for building a comprehensive cloud model A (E) for evaluating the driving level according to the characterization number of the cloud model in each driving state under the same driving level_x,E_n,H_e) The characterization numbers of the comprehensive cloud model are respectively expected values E_xEntropy E_nAnd entropy H_e；

A normal random number calculation unit for calculating a normal random number according to the expected value E_xEntropy E_nAnd entropy H_eAnd a given cloud drop number N to obtain a normal random number

Normal random number

Has a mean value of E_xStandard deviation of H_eAnd a normal random number x whose mean value is E_nStandard deviation of

The quantitative and qualitative unit is used for calculating

Let x be a specific quantization value of the qualitative concept and let y be the certainty of x;

the certainty factor calculation unit is used for outputting the quantitative positions of the cloud drips of the N drivers at the same level in the number domain space and the certainty factors (x, y) of the qualitative concepts when the N cloud drips are generated;

the evaluation cloud picture construction unit is used for constructing the evaluation cloud picture of the driving level according to the quantitative position of the cloud drops of the driving level in the number domain space and the certainty degree of the qualitative concept.

10. The driving level evaluation system according to claim 1, wherein the comparison module includes a similarity calculation unit and a comparison unit;

the similarity calculation unit is used for calculating the similarity between the level evaluation cloud picture of the driver to be evaluated and the evaluation cloud pictures corresponding to different driving levels;

the comparison unit is used for comparing the calculated similarity, and taking the driving level corresponding to the evaluation cloud picture with the maximum similarity as the driving level of the driver to be evaluated.

Images