CN115759880A

CN115759880A - Real-time on-duty adaptability evaluation system and method for bus driver

Info

Publication number: CN115759880A
Application number: CN202211670896.5A
Authority: CN
Inventors: 丁同强; 袁蕾; 李志强; 孙健; 田建; 张华山; 杨帆; 孔永臣; 李加奇; 苗馨宁; 王立强; 刘梓伟; 郑黎黎
Original assignee: Jilin University
Current assignee: Jilin University
Priority date: 2022-12-26
Filing date: 2022-12-26
Publication date: 2023-03-07

Abstract

The invention belongs to the field of driving parameter judgment of a road vehicle driving control system, and particularly relates to a real-time evaluation method and a real-time evaluation system for the on-duty adaptability of a public transport driver, wherein the real-time evaluation method and the real-time evaluation system are used for acquiring video image data, physiological signal data and dynamic track data of a driving vehicle of the driver in real time, identifying the identity of the driver through the data, calling the existing information data of the driver according to an identification result, constructing an on-duty adaptability index system of the public transport driver, generating an index database of the driver, analyzing and processing the data in the index database, calibrating the threshold value of each evaluation index, and evaluating the single-index adaptability of the current driver; if the single-index post adaptability evaluation result is qualified, a comprehensive evaluation model of post adaptability of the public transport driver based on the neural network is brought in, the current post adaptability state of the driver is comprehensively evaluated, and the post adaptability grade of the driver is judged; and correspondingly managing and controlling the driver or the driving vehicle in real time according to different post-fitting grades.

Description

Real-time on-duty adaptability evaluation system and method for bus driver

Technical Field

The invention relates to a method for evaluating the driving style or habit of a driver, belongs to the field of driving parameter judgment of a road vehicle driving control system, and particularly relates to a method and a system for evaluating the on-duty adaptability of a bus driver in real time.

Background

As a core factor of a human-vehicle-road-environment traffic system, a driver plays an important role in the safe operation of public transport. The driving specificity of the bus driver is determined by the occupation characteristics of the bus driver. The bus driver needs to carry out high-strength and high-pressure work in a complex road network environment, and the on-duty state of the bus driver is related to the smooth development of driving work and is more related to the safety of the bus driver and the lives and properties of passengers. Therefore, the on-duty driving state of the bus driver is evaluated in real time, and targeted early warning and intervention are carried out, so that the method has an important role in preventing major bus accidents, guaranteeing the life and property safety of passengers and realizing the dynamic monitoring and management of related departments.

At present, some vehicle-mounted terminal devices, such as advanced auxiliary driving devices and driver behavior monitoring devices, are intended to monitor dangerous driving behaviors of a driver, such as a fatigue state, a distraction state and the like, in a video monitoring mode and the like, and intervene in a real-time voice alarm or vibration mode. In order to improve the operation safety, a public transport company usually installs the bus in a bus cab, but the false alarm rate is very high, so that drivers are extremely repelled. The reason is that the existing system has single consideration factor, mainly detects the unilateral dangerous behavior states such as fatigue and distraction, does not distinguish the grades for intervention, and simultaneously, the targeted objects are all drivers, so that the monitoring result is not targeted enough, and the accuracy requirement required by the early warning of the bus drivers cannot be met.

Most of the related researches are also aimed at evaluating the driving risk state of all drivers. For example, chinese patent CN 110826848A discloses a driver score evaluation method considering driver basic information, history violation information, traffic accident information, and the like, and evaluates driver risk based on an analytic hierarchy process. In other research aiming at the aspect of driving suitability of the driver, for example, chinese patent CN 202198603U discloses a driving suitability detection device based on simulated driving, based on which the moving vision, dark adaptation capability, night vision, depth perception capability and treatment judgment capability of the driver can be detected and the driving suitability thereof can be comprehensively evaluated. However, the evaluation indexes related to the above researches are all non-dynamic, only the influence of the historical static factors and the driving ability difference of the driver is considered, and the evaluation result is only suitable for the screening that the driver can not drive, and can not be further used for evaluating the real-time dangerous driving state of the driver.

In the aspect of evaluating the real-time risk state of the driver, the current research results focus on evaluating the driving state of the driver by means of video monitoring or physiological signals, and for example, the chinese patent CN 105448038B discloses an alarm method for detecting the state of the driver of a taxi based on brain wave signals, which obtains the abnormal state of the driver and gives an alarm by analyzing blood sugar signals and brain wave signals. Chinese patent CN 110338821A discloses a system and method for identifying the driving state of a driver based on facial expressions of the driver. The detection of the abnormal state of the driver by the research focuses on the physiological abnormal state of the driver, and the influence of driving risk behaviors on the driving state of the driver is not considered.

From the existing research, it is not comprehensive to evaluate the driving risk state only by a single index, and particularly for the public traffic driver, the uncomfortable post state and the degree grade of the dangerous driving state are related to the safety attributes of the personal driving habit, the driving ability, the historical violation accidents and the like of the driver, the real-time psychological conditions of the public traffic driver, such as the internal emotion, the working pressure and the like, the driving duration, the working period and the like. Therefore, the real-time dynamic post adaptability evaluation of the bus driver needs to be the result of comprehensive consideration, otherwise, the evaluation result has certain limitation, and the comprehensive post adaptability state of the driver cannot be monitored in real time.

In summary, in the aspect of real-time post adaptability of a bus driver, a post adaptability evaluation method and a system which comprehensively consider dynamic and static attributes and historical real-time characteristics of the bus driver are urgently needed. Therefore, under the support of intelligent monitoring and early warning technology and demonstration (2021 YFC 3001500) of post adaptation state of drivers of commercial vehicles and ships in a national key research and development planning project, the invention provides a comprehensive evaluation method and a comprehensive evaluation system for post adaptation performance of a bus driver, which consider the combination of dynamic and static conditions and historical real-time conditions, so as to fill up the domestic blank and lay a cushion for subsequent related research.

Disclosure of Invention

In order to solve the problems in the background art, the first purpose of the invention is to provide a real-time on-duty fitness evaluation method for a bus driver, and the technical scheme is as follows:

a real-time evaluation method for the on-duty adaptability of a bus driver comprises the following steps:

step S1: after a driver enters a cab to start driving a vehicle, the vehicle-mounted terminal equipment establishes contact with the background main control equipment;

step S2: collecting video image data, physiological signal data and dynamic track data of a driver and a driving vehicle in real time, transmitting the video image data, the physiological signal data and the dynamic track data to a background main control device, and carrying out identity recognition on the driver by processing the video image data;

and step S3: storing video image data, physiological signal data and dynamic track data in real time, and calling the existing information data of the driver according to the identity recognition result;

and step S4: constructing an on-duty adaptive index system of the public transport driver according to the existing information data of the driver and the video image data, the physiological signal data and the dynamic track data which are stored in real time, and generating an index database of the driver;

step S5: preprocessing data in an index database of a driver to screen out repeated, abnormal and missing data, and then normalizing the preprocessed data;

step S6: analyzing the types of the evaluation indexes of the on-duty adaptability of the driver after the normalization processing, calibrating the threshold values of the evaluation indexes, auditing the evaluation indexes of the current driver according to the threshold values of the evaluation indexes, and evaluating the single-index adaptability of the current driver;

step S7: if the single index post adaptability evaluation result is unqualified, the driver or the driven vehicle is directly subjected to corresponding real-time control; if the single index post adaptability evaluation result is qualified, inputting each checked evaluation index into a neural network-based post adaptability comprehensive evaluation model of the bus driver, wherein the model is constructed by training an index database of the driver in the previous year by adopting a neural network algorithm, comprehensively evaluating the current post adaptability state of the driver by utilizing the model, and judging the post adaptability grade of the driver;

step S8: and correspondingly managing and controlling the driver or the driving vehicle in real time according to different duty grades.

Preferably, in step S4, the constructed on-duty adaptability index system includes a four-layer structure, where the highest layer is a target layer and is on-duty adaptability of a bus driver; the second layer is a first-level index layer and comprises five indexes which are an individual characteristic, a driving occupational characteristic, a driving state, a physiological state and a psychological state respectively; the three layers are secondary index layers, and each index of the primary index layer is decomposed respectively; and the fourth layer is a third-level index layer, the indexes of the second-level index layers are refined, and the third-level indexes and the non-refined second-level indexes are used as evaluation indexes of the on-duty adaptability of the driver.

Preferably, the subordinate secondary indexes of the individual features are individual attributes and social attributes; at the level of three-level indexes, the individual attributes comprise sex, age and native place, and the social attributes comprise culture degree and marital state;

the subordinate secondary indexes of the driving occupation characteristics comprise occupation attributes, driving attributes, safety attributes and capability attributes, and on the level of the tertiary indexes, the occupation attributes comprise income, driving age, employment type, occupation value, occupation burnout and whether overtime exists; the driving attributes comprise the current day accumulated driving time, the current day accumulated driving mileage, the driving time period, the station entering and exiting specifications, the lane use specifications, the intersection passing specifications and the driving habit specifications when the sampling time is up; the safety attributes comprise a driving style, a driving attitude, accumulated accident times, the level of the highest liability accident, monthly average safe driving mileage, monthly average safe violation times, monthly average service violation times, monthly average criticizing education times and monthly average punishment times; the capability attributes comprise the professional skill level of a driver, the suitability evaluation level of a road transport driver, the instant memory capability, the complex reaction capability, the communication capability, the anti-interference capability, the compression resistance capability, the anti-monotone capability and the anti-fatigue capability;

subordinate secondary indexes of the driving state comprise drunk driving, emotional driving, fatigue driving, distraction driving and aggressive driving; the subordinate secondary indexes of the physiological state comprise body temperature, pulse, respiration, dynamic blood pressure and diseases; the subordinate secondary indexes of the psychological state comprise personality, mental health, emotional stability and emotional resilience.

Preferably, in step S5, the evaluation index is defined by integers in order from 1 for each category of the classification data; for numerical data, determining values according to specific collected values, and preprocessing data in an index database of a driver, wherein the values are specifically as follows:

for individual features: the gender male was assigned 1 and the gender female was assigned 2; age is taken as its actual value; assigning a value of 1 by a native driver and a value of 2 by an external driver; the culture degree is 1 according to the value of junior middle school and below, 2 according to the value of middle school and high school, 3 according to the value of major specialty, and 4 according to the value of the subject and above; in the marital state, the unmarried value is 1, the married value is 2 and the divorce value is 3;

for driving occupation characteristics: the recruitment type is valued according to the recruitment type of the driver, the internal training value is 1, the external training value is 2, the internal training refers to a self-culture driver and belongs to planned internal recruitment, and the external training refers to a social driver and belongs to planned external recruitment; whether the overtime index is present or not is evaluated as 1 when the overtime index is currently in the overtime state, and the normal working time is evaluated as 2; the driving time periods are classified according to traffic conditions, the value of the driving time period in a flat peak period is 1, the value of the driving time period in a peak period is 2, the time periods in each peak period are different due to different current routes of buses, and the specific time is generally given by a bus company; the driving style index value is obtained by analyzing an index database, the current driving style type of the driver is judged once at each sampling time, and the aggressive type value is 1, the common type value is 2 and the conservative type value is 3; the income and the driving age respectively take the actual numerical values, and the income refers to the average income per month by the current time; the suitability evaluation grade of the road transportation driver, the professional skill grade of the driver, the professional value view, the instantaneous memory capacity, the complex reaction capacity, the communication capacity, the anti-interference capacity, the pressure resistance capacity, the monotonous resistance capacity and the fatigue resistance capacity of the driver are judged by enterprises in each season or each year, and the actual grade value is taken; the driving attitude and the occupation burnout are respectively measured according to a driver safety attitude scale and an MBI-GS measurement scale, and the measurement score of enterprises for drivers in each quarter or each year is a specific value; accumulating the accident frequency, the highest responsible accident grade, the monthly safe driving mileage, the monthly safe violation frequency, the monthly service violation frequency, the monthly criticizing education frequency and the monthly punishment frequency to obtain the actual acquisition value, and updating once a month, wherein the monthly safe driving mileage refers to the average value of the monthly safe driving mileage of the driver in the past year, and other indexes are in the same reason; respectively taking actual accumulated values of the current day accumulated driving time length and the current day accumulated driving mileage which are obtained by cutting the current day to sampling time; the method comprises the following steps of taking an actual value of the accumulated nonstandard times of a driver from the current day to sampling time according to the incoming and outgoing station specification, the lane use specification, the intersection passing specification and the driving habit specification;

for the driving state: the drunk driving judgment method comprises the steps that according to information collected by an alcohol sensor in vehicle-mounted terminal equipment, alcohol concentration in a cab is detected within the first sampling time after a vehicle is started, whether drunk driving is performed or not is determined, the higher alcohol concentration is assigned to be 1, and the normal concentration is assigned to be 0; aiming at emotional driving, fatigue driving and distraction driving, the degree of the bad driving state of the driver is judged by analyzing the index database, and the driver is sequentially assigned with values of 1, 2, 3, 4 and 5 according to the influence of the index database on the driving safety, wherein the larger the value is, the higher the risk level of the driving behavior is, the more unsafe the driving behavior is; the offensive driving behavior is judged according to the driving vehicle driving track analysis result, and is sequentially assigned as 1, 2, 3, 4 and 5 according to the influence of the offensive driving behavior on driving safety, wherein the larger the numerical value is, the higher the offensive driving behavior degree is, and the larger the risk is;

for a physiological state: the information of body temperature, pulse, respiration and dynamic blood pressure is selected as a value; the disease information is used for judging the disease condition of the driver, predicting that the driver has a sudden disease value of 1 and does not have a sudden disease value of 0;

for the psychological state: aiming at personality, mental health, emotional stability and emotional resilience, according to the results of the psychological assessment of the drivers in each season or each year by enterprises, the corresponding scales are taken as specific numerical values.

Preferably, in step S5, the function of the normalization process is:

x′= (x-x _min )/(x _max -x _min )

wherein, the first and the second end of the pipe are connected with each other,x' is some evaluation index data after normalization;xthe data is certain evaluation index data before normalization;x _max the maximum value of all data of a certain evaluation index of the driver before normalization;x _min the minimum value is the minimum value of all data of certain evaluation indexes of the driver before normalization.

Preferably, in step S6, the calibration of each evaluation index threshold is as follows:

aiming at classification data, evaluation indexes without calibration of a threshold value comprise sex, native place, cultural degree, marital state, employment type, overtime condition or not, driving time period, driving style, suitability evaluation level of road transportation drivers, professional skill level of the drivers, professional value view, instantaneous memory capability, complex reaction capability, communication capability, anti-interference capability, compression resistance capability, monotone resistance capability and fatigue resistance capability; the evaluation indexes needing to be calibrated with the threshold value comprise emotional driving, fatigue driving, distracted driving and aggressive driving, and the numerical value of '5' is used as the threshold value T of the evaluation indexes of the on-Shift fitness _max (ii) a For other classification data including drunk driving and diseases, the numerical value '1' is taken as the index threshold T _max ；

For numerical data:

1) For evaluation indexes meeting or being converted into normal distribution, including body temperature, pulse, respiration, dynamic blood pressure, driving attitude, occupation listlessness, personality, mental health, emotional stability and emotional recovery capacity, the evaluation indexes are screened according to the 3 sigma principle, and a mu +3 sigma point is selected as a threshold point T on the evaluation index _max Mu-3. Sigma. Point as the lower threshold Point T of the evaluation index _min ：

T _max ＝μ+3σ

T _min ＝μ-3σ

Wherein x is _i ' is the value obtained after the normalization of the ith sample, mu is the mean value of the normalized samples, sigma is the standard deviation of the samples, n is the number of the samples, T _max As an upper threshold value of the evaluation index, T _min A lower threshold value of the evaluation index;

2) For numerical variables which do not conform to normal distribution, determining the threshold value of the variable as T by an enterprise where the driver is located according to index characteristics of group drivers and industry requirements _max The method specifically comprises the steps of accumulated accident frequency, highest accident grade, monthly safe driving mileage, monthly safe violation frequency, monthly service violation frequency, monthly criticizing education frequency, monthly punishment frequency, daily accumulated driving duration, daily accumulated driving mileage, station entrance and exit specifications, lane use specifications, intersection passing specifications and driving habit specifications.

Preferably, in step S7, the independent variables of the comprehensive on-duty adaptability evaluation model of the public transport driver based on the neural network are each evaluation indexx _i The dependent variable is the post-adaptive state of the driver and is represented by the risk probability of the accident of the driver, and the value range is [0,1 ]]Will beyIf =1 indicates that the driver has an accident:

z ₁ =p(y=1|x ₁ , x ₂ ,x ₃ ,… ,x _n )

wherein the content of the first and second substances,x _i evaluating indexes for the driver;yif the accident variable is determined to be whether the driver has an accident, if the accident variable is determined to be 1, the accident variable is determined to be 0;nexpressing the number of evaluation indexes;z ₁ probability of an accident for the driver;

the learning rule when constructing the model is as follows:

1) Determining number of network input layer nodesnNumber of hidden layer nodesmAnd the number of output layer nodes is 2; determining input variablesx _i Output variabley _k (ii) a Initializing hidden layer node biasesa _j And each node of the output layer is biasedb _k And the connection weight between nodes of each layerω _ij ，ω _jk ；

Wherein the content of the first and second substances,i=1,2,…,n；j=1,2,…,m；k=1,2；ω _ij representation and hidden layer nodesjConnected input layer nodesiA weight;ω _jk presentation and output layer nodeskConnected hidden layer nodesjThe weight of (c);

2) Computing values for nodes in a hidden layerH _j Output layer node valuesz _k ；

3) Calculating the accumulated prediction error of the networkeThe sum of squares of errors between the output value of each node of the output layer and the true value;

4) Judging the network training effect; if it ise≤εIf the network training result is better, determining the neural network structure and parameters of each part; if it ise>εUpdating the network connection weight by gradient descent methodω _ij Andω _jk and network node biasinga _j 、b _k And simultaneously turning to 2) to carry out the next iteration;

5) Determining an output result; each evaluation index updated in real time is brought into the determined model to obtainOutput layer variablesy _k Output value of output layer corresponding to =1z ₁ The probability of the driver accident risk is called as the post-fitting state.

Preferably, in step S7, the suitable duty grade includes five grades, which are:

job-adapted grade 'first grade': z ₁ belongs to [0, 0.4), and shows that the accident risk probability of the driver is lower at present, and the post-fitting state is good;

job-adapted grade "second grade": z ₁ belongs to [0.4, 0.6), and shows that the accident risk probability of the driver is moderate at present, and the post-adaptive state is general;

suitable job grade is 'three-level': z ₁ belongs to [0.6, 0.8) ], which indicates that the accident risk probability of the current driver is higher and the suitable job state is poorer;

the suitable job grade is 'four grades': z ₁ not less than 0.8, which indicates that the driver is easy to have accidents at present, the post-fitting state is extremely poor, and the post-fitting state is not good;

job-adapted grade "five grade": if it isx′≥T _max Orx′≤T _min (ii) a The evaluation of the single-index post adaptability of the driver is unqualified at present, and the driver is in a state of improper post; when the suitable post grade is judged to be a first grade and a second grade, the suitable post state is good, and the driving work of a driver is not interfered; when the suitable post grade is judged to be three-grade, the suitable post state is poor, and the vehicle-mounted terminal equipment broadcasts the grade evaluation result and reason to the driver in a voice mode; when the shift level is judged to be four levels, the shift state is extremely poor, the vehicle-mounted terminal device broadcasts the level evaluation result and reason to the driver in a voice mode, and simultaneously the vehicle is switched to an automatic driving mode; and when the adaptive grade is judged to be five grade, the vehicle-mounted terminal equipment is in an inappropriate state, broadcasts the single index adaptive evaluation result and the inappropriate factors to the driver by voice, and switches the vehicle to an automatic driving mode to terminate the driving work of the driver.

The invention provides a real-time on-duty adaptability evaluation system for a bus driver, which comprises a vehicle-mounted terminal device and a background main control device, wherein the vehicle-mounted terminal device comprises a video image acquisition device, a physical sign parameter acquisition device, a vehicle information acquisition device, an information interaction unit, a terminal early warning module and a terminal control module; the background main control equipment comprises an information interaction unit, an identity recognition module, a data storage module, a data processing module, a single index evaluation module and a post-fitting comprehensive evaluation module;

the video image acquisition equipment is used for acquiring video image data of a driver;

the physical sign parameter acquisition equipment is used for acquiring physiological signal data of a driver;

the vehicle information acquisition equipment is used for acquiring dynamic data of a driving vehicle running track;

the information interaction unit of the vehicle-mounted terminal equipment and the information interaction unit of the background main control equipment are used for mutually transmitting data between the vehicle-mounted terminal equipment and the background main control equipment;

the identity recognition module is used for recognizing the face of the driver so as to determine the identity of the driver;

the data storage module is used for storing existing driver information data and real-time received transmission data input by an enterprise and generating an index database of a driver;

the data processing module is used for preprocessing data in the index database to screen out repeated, abnormal and missing data and then carrying out normalization processing on the preprocessed data;

the single index evaluation module analyzes the types of the evaluation indexes of the on-duty adaptability of the driver after normalization processing, calibrates the threshold values of the evaluation indexes, audits the evaluation indexes of the current driver according to the evaluation index threshold values, and evaluates the on-duty adaptability of the single index of the current driver;

the post-fitting comprehensive evaluation module is based on a neural network bus driver post-fitting comprehensive evaluation model, when a single-index post-fitting comprehensive evaluation result is qualified, each evaluation index is input into the neural network, the model is used for comprehensively evaluating the current post-fitting state of a driver, and the post-fitting grade of the driver is judged;

the terminal early warning module is used for switching the driving mode of driving the vehicle when the suitable duty grade is four-grade or five-grade, namely the suitable duty state is not suitable.

The invention has the following beneficial effects:

1. the invention focuses on the occupational characteristics of the public transport driver, constructs an all-directional on-duty adaptive index system of the driver by combining the dynamic and static indexes and the historical real-time indexes of the public transport driver, improves the problems of single traditional evaluation system and the like, and enhances the hierarchy and comprehensiveness of the evaluation system.

2. The invention relies on the information acquisition equipment on the vehicle-mounted terminal equipment to acquire the video image characteristics and the physical sign characteristics of the driver in real time and transmit the video image characteristics and the physical sign characteristics to the background main control equipment for analysis and processing, thereby improving the timeliness of the on-duty suitability evaluation of the driver; and the implementation process is flexible and simple, the device cost is lower, and the device is suitable for large-scale use.

3. After the driver identity is identified, all-directional historical information of the driver needs to be called, and proper duty state rule mining is carried out. Aiming at different drivers, the excavated post-fitting rules are different, so that the constructed evaluation rules are different; after the evaluation levels of the drivers are determined, different management and control measures are taken according to different levels, so that the design of pertinence and individuation of individual requirements of the drivers is realized, and related enterprises can monitor the on-duty and applicable states of the drivers and can manage the drivers in a targeted manner.

Drawings

Other objects and results of the present invention will become more apparent and more readily appreciated as the same becomes better understood by reference to the following description taken in conjunction with the accompanying drawings. In the drawings:

FIG. 1 is a flow chart of the method of the present invention.

Fig. 2 is a block diagram of the system architecture of the present invention.

FIG. 3 is a post-fitting index system for a bus driver in the invention.

FIG. 4 is a model diagram of the neural network-based on-post adaptive comprehensive evaluation of the public transport driver.

Detailed Description

In order to make the technical solutions and advantages thereof better understood by those skilled in the art, the following detailed description is made with reference to the accompanying drawings, but the present invention is not limited to the scope of the present invention.

Example 1

Referring to FIG. 1: a real-time evaluation method for the on-duty adaptability of a bus driver comprises the following steps:

and step S3: the method comprises the steps of storing video image data, physiological signal data and dynamic track data in real time, and calling existing information data of a driver according to an identity recognition result, wherein the existing information of the driver refers to relevant index information regularly collected by an enterprise during the period of the enterprise's duties, such as individual characteristics of the driver, driving occupation characteristics, psychological state and the like;

step S7: if the single index post adaptability evaluation result is unqualified, the driver or the driven vehicle is directly subjected to corresponding real-time control; if the single index post adaptability evaluation result is qualified, inputting each examined evaluation index into a comprehensive on-post adaptability evaluation model of the public transport driver based on the neural network, wherein the model is constructed by training an index database of the driver in the previous year by adopting a neural network algorithm, and continuously adjusting the weight and the threshold of the network to minimize the square sum of errors of the network through a certain learning rule, thereby determining the internal structure and the related parameters of the neural network, comprehensively evaluating the current post adaptability state of the driver by utilizing the model, and evaluating the post adaptability grade of the driver;

Referring to FIG. 3: in the step S4, the constructed on-duty adaptability index system comprises a four-layer structure, wherein the highest layer is a target layer and is the on-duty adaptability of a bus driver; the second layer is a first-level index layer and comprises five indexes which are an individual characteristic, a driving occupational characteristic, a driving state, a physiological state and a psychological state respectively; the three layers are secondary index layers, and each index of the primary index layer is decomposed respectively; the fourth layer is a third index layer and is used for refining each index of each second index layer; and taking the third-level index and the non-refined second-level index as evaluation indexes of the on-duty adaptability of the driver.

Further, the subordinate secondary indexes of the individual characteristics are individual attributes and social attributes; at the level of three levels of indexes, the individual attributes comprise sex, age and native place, and the social attributes comprise cultural degree and marital state;

the subordinate secondary indexes of the driving occupation characteristics comprise occupation attributes, driving attributes, safety attributes and capability attributes, and on the level of the tertiary indexes, the occupation attributes comprise income, driving age, employment type, occupation value, occupation burnout and whether overtime exists; the driving attributes comprise the current accumulated driving time length up to the sampling time, the current accumulated driving mileage, the driving time period, the station entering and exiting specifications, the lane use specifications, the intersection passing specifications and the driving habit specifications; the safety attributes comprise a driving style, a driving attitude, accumulated accident times, the grade of the highest liability accident, monthly average safe driving mileage, monthly average safe violation times, monthly average service violation times, monthly average criticizing education times and monthly average punishment times; the capability attributes comprise the professional skill level of a driver, the suitability evaluation level of a road transport driver, the instant memory capability, the complex reaction capability, the communication capability, the anti-interference capability, the compression resistance capability, the anti-monotone capability and the anti-fatigue capability;

the subordinate secondary indexes of the driving state comprise drunk driving, emotional driving, fatigue driving, distraction driving and aggressive driving; the subordinate secondary indexes of the physiological state comprise body temperature, pulse, respiration, dynamic blood pressure and diseases; the subordinate secondary indexes of the psychological state comprise personality, mental health, emotional stability and emotional recovery capability.

Further, in step S5, in the evaluation index, for the classification data, integers are sequentially defined from 1 according to the class; aiming at numerical data, determining values according to specific collected values, and preprocessing data in an index database of a driver, wherein the values are specifically as follows:

for individual features: the gender male was assigned 1 and the gender female was assigned 2; age is taken as its actual value; assigning a value of 1 by a native driver and a value of 2 by an external driver; the culture degree is 1 according to the assignment of junior high school and below, 2 according to the assignment of high school and high school, 3 according to the assignment of major specialty, and 4 according to the assignment of this department and above; in the marital state, the unmarried value is 1, the married value is 2, and the divorce value is 3;

for driving occupation characteristics: the recruitment type is valued according to the recruitment type of the driver, the internal training value is 1, the external training value is 2, the internal training refers to a self-culture driver and belongs to planned internal recruitment, and the external training refers to a social driver and belongs to planned external recruitment; whether the overtime index is present or not is evaluated as 1 when the current overtime index is in the overtime state, and the normal working time is evaluated as 2; the driving time periods are classified according to traffic conditions, the value of the driving time period is 1 when the driving is in a flat peak period, the value of the driving time period is 2 when the driving is in a peak period, the time periods of each peak period are different due to different current routes of buses, and the specific time is generally given by a bus company; the driving style index value is obtained by analyzing an index database, the current driving style type of the driver is judged once at each sampling time, and the aggressive type value is 1, the common type value is 2 and the conservative type value is 3; the income and the driving age respectively take the actual numerical values, and the income refers to the average income per month by the current time; the suitability evaluation grade of the road transportation driver, the professional skill grade of the driver, the professional value, the instantaneous memory capacity, the complex reaction capacity, the communication capacity, the anti-interference capacity, the pressure resistance capacity, the monotonous resistance capacity and the fatigue resistance capacity of the driver are judged by enterprises in each season or each year, and the actual grade value is taken; the driving attitude and the occupation burnout are respectively measured according to a driver safety attitude scale and an MBI-GS measurement scale, and the measurement score of enterprises for drivers in each quarter or each year is a specific value; accumulating the accident frequency, the highest responsible accident grade, the monthly safe driving mileage, the monthly safe violation frequency, the monthly service violation frequency, the monthly criticizing education frequency and the monthly punishment frequency to obtain the actual collection value, and updating once per month, wherein the monthly safe driving mileage refers to the average value of the safe driving mileage of the driver in the past year, and other indexes are the same; respectively taking actual accumulated values of the current day accumulated driving duration and the current day accumulated driving mileage, which are obtained by cutting the current day to sampling time; the method comprises the following steps of taking an actual value of the accumulated times of nonstandard times of a driver up to the current day of sampling time according to the specification of an incoming station and an outgoing station, the specification of lane use, the specification of crossing traffic and the specification of driving habits;

for the driving state: the drunk driving judgment method comprises the steps that according to information collected by an alcohol sensor in vehicle-mounted terminal equipment, alcohol concentration in a cockpit is detected within the first sampling time of vehicle starting, whether drunk driving is performed or not is determined, the alcohol concentration is assigned to be 1 when the alcohol concentration is higher, and the alcohol concentration is assigned to be 0 when the alcohol concentration is normal; aiming at emotional driving, fatigue driving and distraction driving, the degree of the bad driving state of the driver is judged by analyzing the index database, and the driver is sequentially assigned with values of 1, 2, 3, 4 and 5 according to the influence of the index database on the driving safety, wherein the larger the value is, the higher the risk grade of the driving behavior is, the more unsafe the driving behavior is; judging the degree of the adverse state of the aggressive driving behavior according to the analysis result of the driving track of the driving vehicle, sequentially assigning values to 1, 2, 3, 4 and 5 according to the influence of the aggressive driving behavior on driving safety, wherein the larger the numerical value is, the higher the degree of the aggressive driving behavior is, the larger the risk is;

for a physiological state: the information of body temperature, pulse, respiration and dynamic blood pressure is selected as a value; the disease information is used for judging the disease condition of the driver, the value of 1 represents that the dangerous event exists, the value of 1 is predicted to be the value of 1 when the sudden disease exists, and the value of 0 is predicted to be the value of 0 when the sudden disease does not exist;

aiming at the psychological state: aiming at personality, mental health, emotional stability and emotional resilience, according to the results of the psychological assessment of the drivers in each season or each year by enterprises, the corresponding scales are taken as specific numerical values.

Further, in step S5, the function of the normalization process is:

x′= (x-x _min )/(x _max -x _min )

wherein the content of the first and second substances,x' is some evaluation index data after normalization;xthe data is certain evaluation index data before normalization;x _max the maximum value of all data of a certain evaluation index of the driver before normalization;x _min the minimum value of all data of certain evaluation indexes of the driver before normalization.

Further, in step S6, the calibration of each evaluation index threshold is specifically as follows:

for the classification data:

evaluation indexes without calibration of a threshold value comprise sex, native place, culture degree, marital state, employment type, whether overtime exists or not, driving time period, driving style, suitability evaluation grade of a road transportation driver, professional skill grade of the driver, professional value, instantaneous memory capacity, complex reaction capacity, communication capacity, anti-interference capacity, compression resistance, anti-monotony capacity and anti-fatigue capacity; the evaluation indexes needing to calibrate the threshold value comprise emotional driving, fatigue driving, distracted driving and aggressive driving, and the numerical value of '5' is used as the threshold value T of the evaluation indexes of the on-duty adaptability _max (ii) a For other classification data including drunk driving and diseases, the numerical value '1' is used as the index threshold T _max ；

For numerical data:

1) For satisfying or convertible toNormally distributed evaluation indexes including body temperature, pulse, respiration, dynamic blood pressure, driving attitude, occupational listlessness, personality, mental health, emotional stability and emotional recovery capability are screened according to the 3 sigma principle, and a mu +3 sigma point is selected as a threshold point T on the evaluation index _max Mu-3. Sigma. Point as the lower threshold Point T of the evaluation index _min ：

T _max ＝μ+3σ

T _min ＝μ-3σ

Wherein x is _i ' is the value obtained after the normalization of the ith sample, mu is the mean value of the normalized sample, sigma is the standard deviation of the sample, n is the number of samples, T _max As an upper threshold value of the evaluation index, T _min A lower threshold value of the evaluation index;

2) For the numerical variable which does not conform to normal distribution, determining the threshold value of the variable as T by the enterprise where the driver is located according to the index characteristics of the group drivers and the industry requirements _max The method specifically comprises the steps of accumulated accident frequency, highest accident grade, monthly safe driving mileage, monthly safe violation frequency, monthly service violation frequency, monthly criticizing education frequency, monthly punishment frequency, daily accumulated driving duration, daily accumulated driving mileage, station entrance and exit specifications, lane use specifications, intersection passing specifications and driving habit specifications.

Further, in step S7, independent variables of the on-duty adaptability comprehensive evaluation model of the bus driver based on the neural network serve as evaluation indexesx _i The dependent variable is the appropriate state of the driver and is represented by the risk probability of the accident of the driver, and the value range is [0,1 ]]Will bey=1 indicates that the driver has an accident, then:

z ₁ =p(y=1|x ₁ , x ₂ ,x ₃ ,… ,x _n )

wherein, the first and the second end of the pipe are connected with each other,x _i evaluating indexes for the driver;yif the accident variable is determined to be whether the driver has an accident, if the accident variable is determined to be 1, the accident variable is determined to be 0;nexpressing the number of evaluation indexes;z ₁ probability of accident for the driver;

referring to fig. 4, the learning rule when constructing the model is as follows:

5) Determining an output result; each evaluation index updated in real time is brought into the determined model to obtain the output layer variabley _k Output value of output layer corresponding to =1z ₁ The probability of the accident risk of the driver is called as the suitable post state.

Further, in step S7, the suitable duty level includes five levels, which are:

post fitting grade is 'first grade': z ₁ belongs to [0, 0.4), and shows that the accident risk probability of the driver is lower at present, and the post-fitting state is good;

suitable job grade is 'three-level': z ₁ belongs to [0.6, 0.8), and shows that the accident risk probability of the driver is higher at present, and the post-adaptive state is poorer;

job-adapted grade "five grade": if it isx′≥T _max Orx′≤T _min (ii) a The evaluation of the single-index post adaptability of the driver is unqualified at present and the driver is in a state of no good job;

when the post adaptation grade is first grade or second grade, namely the current post adaptation state of the bus driver is good, and the system does not take any intervention measures on the bus driver at the moment;

when the suitable post grade is judged to be three-level, the suitable post state is poor, the driver needs to be warned, and the connection weight between the neural network input layer and the hidden layer is constructed according toω _ij Arranging the absolute values of the input layer nodes according to the sequence from large to small, selecting the first three numbers with the largest absolute value, and setting the corresponding input layer nodes as three indexes which have the largest influence on the on-duty adaptability of the bus driver; for example, if the connection weightsω _ij The node indexes of the input layer corresponding to the first three numbers with the maximum absolute value are fatigue driving, the highest continuous working time of the day and the lane use specification respectively, and then the voice alarm for the driver is as follows: "you are good, your current suitable duty state is three-level, the system has detected you are in low fatigue driving state at present, the highest continuous working time of the day has reached 3h, the number of times of using the lane is not standard of the day has reached 5 times, you need to adjust your current driving state; the system can monitor the state change of your in real time and please drive cautiously;

when the post adaptation level is judged to be level four, the post adaptation state is extremely poor, namely the bus driver is not suitable for continuously driving the bus, the system can automatically convert the bus into an automatic driving mode, early warning is carried out at the same time, the three index determining methods with the largest influence on post adaptation are consistent with the three-level result processing method, and the voice warning result of the driver is changed as follows: "you are good, your current suitable duty state is four levels, the system detects that you are currently in a high fatigue driving state, the highest continuous working time of the day has reached 3h, the nonstandard times of lane use of the day has reached 5 times, you are required to stop driving work immediately, and the system is converted into an automatic driving mode";

when the suitable post grade is judged to be five grade, the bus is in a state of being not suitable for post, namely the bus driver is suitable for the single indexIf the post evaluation is not qualified, the system automatically converts the vehicle into an automatic driving mode and simultaneously warns a bus driver of unqualified indexes; for example, if the evaluation index exceeding the threshold range is the body temperature, the voice broadcast to the driver is as follows: "you are good, your current suitable duty state is five grades, and the system detects your current high temperature, is 38.5 degrees centigrade, needs you stop driving work soon, and the system will change into automatic driving mode".

Example 2

Referring to FIG. 2: a real-time on-duty adaptability evaluation system for a bus driver comprises a vehicle-mounted terminal device and a background main control device, wherein the vehicle-mounted terminal device comprises a video image acquisition device, a sign parameter acquisition device, a vehicle information acquisition device, an information interaction unit, a terminal early warning module and a terminal control module; the background main control equipment comprises an information interaction unit, an identity recognition module, a data storage module, a data processing module, a single index evaluation module and a post-fitting comprehensive evaluation module;

the data storage module is used for storing the existing driver information data and the real-time received transmission data input by an enterprise and generating an index database of the driver;

the post-fitting comprehensive evaluation module is based on a neural network-based post-fitting comprehensive evaluation model of a bus driver, when a single-index post-fitting evaluation result is qualified, each evaluation index is input into the neural network-based post-fitting comprehensive evaluation model, the current post-fitting state of the driver is comprehensively evaluated by using the model, and the post-fitting grade of the driver is judged;

the terminal early warning module is used for switching the driving mode of the vehicle when the suitable duty grade is four-grade or five-grade, namely the suitable duty state is not suitable duty.

The above description is only for the specific embodiments of the present invention, but the scope of the present invention is not limited thereto, and any person skilled in the art can easily conceive of the changes or substitutions within the technical scope of the present invention, and all the changes or substitutions should be covered within the scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the appended claims.

Claims

1. A real-time evaluation method for the on-duty adaptability of a bus driver is characterized by comprising the following steps:

step S6: analyzing the types of all evaluation indexes of the on-duty adaptability of the driver after normalization processing, calibrating the threshold value of each evaluation index, auditing all evaluation indexes of the current driver according to all evaluation index threshold values, and evaluating the single-index adaptability of the current driver;

step S7: if the single index post adaptability evaluation result is unqualified, the driver or the driven vehicle is directly subjected to corresponding real-time control; if the single-index post adaptability evaluation result is qualified, inputting each examined evaluation index into a neural network-based public transport driver post adaptability comprehensive evaluation model, wherein the model is constructed by training an index database of the driver in the previous year by adopting a neural network algorithm, comprehensively evaluating the current post adaptability state of the driver by utilizing the model, and evaluating the post adaptability grade of the driver;

2. The method for evaluating the on-duty adaptability of the bus driver in real time according to the claim 1 is characterized in that: in the step S4, the constructed on-duty adaptability index system comprises a four-layer structure, wherein the highest layer is a target layer and is the on-duty adaptability of a bus driver; the second layer is a first-level index layer and comprises five indexes which are an individual characteristic, a driving occupational characteristic, a driving state, a physiological state and a psychological state respectively; the three layers are secondary index layers, and each index of the primary index layer is decomposed respectively; the fourth layer is a third index layer which is used for refining each index of each second index layer; and taking the third-level index and the non-refined second-level index as evaluation indexes of the on-duty adaptability of the driver.

3. The method for evaluating the on-duty adaptability of the bus driver in real time according to claim 2 is characterized in that:

the subordinate secondary indexes of the individual characteristics are individual attributes and social attributes; at the level of three levels of indexes, the individual attributes comprise sex, age and native place, and the social attributes comprise cultural degree and marital state;

the subordinate secondary indexes of the driving occupation characteristics comprise occupation attributes, driving attributes, safety attributes and capability attributes, and on the level of the tertiary indexes, the occupation attributes comprise income, driving age, employment type, occupation value, occupation burnout and whether overtime happens or not; the driving attributes comprise the current day accumulated driving time, the current day accumulated driving mileage, the driving time period, the station entering and exiting specifications, the lane use specifications, the intersection passing specifications and the driving habit specifications when the sampling time is up; the safety attributes comprise a driving style, a driving attitude, accumulated accident times, the grade of the highest liability accident, monthly average safe driving mileage, monthly average safe violation times, monthly average service violation times, monthly average criticizing education times and monthly average punishment times; the capability attributes comprise the professional skill level of a driver, the suitability evaluation level of a road transport driver, the instant memory capability, the complex reaction capability, the communication capability, the anti-interference capability, the compression resistance capability, the monotonous resistance capability and the fatigue resistance capability;

4. The method for evaluating the on-duty adaptability of the bus driver in real time according to claim 3, characterized in that: in the evaluation index, sequentially defining integers from 1 according to the category aiming at the classified data; aiming at numerical data, values are determined according to specific collected values, and data in an index database of a driver are preprocessed, wherein the values are specifically as follows:

for individual features: the sex male is assigned as 1 and the sex female is assigned as 2; age is taken as its actual value; native drivers assign a value of 1 and foreign drivers assign a value of 2; the culture degree is 1 according to the value of junior middle school and below, 2 according to the value of middle school and high school, 3 according to the value of major specialty, and 4 according to the value of the subject and above; in the marital state, the unmarried value is 1, the married value is 2, and the divorce value is 3;

for driving occupation characteristics: the recruitment type is valued according to the recruitment type of the driver, the internal training value is 1, the external training value is 2, the internal training refers to a self-culture driver and belongs to planned internal recruitment, and the external training refers to a social driver and belongs to planned external recruitment; whether the overtime index is present or not is evaluated as 1 when the overtime index is currently in the overtime state, and the normal working time is evaluated as 2; the driving time periods are classified according to traffic conditions, the value of the driving time period in a flat peak period is 1, the value of the driving time period in a high peak period is 2, the time periods in each peak period are different due to different current routes of buses, and the specific time is generally given by a bus company; the driving style index value is obtained by analyzing an index database, the current driving style type of the driver is judged once at each sampling time, and the aggressive type value is 1, the normal type value is 2 and the conservative type value is 3; the income and the driving age respectively take the actual numerical values, and the income refers to the average monthly income at the current time; the suitability evaluation grade, the professional skill grade, the professional value of the driver, the instantaneous memory capacity, the complex reaction capacity, the communication capacity, the anti-interference capacity, the compression resistance capacity, the anti-monotony capacity and the anti-fatigue capacity of the road transportation driver are evaluated according to the evaluation of enterprises on the driver every season or every year, and the actual grade value is obtained; the driving attitude and the occupation burnout are respectively measured according to a driver safety attitude scale and an MBI-GS measurement scale, and the measurement score of enterprises for drivers in each quarter or each year is a specific value; accumulating the accident frequency, the highest responsible accident grade, the monthly safe driving mileage, the monthly safe violation frequency, the monthly service violation frequency, the monthly criticizing education frequency and the monthly punishment frequency to obtain the actual acquisition value, and updating once a month, wherein the monthly safe driving mileage refers to the average value of the monthly safe driving mileage of the driver in the past year, and other indexes are in the same reason; respectively taking actual accumulated values of the current day accumulated driving duration and the current day accumulated driving mileage, which are obtained by cutting the current day to sampling time; the method comprises the following steps of taking an actual value of the accumulated nonstandard times of the driver up to the current day of sampling time according to the standard of the station entering and exiting, the standard of the lane use, the standard of the crossing traffic and the standard of the driving habit;

for the driving state: the drunk driving judgment method comprises the steps that according to information collected by an alcohol sensor in vehicle-mounted terminal equipment, the alcohol concentration in a cab is detected within the first sampling time after a vehicle is started, whether drunk driving is carried out or not is determined, the alcohol concentration is assigned to be 1 when the alcohol concentration is higher than a set value, and the alcohol concentration is assigned to be 0 when the alcohol concentration is normal; aiming at emotional driving, fatigue driving and distraction driving, the degree of the bad driving state of the driver is judged by analyzing the index database, and the driver is sequentially assigned with values of 1, 2, 3, 4 and 5 according to the influence of the index database on the driving safety, wherein the larger the value is, the higher the risk level of the driving behavior is, the more unsafe the driving behavior is; the degree of the bad state of the aggressive driving behavior is judged according to the analysis result of the driving track of the driving vehicle, and is sequentially assigned as 1, 2, 3, 4 and 5 according to the influence of the aggressive driving behavior on the driving safety, and the higher the numerical value is, the higher the degree of the aggressive driving behavior is, the higher the risk is;

for the psychological state: aiming at personality, mental health, emotional stability and emotional resilience, according to the results of the enterprise psychology evaluation on the drivers every season or every year, the corresponding scale scores are taken as specific numerical values.

5. The method for evaluating the on-duty adaptability of the bus driver in real time according to claim 4, characterized by comprising the following steps of:

in step S5, the function of the normalization process is:

x′= (x-x _min )/(x _max -x _min )

wherein the content of the first and second substances,x' is some evaluation index data after normalization;xis made ofSome evaluation index data before normalization;x _max the maximum value of all data of a certain evaluation index of the driver before normalization;x _min the minimum value is the minimum value of all data of certain evaluation indexes of the driver before normalization.

6. The method for evaluating the on-duty adaptability of the bus driver in real time according to claim 5, characterized in that: in step S6, the evaluation index threshold values are calibrated as follows:

for the classification data:

evaluation indexes without calibration of a threshold value comprise sex, native place, culture degree, marital state, employment type, overtime, driving time period, driving style, road transportation driver suitability evaluation level, driver professional skill level, professional value view, instant memory capability, complex reaction capability, communication capability, anti-interference capability, compression resistance capability, anti-monotony capability and anti-fatigue capability; the evaluation indexes needing to calibrate the threshold value comprise emotional driving, fatigue driving, distracted driving and aggressive driving, and the numerical value of '5' is used as the threshold value T of the evaluation indexes of the on-duty adaptability _max (ii) a For other classification data including drunk driving and diseases, the numerical value '1' is taken as the index threshold T _max ；

For numerical data:

1) For the evaluation indexes meeting or being converted into normal distribution, including body temperature, pulse, respiration, dynamic blood pressure, driving attitude, occupation burnout, personality, mental health, emotional stability and emotional resilience, the evaluation indexes are screened according to the 3 sigma principle, and the mu +3 sigma point is selected as the upper threshold point T of the evaluation index _max Mu-3. Sigma. Point as the lower threshold Point T of the evaluation index _min ：

T _max ＝μ+3σ

T _min ＝μ-3σ

Wherein x is _i ^′ Taking the value of the ith sample after normalization, wherein mu is the mean value of the normalized sample, sigma is the standard deviation of the sample, n is the number of the samples, and T is the standard deviation of the samples _max As an upper threshold value of the evaluation index, T _min Is the lower threshold of the evaluation index;

2) For the numerical variable which does not conform to normal distribution, determining the threshold value of the variable as T by the enterprise where the driver is located according to the index characteristics and the industry requirements of the group drivers _max The method specifically comprises the steps of accumulated accident frequency, highest accident grade, monthly safe driving mileage, monthly safe violation frequency, monthly service violation frequency, monthly criticizing education frequency, monthly punishment frequency, daily accumulated driving duration, daily accumulated driving mileage, station entrance and exit specifications, lane use specifications, intersection passing specifications and driving habit specifications.

7. The method for evaluating the on-duty adaptability of the bus driver in real time according to the claim 6, characterized in that: in step S7, independent variables of the on-duty adaptability comprehensive evaluation model of the bus driver based on the neural network are all evaluation indexesx _i The dependent variable is the post-adaptive state of the driver and is represented by the risk probability of the accident of the driver, and the value range is [0,1 ]]Will bey=1 indicates that the driver has an accident, then:

z ₁ =p(y=1|x ₁ , x ₂ ,x ₃ ,… ,x _n )

the learning rule when constructing the model is as follows:

1) Determining number of network input layer nodesnNumber of hidden layer nodesmAnd the number of output layer nodes is 2; determining input variablesx _i Output variabley _k (ii) a Initializing hidden layer node biasesa _j And each node of the output layer is biasedb _k And connection weight between nodes of each layerω _ij ，ω _jk ；

Wherein, the first and the second end of the pipe are connected with each other,i=1,2,…,n；j=1,2,…,m；k=1,2；ω _ij representation and hidden layer nodesjConnected input layer nodesiA weight;ω _jk presentation and output layer nodeskConnected hidden layer nodesjThe weight of (c);

5) Determining an output result(ii) a Each evaluation index updated in real time is brought into the determined model to obtain the output layer variabley _k Output value of output layer corresponding to =1z ₁ The probability of the driver accident risk is called as the post-fitting state.

8. The method as claimed in claim 7, wherein said method comprises the steps of,

in step S7, the post fitting grade includes five grades, which are respectively:

post fitting grade is 'two-stage': z ₁ belongs to [0.4,0.6 ]), which indicates that the accident risk probability of the current driver is medium and the suitable job state is general;

the suitable post grade is 'three-level': z ₁ belongs to [0.6, 0.8), and shows that the accident risk probability of the driver is higher at present, and the post-adaptive state is poorer;

the suitable job grade is 'four grades': z ₁ more than or equal to 0.8, which indicates that the driver is very easy to have accidents at present, and the state of being suitable for posts is very poor and is not suitable for posts;

job-adapted grade "five grade": if it isx′≥T _max Orx′≤T _min (ii) a The evaluation of the single-index post adaptability of the driver is unqualified at present, and the driver is in a state of improper post; when the suitable post grade is judged to be a first grade and a second grade, the suitable post state is good, and the driving work of a driver is not interfered; when the suitable post grade is judged to be three-grade, the suitable post state is poor, and the vehicle-mounted terminal equipment broadcasts the grade evaluation result and reason to the driver in a voice mode; when the shift level is judged to be four levels, the shift state is extremely poor, the vehicle-mounted terminal device broadcasts the level evaluation result and reason to the driver in a voice mode, and simultaneously the vehicle is switched to an automatic driving mode; when the suitable duty grade is judged to be five grades, the vehicle-mounted terminal equipment is in a state of being not suitable for duty, broadcasts a single-index suitable duty evaluation result and factors of the not suitable duty to a driver in a voice mode, and switches the vehicle to an automatic driving mode so as to enable the vehicle to be in the automatic driving modeThe driving work of the driver is terminated.

9. The utility model provides a public transit driver is at real-time evaluation system of suitable post nature on duty which characterized in that: the system comprises vehicle-mounted terminal equipment and background main control equipment, wherein the vehicle-mounted terminal equipment comprises video image acquisition equipment, sign parameter acquisition equipment, vehicle information acquisition equipment, an information interaction unit, a terminal early warning module and a terminal control module; the background main control equipment comprises an information interaction unit, an identity recognition module, a data storage module, a data processing module, a single index evaluation module and a post-fitting comprehensive evaluation module;

the sign parameter acquisition equipment is used for acquiring physiological signal data of a driver;

the data processing module is used for preprocessing data in the index database to screen out repeated, abnormal and missing data, and then normalizing the preprocessed data;

the single index evaluation module analyzes the types of the evaluation indexes of the on-duty adaptability of the driver after normalization processing, calibrates the threshold values of the evaluation indexes, verifies the evaluation indexes of the current driver according to the evaluation index threshold values, and evaluates the on-duty adaptability of the current driver;

the terminal early warning module is used for receiving the information of the qualified grade and carrying out voice early warning on a driver;

and the terminal control module is used for switching the driving mode of the driving vehicle when the suitable duty grade is four-grade or five-grade, namely the suitable duty state is not suitable duty.