Multi-dimensional driver capability assessment and intelligent matching scheduling system

Abstract

The invention provides a multi-dimensional driver capability evaluation and intelligent matching scheduling system, and the system comprises a data collection module which is used for obtaining driver driving behavior data, vehicle state data and environment data in real time; the preprocessing module is used for carrying out noise filtering, missing value filling and standardization on the acquired data; the multi-dimensional capability evaluation module is used for calculating a driving safety score, an efficiency score and an emergency response score of the driver through a dynamic weight distribution algorithm based on the preprocessed data; the demand analysis module is used for analyzing the route complexity, the time sensitivity and the special service demand of the passenger order; the matching scheduling module is used for generating a matching result according to the driver ability score and the passenger demand and outputting a scheduling instruction; the dynamic optimization module monitors the driver state and the road condition change in real time and adjusts the matching weight; and the interaction module is used for pushing real-time scheduling information and abnormal event early warning to the driver and the passenger. The scheduling efficiency and safety can be improved, and the passenger travel experience and the operation management level are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent transportation and logistics scheduling, and more particularly, to a multi-dimensional driver capability evaluation and intelligent matching scheduling system. BACKGROUND

[0002] In the field of modern transportation, with the acceleration of urbanization and the increasing demand for travel efficiency, traditional driver scheduling systems face many challenges. Existing scheduling systems mainly rely on simple task allocation logic, usually only considering the availability and geographical location of drivers, lacking comprehensive evaluation of driver driving ability, safety record and ability to cope with complex road conditions. This single-dimensional scheduling method cannot meet the needs of passengers for safe, efficient and personalized services. At the same time, traditional systems lack the ability to adjust and optimize in real time when facing dynamic road conditions, resulting in inaccurate scheduling results and easy mismatches between drivers and passenger needs.

[0003] In the process of implementing the embodiments of the present application, the inventors found that the existing technology at least has the following problems or defects: the existing scheduling system cannot comprehensively evaluate the comprehensive ability of the driver, and it is difficult to accurately match according to the specific needs of the passengers; during the order execution process, there is a lack of real-time monitoring and dynamic adjustment mechanism for the driver's state and road condition changes, resulting in insufficient stability and reliability of the scheduling results; in addition, the interaction function between the system and the driver and the passenger is limited, and it is not possible to timely push the scheduling information or warn abnormal events, affecting user experience and operational efficiency. SUMMARY

[0004] The present application provides a multi-dimensional driver capability evaluation and intelligent matching scheduling system, comprising: a data acquisition module for real-time acquisition of driver driving behavior data, vehicle state data and environmental data; a preprocessing module for noise filtering, missing value filling and data standardization of the output of the data acquisition module; a multi-dimensional capability evaluation module for calculating the driver's driving safety score, efficiency score and emergency response score based on the preprocessed data through a dynamic weight allocation algorithm; a demand analysis module for analyzing the route complexity, time sensitivity and special service needs of the passenger order; a matching scheduling module for generating a matching result according to the multi-objective optimization function of the driver capability score and the passenger demand, and outputting a scheduling instruction; a dynamic optimization module for real-time monitoring of driver state and road condition changes during order execution, and adjusting the matching weight through a feedback mechanism; an interaction module for pushing real-time scheduling information and abnormal event warnings to the driver and the passenger.

[0005] Further, the data collection module comprises: a driving behavior collection unit for collecting steering wheel angular rate of change (degree / second), brake frequency (number of brake times per unit time), and acceleration fluctuation variance through vehicle-mounted sensors; an environmental data collection unit for obtaining real-time road congestion index (01 standardized value), weather visibility (meters), and road curvature (meters) through a GPS and a weather interface; a physiological index collection unit for obtaining driver fatigue index (01) and heart rate variability HRV (milliseconds) through a camera and a heart rate sensor;

[0006] Further, the multi-dimensional capability evaluation module comprises: a dynamic weight allocation unit for generating weight coefficients based on road congestion index and weather visibility in the environmental data, satisfying ; a safety score calculation subunit for calculating a safety score according to the formula

[0007] , wherein is the steering wheel angular rate of change (degree / second), is the acceleration fluctuation variance (), is the number of brake times per unit time, is the brake frequency normalization factor for normalizing to the interval [0, 1]; an efficiency score calculation subunit for calculating an efficiency score according to the formula

[0008] , wherein is the historical average travel delay time (seconds), is the road congestion index (01 standardized value), is the weather visibility (meters), is the road curvature radius (meters); a comprehensive score generation unit for calculating a driver comprehensive capability score according to the formula

[0009] , wherein​​​​​​​​​​​​ score an emergency response.

[0010] Further, the emergency response score is calculated by: a historical accident correlation unit, according to the driver's historical emergency braking times and accident rate , according to the formula

[0011] , where is the historical emergency braking times, is the historical accident probability; a simulation test unit, based on virtual scene test to generate reaction time ( seconds) and decision accuracy (%), according to the formula

[0012] , where is the scene complexity coefficient; a real-time correction unit, according to the current heart rate variability HRV ( milliseconds) and fatigue index (01), according to the formula

[0013] to update the score.

[0014] Further, the matching scheduling module performs the following steps: Step 1: decompose the passenger demand into time constraints , route risk threshold and service level ; Step 2: construct the objective function

[0015] , where is the estimated travel time ( seconds), is the service level, is the actual route risk, is the route risk threshold, is the demand type correlation coefficient, used to balance the importance of different demands; Step 3: use the tabu search algorithm to traverse the driver queue, select the driver order combination that makes Q the largest; Step 4: when Q is lower than the preset threshold, trigger the dynamic optimization module to adjust the weight.

[0016] Further, the calculation of the route risk includes: The road segmentation unit divides the order route into sub-segments and extracts the road curvature radius of each segment. (meters), congestion index (01 standardized value) and accident frequency ; Risk stacking unit, according to formula

[0017] Calculate, where N is the total number of sub-segments; The dynamic correction unit adjusts the temperature based on real-time meteorological data. (Visibility, unit: meters), according to the formula

[0018] Update risk values, among which This is the maximum visibility constant (set to 10,000 meters).

[0019] Furthermore, the dynamic optimization module includes: Deviation detection unit, according to formula

[0020] Calculate the execution deviation, where This represents the actual travel time (in seconds). For actual risks; Weight adjustment unit, if (Deviation threshold), then according to the formula

[0021] Adjust the weights of the objective function, where As the current weight, The adjusted weights, Adjust the step size for weights. For deviation direction identifier; The matching trigger unit will re-execute the matching scheduling module and freeze driver score updates when D continues to exceed the threshold.

[0022] Furthermore, the determination of the deviation direction identifier k includes: like and If k=1, then k=1; like and If k=2, then k=2; In other cases, the multi-dimensional weight balancing strategy is triggered when k=0.

[0023] Furthermore, the interaction module includes: The driver end pushing unit displays high matching degree orders according to priority and marks route risk hotspot areas; The passenger end early warning unit generates a delay compensation scheme and a list of alternative drivers when the dynamic optimization module triggers re-matching; The emergency takeover unit automatically switches to the automatic driving mode and notifies the passenger when the driver fatigue index exceeds the threshold value and the heart rate variability HRV is abnormal.

[0024] Further, the dynamic weight distribution algorithm satisfies:

[0025]

[0026]

[0027] wherein, is the road congestion index (01 standardized value), is the weather visibility (meters), is the historical average travel delay time (seconds), is the road curvature radius (meters); and when (minimum visibility threshold value), is forcibly set .

[0028] The above-mentioned embodiments of the present application have at least the following beneficial effects: The multi-dimensional driver capability evaluation and intelligent matching dispatching system of the present application can realize comprehensive collection and analysis of driver driving behavior, vehicle state and environmental factors, accurately calculate the safety, efficiency and emergency response scores of the driver through the dynamic weight distribution algorithm, and thus provide more personalized and accurate matching services for passengers. The system can optimize the dispatching instructions according to the route complexity, time sensitivity and special service needs of the passenger order, and improve the dispatching efficiency and satisfaction.

[0029] In addition, the system can monitor the driver state and road condition changes in the order execution process in real time, and adjust the matching weight through the dynamic optimization module to ensure the stability and reliability of the dispatching result. When the driver is tired or in an emergency, the system can automatically switch to the automatic driving mode and notify the passenger, ensuring the safety of travel. At the same time, the interactive module can timely push the dispatching information and abnormal early warning to the driver and the passenger, further improving the user experience and the operation management level. BRIEF DESCRIPTION OF DRAWINGS

[0030] The above and other objects, features and advantages of the exemplary embodiments of the present application will be readily understood through reading the detailed description of the exemplary embodiments of the present application below, with reference to the accompanying drawings. In the drawings, several embodiments of the present application are illustrated by way of example and not limitation, in which: Figure 1 A structural schematic diagram of a multi-dimensional driver capability evaluation and intelligent matching scheduling system provided by an embodiment of the present application is shown in FIG. 1. DETAILED DESCRIPTION

[0031] The principles and spirits of the present application will be described below with reference to several exemplary embodiments. It should be understood that these embodiments are given only to enable those skilled in the art to better understand and implement the present application, and do not limit the scope of the present application in any way. On the contrary, these embodiments are provided to make the present application more thorough and complete, and to fully convey the scope of the present application to those skilled in the art.

[0032] Those skilled in the art understand that the embodiments of the present application can be implemented as a system, device, apparatus, method or computer program product. Therefore, the present application can be embodied in the form of a complete hardware, a complete software (including firmware, resident software, microcode, etc.), or a combination of hardware and software.

[0033] It should be noted that any number of elements in the drawings is used for example only and not limitation, and any naming is only for distinction and does not have any limiting meaning.

[0034] The principles and spirits of the present application will be described below with reference to several exemplary embodiments. It should be understood that these embodiments are given only to enable those skilled in the art to better understand and implement the present application, and do not limit the scope of the present application in any way. On the contrary, these embodiments are provided to make the present application more thorough and complete, and to fully convey the scope of the present application to those skilled in the art. Figure 1 , Figure 1 A structural schematic diagram of a multi-dimensional driver capability evaluation and intelligent matching scheduling system provided by an embodiment of the present application is shown in FIG. 1. As shown in FIG. 1, a multi-dimensional driver capability evaluation and intelligent matching scheduling system 100 includes: Figure 1 a data acquisition module 101 for acquiring driver driving behavior data, vehicle state data and environmental data in real time; a preprocessing module 102 for performing noise filtering, missing value filling and data standardization on the output of the data acquisition module; a multi-dimensional capability evaluation module 103 for calculating a driving safety score, an efficiency score and an emergency response score of a driver based on the preprocessed data through a dynamic weight distribution algorithm; a demand analysis module 104 for analyzing the route complexity, time sensitivity and special service demand of a passenger order; a matching scheduling module 105 for generating a matching result according to a multi-objective optimization function of driver capability score and passenger demand, and outputting a scheduling instruction; a dynamic optimization module 106 for monitoring driver state and road condition changes in real time during order execution, and adjusting matching weight through a feedback mechanism; an interaction module 107 for pushing real-time scheduling information and abnormal event warning to drivers and passengers.

[0035] ​It should be noted that the system includes a data acquisition module for real-time acquisition of driver driving behavior data, vehicle state data and environmental data. The data acquisition module is the basic part of the system, which collects information related to driver driving through various sensors and interfaces, such as steering wheel angular rate of change, brake frequency, acceleration fluctuation variance and other driving behavior data. These data reflect the driver's operating habits and driving style. Vehicle state data includes vehicle speed, fuel consumption and other information, while environmental data involves road congestion index, weather visibility and road curvature and other external conditions, which directly affect the safety and efficiency of driving. The preprocessing module filters noise, fills missing values and standardizes data to ensure data quality and usability. Noise filtering can remove outliers in the data, missing value filling can supplement incomplete data, and data standardization converts data from different sources to a unified format and range for subsequent module processing.

[0036] Specifically, the driving behavior acquisition unit in the data acquisition module collects the steering wheel angular rate of change (unit: degree / s) through the vehicle-mounted sensor, which reflects the driver's control frequency and amplitude of the direction during driving; brake frequency (number of brake times per unit time) is used to evaluate the driver's braking habits, and frequent braking may indicate a more aggressive driving style or complex road conditions; acceleration fluctuation variance (unit: m² / s 4 ) measures the stability of vehicle speed changes. The environmental data acquisition unit obtains real-time road congestion index (0-1 standardized value) through GPS and weather interface, which reflects the degree of road smoothness; weather visibility (unit: meters) and road curvature (unit: meters) describe the visual range of the driving environment and the degree of road curvature, respectively. The physiological index acquisition unit obtains the driver fatigue index (0-1) and heart rate variability (unit: milliseconds) through the camera and heart rate sensor, which are used to evaluate the driver's physical state. The higher the fatigue index, the more tired the driver is, and the heart rate variability reflects the physiological stress level of the driver. When processing these data, the preprocessing module will select appropriate noise filtering algorithms according to the specific data type and application scenario, such as median filtering algorithm for smoothing processing of abnormal acceleration values in driving behavior data; missing value filling can be achieved through interpolation method or prediction model based on historical data; data standardization usually uses Z-score standardization or Min-Max standardization method to convert data to the range of [0, 1] or [-1, 1] for unified processing by subsequent modules.

[0037] Preferably, in the data acquisition module, the driving behavior acquisition unit can further extend the acquisition parameters, such as increasing the lateral acceleration and longitudinal acceleration of the vehicle, to more comprehensively reflect the dynamic behavior of the vehicle. For the preprocessing module, the noise filtering algorithm can select a higher-level algorithm such as wavelet transform filtering according to the distribution characteristics of the data to more effectively remove high-frequency noise. In the data standardization process, an adaptive standardization method can be introduced to dynamically adjust the standardization parameters according to the real-time statistical characteristics of the data, to improve the flexibility and accuracy of data processing. In addition, the preprocessing module can also increase the data fusion function to integrate data from different sources, such as combining driving behavior data with environmental data to generate more rich feature vectors, providing more comprehensive input information for the subsequent multi-dimensional ability evaluation module.

[0038] In some embodiments, the data acquisition module comprises: a driving behavior acquisition unit for acquiring steering wheel angle change rate (deg / s), brake frequency (number of brake times per unit time), and acceleration fluctuation variance ; an environmental data acquisition unit for acquiring real-time road congestion index (01 standardized value), weather visibility (meters), and road curvature (meters); a physiological index acquisition unit for acquiring driver fatigue index (01) and heart rate variability HRV (milliseconds) through a camera and a heart rate sensor.

[0039] It should be noted that the data acquisition module in the system includes a driving behavior acquisition unit, an environmental data acquisition unit, and a physiological index acquisition unit. The driving behavior acquisition unit acquires parameters such as steering wheel angle change rate, brake frequency, and acceleration fluctuation variance through vehicle-mounted sensors, which reflect the driver's driving habits and driving style during driving. The environmental data acquisition unit uses GPS and weather interface to acquire real-time road congestion index, weather, visibility, and road curvature, etc. These data are used to evaluate the complexity and safety of the driving environment. The physiological index acquisition unit acquires driver fatigue index and heart rate variability through a camera and a heart rate sensor, which is used to monitor the driver's physical state and concentration. The design of these acquisition units aims to comprehensively collect multi-dimensional data related to the driver's driving ability, providing basic support for subsequent evaluation and scheduling.

[0040] Specifically, the steering wheel angle change rate (unit: degree / second) in the driving behavior collection unit is an indicator to measure the sensitivity and frequency of the driver's control over the vehicle's direction. For example, in urban road driving, frequent angle changes may imply frequent lane changes or avoiding obstacles, while on the highway, a smaller angle change rate indicates a more stable driving style. Brake frequency (number of brake applications per unit time) reflects the driver's braking habits, and frequent braking may be related to complex road conditions or aggressive driving style. The acceleration fluctuation variance (unit: m² / s 4 ) is used to evaluate the stability of vehicle speed changes, and a larger variance may mean frequent acceleration and deceleration. The road congestion index in the environmental data collection unit is a 0-1 normalized value, where 0 represents complete smoothness and 1 represents severe congestion. Weather visibility (unit: meters) describes the visibility range in the driving environment, and lower visibility may be related to adverse weather conditions such as fog, rain, or snow. Road curvature (unit: meters) reflects the degree of road curvature, and smaller curvature indicates more curved roads. The fatigue index in the physiological indicator collection unit is a 0-1 normalized value, and the driver's fatigue level is evaluated by analyzing their facial expressions and eye movements through a camera. Heart rate variability (unit: milliseconds) is measured through a heart rate sensor, reflecting the driver's physiological stress level, and lower heart rate variability may be related to fatigue or stress.

[0041] Preferably, the driving behavior collection unit can further expand the collection parameters, such as adding the vehicle's lateral and longitudinal acceleration, to more comprehensively reflect the vehicle's dynamic behavior. For the environmental data collection unit, the road congestion index can be obtained through a real-time data interface with the traffic management department, ensuring the accuracy and timeliness of the data. Weather visibility can be more accurately measured through weather radar or satellite image data. Road curvature can be extracted through high-precision map data, combined with real-time road conditions for dynamic updates. The fatigue index in the physiological indicator collection unit can be further optimized through machine learning algorithms, combining the driver's driving duration, historical fatigue data, and other multi-dimensional information for comprehensive evaluation. In addition, electroencephalogram sensors can be introduced as a supplement to the physiological indicator collection unit, monitoring the driver's brain activity to more accurately assess their fatigue state and attention level.

[0042] In some embodiments, the multi-dimensional capability assessment module includes: a dynamic weight allocation unit based on the road congestion index and weather visibility in the environmental data to generate weight coefficients ; a safety score calculation sub-unit according to the formula

[0043] calculating a safety score, wherein is the steering wheel angle change rate (degree / s), is the acceleration fluctuation variance (m² / s ), is the number of brake times per unit time, is the brake frequency normalization factor, used to normalize to the [0, 1] interval; an efficiency score calculation subunit, according to the formula

[0044] calculating an efficiency score, wherein is the historical average trip delay time (seconds), is the road congestion index (01 standardized value), is the weather visibility (meters), is the road curvature radius (meters); a comprehensive score generation unit, according to the formula

[0045] calculating a driver's comprehensive ability score, wherein is the emergency response score.

[0046] It should be noted that the multi-dimensional ability evaluation module in the system is used to calculate the driver's driving safety score, efficiency score and emergency response score based on pre-processed data through a dynamic weight distribution algorithm. The dynamic weight distribution unit generates a weight coefficient according to the road congestion index and weather visibility in the environmental data, ensuring that the score result can be dynamically adjusted according to the real-time environmental conditions. The safety score calculation subunit calculates the safety score through a specific formula, mainly considering the steering wheel angle change rate, acceleration fluctuation variance and brake frequency, etc. These parameters reflect the stability and caution of the driver during driving. The efficiency score calculation subunit combines historical average trip delay time, road congestion index, weather visibility and road curvature radius, etc. to evaluate the driving efficiency of the driver in complex environments. The comprehensive score generation unit combines the safety score, efficiency score and emergency response score according to the weight to form the comprehensive ability score of the driver, providing a basis for subsequent dispatching decisions.

[0047] Specifically, the dynamic weight distribution unit generates a weight coefficient according to the road congestion index and weather visibility, for example, when the road congestion index is high, the weight of the efficiency score may be appropriately reduced, and the weight of the safety score will be correspondingly increased, to ensure that driving safety is prioritized in complex road conditions. In the safety score calculation formula, the steering wheel angle change rate (unit: degree / s) reflects the frequency and amplitude of the driver's control of the vehicle's direction, and the acceleration fluctuation variance (unit: m² / s 4The stability of the vehicle speed change is measured, and the brake frequency (the number of times of braking per unit time) reflects the driver's braking habits. These parameters, after normalization, can quantify the driver's safe driving level by combining specific calculation formulas. In the efficiency score calculation formula, the historical average trip delay time (unit: seconds) reflects the driver's time management ability in the historical trip, and the road congestion index and weather visibility are used as adjustment coefficients of environmental factors. The road curvature radius (unit: meters) reflects the influence of road curvature on driving efficiency. Through the comprehensive calculation of these parameters, the driving efficiency of the driver in different environments can be accurately evaluated.

[0048] Preferably, the dynamic weight distribution algorithm can be optimized according to different application scenarios. For example, in severe weather conditions, the weight coefficient of the safety score can be forcibly set to not less than 0.6 to further highlight the importance of safe driving. In the calculation of the safety score, the brake frequency normalization factor can be dynamically adjusted according to historical data to adapt to the driving habits of different drivers. For the calculation of the efficiency score, more environmental factors such as wind speed and temperature can be introduced as adjustment coefficients to more comprehensively reflect the influence of the driving environment on efficiency. In addition, the comprehensive score generation unit can introduce a machine learning algorithm to dynamically adjust the weight coefficient according to historical scheduling data and driver performance, further improving the accuracy and adaptability of the score.

[0049] In some embodiments, the emergency response score is calculated by: a historical accident correlation unit, according to the number of historical emergency braking times of the driver and the accident rate , according to the formula

[0050] , wherein is the number of historical emergency braking times, is the historical accident probability; a simulation test unit, based on virtual scene test to generate reaction time (seconds) and decision accuracy (%), according to the formula

[0051] , wherein is the scene complexity coefficient; a real-time correction unit, according to the current heart rate variability HRV (milliseconds) and fatigue index (01), according to the formula

[0052] to update the score.

[0053] It should be noted that the emergency response score calculation in the system is an important part of the multi-dimensional ability evaluation module, and its purpose is to comprehensively measure the driver's reaction ability and decision accuracy in emergency situations. The emergency response score is calculated by three parts: historical accident correlation unit, simulation test unit and real-time correction unit. The historical accident correlation unit calculates the basic score based on the driver's historical emergency braking times and accident rate, reflecting the driver's past performance in emergency situations; the simulation test unit tests the driver's reaction time and decision accuracy rate through virtual scenarios to evaluate their emergency response ability in simulated environments; the real-time correction unit dynamically adjusts the score according to the driver's current physiological state (such as fatigue and heart rate variability) to reflect their emergency response ability in actual driving. This comprehensive evaluation method can more accurately reflect the driver's emergency handling ability in complex environments and provide an important basis for dispatching decisions.

[0054] Specifically, in the calculation formula of the historical accident correlation unit, the number of emergency braking times refers to the number of braking operations performed by the driver due to emergency situations within a certain period of time, and the accident rate refers to the frequency of accidents of the driver. The basic score calculated by the formula can reflect the driver's emergency performance in real driving scenarios. In the simulation test unit, the reaction time refers to the time interval from recognizing danger to taking action in a virtual scenario, the decision accuracy rate refers to the proportion of correct decisions made by the driver in the simulation scenario, and the scenario complexity coefficient is adjusted according to the difficulty of the test scenario. In the real-time correction unit, the fatigue index is a standardized value of 0-1, reflecting the driver's current fatigue level; the heart rate variability (unit: milliseconds) is measured by a heart rate sensor, reflecting the driver's physiological stress level. The emergency response score calculated by these parameters in combination with a specific formula can dynamically reflect the driver's emergency ability, ensuring accurate assessment of their ability to respond to unexpected situations in actual driving.

[0055] Preferably, the historical accident correlation unit can be further refined, for example, by introducing a time decay factor, so that recent emergency braking and accidents have a greater impact on the score, thus more accurately reflecting the driver's current driving state. In the simulation test unit, more virtual scenario types can be added, such as urban road, highway and driving scenarios under adverse weather conditions, to more comprehensively evaluate the driver's emergency ability. In addition, more physiological indicators can be introduced in the real-time correction unit, such as electroencephalogram data, to more accurately assess the driver's attention and fatigue state. Machine learning algorithms can also be used to analyze historical data to dynamically adjust the scenario complexity coefficient and correction factor, making the emergency response score more scientific and personalized.

[0056] In some embodiments, the matching dispatching module performs the following steps: Step 1: Break down the passenger demand into time constraints route risk threshold and service level ; Step 2: Construct the objective function

[0057] wherein is the estimated travel time (seconds), is the service level, is the actual route risk, is the route risk threshold, is the demand type correlation coefficient, used to balance the importance of different demands; Step 3: Use the tabu search algorithm to traverse the driver queue and select the driver order combination that maximizes Q; Step 4: When Q is lower than the preset threshold, trigger the dynamic optimization module to adjust the weights.

[0058] It should be noted that the matching dispatch module is the core part of the system, which functions to generate the optimal matching result and output the dispatch instructions based on the driver's ability score and the passenger's order demand through a multi-objective optimization function. This module first disassembles the passenger demand into time constraints, route risk threshold, and service level, then constructs the objective function, uses the tabu search algorithm to traverse the driver queue, and selects the driver order combination that maximizes the objective function value. When the objective function value is lower than the preset threshold, the dynamic optimization module will be triggered to adjust the weights to ensure the rationality and adaptability of the dispatch result. The design of this module aims to achieve precise matching of drivers and passenger demands, while also considering the efficiency and dynamic adjustment capability of the dispatch.

[0059] Specifically, in the passenger demand disassembly part, the time constraint refers to the passenger's expected range of travel time, for example, the passenger may hope to arrive at the destination within a certain specific time; the route risk threshold refers to the maximum value of route risk that the passenger can accept, which may be related to road conditions, weather, and other factors; the service level reflects the passenger's requirements for service quality and comfort. In the construction of the objective function, the estimated travel time is calculated based on historical data and real-time traffic conditions, the service level is usually quantified according to the passenger's selected service type (such as ordinary, comfortable, or luxurious), and the actual route risk is calculated based on the route risk evaluation model. The tabu search algorithm is an efficient optimization algorithm that avoids local optimal solutions by setting a tabu list and candidate solution set, thereby searching for the globally optimal driver order combination in the driver queue. When the objective function value is lower than the preset threshold, it indicates that the current matching result may not meet the passenger's demand or there may be potential risks, at which point the dynamic optimization module will intervene to adjust the weights to re-optimize the dispatch result.

[0060] Preferably, the time constraint can be further refined as a priority setting of the departure time and the arrival time, for example, for an urgent order, the weight of the arrival time can be higher. The route risk threshold can be dynamically adjusted according to the historical preferences of the passengers, for example, passengers who often choose highways may have a higher tolerance for route risk. In the construction of the objective function, more demand type correlation coefficients can be introduced, such as the coefficient of passenger price sensitivity, to more comprehensively reflect passenger demand. For the tabu search algorithm, dynamic tabu length and candidate solution set size can be set to adjust according to the complexity of the real-time scheduling scene. In addition, when the objective function value is lower than the threshold, in addition to adjusting the weight, multiple rounds of optimization iterations can be introduced, or combined with other optimization algorithms (such as genetic algorithm) for hybrid optimization, to improve the accuracy and adaptability of the scheduling result.

[0061] In some embodiments, the route risk is calculated by: a route segment segmentation unit that segments the order route into sub-routes, extracts the road curvature radius (meters), congestion index (01 standardized value), and accident frequency of each segment; a risk superposition unit that calculates the route risk according to the formula

[0062] where N is the total number of sub-routes; a dynamic correction unit that updates the risk value according to real-time meteorological data (visibility, unit: meters) according to the formula

[0063] where is the maximum visibility constant (set to 10000 meters).

[0064] It should be noted that the calculation of route risk is an important part of the matching scheduling module, and its purpose is to quantify the risk level of the order route to provide a scientific basis for scheduling decisions. The calculation of route risk includes the route segment segmentation unit, the risk superposition unit, and the dynamic correction unit. The route segment segmentation unit segments the order route into multiple sub-routes and extracts the road curvature radius, congestion index, and accident frequency of each segment; the risk superposition unit calculates the overall route risk according to these features; the dynamic correction unit updates the risk value according to real-time meteorological data. This step-by-step calculation method can ensure the accuracy and dynamics of route risk evaluation, thereby better supporting scheduling decisions.

[0065] Specifically, the route segment division unit divides the order route into several sub-segments according to road type, length, or other preset conditions. The road curvature radius (unit: meters) reflects the degree of road curvature, and the smaller the curvature radius, the higher the risk; the congestion index is a 0-1 standardized value representing the degree of congestion of the road segment, and the higher the congestion index, the higher the risk; the accident frequency refers to the number of historical accidents on the road segment, and the higher the accident frequency, the greater the risk. The risk superposition unit calculates the overall route risk by weighting and summing these characteristic values according to the formula. The dynamic correction unit corrects the risk value according to the visibility (unit: meters) in real-time weather data, for example, when the visibility is low, the route risk will increase accordingly. The maximum visibility constant is usually set to a high value (such as 10000 meters) to standardize the correction coefficient.

[0066] Preferably, the route segment division unit can select different segmentation strategies according to different application scenarios. For example, in urban roads, segmentation can be performed according to traffic lights or intersections; on highways, segmentation can be performed according to service areas or toll stations. For the risk superposition unit, more factors that affect risk can be introduced, such as road slope or construction conditions, and reasonable weights can be set for each factor. In the dynamic correction unit, in addition to visibility, other weather factors such as rainfall, wind speed, etc. can be introduced to further refine the risk correction formula. In addition, machine learning algorithms can be combined to analyze historical accident data and dynamically adjust the weights of each factor to improve the accuracy and adaptability of route risk calculation.

[0067] In some embodiments, the dynamic optimization module comprises: a deviation detection unit according to the formula

[0068] Calculate the deviation, where is the actual driving time (seconds), is the actual risk; a weight adjustment unit, if (deviation threshold), then according to the formula

[0069] Adjust the target function weight, where is the current weight, is the adjusted weight, is the weight adjustment step, is the deviation direction identifier; a re-matching trigger unit that re-executes the matching scheduling module and freezes the driver score update when D continues to exceed the threshold.

[0070] It should be noted that the dynamic optimization module is a key part of the system for real-time adjustment of the scheduling strategy, and its purpose is to dynamically adjust the matching weight by monitoring the deviation during the order execution process, so as to ensure the stability and adaptability of the scheduling result. The deviation detection unit is used to calculate the deviation between the actual driving time and the estimated time, and the actual risk and the estimated risk. When the deviation exceeds the preset threshold, the weight adjustment unit adjusts the weight of the objective function according to the deviation direction to optimize the scheduling result. The re-matching trigger unit re-executes the matching scheduling module when the deviation continues to exceed the threshold, and freezes the driver score update to ensure the accuracy and reliability of the scheduling decision.

[0071] Specifically, the deviation detection unit quantifies the deviation by calculating the difference between the actual driving time and the estimated time, and the difference between the actual route risk and the estimated risk. The actual driving time refers to the actual time taken by the driver to complete the order, and the estimated time is calculated based on historical data and real-time traffic conditions; the actual route risk is calculated based on the dynamically corrected risk value, and the estimated risk is based on the route risk assessment result at the initial scheduling. The deviation threshold is a preset parameter for determining whether the deviation is within an acceptable range. The weight adjustment unit adjusts the weight according to the deviation direction, for example, when the actual driving time exceeds the estimated time and the actual risk is higher than the estimated risk, the safety weight is increased and the efficiency weight is reduced. The weight adjustment step is a dynamic parameter used to control the amplitude of weight adjustment, ensuring the stability of the adjustment process.

[0072] Preferably, the deviation detection unit can introduce a time window mechanism to perform a sliding average calculation on the deviation to smooth short-term fluctuations and avoid unnecessary optimization triggered by instantaneous deviation. The weight adjustment unit can dynamically adjust the weight adjustment step based on historical scheduling data and deviation, making it more adaptive. For example, in the case of frequent deviation, the step can be appropriately increased to speed up the optimization. The re-matching trigger unit can introduce a priority queue mechanism when re-executing the matching scheduling, prioritizing high-risk or high-deviation orders to improve the response efficiency of the system. In addition, when freezing the driver score update, a time threshold can be set to automatically unfreeze the score update when the deviation returns to normal, to maintain the flexibility of the system dynamic adjustment.

[0073] In some embodiments, the determination of the deviation direction identifier k includes: If and , then k = 1; If and , then k = 2; In other cases, k = 0 triggers a multi-dimensional weight balancing strategy.

[0074] It should be noted that the determination of the deviation direction identifier is an important link in the dynamic optimization module, which determines the direction and type of deviation according to the changes of actual driving time and actual risk, so as to provide a basis for weight adjustment. The deviation direction identifier has three possible values: when the actual driving time is greater than the estimated time and the actual risk is higher than the estimated risk, the identifier takes the value of 1, indicating that both time and risk deviate; when the actual driving time is less than the estimated time and the actual risk is lower than the estimated risk, the identifier takes the value of 2, indicating that both time and risk are better than expected; otherwise, the value is 0, triggering the multi-dimensional weight balancing strategy. This determination mechanism can help the system accurately identify the deviation and take corresponding optimization measures.

[0075] Specifically, the actual driving time refers to the actual time taken by the driver to complete the order, and the estimated time is calculated based on historical data and real-time traffic conditions; the actual risk is calculated based on the dynamically corrected risk value, and the estimated risk is based on the route risk assessment result at the initial dispatch. The determination of the deviation direction identifier is based on the comparison of the two sets of data. For example, when the actual driving time exceeds the estimated time, it may be due to traffic congestion or slow driving speed of the driver; when the actual risk is higher than the estimated risk, it may be due to unexpected bad weather or road construction. Through this determination mechanism, the system can adjust the weight distribution strategy according to different deviation situations to optimize the dispatch result. For example, when the identifier is 1, the system will increase the safety weight and reduce the efficiency weight; when the identifier is 2, the system may appropriately reduce the safety weight and increase the efficiency weight.

[0076] Preferably, the determination of the deviation direction identifier can introduce more dimensions and conditions. For example, in addition to time and risk, physiological indicators such as driver fatigue index and heart rate variability can also be considered to further refine the classification of deviation situations. In the comparison of actual driving time, the concept of time deviation rate can be introduced, i.e. the ratio of actual time to estimated time, when the time deviation rate exceeds a certain threshold, combined with the risk situation to determine the deviation direction. In addition, for the multi-dimensional weight balancing strategy (identifier 0), machine learning algorithms can be introduced to dynamically adjust the weight distribution based on historical data to adapt to complex deviation situations. For example, the system can train a regression model based on historical deviation data to predict the optimal weight distribution scheme, thereby improving the adaptability and optimization ability of the system to deviation situations.

[0077] In some embodiments, the interaction module comprises: The driver side push unit displays high matching degree orders according to priority and marks the route risk hot spot area; The passenger side early warning unit generates a delay compensation scheme and a list of alternative drivers when the dynamic optimization module triggers re-matching; An emergency takeover unit automatically switches to the autonomous driving mode and notifies the passenger when the driver fatigue index exceeds a threshold and the heart rate variability (HRV) is abnormal. An emergency takeover unit automatically switches to the autonomous driving mode and notifies the passenger when the driver fatigue index exceeds a threshold and the heart rate variability (HRV) is abnormal.

[0078] It should be noted that the interaction module is an important component of the system for interacting with the driver and the passenger. Its functions include pushing high-matching orders to the driver, warning abnormal events to the passenger, and switching to the autonomous driving mode when necessary to ensure safety. The driver-side push unit displays high-matching orders according to the results of the matching scheduling module, with priority and risk hot spot areas marked on the route, helping the driver quickly understand task information and potential risks. The passenger-side warning unit generates a delay compensation plan and a list of alternative drivers when the dynamic optimization module triggers re-matching, to alleviate the inconvenience caused by scheduling adjustments. The emergency takeover unit automatically switches to the autonomous driving mode and notifies the passenger when the driver fatigue index exceeds a threshold and the heart rate variability (HRV) is abnormal, ensuring travel safety.

[0079] Specifically, the driver-side push unit displays order information to the driver through the system interface or mobile device. The priority of high-matching orders is based on the matching degree between the driver's ability score and the order demand, for example, orders that better match the driver's ability will be displayed in a more prominent position. The marking of risk hot spot areas on the route is based on the system's risk assessment results for the route, for example, high-traffic road segments or accident-prone areas will be specially marked. When the dynamic optimization module triggers re-matching, the passenger-side warning unit generates a delay compensation plan according to pre-set rules, for example, providing passengers with coupons or points compensation; at the same time, the system selects a list of alternative drivers from the standby driver pool for the passenger to choose from. The triggering condition of the emergency takeover unit is that the driver fatigue index exceeds a pre-set threshold (such as 0.8) and the heart rate variability (HRV) is abnormal (such as below the normal range), at which point the system automatically switches to the autonomous driving mode and notifies the passenger through SMS or application, ensuring travel safety.

[0080] Preferably, the driver-side push unit can further optimize the order display method, for example, by combining map visualization technology to visually display the order starting point, end point, and risk hot spot areas. For the passenger-side warning unit, the delay compensation plan can be dynamically adjusted according to the passenger's membership level or order amount, for example, high-value orders or high-level members can receive higher compensation. The selection of the alternative driver list can introduce more dimensions of evaluation, such as the real-time state and historical service evaluation of the driver. When switching to the autonomous driving mode, the emergency takeover unit can issue an early warning to the driver, reminding them to prepare for the handover, and at the same time, explain the situation to the passenger through voice broadcast, improving user experience. In addition, the system can also be deeply integrated with the vehicle's autonomous driving system to ensure the smoothness and safety of the switching process.

[0081] In some embodiments, the dynamic weight distribution algorithm satisfies:

[0082]

[0083]

[0084] wherein, is the road congestion index (0-1 normalized value), is the weather visibility (meters), is the historical average travel delay time (seconds), is the road curvature radius (meters); and when (minimum visibility threshold), the safety weight coefficient is forcibly set to .

[0085] It should be noted that the dynamic weight distribution algorithm is the core part of the multi-dimensional capability evaluation module, and its purpose is to dynamically adjust the weight distribution of the safety score, the efficiency score and the emergency response score according to the real-time road conditions and driving environment. This algorithm calculates the weight coefficient through parameters such as road congestion index, weather visibility, historical average travel delay time and road curvature radius, to ensure that the driver's capability score can accurately reflect his driving performance under different driving conditions. When the weather visibility is lower than the minimum visibility threshold, the algorithm forcibly sets the safety weight coefficient to not less than 0.6 to prioritize driving safety. This mechanism can effectively cope with complex and variable driving environments and improve the scientificity and safety of dispatching decisions.

[0086] Specifically, the road congestion index in the dynamic weight distribution algorithm is a 0-1 normalized value, which is used to measure the smoothness of the road; the weather visibility is in meters, which reflects the visibility range of the driving environment; the historical average travel delay time is in seconds, which represents the average delay of the driver in the historical journey; the road curvature radius is in meters, which describes the degree of road curvature. The calculation formula of the weight coefficient is: safety weight coefficient:

[0087] efficiency weight coefficient:

[0088] emergency response weight coefficient:

[0089] When the weather visibility is below the minimum visibility threshold (such as 100 meters), the safety weight coefficient is forced to be no less than 0.6 to ensure that safety factor dominates in low visibility conditions. This dynamic adjustment mechanism can flexibly allocate weights according to real-time environmental conditions, making the driver's ability score more targeted and adaptive.

[0090] Preferably, the dynamic weight allocation algorithm can be further refined according to specific application scenarios. For example, for urban roads and highways, different weight allocation strategies can be set respectively. In urban roads, due to complex road conditions and frequent congestion, the efficiency weight coefficient can be appropriately increased; while on highways, due to faster speed and greater visibility changes, more attention can be paid to the safety weight coefficient. In addition, the minimum visibility threshold can be dynamically adjusted according to the climate conditions of different regions, such as setting it to 150 meters in foggy areas and 200 meters in snowy areas. The algorithm can also introduce a machine learning model to dynamically adjust the weight allocation formula based on historical data and real-time feedback, further optimizing the accuracy and adaptability of the score. For example, by training a neural network model to automatically learn the complex relationships between parameters, more accurate weight allocation can be achieved.

[0091] The above-mentioned various embodiments of the present application have the following beneficial effects: The multi-dimensional driver ability evaluation and intelligent matching dispatching system of the present application can obtain driver driving behavior, vehicle state and environmental data in real time through the data acquisition module, and can perform noise filtering, missing value filling and data standardization through the preprocessing module, so as to ensure the accuracy and reliability of the input data. The multi-dimensional ability evaluation module calculates the safety, efficiency and emergency response scores of the driver based on the dynamic weight allocation algorithm, and the demand analysis module analyzes the passenger order in detail, so that the matching and dispatching module can generate accurate dispatching instructions to realize efficient matching of drivers and passenger demand. The dynamic optimization module can monitor the driver state and road condition changes in the order execution process in real time, and adjust the matching weight through the feedback mechanism, which can effectively cope with the uncertainty in dynamic environment and ensure the stability and adaptability of the dispatching result. The interaction module can push real-time dispatching information and abnormal event warning to the driver and passenger, improving user experience and operation efficiency.

[0092] In addition, the calculation of emergency response scores in combination with historical accident correlation, simulation testing, and real-time correction can more comprehensively evaluate the driver's response ability in emergency situations. The matching scheduling module uses a tabu search algorithm to optimize the combination of driver orders, and when the deviation exceeds the threshold, triggers the dynamic optimization module to adjust the weight, which can further improve the accuracy and flexibility of scheduling. Dynamic calculation and correction of route risk can update the risk value according to real-time weather data and road condition information, ensuring the scientificity and safety of scheduling decisions. The determination of the deviation direction identifier and the weight adjustment mechanism can achieve multi-dimensional weight balance and avoid the excessive influence of a single factor on the scheduling result. The driver-side push unit, passenger-side warning unit, and emergency takeover unit of the interaction module can ensure the right to know and safe travel of drivers and passengers, further improving the intelligent level and user experience of the system.

[0093] Further, the storage medium of the embodiments of the present application stores program instructions capable of implementing all the above methods, wherein the program instructions can be stored in the above storage medium in the form of a software product, including a number of instructions to make a computer device (which can be a personal computer, a server, or a network device, etc.) or a processor execute all or part of the steps of the methods described in various embodiments of the present application. The aforementioned storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), a magnetic disk or an optical disk, and various media that can store program codes, or a computer, a server, a mobile phone, a tablet, and other terminal devices.

[0094] The above description is only some of the preferred embodiments of the present application and the explanation of the applied technical principles. Those skilled in the art should understand that the scope of the application involved in the embodiments of the present application is not limited to the technical solutions formed by the specific combinations of the above technical features, and should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the above inventive concept. For example, the above features are replaced with technical features with similar functions disclosed in the embodiments of the present application (but not limited to) to form technical solutions.

Claims

1. A multi-dimensional driver capability assessment and intelligent matching dispatch system, characterized in that, The system comprises the following modules: a data acquisition module for acquiring driver driving behavior data, vehicle state data and environmental data in real time; a preprocessing module for noise filtering, missing value filling and data standardization of the output of the data acquisition module; a multi-dimensional capability evaluation module for calculating the driving safety score, efficiency score and emergency response score of the driver based on the preprocessed data through a dynamic weight distribution algorithm; a demand analysis module for analyzing the route complexity, time sensitivity and special service demand of the passenger order; a matching and scheduling module for generating a matching result according to a multi-objective optimization function of the driver capability score and the passenger demand, and outputting a scheduling instruction; a dynamic optimization module for monitoring the driver state and road condition changes in real time during the order execution process, and adjusting the matching weight through a feedback mechanism; an interaction module for pushing real-time scheduling information and abnormal event warning to the driver and the passenger.

2. The system of claim 1, wherein, The data acquisition module comprises: A driving behavior collecting unit is configured to collect a steering wheel angular rate change , a brake frequency , and an acceleration fluctuation variance by a vehicle-mounted sensor. An environmental data collection unit for acquiring real-time traffic congestion index via GPS and weather interface , weather visibility , and road curvature ; The physiological index acquisition unit is used to obtain the driver fatigue index through a camera and a heart rate sensor. and heart rate variability (HRV).

3. The system of claim 1, wherein, The multi-dimensional capability evaluation module comprises: a dynamic weight allocation unit based on a road congestion index in the environment data and weather visibility generating weight coefficients satisfying ; a safety score calculation subunit for calculating the safety score according to the following formula, ； wherein, is the steering wheel angle rate of change, is the acceleration fluctuation variance, is the number of braking per unit time, is the braking frequency normalization factor, used to normalize to the interval [0, 1]; an efficiency score calculation subunit for calculating the efficiency score according to the following formula, ； wherein, is a historical average travel delay time, is a road congestion index, is a weather visibility, is a road curvature radius; a comprehensive score generation unit according to the formula ； calculating a driver overall ability score, wherein score for emergency response.

4. The system of claim 3, wherein, The emergency response score The calculation of the emergency response score is performed by the following units, comprising: a historical accident correlation unit that correlates the number of times the driver has historically braked hard with the accident rate according to the formula: ； Compute, where is the number of historical emergency braking events, is the historical accident probability; Analog test unit, based on virtual scenario test generates reaction time (seconds) and decision accuracy rate , according to the formula: ； Computations, wherein, is a scene complexity coefficient; a real-time correction unit for correcting the fatigue index according to the current heart rate variability HRV according to the formula: ; update score.

5. The system of claim 1, wherein, The matching and scheduling module performs the following steps: Step 1: Decompose passenger demand into time constraints , route risk threshold , and service level ; Step 2: Construct the objective function: ； wherein, is a predicted travel time, is a service level, is an actual route risk, is a route risk threshold, is a demand type correlation coefficient for balancing the importance of different demands; Step 3: Use the tabu search algorithm to traverse the driver queue and select the driver order combination that maximizes Q; Step 4: When Q is lower than a preset threshold, trigger the dynamic optimization module to adjust the weights. weights.

6. The system of claim 5, wherein, The route risk comprises calculating: A road section segmentation unit segments the order route into sub-road sections, and extracts the road curvature radius of each section , congestion index , and accident frequency ; a risk superposition unit for calculating the risk value RA according to the following formula, ； where N is the total number of sub-sections; a dynamic correction unit which calculates a corrected risk value from real-time weather data according to the following formula ​ ； wherein is the maximum visibility constant.

7. The system of claim 1, wherein, The dynamic optimization module comprises: a deviation detection unit for calculating the execution deviation degree according to the following formula, ； wherein is the actual driving time (seconds), is the actual risk; a weight adjustment unit, if then adjusts the objective function weight according to the following formula, ; wherein, is the current weight, is the adjusted weight, is the weight adjustment step size, is the deviation direction identifier; a re-matching trigger unit for re-executing the matching and scheduling module and freezing the driver score update when D continuously exceeds the threshold value.

8. The system of claim 7, wherein, The determination of the deviation direction identifier k comprises: If and then k = 1; If and then k = 2; In other cases, k=0 triggers the multi-dimensional weight balancing strategy.

9. The system of claim 1, wherein, The interaction module comprises: a driver end pushing unit for displaying high matching degree orders according to the priority and marking the route risk hotspot area; a passenger end warning unit for generating a delay compensation scheme and a list of alternative drivers when the dynamic optimization module triggers re-matching; An emergency takeover unit, when the driver fatigue index exceeds a threshold and the heart rate variability, HRV, is abnormal, automatically switches to an autonomous driving mode and notifies the passenger.

10. The system of claim 3, wherein, The dynamic weight distribution algorithm satisfies: ； ； ； wherein, is a road congestion index, is a weather visibility, is a historical average travel delay time, is a road curvature radius; and when is forcibly set .

Images