Intelligent dynamic scheduling method and system for freight vehicles

Abstract

The present application relates to the technical field of logistics, in particular to a freight vehicle intelligent dynamic scheduling method and system, comprising: determining a risk index of each road section unit according to the change between connected road section units, updating the risk index in combination with the connection between the road section unit and the node, obtaining an updated risk index of each road section unit, adjusting the updated risk index in combination with the road section unit in the basic road network model at each moment according to the driving speed of the current freight driver at each moment in each historical transportation process and the number of emergency brakes in each preset time period, obtaining the final risk index of each road section unit, thereby constructing a weighted road network model, and outputting the optimal path from the starting point to the terminal of the current freight driver in combination with a graph neural network model. The present application determines the optimal path to make the scheduling result safer.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of logistics, in particular to a freight vehicle intelligent dynamic scheduling method and system. BACKGROUND

[0002] With the development of e-commerce logistics, cross-border trade, etc., the number of freight vehicles and transportation demand continues to develop, and there are generally many goods, many vehicles and complex scheduling in the freight industry. At present, most freight companies still rely on manual scheduling, or use simple algorithms with "shortest path or lowest cost" as the goal, and the scheduling results often ignore the driver's driving behavior and road safety risk conditions, etc., and generally still rely on navigation recommendations, lacking real "dynamic risk perception and personalized scheduling" of actual drivers during driving.

[0003] The existing problem is that when performing dynamic scheduling of freight vehicles, only the Dijkstra shortest path algorithm is used to determine the current freight path, without considering the actual driving conditions of the driver and the possible risks around the road conditions. When recommending the driving route of the freight vehicle, there may be high-risk situations, which is difficult to meet the actual needs of the freight industry for high safety, high reliability and personalized scheduling. SUMMARY

[0004] The present application provides a freight vehicle intelligent dynamic scheduling method and system to solve the existing problems.

[0005] The freight vehicle intelligent dynamic scheduling method and system of the present application adopts the following technical solutions:

[0006] An embodiment of the present application provides a freight vehicle intelligent dynamic scheduling method, which comprises the following steps:

[0007] A basic road network model is constructed based on the historical freight driving routes of the current freight driver, and the basic road network model is composed of a plurality of nodes and road segment units. The driving speed of the current freight driver at each time, the road segment unit in which the current freight driver is located at each time, and the number of sudden brakes in each predetermined time period during each historical transportation process are obtained.

[0008] According to the change between the connected road segment units, the risk index of each road segment unit is determined. According to the connection between the road segment units and the connection between the road segment units and the nodes, the risk index is updated to obtain the updated risk index of each road segment unit.

[0009] According to the driving speed of the current freight driver at each moment in each historical transportation process and the number of emergency braking in each preset time period, and in combination with the road section unit in which each moment is located in the basic road network model, the updated risk index is adjusted to obtain the final risk index of each road section unit;

[0010] According to the final risk index of all road section units in the basic road network model, a weighted road network model is constructed, and in combination with a graph neural network model, an optimal path of the current freight driver from the starting point to the terminal point is output.

[0011] Further, the specific steps of determining the risk index of each road section unit include the following:

[0012] The nodes include turning nodes; the road section units correspond to a traffic flow, an accident rate and a slope, and there is a turning angle between the connected road section units;

[0013] For any road section unit, the normalized value of the product of the sine function value of the slope, the traffic flow and the accident rate is recorded as the basic risk of the road section unit;

[0014] In the basic road network model, if the number of road section units connected with the yth road section unit is a preset constant, and the yth road section unit is connected with a turning node, the yth road section unit is recorded as a candidate road section unit, and all the connected candidate road section units form a candidate continuous turning path;

[0015] In the xth candidate continuous turning path, when the number of candidate road section units is greater than a preset number threshold, and the average value of the turning angles between all the connected candidate road section units is greater than a preset angle threshold, the xth candidate continuous turning path is recorded as a continuous turning path;

[0016] According to the turning angles between all the connected candidate road section units in each continuous turning path, the driving risk of each continuous turning path is determined;

[0017] According to the basic risk of each road section unit and the driving risk of each continuous turning path, the risk index of each road section unit is determined.

[0018] Further, the specific steps of determining the driving risk of each continuous turning path include the following:

[0019] In each continuous turning path, the inverse proportional normalized value of the average value of the turning angles between all the connected candidate road section units is obtained, which is recorded as the driving risk of each continuous turning path.

[0020] Further, the specific steps of determining the risk index of each road section unit according to the basic risk of each road section unit and the driving risk of each continuous turning path include the following:

[0021] For the road segment units in the continuous turning path, the product of the basic risk of each road segment unit and the driving risk of the continuous turning path in which each road segment unit is located is obtained, denoted as a first product, and the sum value of the first product and the basic risk of each road segment unit is denoted as the risk index of each road segment unit;

[0022] For the road segment units not in the continuous turning path, the basic risk of each road segment unit is denoted as the risk index of each road segment unit.

[0023] Further, the obtaining of the updated risk index of each road segment unit comprises the following specific steps:

[0024] The node comprises an intersection node.

[0025] The number of road segment units connected with the yth road segment unit is obtained as B1, the number of intersection nodes connected with the yth road segment unit is obtained as B2, and the product of B1 and B2 is denoted as the complexity of the yth road segment unit.

[0026] A preset range threshold M is set, a circular region is constructed with the center point of the yth road segment unit as a circle point and M as a radius, and the circular region is used as the surrounding region of the yth road segment unit.

[0027] In the surrounding region of the yth road segment unit, the number of road segment units B3 and the number of intersection nodes B4 are counted.

[0028] The ratio of B1 to B3 is denoted as a first ratio, the ratio of B2 to B4 is denoted as a second ratio, and the product of the first ratio and the second ratio is denoted as the central degree condition of the yth road segment unit for the surrounding region.

[0029] According to the complexity of the yth road segment unit and the central degree condition of the yth road segment unit for the surrounding region, the risk index of the yth road segment unit is updated to determine the updated risk index of the yth road segment unit.

[0030] Further, the updating of the risk index of the yth road segment unit according to the complexity of the yth road segment unit and the central degree condition of the yth road segment unit for the surrounding region to determine the updated risk index of the yth road segment unit comprises the following specific steps:

[0031] The normalized value of the product of the complexity of the yth road segment unit and the central degree condition of the yth road segment unit for the surrounding region is obtained as a risk update coefficient, the product of the risk index of the yth road segment unit and the risk update coefficient is obtained as a second product, and the sum value of the second product and the risk index of the yth road segment unit is denoted as the updated risk index of the yth road segment unit.

[0032] Further, the specific steps of obtaining the final risk index of each road section unit include the following:

[0033] In the current freight driver, in each historical transportation process, the normalized value of the ratio of the average of the driving speed at all times in each preset time period to the number of emergency brakes in each preset time period is obtained, which is recorded as the driving behavior stability of each preset time period;

[0034] According to the time sequence, the stability sequence composed of the driving behavior stability of all preset time periods is obtained;

[0035] The least square method is used to curve fit the stability sequence to obtain a stability change curve;

[0036] On the stability change curve, the preset time period corresponding to the data point with the maximum curvature is recorded as the split preset time period;

[0037] The time period composed of all preset time periods before the split preset time period is recorded as the normal driving time period;

[0038] All road section units in which all times in the normal driving time period are located in the basic road network model are recorded as target road section units;

[0039] According to the driving behavior stability of all preset time periods in the normal driving time period and the driving speed of all times corresponding to each target road section unit, the initial adjustment value of each target road section unit is determined;

[0040] According to the initial adjustment value of each target road section unit and the updated risk index of each road section unit, the final risk index of the road section unit is determined.

[0041] Further, the specific steps of determining the initial adjustment value of each target road section unit include the following:

[0042] The average of the driving behavior stability of all preset time periods in the normal driving time period is obtained, which is recorded as the driver behavior state reliability;

[0043] The ratio of the average to the variance of the driving speed of all times corresponding to each target road section unit is recorded as the driving speed stability value of each target road section unit;

[0044] The inverse proportional normalized value of the product of the driver behavior state reliability and the driving speed stability value of each target road section unit is obtained, which is recorded as the initial adjustment value of each target road section unit.

[0045] Further, the specific steps of determining the final risk index of each road section unit according to the initial adjustment value of each target road section unit and the updated risk index of each road section unit include the following:

[0046] In the basic road network model, if the kth road segment unit is not the target road segment unit in all historical transportation processes of the current freight driver, the update risk index of the kth road segment unit is recorded as the final risk index;

[0047] If the kth road segment unit is the target road segment unit in any historical transportation process of the current freight driver, the average value of the initial adjustment value of the kth road segment unit in all historical transportation processes of the kth road segment unit as the target road segment unit is recorded as the final adjustment value of the kth road segment unit, the product of the final adjustment value and the update risk index of the kth road segment unit is recorded as the third product, and the sum of the third product and the update risk index of the kth road segment unit is recorded as the final risk index.

[0048] The application further provides a freight vehicle intelligent dynamic scheduling system, which comprises a memory, a processor, and a computer program stored in the memory and capable of running on the processor, and the processor executes the computer program stored in the memory to realize the steps of the freight vehicle intelligent dynamic scheduling method.

[0049] The technical scheme of the application has the following beneficial effects:

[0050] In the embodiment of the application, the risk index of each road segment unit is determined according to the change between the connected road segment units, the risk index is updated according to the connection between the road segment units and the connection between the road segment units and the nodes, the update risk index of each road segment unit is obtained, the update risk index is adjusted according to the driving speed of the current freight driver at each moment in each historical transportation process and the number of emergency brakes in each preset time period, in combination with the road segment unit in which the current freight driver is located in the basic road network model, the final risk index of each road segment unit is obtained, the basic risk condition of the road segment unit is generated according to the road condition, the basic risk condition of the road segment unit is updated according to the complex condition of the surrounding road condition, the driving habit and risk preference of the driver are further considered in the road risk assessment, the risk condition of the road segment unit is further updated according to the behavior state of the driver in the normal driving condition, and the final weighted road network with the risk attribute is constituted, the graph neural network model is combined, and the optimal path of the current freight driver from the starting point to the terminal point is output. Therefore, the optimal path is determined when the road segment driving recommendation is performed based on the behavior performance of the current driver in the driving process and the local complex condition of each road network, the overall adaptation relationship of "person-vehicle-road" is determined, the scheduling result is safer and more dynamic. BRIEF DESCRIPTION OF DRAWINGS

[0051] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings needed in the embodiments or prior art description. Obviously, the drawings in the following description only constitute some embodiments of the present application, and for those skilled in the art, other drawings can also be obtained from these drawings without creative labor.

[0052] Figure 1 A step flow chart of a freight vehicle intelligent dynamic scheduling method of the present application;

[0053] Figure 2 A stability change curve schematic diagram. DETAILED DESCRIPTION

[0054] In order to further illustrate the technical means and effects adopted by the present application to achieve the predetermined inventive purpose, the following will combine the drawings and the preferred embodiments to specifically describe the specific implementation, structure, features and effects of the freight vehicle intelligent dynamic scheduling method and system according to the present application. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. In addition, the specific features, structures or characteristics in one or more embodiments can be combined in any suitable form.

[0055] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the present application belongs.

[0056] The following will specifically describe the specific scheme of the freight vehicle intelligent dynamic scheduling method and system provided by the present application in combination with the drawings.

[0057] Please refer to Figure 1 , which shows a step flow chart of a freight vehicle intelligent dynamic scheduling method provided by an embodiment of the present application. The method comprises the following steps:

[0058] Step S001: Obtain a basic road network model constructed based on historical driving routes of the current freight driver. The basic road network model is composed of a plurality of nodes and road segment units. Obtain the driving speed of the current freight driver at each time in each historical transportation process, the road segment unit in which the current freight driver is located at each time in the basic road network model, and the number of emergency brakes in each preset time length period.

[0059] In the present embodiment, when the freight route is recommended, the risk index of different road segments is determined based on the behavior performance of the current drivers in the driving process and the local complex conditions of different road segments in each road network, and then the recommended route of the optimal path is determined when the line driving recommendation is performed.

[0060] It is required to be explained: first, acquire the map situation and the driving path of the driver, determine the historical driving path situation of the driver on the map. Specifically: prepare the map, that is, register the developer account, enter the URL (URL is the unique address of resources on the Internet) of the map software, create Web (network) service to obtain publicly available map API (map API is a programming interface). Obtain the basic information of each section, such as traffic flow, accident rate, slope, etc. Open the trajectory service, allow to determine the path trajectory mark in the map. Install the vehicle GPS terminal of the freight car, and transmit the vehicle GPS original file of the historical trajectory of the freight car to the map cloud. Use the front and back effective points and the hidden Markov model to complete the route, determine the complete path of each historical path. Draw the complete historical route on the map, use the map polyline to draw a line. Get the historical path route map of the freight running on the map. The above are all known operations.

[0061] Thus, a basic road network model constructed by the historical freight running route of the current freight driver can be obtained, the basic road network model is composed of a plurality of nodes and road section units, wherein the nodes include intersection nodes and turning nodes, each road section unit corresponds to a traffic flow, an accident rate and a slope, and there is a turning angle between the connected road section units.

[0062] It is required to be explained: in this embodiment, the historical path route map (all historical driving paths) of the current freight driver's freight driving on the acquired map is taken as the basic road network model, which is composed of nodes (intersection nodes, turning nodes) and edges (road segments). The road segment division rules are set: including distance rule, structure rule and special rule. The preset distance value is 300 meters, which is taken as an example for description. The distance rule is that the vehicle driving distance on the continuous road exceeds 300 meters, and the road is automatically divided into a road segment unit. The structure rule is to divide from the existing nodes on the road to form a new road segment unit. The special rule is that the bridge, tunnel and other special road segments can be regarded as a separate road segment unit. The road network division is executed: determine any driving path, start from the starting point and cut along the road, when the cumulative length reaches 300 meters, a short road unit is generated, if a node is encountered, the current road unit is forced to end and a new unit is started, if a special road segment is encountered, the current road unit is also forced to end and a new unit is started. Thus, the road segment unit is generated, and the information contained in each road segment unit includes: traffic flow, accident rate, road length, slope condition, intersection or turning number, historical truck driving frequency and driver driving information. Among them, the traffic flow of each road segment unit is collected by using the inductive coil installed on the road surface, that is, the total number of vehicles in the measurement period is divided by the length of the time period. Access the accident record database of the traffic management department to obtain the accident data of each road segment unit, divide the number of accidents by the traffic flow and the time period to obtain the accident rate. The slope of each road segment unit is calculated through the GIS software, and the angle between the connected road segment units, that is, the turning angle, is calculated through the spatial analysis tool.

[0063] Then, the driving speed of the current freight driver at each time in each historical transportation process, the road segment unit in the basic road network model at each time, and the number of emergency brakes in each preset time period are obtained.

[0064] It is required to be explained: in this embodiment, the preset time period is 15 minutes, which is taken as an example for description, that is, every 15 minutes during the driving of the freight vehicle, the number of emergency brakes in 15 minutes is obtained, if the last preset time period does not meet 15 minutes, it is also a preset time period, the driving speed of the freight vehicle at each time is obtained in real time through the vehicle-mounted GPS, the collection frequency is once per second, and then the threshold value is set to identify the emergency brake event. For example, when the reduction of driving speed per unit time exceeds the set value, it can be considered that an emergency brake has occurred.

[0065] Step S002: determining the risk index of each road segment unit according to the change between the connected road segment units; updating the risk index according to the connection between the road segment units and the connection between the road segment units and the nodes, to obtain the updated risk index of each road segment unit.

[0066] It needs to be explained that further, the driving risk degree of each path area needs to be determined. That is, the risk degree of the current road section is determined according to the road section condition and the road section complexity of the road section unit. After the information of each road section unit in the road network is formed, the basic risk condition of the current road section can be determined according to the road section information in each road network. In the actual road network, there are some paths that are divided into different road section units by multiple turning points due to continuous turning. For such paths, they should be regarded as a complete continuous turning path for analysis. Since the continuous turning path usually has higher driving risk, after the basic risk assessment of each road section is completed, it is necessary to further identify the continuous turning path and update and correct the overall risk degree thereof.

[0067] Preferably, in an embodiment of the present application, the method for obtaining the updated risk index of each road section unit includes:

[0068] For any one road section unit, the normalized value of the product of the sine function value of the slope, the traffic flow and the accident rate is denoted as the basic risk of the road section unit.

[0069] Specifically, the basic risk of the road section unit is wherein, , and are the traffic flow, the accident rate and the slope of the road section unit, respectively, is the sine function value, is a linear normalization function for normalizing the data value to between 0 and 1. The greater the traffic flow and the accident rate, the higher the basic risk, and the closer the slope to 90 degrees, the higher the basic risk.

[0070] The preset constant is 2, the preset number threshold is 3, and the preset angle threshold is 10 degrees, which are taken as examples for description.

[0071] In the basic road network model, if the number of road section units connected with the yth road section unit is a preset constant, and the yth road section unit is connected with a turning node, the yth road section unit is denoted as a candidate road section unit, and all the connected candidate road section units form a candidate continuous turning path.

[0072] It needs to be explained that when the number of road section units connected with the road section unit is 2, that is, there are only turning-in paths and turning-out paths, so the candidate road section units in the candidate continuous turning path are connected two by two.

[0073] In the xth candidate continuous turning path, when the number of candidate road section units is greater than the preset number threshold, and the average of the turning angles between all the connected candidate road section units is greater than the preset angle threshold, the xth candidate continuous turning path is denoted as a continuous turning path.

[0074] In each continuous turning path, the average of the turning angles between all the connected candidate road segment units is obtained The inverse proportional normalized value of the driving risk of each continuous turning path is recorded as

[0075] It is necessary to explain that: taking The inverse proportional normalized value of the driving risk of each continuous turning path is recorded as When the continuous turning angles of the road are small, it means that the vehicle needs to make multiple sharp turns in a short time. This increases the difficulty of the driver's operation because the steering wheel needs to be adjusted frequently, that is, the greater the driving risk. Therefore, for any road segment unit, update the risk index of the road segment unit according to the risk situation of the continuous turning path. If the road segment belongs to the continuous turning path, update the risk index of the road segment unit, otherwise keep the risk index of the road segment unit.

[0076] In the basic road network model, for the road segment units in the continuous turning path, the product of the basic risk of each road segment unit and the driving risk of the continuous turning path where each road segment unit is located is obtained, recorded as the first product. The sum of the first product and the basic risk of each road segment unit is recorded as the risk index of each road segment unit. For road segment units not in the continuous turning path, the basic risk of each road segment unit is recorded as the risk index of each road segment unit. Thus, the risk index of each road segment unit is obtained.

[0077] It is necessary to explain that: Because the traffic information of the road segment is flowing, vehicles will enter and exit, it is necessary to further determine the risk degree of the current road segment in combination with the surrounding path complexity of the current road segment. If the road segment unit is the center position of the current path area, the road segment unit has more connected road segments and intersection nodes on the map (since the basic road network model is only the route area of the freight trajectory on the map, when determining the connected road segments of the road segment unit, the connected road segments are the connected route conditions of the road segment on the actual map, and the intersection nodes are also the intersection nodes of the road segment unit on the actual map). The risk degree of the road segment is higher.

[0078] The number of road segment units connected to the yth road segment unit B1 is obtained, and the number of intersection nodes connected to the yth road segment unit B2 is obtained. The product of B1 and B2 is recorded as the complexity of the yth road segment unit.

[0079] The preset range threshold M is 900 meters, which is used as an example for description.

[0080] Taking the center point of the yth road segment unit as the circle point and M as the radius, a circular area is constructed as the surrounding area of the yth road segment unit.

[0081] In the surrounding area of the yth road segment unit, the number of road segment units B3 (only complete road segment units are counted) and the number of intersection nodes B4 are counted.

[0082] The ratio of B1 to B3 is recorded as the first ratio, the ratio of B2 to B4 is recorded as the second ratio, and the product of the first ratio and the second ratio is recorded as the center degree condition of the yth road segment unit as the surrounding area.

[0083] It should be noted that the number of road segment units connected to the yth road segment unit, the number of intersection nodes, and the number of road segment units and intersection nodes in the surrounding area are counted on the actual road map, i.e., the map derived from the map software. Therefore, the yth road segment unit is the current freight driver's historical driving road segment, and the counted number of road segment units and intersection nodes may include road segment units and intersection nodes that the current freight driver has not driven but exist on the map.

[0084] The product of the complexity of the yth road segment unit and the center degree condition of the yth road segment unit as the surrounding area is obtained The normalized value of the product is recorded as the risk update coefficient, the product of the risk index of the yth road segment unit and the risk update coefficient is recorded as the second product, and the sum of the second product and the risk index of the yth road segment unit is recorded as the updated risk index of the yth road segment unit.

[0085] Wherein, is the normalized value of .

[0086] Step S003: According to the driving speed of the current freight driver at each moment in each historical transportation process and the number of sudden brakes in each predetermined time period, and combining the road segment unit in which each moment is located in the basic road network model, the updated risk index is adjusted to obtain the final risk index of each road segment unit.

[0087] It should be noted that there is a certain limitation in the road risk assessment only considering the surrounding road segment. Since different drivers have different driving habits and risk preferences on the same road segment, relying solely on road segment information may ignore the impact of driver behavior on risk. Therefore, in order to realize the unity of "man-road-vehicle", it is necessary to introduce the driver behavior state to dynamically update the road risk and reflect the actual risk situation. Since drivers have fatigue accumulation effect, the longer the driving time, the higher the possibility of abnormal behavior. However, different drivers have different fatigue tolerance levels and behavior performances, so it is necessary to analyze the historical driving data of each driver, such as the speed change trend or the statistics of sudden events (such as sudden brakes) during driving, to determine the normal driving time, so as to update the road risk.

[0088] Preferably, in one embodiment of the present application, the method for obtaining the final risk index of each road section unit comprises:

[0089] In each historical transportation process of the current freight driver, the ratio of the average of driving speed at all time points in each preset time length period to the number of emergency braking in each preset time length period is obtained , denoted as the driving behavior stability of each preset time length period. In chronological order, the stability sequence composed of the driving behavior stability of all preset time length periods is obtained, the least squares method is used to curve fit the stability sequence, the stability change curve is obtained, the curvature of each data point on the stability change curve is obtained, the preset time length period corresponding to the data point corresponding to the maximum curvature is denoted as the split preset time length period, and the time period composed of all preset time length periods before the split preset time length period (including the split preset time length period) is denoted as the normal driving period, i.e. the normal driving period of the current freight driver in each historical transportation process.

[0090] It should be noted that: is the normalized value of . The least squares method and the calculation of the curvature of the data points on the curve are both known technologies, and the specific method is not introduced here. The data point corresponding to the maximum curvature, i.e. the point where the stability changes the most, so the period before the split preset time length period is the normal driving period. The stability change curve is shown in Figure 2 . Figure 2 The horizontal axis is time and the vertical axis is stability, i.e. the driving behavior stability changes with time, and a data point is determined every 15 minutes.

[0091] Further, in order to avoid the influence of the driver's fatigue driving, the driving section in the normal driving time of the driver can be used to determine the risk index in the current road network, wherein if the average speed of the driver in the section is relatively high and the volatility of the speed is relatively small, it indicates that the risk index of the road network is relatively small. At the same time, if the driving stability of the driver is relatively high, it indicates that the behavior state credibility of the driver is relatively high.

[0092] In each historical transportation process of the current freight driver, the average of the driving behavior stability of all preset time length periods in the normal driving period is obtained, denoted as the driver behavior state credibility.

[0093] In the current freight driver in each historical transportation process, all the time in the normal driving period in the basic road network model The road segment unit where the driver is located is recorded as the target road segment unit. Obtain the mean of the driving speed of all times corresponding to each target road segment unit, and then obtain the variance of the driving speed of all times corresponding to each target road segment unit. The ratio of the mean and variance of the driving speed of all times corresponding to each target road segment unit is recorded as the driving speed stability value of each target road segment unit.

[0094] It should be noted that: the faster the driving speed, and the smaller the fluctuation of the driving speed, that is, the smaller the variance, the more stable the driving speed. Therefore, according to the driving speed stability value of the driver on this road segment unit and the behavior state credibility of the driver, the risk index of this road segment unit can be updated. If the driving speed stability value of the current road segment unit is relatively strong, the risk index of the current road segment unit is reduced, if the driving stability of the current road segment unit is relatively weak, the risk index of the current road segment unit is increased, and the greater the behavior state credibility of the driver, the more credible the risk adjustment.

[0095] In the current freight driver in each historical transportation process, the product of the driver behavior state credibility and the driving speed stability value of each target road segment unit is obtained The inverse proportional normalized value of

[0096] is recorded as the initial adjustment value of each target road segment unit. The inverse proportional normalized value of

[0097] In the basic road network model, if the kth road segment unit is not a target road segment unit in all historical transportation processes of the current freight driver, the update risk index of the kth road segment unit is recorded as the final risk index.

[0098] If the kth road segment unit is a target road segment unit in any historical transportation process of the current freight driver, the mean of the initial adjustment value of the kth road segment unit in all historical transportation processes of the kth road segment unit as a target road segment unit is recorded as the final adjustment value of the kth road segment unit. The product of the final adjustment value and the update risk index of the kth road segment unit is recorded as the third product, and the sum of the third product and the update risk index of the kth road segment unit is recorded as the final risk index.

[0099] It should be noted that: if the kth road segment unit is a target road segment unit in a historical transportation process, the initial adjustment value of the kth road segment unit in this process is taken as the final adjustment value. Thus, the final risk index of each road segment unit in the basic road network model can be obtained.

[0100] ​Step S004: according to the final risk index of all road segment units in the basic road network model, a weighted road network model is constructed, combined with a graph neural network model, and the optimal path of the current freight driver from the starting point to the ending point is output.

[0101] According to the final risk index of all road segment units in the basic road network model, a weighted road network model is constructed, combined with a graph neural network model, and the optimal path of the current freight driver from the starting point to the ending point is output.

[0102] It should be noted that after determining the final risk index of each road segment unit in the basic road network model, in the map software, each road segment unit in the basic road network model on the map is assigned a label final risk index, and each road segment unit outside the basic road network model on the map is assigned a label basic risk (only according to the traffic flow, accident rate and slope of the road segment unit). Thus, the label value of each road segment unit on the map is obtained, and a weighted road network with risk attributes is formed. Further, combined with a graph neural network (GNN) model, the road network with risk weight is trained and predicted, and finally the optimal path of the current freight driver from the starting point to the ending point is output, providing more accurate, safe and efficient scheduling and navigation suggestions for freight vehicles. The graph neural network (GNN) model is a known technology, and the specific method is not described here.

[0103] The application also provides a freight vehicle intelligent dynamic scheduling system, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor. The processor executes the computer program stored in the memory to realize the steps of the aforementioned freight vehicle intelligent dynamic scheduling method.

[0104] Thus, the application is completed.

[0105] In summary, in the embodiment of the application, the risk index of each road segment unit is determined according to the change between connected road segment units, the risk index is updated according to the connection between road segment units and the connection between road segment units and nodes, the updated risk index of each road segment unit is obtained, the updated risk index is adjusted according to the driving speed of the current freight driver at each moment in each historical transportation process and the number of emergency brakes in each predetermined time period, combined with the road segment unit in the basic road network model at each moment, the final risk index of each road segment unit is obtained, a weighted road network model is constructed according to the final risk index of all road segment units in the basic road network model, combined with a graph neural network model, and the optimal path of the current freight driver from the starting point to the ending point is output. The application determines the optimal path to make the scheduling result safer.

[0106] The above merely provides the preferred embodiment of the present application, and is not used to limit the present application, and any modification, equivalent replacement, improvement, etc. made within the principle of the present application should be included in the protection scope of the present application.

Claims

1. A freight vehicle intelligent dynamic scheduling method, characterized in that, The method comprises the following steps: Obtain a basic road network model constructed based on historical freight driving routes of a current freight driver, the basic road network model being composed of a plurality of nodes and road segment units; obtain driving speed of the current freight driver at each time during each historical transportation process, a road segment unit in which the current freight driver is located at each time, and the number of emergency brakes within each preset time period; Determine a risk index of each road segment unit according to changes between connected road segment units; update the risk index according to connection between road segment units and connection between road segment units and nodes, to obtain an updated risk index of each road segment unit; Adjust the updated risk index according to the driving speed of the current freight driver at each time during each historical transportation process, the number of emergency brakes within each preset time period, and the road segment unit in which the current freight driver is located at each time, to obtain a final risk index of each road segment unit; Construct a weighted road network model according to the final risk index of all road segment units in the basic road network model, and output an optimal path from a starting point to a terminal point of the current freight driver in combination with a graph neural network model; The method comprises the following specific steps for determining the risk index of each road segment unit: The nodes include turning nodes; each road segment unit corresponds to a traffic flow, an accident rate and a slope, and there is a turning angle between connected road segment units; For any road segment unit, a normalized value of a product of a sine function value of the slope, the traffic flow and the accident rate is recorded as a basic risk of the road segment unit; In the basic road network model, if the number of road segment units connected to the yth road segment unit is a preset constant, and the yth road segment unit is connected to a turning node, the yth road segment unit is recorded as a candidate road segment unit, and all connected candidate road segment units form a candidate continuous turning path; In the xth candidate continuous turning path, when the number of candidate road segment units is greater than a preset number threshold, and the average value of the turning angles between all connected candidate road segment units is greater than a preset angle threshold, the xth candidate continuous turning path is recorded as a continuous turning path; Determine a driving risk of each continuous turning path according to the turning angles between all connected candidate road segment units in each continuous turning path; Determine the risk index of each road segment unit according to the basic risk of each road segment unit and the driving risk of each continuous turning path; The method comprises the following specific steps for obtaining the updated risk index of each road segment unit: The nodes include intersection nodes; Obtain the number B1 of road segment units connected to the yth road segment unit, and obtain the number B2 of intersection nodes connected to the yth road segment unit; a product of B1 and B2 is recorded as the complexity of the yth road segment unit; A preset range threshold M is used to construct a circular region with the center point of the yth road segment unit as a circle point and M as a radius, as a surrounding region of the yth road segment unit; In the surrounding region of the yth road segment unit, count the number B3 of road segment units and the number B4 of intersection nodes; The ratio of B1 to B3 is denoted as a first ratio, the ratio of B2 to B4 is denoted as a second ratio, and the product of the first ratio and the second ratio is denoted as a center degree condition of the yth road segment unit for the surrounding area; According to the complexity of the yth road segment unit and the center degree condition of the yth road segment unit for the surrounding area, the risk index of the yth road segment unit is updated to determine an updated risk index of the yth road segment unit; The specific steps of obtaining the final risk index of each road segment unit include the following: In each historical transportation process of the current freight driver, a normalized value of the ratio of the average driving speed at all times in each preset time period to the number of emergency braking in each preset time period is obtained, and the normalized value is denoted as the driving behavior stability of each preset time period; In time sequence, a stability sequence composed of the driving behavior stability of all preset time periods is obtained; A stability change curve is obtained by using the least square method to perform curve fitting on the stability sequence; On the stability change curve, a preset time period corresponding to a data point corresponding to the maximum curvature is denoted as a split preset time period; A time period composed of all preset time periods before the split preset time period is denoted as a normal driving time period; All road segment units in which all times in the normal driving time period are located in the basic road network model are denoted as target road segment units; According to the driving behavior stability of all preset time periods in the normal driving time period and the driving speed of all times corresponding to each target road segment unit, an initial adjustment value of each target road segment unit is determined; According to the initial adjustment value of each target road segment unit and the updated risk index of each road segment unit, the final risk index of the road segment unit is determined.

2. The intelligent dynamic dispatching method for freight vehicles according to claim 1, characterized in that, The specific steps of determining the driving risk of each continuous turning path include the following: In each continuous turning path, an inverse proportional normalized value of the average of the turning angles between all connected candidate road segment units is obtained, and the inverse proportional normalized value is denoted as the driving risk of each continuous turning path.

3. The intelligent dynamic dispatching method for freight vehicles according to claim 1, characterized in that, The specific steps of determining the risk index of each road segment unit according to the basic risk of each road segment unit and the driving risk of each continuous turning path include the following: For the road segment units in the continuous turning path, a product of the basic risk of each road segment unit and the driving risk of the continuous turning path in which each road segment unit is located is obtained, and the product is denoted as a first product; a sum value of the first product and the basic risk of each road segment unit is denoted as the risk index of each road segment unit; For the road segment units not in the continuous turning path, the basic risk of each road segment unit is denoted as the risk index of each road segment unit.

4. The intelligent dynamic dispatching method for freight vehicles according to claim 1, characterized in that, The specific steps of updating the risk index of the yth road segment unit according to the complexity of the yth road segment unit and the center degree condition of the yth road segment unit for the surrounding area to determine an updated risk index of the yth road segment unit include the following: The normalized value of the product of the complexity of the yth road section unit and the degree of the yth road section unit being the center of the surrounding area is denoted as a risk updating coefficient, the product of the risk index of the yth road section unit and the risk updating coefficient is denoted as a second product, and the sum of the second product and the risk index of the yth road section unit is denoted as an updated risk index of the yth road section unit.

5. The intelligent dynamic dispatching method for freight vehicles according to claim 1, characterized in that, The specific steps of determining the initial adjustment value of each target road section unit include the following steps. An average value of the driving behavior stability of all preset time length periods in the normal driving period is denoted as a driver behavior state credibility. A ratio of the average value of the driving speed of each target road section unit at all time points to the variance is denoted as a driving speed stability value of each target road section unit. An inverse proportional normalized value of the product of the driver behavior state credibility and the driving speed stability value of each target road section unit is denoted as an initial adjustment value of each target road section unit.

6. The intelligent dynamic dispatching method for freight vehicles according to claim 1, characterized in that, The specific steps of determining the final risk index of the road section unit according to the initial adjustment value of each target road section unit and the updated risk index of each road section unit include the following steps. In the basic road network model, if the kth road section unit is not a target road section unit in all historical transportation processes of the current freight driver, the updated risk index of the kth road section unit is denoted as a final risk index. If the kth road section unit is a target road section unit in any historical transportation process of the current freight driver, an average value of the initial adjustment value of the kth road section unit in all historical transportation processes in which the kth road section unit is a target road section unit is denoted as a final adjustment value of the kth road section unit, a product of the final adjustment value of the kth road section unit and the updated risk index of the kth road section unit is denoted as a third product, and the sum of the third product and the updated risk index of the kth road section unit is denoted as the final risk index.

7. A smart dynamic dispatching system for freight vehicles, comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, The computer program is executed by the processor to implement the steps of the intelligent dynamic scheduling method of the freight vehicle according to any one of claims 1-6.

Images