Cooperative control method and system based on automatic driving mixed driving scene of port

Abstract

The application provides a kind of based on port automatic driving mixed scene's cooperative control method and system, it is related to port traffic management technical field, first, the dynamic interaction data set containing automatic driving vehicle, manual driving equipment and other aspects of information in port operation area is acquired, then the dynamic interaction data set is associated feature extraction, generates the dynamic association representation data containing the features such as traffic participant association, based on the dynamic association representation data, cooperative decision analysis is executed, determines the priority sequence of right of way and cooperative path planning scheme, generates the cooperative control instruction set containing speed cooperative parameter etc. according to cooperative decision analysis result, finally, the cooperative control instruction set is distributed to corresponding control system, realizes the dynamic cooperative driving control under the port mixed scene, improves port operation efficiency and safety.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of port traffic management, in particular to a collaborative control method and system based on a port automatic driving mixed driving scenario. BACKGROUND

[0002] In the port operation scenario, with the continuous development of automation technology, automatic driving vehicles are gradually introduced, forming a complex situation of automatic driving vehicles and manually driven equipment mixed driving. However, the existing port traffic management mode is mainly designed for traditional manually driven equipment, and it is difficult to meet the needs of the above mixed driving scenario.

[0003] At present, the management of traffic participants in the port mostly relies on manual command or simple signal control system, lacking comprehensive acquisition and accurate analysis of dynamic information of various traffic participants. When automatic driving vehicles and manually driven equipment work together, it is difficult to accurately determine the correlation between traffic participants and the coordination needs due to the inability to obtain real-time running data of automatic driving vehicles, operation state data of manually driven equipment, and behavior intention data of traffic participants, etc. in real time, leading to unreasonable in task scheduling and right of way allocation. At the same time, lacking effective evaluation and early warning mechanism for possible conflict risks, it is easy to cause traffic accidents, affecting the operation efficiency and safety of the port. In addition, the existing control method cannot provide accurate collaborative path planning and conflict avoidance operation rules for each traffic participant, making it difficult to realize efficient, safe and collaborative driving control in the port mixed driving scenario. SUMMARY

[0004] In view of the above-mentioned problems, in combination with the first aspect of the present application, the embodiments of the present application provide a collaborative control method based on a port automatic driving mixed driving scenario, the method comprising:

[0005] acquiring a set of dynamic interaction data in the port operation area, the set of dynamic interaction data comprising real-time running data of automatic driving vehicles, operation state data of manually driven equipment, port operation task scheduling information and behavior intention data of traffic participants;

[0006] performing correlation feature extraction processing on the set of dynamic interaction data to generate dynamic correlation representation data of the port mixed driving scenario, the dynamic correlation representation data comprising traffic participant correlation relationship features, task coordination demand parameters and conflict risk evaluation indexes;

[0007] performing collaborative decision analysis based on the dynamic correlation representation data to determine a right of way priority sequence of each traffic participant in the port operation area and a collaborative path planning scheme, the right of way priority sequence being generated by sorting node importance indexes, and the collaborative path planning scheme comprising recommended driving paths, predicted passing times and path change permission conditions;

[0008] generating a set of cooperative control instructions according to the priority sequence of the right-of-way and the cooperative path planning scheme, the set of cooperative control instructions including a speed coordination parameter, a path adjustment sequence and a conflict avoidance operation rule;

[0009] distributing the set of cooperative control instructions to corresponding traffic participant control systems to perform dynamic cooperative driving control in the port mixed driving scene.

[0010] In still another aspect, the embodiment of the present application also provides a cooperative control system based on a port automatic driving mixed driving scene, comprising a processor and a machine readable storage medium, the machine readable storage medium is connected with the processor, the machine readable storage medium is used for storing programs, instructions or codes, and the processor is used for executing the programs, instructions or codes in the machine readable storage medium to realize the above method.

[0011] Based on the above aspects, the embodiment of the present application obtains a set of dynamic interaction data in a port operation area, which covers information such as automatic driving vehicles, manually driven devices, operation task scheduling and traffic participant behavior intention, then performs correlation feature extraction processing on the set of dynamic interaction data, and the generated dynamic correlation representation data can accurately depict the correlation between traffic participants, task coordination demand and conflict risk, the cooperative decision analysis performed based on the dynamic correlation representation data can reasonably determine the priority sequence of the right-of-way and the cooperative path planning scheme of each traffic participant, the priority sequence of the right-of-way is generated by sorting the node importance index, which ensures the fairness and rationality of the right-of-way allocation; the cooperative path planning scheme includes a recommended driving path, an expected passing time and a path change permission condition, thereby providing clear driving guidance. The set of cooperative control instructions generated according to the priority sequence of the right-of-way and the cooperative path planning scheme includes a speed coordination parameter, a path adjustment sequence and a conflict avoidance operation rule, which can effectively guide the traffic participants to perform dynamic cooperative driving control. Finally, the set of cooperative control instructions is distributed to the corresponding traffic participant control systems, realizing efficient, safe and orderly dynamic cooperative driving in the port mixed driving scene, and significantly improving the operation efficiency and safety of the port. BRIEF DESCRIPTION OF DRAWINGS

[0012] Figure 1 is an execution flow schematic diagram of the cooperative control method based on the port automatic driving mixed driving scene provided by the embodiment of the present application;

[0013] Figure 2 is a schematic diagram of exemplary hardware and software components of the cooperative control system based on the port automatic driving mixed driving scene provided by the embodiment of the present application. DETAILED DESCRIPTION

[0014] The application will be described in detail below with reference to the accompanying drawings, Figure 1 is a flowchart of a collaborative control method based on a mixed driving scene of port automatic driving provided by an embodiment of the application. The collaborative control method based on the mixed driving scene of port automatic driving will be described in detail below.

[0015] Step S110: Obtain a dynamic interaction data set in the port operation area, which includes real-time running data of an automatic driving vehicle, operation state data of a manually driven device, port operation task scheduling information, and behavior intention data of a traffic participant.

[0016] In this embodiment, the port operation area includes multiple container stacking areas, cargo loading and unloading areas, transportation channels, and multiple operation berths. Various traffic participants exist in the port operation area, including an automatic driving container transport vehicle, an automatic driving tractor trailer, a manually driven forklift, a manually driven truck, a shore-based container crane, a yard crane, etc.

[0017] For the real-time running data of the automatic driving vehicle, a multi-sensor installed on the automatic driving vehicle can be used for collection. The multi-sensor can include a laser radar, a millimeter wave radar, a high-definition camera, a global positioning system receiver, an inertial measurement unit, etc. The laser radar is used to obtain three-dimensional point cloud data of the environment around the vehicle, which can accurately perceive the position, shape and distance of the surrounding objects; the millimeter wave radar can work in various weather conditions, continuously monitor the dynamic targets in front of and around the vehicle, and provide the relative speed and distance information of the targets; the high-definition camera is used to collect image information of the surrounding environment, which assists in identifying traffic signs, lane lines and other appearance characteristics of traffic participants; the global positioning system receiver, in combination with the positioning base station deployed inside the port, obtains the accurate position coordinates of the vehicle in the port coordinate system; the inertial measurement unit is used to collect motion state data such as acceleration and angular velocity of the vehicle. The real-time running data can specifically include the current position coordinates of the vehicle (based on the three-dimensional coordinate system set in the port operation area), the instantaneous driving speed, the driving direction (represented by the heading angle), the acceleration, the current lane position, the size parameters (length, width, height) of the vehicle itself, the remaining power reserve (electricity or fuel), the vehicle identification number, etc. The collected data is transmitted to the port central data processing center through the built-in wireless communication module of the vehicle according to the set transmission period.

[0018] Further, the operation state data of the manually driven equipment can be acquired by a data acquisition device installed on the manually driven equipment. For example, an operation state sensor is installed on the manually driven forklift truck to acquire data such as the steering operation amount of the steering wheel, the depression depth of the accelerator pedal, the depression force of the brake pedal, the lifting height and the inclination angle of the forklift truck forks, and the like of the driver; and corresponding sensors are installed on the manually driven truck to acquire data such as the driving speed, the steering angle, the braking state, the gear information, and the like of the vehicle. At the same time, the real-time position information of the manually driven equipment is acquired by a vehicle positioning device, and the running state of the equipment, such as the engine speed, the equipment fault information, and the like, is acquired by a state monitoring module on the equipment. These operation state data are also transmitted in real time to the central data processing center through a wireless communication mode.

[0019] Further, the port operation task scheduling information can be provided by a port operation management system, which contains the basic content of each operation task, such as the task number, the task type (container loading and unloading, cargo transfer, site arrangement, etc.), the starting point and the target point involved in the task, the required completion time range of the task, the traffic participants (the automatic driving vehicles or manually driven equipment responsible for executing the task) involved in the task, the cargo information (such as the container number, the cargo type, the weight, the size, etc.) associated with the task. In addition, the priority setting of the task, the dependency relationship between tasks (such as a transfer task must be started after another loading and unloading task is completed), and the like can also be included, and the above information is synchronized in real time through the interface between the port operation management system and the central data processing center.

[0020] The behavior intention data of the traffic participants can be obtained by analyzing and processing the acquired real-time running data and operation state data. For example, for the automatic driving vehicle, the behavior intention can be extracted from the navigation planning information and the real-time control instruction of the vehicle, such as the preset driving route of the vehicle, the upcoming steering operation, the planned stop point, and the like; for the manually driven equipment, the behavior intention is inferred by analyzing the operation behavior mode of the driver and the movement trend of the equipment, such as whether the manually driven truck is ready to turn, change lanes, or stop by continuously monitoring the steering operation and the speed change. At the same time, the port operation task scheduling information is combined to assist in judging the behavior intention of the traffic participants, such as the approximate driving direction and the destination of the traffic participants can be inferred according to the task target point.

[0021] It is worth mentioning that in the process of obtaining the above data, for the data involving sensitive data such as the identity information of the driver, the detailed operation log of the equipment, etc., the data desensitization processing technology is adopted. Through encryption conversion of sensitive fields, the information that can identify individuals or specific devices is removed, and at the same time, an encrypted transmission protocol is adopted to ensure the security of the data in the transmission process and prevent data leakage. After all the collected data is summarized, a dynamic interaction data set is formed and stored in the database of the central data processing center.

[0022] Step S120: performing correlation feature extraction processing on the dynamic interaction data set to generate dynamic correlation representation data of the port mixed traffic scene, wherein the dynamic correlation representation data includes traffic participant correlation relationship features, task coordination demand parameters and conflict risk evaluation indexes.

[0023] Step S121: analyzing the port operation task scheduling information in the dynamic interaction data set to extract task attribute parameters of each traffic participant, wherein the task attribute parameters include task urgency, cargo type features and operation area correlation.

[0024] In this embodiment, when analyzing the port operation task scheduling information in the dynamic interaction data set, first, the task items associated with each traffic participant are selected from the task scheduling information. Each traffic participant may correspond to one or more operation tasks, which need to be analyzed one by one.

[0025] For the extraction of the task urgency, the completion time range required by the task in the task scheduling information and the current time are analyzed. If the interval between the completion time required by the task and the current time is short, it means that the task needs to be executed as soon as possible, and the urgency is high; if the interval is long, the urgency is relatively low. At the same time, combined with the priority setting of the task, the task with high priority has relatively high urgency. Through the above analysis, the corresponding task urgency description is determined for each task executed by the traffic participant.

[0026] The extraction of the cargo type features is based on the cargo information associated in the task scheduling information. The cargo type features include the physical properties and category properties of the cargo, such as for container cargo, the features include the material of the container, whether it is a dangerous goods (if it is a dangerous goods, the category of the dangerous goods also needs to be specified), the use classification of the cargo (such as ordinary cargo, refrigerated cargo, special cargo, etc.); for bulk cargo, the features include the particle size, density, and whether it is easy to dust, etc. Thus, the above features are sorted to form the cargo type features of the tasks involved by each traffic participant.

[0027] The extraction of the work area relevance is achieved by analyzing the relationship between the starting point and target point of the task and the current work area of the traffic participant. If the current area of the traffic participant is the same or adjacent to the starting point or target point of the task, the work area relevance is strong; if the distance is far, the relevance is weak. At the same time, the connection relationship between the work processes involved in the task in different areas is considered, such as the target point of a task is the starting point of another task, and the two tasks are executed by different traffic participants, so the work area relevance of the two traffic participants is enhanced due to the connection of the task process. Through the above analysis, the work area relevance parameter of each traffic participant is determined.

[0028] Step S122: constructing a task association matrix based on the task attribute parameters, the task urgency level is positively correlated with the urgency weight value, and the work area relevance value is positively correlated with the correlation value between the traffic participants.

[0029] Step S1221: dividing the task urgency in the task attribute parameters into different levels, and assigning a corresponding urgency weight value to each level, and the urgency weight value corresponding to the task urgency level increases in order of level.

[0030] In this embodiment, when the task urgency is graded, the completion time interval and task priority required by the task are considered comprehensively. The task urgency is divided into multiple different levels, for example, four levels from low to high.

[0031] For each level, a corresponding urgency weight value is assigned according to the high and low of its urgency. The task urgency with the lowest level corresponds to the smallest urgency weight value, and the urgency weight value increases in turn as the level rises, and the task urgency with the highest level corresponds to the largest weight value. The above setting ensures that the task urgency level and the urgency weight value present an increasing relationship in order of level, so that in the subsequent association matrix construction, the task with high urgency can have a greater impact in the correlation calculation.

[0032] For example, level one (the lowest urgency) corresponds to a smaller urgency weight value, level two corresponds to a larger weight value than level one, the weight value of level three is greater than that of level two, and the weight value of level four (the highest urgency) is the largest. Through the above level division and weight value assignment, the task urgency attribute parameter is quantified

[0033] Step S1222: classifying the cargo type characteristics in the task attribute parameters to determine the cargo type weight coefficient.

[0034] In this embodiment, when classifying the cargo type characteristics, the processing specifications and priority requirements for different cargos in port operations are followed. First, the cargos are classified into categories, such as container cargos, bulk cargos, and piece cargos. Then, each category is further classified, for example, container cargos can be classified into ordinary container cargos, refrigerated container cargos, and dangerous container cargos; dangerous container cargos can be further classified according to the danger levels of dangerous goods, such as explosives, flammable liquids, and corrosive goods.

[0035] For each classified cargo type, a corresponding cargo type weight coefficient is determined according to the processing difficulty, safety requirements, and transportation priority of the cargo type in port operations. For example, the corresponding cargo type weight coefficient of dangerous container cargos is relatively large because of the higher safety and professionalism required for the transportation and processing of dangerous goods; the weight coefficient of ordinary container cargos is relatively small; and the weight coefficient of refrigerated container cargos is between that of dangerous container cargos and ordinary container cargos because of the need to maintain a specific temperature. Through the above classification and weight coefficient determination, the cargo type characteristics are quantified to reflect the influence of different cargo types on the correlation degree in the task correlation matrix construction.

[0036] Step S1223: Calculate the operation area correlation in the task attribute parameters, which is determined by the distance normalized value between the current operation area and the target operation area of the traffic participant and the operation process dependency.

[0037] In this embodiment, when calculating the operation area correlation, the current operation area of the traffic participant and the target operation area of the task to be performed are first determined. Through the digital map of the port operation area, the coordinate ranges of the two areas are obtained, and the straight-line distance between the centers of the two areas is calculated. The closer the distance, the stronger the correlation in the spatial position.

[0038] At the same time, the operation process dependency is analyzed. If the output of the task currently performed by the traffic participant is the input of another task (performed by another traffic participant), and the operation area of the other task is the target operation area of the current task, then the operation areas of the two traffic participants have a strong dependency due to the connection of the operation process, and the operation area correlation is correspondingly high.

[0039] Both the distance factor and the operation process dependency are considered, and a set of influence weights are assigned to them respectively. The correlation values corresponding to the distance and the correlation values corresponding to the operation process dependency are weighted and processed to obtain the final operation area correlation value. For example, the closer the distance, the larger the correlation value corresponding to the distance; the stronger the operation process dependency, the larger the correlation value corresponding to the operation process dependency, and the comprehensive operation area correlation is obtained through weighted calculation.

[0040] Step S1224: constructing a task attribute comprehensive weight based on the emergency degree weight value, the cargo type weight coefficient and the operation area correlation score, the task attribute comprehensive weight being a weighted sum result of the emergency degree weight value, the cargo type weight coefficient and the operation area correlation score.

[0041] In this embodiment, when constructing the task attribute comprehensive weight, firstly, an weight coefficient is set for the emergency degree weight value, the cargo type weight coefficient and the operation area correlation respectively, and the weight coefficient is determined according to the importance of the three parameters in the task correlation. For example, if the task emergency degree is more important in the task correlation, a larger weight coefficient is set for it; if the importance of the operation area correlation is relatively low, a smaller weight coefficient is set.

[0042] Then, the emergency degree weight value is multiplied by the corresponding weight coefficient to obtain the contribution value of the emergency degree in the comprehensive weight; similarly, the cargo type weight coefficient is multiplied by the corresponding weight coefficient to obtain the contribution value of the cargo type in the comprehensive weight; and the operation area correlation is multiplied by the corresponding weight coefficient to obtain the contribution value of the operation area correlation in the comprehensive weight.

[0043] Finally, the three contribution values are summed to obtain the result, which is the task attribute comprehensive weight. Through the above weighted sum method, the three different task attribute parameters are integrated into a comprehensive weight value, which is used in subsequent construction of the task correlation matrix.

[0044] Step S1225: taking the traffic participants in the port operation area as the matrix row and the matrix column, and taking the product of the task attribute comprehensive weights of any two traffic participants as the matrix element value to construct an initial task correlation matrix.

[0045] In this embodiment, when constructing the initial task correlation matrix, firstly, all the traffic participants in the port operation area are listed, and each traffic participant is assigned a unique identification number. The identification numbers of the traffic participants are taken as the row index and the column index of the initial task correlation matrix, that is, each row and each column of the matrix corresponds to a traffic participant.

[0046] For each element in the initial task correlation matrix, the traffic participant corresponding to the row is A, and the traffic participant corresponding to the column is B, and the value of the element is obtained by calculating the product of the task attribute comprehensive weight of the traffic participant A and the task attribute comprehensive weight of the traffic participant B. According to the above method, the products between the traffic participants corresponding to all rows and columns are calculated and filled into the corresponding positions of the matrix, so as to construct the initial task correlation matrix.

[0047] For example, if the task attribute comprehensive weight of traffic participant 1 is W1 and the task attribute comprehensive weight of traffic participant 2 is W2, the element value of the first row and the second column in the matrix is W1 multiplied by W2, the element value of the second row and the first column is W2 multiplied by W1, and so on, so as to complete the construction of the entire initial task association matrix.

[0048] Step S1226: The initial task association matrix is normalized, and the normalized task association matrix is adjusted according to the port operation rules and historical coordination data, the association degree value between traffic participants with direct operation process association is increased, and the association degree value between traffic participants without operation association is reduced, to obtain a final task association matrix.

[0049] In this embodiment, when the initial task association matrix is normalized, first, the maximum value and the minimum value of all elements in the initial task association matrix are found. Then, for each element value in the matrix, the element value is converted to a set numerical range (such as between 0 and 1) according to the normalization formula. The purpose of normalization is to eliminate the influence of different magnitudes, so that the element values in the initial task association matrix have comparability.

[0050] After the normalization is completed, the matrix is adjusted according to the port operation rules. For example, according to the port operation rules, there is a fixed operation process connection between certain traffic participants, such as the shore-based container crane hoisting the container to the automatic driving container transport vehicle, so there is a direct operation process association between the two, and the corresponding element value in the initial task association matrix needs to be appropriately increased to reflect the stronger association.

[0051] At the same time, referring to the historical coordination data, for the traffic participant pairs that often work together and have close association in historical operation, the matrix element values of the traffic participant pairs are also appropriately increased; and for the traffic participant pairs that have no association in the operation process and rarely have coordination interaction in history, the matrix element values of the traffic participant pairs are reduced.

[0052] Therefore, through the above adjustment, the task association matrix can more accurately reflect the actual association between traffic participants, and the final task association matrix is obtained.

[0053] Step S123: Extract traffic participant behavior intention data in the dynamic interaction data set, and identify the driving intention categories of each traffic participant, the driving intention categories including straight-through, turning operation, parking operation and emergency avoidance.

[0054] In this embodiment, after extracting the traffic participant behavior intention data, the traffic participant behavior intention data is analyzed to identify the driving intention category. For an autonomous vehicle, its behavior intention data is contained in the preset navigation path and real-time control instructions. By analyzing the above behavior intention data, if the vehicle is straight driving according to the current path and there is no steering instruction, it is identified as straight driving. If the vehicle is detected to have a steering control signal and the navigation path shows that the direction will be changed soon, it is identified as a steering operation. If the vehicle receives an instruction to stop at a specific location and is gradually slowing down to approach the location, it is identified as a parking operation. If the vehicle detects an emergency obstacle and has control actions such as sudden deceleration and sharp steering, it is identified as emergency avoidance.

[0055] For a manually driven device, by analyzing its operation state data and motion trajectory, if the device maintains a straight driving state and the steering operation amount is zero or very small, it is identified as straight driving. If a large steering operation amount is detected and the driving direction of the device is changing, it is identified as a steering operation. If the device is gradually decelerating and finally stops at a specified work point (such as a loading and unloading area or a parking space), and combined with the parking requirements of the device in the task scheduling information, it is identified as a parking operation. If the device suddenly decelerates greatly and frequently changes direction during driving, and there are sudden obstacles around it (such as other devices suddenly entering or falling goods), it is identified as emergency avoidance.

[0056] In the identification process, for a manually driven device, it also needs to be combined with its historical driving behavior pattern for auxiliary judgment. For example, a certain manually driven forklift usually slows down before turning to park when approaching a specific loading and unloading area in past operations. When similar speed and steering changes are detected again, it can be more accurately identified as a parking operation. Through the above method, the driving intention categories of all traffic participants are identified, and the identification results are stored as structured data and associated with the identification information of the traffic participants.

[0057] Step S124: determining the potential interaction relationship between the traffic participants according to the driving intention category, the potential interaction relationship including cross driving, converging driving, following driving and parallel driving.

[0058] Step S1241: constructing a driving intention interaction rule library, the driving intention interaction rule library including potential interaction relationship judgment rules corresponding to different driving intention category combinations.

[0059] In this embodiment, when constructing the driving intention interaction rule library, all possible driving intention category combinations are first shared. Since the driving intention category includes straight-through, turning operation, parking operation and emergency avoidance, the possible combinations include straight-through and straight-through, straight-through and turning operation, straight-through and parking operation, straight-through and emergency avoidance, turning operation and turning operation, turning operation and parking operation, turning operation and emergency avoidance, parking operation and parking operation, parking operation and emergency avoidance, and emergency avoidance and emergency avoidance.

[0060] For each driving intention category combination, the corresponding potential interaction relationship judgment rule can be constructed in combination with the driving characteristics and interaction rules of the traffic participants in the port operation scene. For example, for the combination of straight-through and straight-through, if the driving routes of the two traffic participants intersect and can reach the intersection within a preset time, it is judged that there is a potential interaction relationship of crossing driving; if the driving routes are parallel and in the same direction, and the speeds are similar, it is judged that there is a potential interaction relationship of parallel driving; if the driving routes are in the same direction and follow each other, it is judged that there is a potential interaction relationship of following driving.

[0061] For the combination of straight-through and turning operation, if the turning route of the traffic participant of the turning operation intersects with the driving route of the traffic participant of the straight-through, it is judged that there is a potential interaction relationship of crossing driving; if the driving routes of the two after turning tend to be consistent, and there is a merging point, it is judged that there is a potential interaction relationship of confluence driving.

[0062] For the combination of turning operation and turning operation, if the turning directions of the two make the driving routes intersect, it is judged that there is a potential interaction relationship of crossing driving; if the driving routes are in the same direction after turning and form front and rear following, it is judged that there is a potential interaction relationship of following driving.

[0063] For other combinations, such as parking operation and straight-through, if the driving route of the traffic participant of the straight-through passes through the vicinity of the parking area where the traffic participant of the parking operation is located, there may be a potential interference interaction relationship, which needs to be further judged according to the specific distance and speed.

[0064] These rules are systematically arranged to clearly define the judgment conditions of different potential interaction relationships under each combination (such as route intersection condition, distance range, speed relationship, etc.), forming a structured driving intention interaction rule library, which is stored in the rule database of the central data processing center for subsequent potential interaction relationship judgment.

[0065] Step S1242: For any two traffic participants in the port operation area, the driving intention category combination thereof is obtained.

[0066] In this embodiment, when obtaining the driving intention category combination of any two traffic participants, first, all traffic participants in the port operation area are traversed to form a traffic participant pair list, ensuring that each pair of traffic participants is covered and not repeated.

[0067] For each pair of traffic participants (such as traffic participant E and traffic participant F) in the traffic participant pair list, the driving intention categories of E and F are extracted from the identified driving intention category data, respectively. For example, the driving intention category of E is straight-through, and the driving intention category of F is turning operation, and the combination of the two is straight-through and turning operation.

[0068] The extracted driving intention category combination is stored in association with the identification information of the traffic participant pair (such as the identification numbers of E and F) to form a data set containing all traffic participant pairs and their corresponding driving intention category combinations.

[0069] Step S1243: Matching the driving intention category combination with the driving intention interaction rules in the driving intention interaction rule library to determine the preliminary judgment result of the corresponding potential interaction relationship.

[0070] In this embodiment, when performing the matching operation, for the driving intention category combination of each pair of traffic participants, the corresponding judgment rule is retrieved from the driving intention interaction rule library. For example, for the combination of straight-through and turning operation, the corresponding rule is retrieved as “if the turning route of the traffic participant performing the turning operation intersects with the driving route of the traffic participant performing the straight-through, it is determined that there is a potential interaction relationship of intersecting driving; if the routes after turning converge, it is determined that there is a potential interaction relationship of converging driving”.

[0071] Then, according to the judgment conditions in the rule, the real-time driving route information of the pair of traffic participants (extracted from the dynamic interaction data set) is combined to perform condition verification. If the turning route of traffic participant F (turning operation) intersects with the driving route of traffic participant E (straight-through), it is preliminarily determined that there is a potential interaction relationship of intersecting driving between the two; if the route after F turns converges with the route of E at a certain point, it is preliminarily determined that there is a potential interaction relationship of converging driving.

[0072] The potential interaction relationship type obtained by matching is recorded as the preliminary judgment result, and the rule item and the verified condition information in the matching process are recorded for reference when verifying the authenticity of the potential interaction relationship subsequently.

[0073] Step S1244: Extracting the real-time position data and motion direction data of the traffic participants in the dynamic interaction data set, combining the preliminary judgment result of the potential interaction relationship, and verifying the authenticity of the potential interaction relationship.

[0074] In this embodiment, after extracting the real-time position data (such as coordinate values in the port coordinate system) and the motion direction data (such as the heading angle) of the traffic participants, the authenticity of the potential interaction relationship is verified for each pair of traffic participants with the preliminary judgment result.

[0075] For example, for traffic participants E and F preliminarily judged to have a potential interaction relationship of cross driving, the current position coordinates of the two are calculated through the real-time position data, and the driving trajectories of the two in a future period of time are simulated in combination with the motion direction data. If the simulated trajectories show that the two will indeed meet at the intersection, and the meeting time is within a reasonable range (such as the time difference between the arrival of the two at the intersection being within a preset time threshold), it is indicated that the preliminary judgment of the potential interaction relationship of cross driving has high authenticity; if the simulated trajectories show that the two will not arrive at the same intersection, or the time difference is too large, it is indicated that the preliminary judgment result may be inaccurate and needs to be further verified.

[0076] For traffic participants preliminarily judged to have a potential interaction relationship of converging driving, it is verified whether the driving trajectories of the two will converge at a preset merging point, and whether the speed and distance at the time of converging are consistent with the characteristics of converging; for a preliminary judgment of following driving, it is verified whether the two maintain the same direction of driving, and whether the distance between the following vehicle and the leading vehicle is within the following distance range; for a preliminary judgment of parallel driving, it is verified whether the driving routes of the two are parallel, the directions are consistent, and the speeds are similar.

[0077] Through the above verification process, the preliminary judgment results that do not conform to the actual driving situation are eliminated, and the preliminary judgment results of the real and effective potential interaction relationship are retained.

[0078] Step S1245: If the real-time position data and the motion direction data of the two traffic participants indicate that they will be in the same area within a preset time, it is confirmed that the potential interaction relationship is established.

[0079] In this embodiment, when judging whether the two traffic participants will be in the same area within a preset time, the driving trajectory ranges of the two within the preset time are first predicted according to their real-time position data and motion direction data. The preset time is set according to the general driving speed of the traffic participants in the port operation area and the size of the area, and the principle is to effectively judge whether there is an interaction possibility.

[0080] Then, it is calculated whether the predicted trajectory ranges of the two traffic participants have an overlapping area, i.e., the same area. If there is an overlapping area, and the time when the two arrive at the area is within the preset time, it is confirmed that the preliminary judgment of the potential interaction relationship is established. For example, the predicted trajectories of traffic participants G and H both cover a certain specific area (such as the range of an intersection) within the preset time, and it is confirmed that the potential interaction relationship between them is established.

[0081] If the predicted trajectory range has no overlap, or one of the parties reaches the overlap area beyond a preset time, the potential interaction relationship is not confirmed to exist. Through the above judgment, the potential interaction relationship that actually exists is further screened out.

[0082] Step S1246: The potential interaction relationship confirmed to exist is evaluated for interaction strength by the ratio of the distance between the traffic participants to the relative speed, and an interaction strength value is obtained.

[0083] In this embodiment, when the interaction strength is evaluated, the real-time distance data (extracted from the dynamic interaction data set, which is the straight-line distance between the current positions of the two traffic participants) and the relative speed data (calculated according to the motion direction and the driving speed of the two, that is, the size of the speed vector difference of the two) of the two traffic participants involved in the potential interaction relationship confirmed to exist are first obtained.

[0084] Then, the ratio of the distance between the two to the relative speed is calculated, which reflects the approximate time required for the two traffic participants to reach each other's current position. The smaller the ratio, the closer the interaction occurs, and the higher the interaction strength. The larger the ratio, the farther the interaction occurs, and the lower the interaction strength. The ratio is taken as the interaction strength value, and the smaller the value, the higher the interaction strength.

[0085] For example, the distance between traffic participants I and J is D, and the relative speed is V (V is not zero), and the interaction strength value is D divided by V. If D is small and V is large, the interaction strength value is small, indicating that the interaction strength of the two is high.

[0086] Step S1247: The potential interaction relationship with an interaction strength value greater than a preset threshold is marked as a primary potential interaction relationship, and the potential interaction relationship with an interaction strength value less than or equal to the preset threshold is marked as a secondary potential interaction relationship.

[0087] In this embodiment, the preset threshold is set based on the average driving speed of the traffic participants in the port operation area, the distance characteristics of the typical interaction scenarios, and the safety operation requirements. By analyzing the interaction cases in the historical interaction data that lead to conflicts or require key coordination, a reasonable threshold range is determined, so that the potential interaction relationship with an interaction strength value less than or equal to the preset threshold is usually the one that needs to be given priority to and handled.

[0088] For example, after analysis, it is determined that the preset threshold is T. For a potential interaction relationship with an interaction strength value S, if S is less than or equal to T, it is marked as a primary potential interaction relationship, indicating that the interaction relationship is urgent and needs to be considered in the subsequent collaborative decision-making; if S is greater than T, it is marked as a secondary potential interaction relationship, indicating that the interaction relationship is relatively moderate and can be considered as a secondary factor in the collaborative decision-making.

[0089] After the marking is completed, the marking result of each potential interaction relationship is recorded and stored in association with information such as the pair of traffic participants, the interaction type, and the like.

[0090] Step S1248: Record the main potential interaction relationship and the secondary potential interaction relationship between all traffic participants in the port operation area, and establish a potential interaction relationship list.

[0091] In this embodiment, when the potential interaction relationship list is established, all the marked main potential interaction relationship and the secondary potential interaction relationship are summarized. Each item in the potential interaction relationship list contains information such as the identification of the pair of traffic participants (such as the identification numbers of the two traffic participants), the type of potential interaction relationship (such as cross driving, converging driving, and the like), the interaction intensity value, the marking type (main or secondary), the basis for confirming the establishment (such as the verification result of the real-time position and motion direction data), and the like.

[0092] The list is structured and organized, which can be arranged in order from large to small or from small to large according to the interaction intensity value, or can be classified and arranged according to the interaction type, so that the relevant information can be quickly queried and used when the traffic participant correlation feature is extracted subsequently. At the same time, an index mechanism of the potential interaction relationship list is established, and all potential interaction relationships involving a traffic participant can be quickly retrieved through the identification number of the traffic participant, thereby improving the data query efficiency.

[0093] Step S125: Generate a traffic participant correlation feature in combination with the task correlation matrix and the potential interaction relationship, the traffic participant correlation feature being used to describe the interaction closeness of different traffic participants in the task execution process.

[0094] In this embodiment, when the traffic participant correlation feature is generated in combination with the task correlation matrix and the potential interaction relationship, first, the correlation degree value of each pair of traffic participants (i.e., the corresponding element value in the matrix) is extracted from the task correlation matrix, which reflects the correlation closeness of the two in the task attribute.

[0095] Then, the potential interaction relationship information of the pair of traffic participants is extracted from the potential interaction relationship list, including the interaction type, the interaction intensity value, and the marking type (main or secondary). For the main potential interaction relationship, a higher interaction influence weight is given; for the secondary potential interaction relationship, a lower interaction influence weight is given.

[0096] The correlation degree value of the task correlation matrix and the interaction influence weight of the potential interaction relationship are weighted and fused (such as multiplying the correlation degree value by the corresponding interaction influence weight), to obtain a comprehensive correlation value. At the same time, the interaction intensity value is considered, the smaller the interaction intensity value (the more urgent the interaction), the larger the correction coefficient of the comprehensive correlation value, that is, the comprehensive correlation value is further improved; otherwise, the correction coefficient is smaller.

[0097] In addition, the comprehensive correlation value can be further adjusted in combination with the historical interaction frequency and interaction quality (such as whether a conflict has occurred and how the coordination efficiency is) between the traffic participants. The comprehensive correlation value of traffic participants with high historical interaction frequency and good coordination efficiency is appropriately increased, and vice versa.

[0098] The finally obtained comprehensive correlation value and the corresponding interaction type, marker type and the like information jointly constitute a traffic participant correlation relationship feature, which comprehensively reflects the interaction closeness of different traffic participants in the task execution process, including two aspects of factors of task attribute correlation and actual driving interaction.

[0099] Step S126: analyzing the automatic driving vehicle real-time running data and the manual driving device operation state data in the dynamic interaction data set to extract a task coordination demand parameter, the task coordination demand parameter including a task completion time requirement, a path coincidence degree and a resource sharing demand.

[0100] In the embodiment, when analyzing the related data in the dynamic interaction data set to extract the task coordination demand parameter, first, the required completion time of the task executed by each traffic participant is obtained from the port operation task scheduling information according to the task completion time requirement, and the remaining task completion time is calculated in combination with the current time, which is taken as the parameter value of the task completion time requirement. The smaller the parameter value is, the more urgent the task is and the higher the coordination demand is.

[0101] For the path coincidence degree, the planned driving path of each traffic participant (extracted from the navigation information for the automatic driving vehicle and inferred according to the task starting point, target point and historical driving route for the manual driving device) is decomposed into a plurality of continuous road segments (such as the road segments divided by taking the landmark points in the port operation area as nodes). Then, the proportion of the length of the overlapping road segment in the path of any two traffic participants to the total length of the path is calculated, and the smaller value or the average value of the two proportions is taken as the path coincidence degree parameter value of the two traffic participants. The larger the path coincidence degree parameter value is, the higher the overlap degree of the two in the driving path is, and the more intense the demand for coordinated path planning is.

[0102] For resource sharing needs, the cargo information associated with the task and the resource capabilities of the traffic participants are analyzed. For example, if the tasks performed by two traffic participants involve different links of the same batch of goods (such as one responsible for transportation and one responsible for loading and unloading), and the use of loading and unloading equipment needs to be coordinated with the arrival time of the transportation vehicle, there is a need for equipment resource sharing; if multiple traffic participants need to use the same work area (such as the same loading and unloading position), there is a need for site resource sharing. According to the necessity and closeness of resource sharing, the resource sharing needs are quantitatively described, such as represented by a resource sharing coefficient. The larger the resource sharing coefficient, the higher the demand for resource sharing.

[0103] The three parameters of task completion time requirement, path coincidence degree and resource sharing demand are integrated to form a task coordination demand parameter, and each parameter is associated with the corresponding traffic participant or traffic participant pair, so as to be used in subsequent coordination decision analysis.

[0104] Step S127: Based on the traffic participant association relationship features and the task coordination demand parameters, a conflict risk assessment model is constructed, and a conflict risk assessment index between the traffic participants is calculated through the conflict risk assessment model. The conflict risk assessment index is used to represent the possibility and severity of conflict between the traffic participants in the driving process.

[0105] In this embodiment, when constructing the conflict risk assessment model, first, the input variables of the model are determined as the traffic participant association relationship features (such as the comprehensive association value, the interaction type, etc.) and the task coordination demand parameters (such as the task completion time requirement, the path coincidence degree, etc.). Then, a suitable model structure is selected, and a machine learning model (such as a neural network model) trained based on historical data can be used. The machine learning model learns the case data of conflicts between traffic participants in history to establish a mapping relationship between the input variables and the conflict risk.

[0106] The training data of the conflict risk assessment model is derived from the historical conflict records in the port operation area and the corresponding traffic participant association relationship features and task coordination demand parameter data. In the training process, the possibility of conflict (such as whether a conflict occurs) and the severity (such as the size of the loss caused by the conflict, the range of influence, etc.) are taken as labels, and the parameters of the model are adjusted to enable the model to accurately predict the conflict risk.

[0107] When calculating conflict risk assessment indicators using a conflict risk assessment model, the relationship characteristics and task coordination requirements of the traffic participants to be assessed are input into the trained conflict risk assessment model. The model then outputs the corresponding conflict risk assessment indicators. These indicators can consist of two parts: a probability value indicating the likelihood of conflict and a severity level or quantitative value indicating the severity of the conflict. The combination of these two parts comprehensively characterizes the conflict risk among traffic participants.

[0108] For example, for traffic participants K and L, their comprehensive correlation values, path overlap, task completion time requirements, etc. are input into the conflict risk assessment model. The conflict risk assessment model outputs the probability of conflict as P and the severity as S. Then the conflict risk assessment index between K and L is (P, S). These two values ​​can be used to determine the risk status of conflict between the two during the driving process.

[0109] Step S128: Integrate the relationship characteristics of the traffic participants, the task coordination requirement parameters, and the conflict risk assessment indicators to generate dynamic relationship representation data.

[0110] In this embodiment, when generating dynamic association representation data by integrating the data from the above three parts, the structure and feature dimensions of each part of the data are first determined. The association characteristics of traffic participants exist in the form of matrices and lists. The matrix reflects the degree of association between all traffic participants, while the list clarifies the primary and secondary potential interaction relationships. The task coordination requirement parameters contain multiple sub-parameters, each with its corresponding description and quantitative value. The conflict risk assessment index also consists of multiple assessment items, each corresponding to the risk situation of different combinations of traffic participants.

[0111] Next, feature alignment processing is performed on the above three parts of data. Based on the unique identifier of traffic participants, the correlation degree value and potential interaction relationship information related to each traffic participant in the traffic participant correlation feature are matched with the task collaboration requirement parameters and conflict risk assessment indicators corresponding to that traffic participant, to ensure that the relevant data of the same traffic participant can be accurately matched.

[0112] Then, feature splicing is used for fusion. The matrix data in the traffic participant relationship features is converted into vector form, with each traffic participant corresponding to a relationship vector. The elements of the vector contain the relationship degree value and interaction relationship type identifier between the traffic participant and all other traffic participants. The sub-parameters in the task coordination requirement parameters are arranged in a preset order to form the task coordination requirement vector. Similarly, the evaluation items in the conflict risk assessment indicators are arranged in order to form the conflict risk vector.

[0113] Then, the association relationship vector, the task coordination demand vector and the conflict risk vector of each traffic participant are spliced to form a comprehensive feature vector of the traffic participant. For example, for traffic participant A, the association relationship vector is V1, the task coordination demand vector is V2, and the conflict risk vector is V3. After splicing, the comprehensive feature vector formed is [all elements of V1, all elements of V2, and all elements of V3].

[0114] In the splicing process, it is necessary to ensure that the dimensions of the vectors can be correctly connected and there is no dimension conflict. At the same time, in order to facilitate the subsequent processing of the data by the collaborative decision analysis, the comprehensive feature vector after splicing can be standardized in format, and the storage format and coding method of the data can be unified.

[0115] Finally, the comprehensive feature vectors of all traffic participants are summarized to form a set containing the dynamic association information of all traffic participants in the entire port operation area, i.e. dynamic association representation data. The dynamic association representation data completely retains the association relationship, coordination demand and conflict risk information between traffic participants, and the information is interrelated, which can fully reflect the dynamic characteristics of the port mixed running scene.

[0116] Step S130: performing collaborative decision analysis based on the dynamic association representation data to determine the priority sequence of the traffic right of each traffic participant in the port operation area and the collaborative path planning scheme, the priority sequence of the traffic right being generated by sorting the node importance index, and the collaborative path planning scheme including the recommended driving path, the expected passing time and the path change permission condition.

[0117] Step S131: analyzing the traffic participant association relationship characteristics in the dynamic association representation data to determine the association influence range of each traffic participant, the association influence range including other traffic participants having main potential interaction relationship with the traffic participant.

[0118] In this embodiment, when analyzing the traffic participant association relationship characteristics in the dynamic association representation data, the main potential interaction relationship list and the association degree value are focused on. For each traffic participant, all other traffic participants having main potential interaction relationship with the traffic participant are selected from the main potential interaction relationship list.

[0119] Then, taking the real-time position distribution of the other traffic participants as a reference, and in combination with the geographical layout of the port operation area, the associated influence range of the traffic participant is determined. The associated influence range not only includes the area where the other traffic participants are currently located, but also includes the area that they can reach within a preset time period. For example, if traffic participant B has a main potential interaction relationship with traffic participant A, and traffic participant B is driving in a certain direction at a certain speed, the associated influence range of traffic participant A will cover the current position of traffic participant B and the area that traffic participant B can travel to in the future.

[0120] Meanwhile, the size of the associated influence range is also related to the correlation degree value. The higher the correlation degree value of the other traffic participant, the greater the weight of the other traffic participant in the associated influence range, and the greater the proportion of the corresponding area in the associated influence range. Through the above-mentioned manner, the associated influence range of each traffic participant is accurately determined, and the other traffic participants that need to be considered in the collaborative decision-making are clearly determined.

[0121] Step S132: Construct a traffic participant collaborative network based on the associated influence range, wherein the nodes in the traffic participant collaborative network are traffic participants, the edges between the nodes represent the correlation relationships between the traffic participants, and the weights of the edges are the correlation degree values.

[0122] Step S1321: Take each traffic participant in the port operation area as an independent node, assign a unique identifier to each node, and implement the identifier to contain traffic participant type information and task number information.

[0123] In this embodiment, when assigning a unique identifier to each traffic participant, a hierarchical coding structure is adopted. The first two characters of the identifier represent the type of the traffic participant, such as “AV” representing an autonomous vehicle, “MV” representing a manually driven device, “QC” representing a shore-based container crane, and “YG” representing a yard bridge; the middle four digits represent the device number of the traffic participant, which is used to distinguish different devices under the same type; and the last six digits represent the task number of the current execution, which is consistent with the task number in the port operation task scheduling information. For example, the identifier “AV0001T202305” indicates that the autonomous vehicle numbered 0001 is executing the operation task numbered T202305. Through this coding method, the core attribute information of the traffic participant can be intuitively obtained from the identifier, which is convenient for subsequent management and analysis of the collaborative network.

[0124] Step S1322: For each traffic participant node, find other traffic participant nodes that have a main potential interaction relationship within the associated influence range of the traffic participant node.

[0125] In this embodiment, when looking for main potential interaction relationship nodes, first, the potential interaction relationship list established in step S1248 is called to screen out entries marked as "main potential interaction relationship" with the identifier of the current traffic participant node as the index. Then, in combination with the associated influence range geographical boundary determined in step S131, the screened entries are spatially verified to eliminate traffic participant nodes marked as main potential interaction relationship but actually located beyond the associated influence range. For example, the associated influence range of traffic participant A is a circular area with a radius of 50 meters centered at its current location. By comparing the position coordinates, traffic participant nodes located outside the area in the potential interaction relationship list are excluded, and only nodes within the area are retained as associated nodes.

[0126] Step S1323: A connection edge is established between the traffic participant node and other traffic participant nodes within its associated influence range, and the direction of the connection edge represents the direction of interaction influence.

[0127] In this embodiment, a directed edge is used to represent the direction of interaction influence when establishing the connection edge. If the driving trajectory of traffic participant B will influence the driving decision of traffic participant A (e.g., B is located in front of A), a directed edge is established from B to A. If there is mutual influence between the two (e.g., two-way traffic at an intersection), a bidirectional directed edge is established. The establishment of the connection edge is based on real-time motion state analysis, and the influence direction is determined by calculating the relative position vector and motion trend prediction. For example, when A is approaching the parking area of B, the parking state of B will affect the traffic decision of A, so a directed edge from B to A is established. When A and B are driving in the same lane in the same direction and are close to each other, there is mutual influence between them, and a bidirectional directed edge is established.

[0128] Step S1324: The weight value of the connection edge is determined according to the correlation value between the traffic participants.

[0129] In this embodiment, the weight value of the connection edge directly uses the comprehensive correlation value in the traffic participant correlation relationship feature generated in step S125. This value has fused the correlation degree value of the task correlation matrix and the interaction strength of the potential interaction relationship, and the value range has been normalized to between 0 and 1. The larger the comprehensive correlation value, the higher the weight value of the connection edge, indicating that the correlation between the two is closer. For example, the comprehensive correlation value of traffic participants C and D is 0.85, so the weight value of the connection edge between them is set to 0.85. The comprehensive correlation value of traffic participants E and F is 0.32, so the weight value of the connection edge is 0.32. The weight value is accurate to two decimal places, ensuring that the quantization accuracy meets the needs of collaborative decision-making.

[0130] Step S1325: Adding attribute labels to the connection edges in the traffic actor collaborative network, the attribute labels including interaction relationship type, interaction intensity, and interaction duration.

[0131] In this embodiment, the addition of attribute labels is based on the list of potential interaction relationships and real-time monitoring data. The interaction relationship type label directly adopts the type determined in step S124, such as "crossing driving", "converging driving", "following driving", "parallel driving", etc.; the interaction intensity label adopts the interaction intensity value calculated in step S1246, and the original calculation result is retained to reflect the real interaction urgency; the interaction duration label is determined by predicting the time difference between the entry and exit of the traffic actor into the interaction area, for example, the time interval from the entry into the interaction influence range of A and B to the exit is 3 minutes, and the interaction duration label is "180 seconds". All attribute labels are stored in the attribute field of the connection edge in the form of key-value pairs, facilitating the quick extraction of key information by network analysis algorithms.

[0132] Step S1326: Topological structure analysis of the traffic actor collaborative network to identify key nodes and key paths, the key nodes being the traffic actor nodes having connection edges with multiple other nodes, and the key paths being the connection edge sequences connecting multiple key nodes.

[0133] In this embodiment, the topological structure analysis adopts complex network analysis algorithms. The identification of key nodes is achieved by calculating the degree centrality of the nodes, and the top 20% of nodes with the highest degree centrality are marked as key nodes. For example, if there are 50 nodes in the network, the top 10 nodes with the highest degree centrality are selected as key nodes, which are usually the core equipment or hub traffic actors in the busy area. The identification of key paths adopts the shortest path algorithm, taking the connection edge weight value as the path cost, searching for the path sequence connecting the most key nodes, and the more key nodes included in the path and the larger the total weight value, the more critical the path. For example, the path sequence connecting the shore crane, the main transport channel, and the yard bridge is identified as the key path if it contains 3 key nodes and has the highest total weight value.

[0134] Step S1327: After spatial constraint processing of the traffic actor collaborative network according to the physical layout of the port operation area, visual processing of the constructed traffic actor collaborative network is performed to generate the traffic actor collaborative network.

[0135] In this embodiment, the spatial constraint processing maps the abstract positions of the network nodes to the actual geographic coordinates of the port. By associating the identifier of each traffic participant node with its real-time GPS coordinates through coordinate conversion of the electronic map of the port, the relative positions of the nodes in the network are ensured to be consistent with the physical layout of the actual operation area. For example, the traffic participant nodes located in the A area of the container yard are positioned within the coordinate range of the A area of the map, and the nodes located in the transportation channel are distributed along the axis of the channel. The visualization processing adopts a layered rendering technique, taking the electronic map of the port as the base map, highlighting the key nodes in red, displaying the ordinary nodes in blue, and thickening the line width of the connection edges as the weight value increases, and using arrows to indicate the direction of the directed edges. At the same time, through the interactive controls, the detailed attribute tags of any node or connection edge can be viewed, and the generated visual network can be refreshed in real time, supporting the management personnel to intuitively master the port traffic coordination state.

[0136] Step S133: Analyzing the task coordination demand parameters in the dynamic correlation representation data, extracting the task completion time requirement and the path coincidence degree, and determining the traffic participant combination whose task completion time requirement value is less than the first set value and whose path coincidence degree value is greater than the second set value as the key object of collaborative decision-making.

[0137] In this embodiment, when analyzing the task coordination demand parameters in the dynamic correlation representation data, the task completion time requirement and the path coincidence degree are first extracted from the parameter set. The task completion time requirement reflects the urgency of the task to be completed, and the path coincidence degree reflects the degree of overlap between the driving paths of different traffic participants.

[0138] For each traffic participant combination (composed of two or more traffic participants with a correlation relationship), the task completion time requirement of each traffic participant in the combination and the path coincidence degree between them are obtained. Then, the value of the task completion time requirement is compared with the first set value, and the value of the path coincidence degree is compared with the second set value.

[0139] If the task completion time requirement value of a certain traffic participant combination is less than the first set value, it means that the task completion time of the combination is relatively urgent; and if the path coincidence degree value is greater than the second set value, it means that their driving paths overlap more and are prone to conflict. The traffic participant combination that meets these two conditions needs to be focused on and handled in collaborative decision-making, so it is determined as the key object of collaborative decision-making.

[0140] Through the above screening, the limited decision-making resources can be concentrated on the traffic participant combinations that most need collaborative handling, improving the efficiency and pertinence of collaborative decision-making.

[0141] Step S134: importance evaluation is performed on the nodes in the traffic participant coordination network, and a preliminary determination of the traffic right priority ranking of the traffic participants is performed according to the importance evaluation result of the nodes, to obtain a first traffic right priority ranking result.

[0142] In this embodiment, when the importance of the nodes in the traffic participant coordination network is evaluated, a combination of multiple evaluation indexes is adopted. Firstly, the degree centrality of the node is considered, that is, the number of connections between the node and other nodes. The more the number of connections, the more extensive the association between the traffic participant and other traffic participants, and the higher the degree centrality.

[0143] Secondly, the betweenness centrality is considered. The betweenness centrality reflects the degree of the node as an intermediary between other nodes in the entire network. The node with high betweenness centrality plays a key intermediary role in network information transmission and interaction.

[0144] Thirdly, the closeness centrality is considered. The closeness centrality measures the average distance of the node to all other nodes in the network. The shorter the average distance, the higher the closeness centrality, indicating that the traffic participant can quickly interact with other traffic participants.

[0145] In addition, the task attribute comprehensive weight of the node corresponding to the traffic participant is also combined. The node with a large task attribute comprehensive weight has relatively high importance.

[0146] The evaluation indexes are respectively given a set weight, and the comprehensive importance index of each node is obtained through weighted calculation. Then, all the traffic participants are sorted in descending order of the comprehensive importance index, to obtain the first traffic right priority ranking result.

[0147] For example, the degree centrality, betweenness centrality, closeness centrality and task attribute comprehensive weight of node A are weighted and calculated to obtain the highest comprehensive importance index. Therefore, in the first traffic right priority ranking result, the position of traffic participant A is the most advanced, and so on.

[0148] Step S135: the first traffic right priority ranking result is adjusted in combination with the conflict risk evaluation index in the dynamic correlation representation data, to obtain a second traffic right priority ranking result.

[0149] In this embodiment, when the first traffic right priority ranking result is adjusted in combination with the conflict risk evaluation index in the dynamic correlation representation data, the conflict risk evaluation indexes between the traffic participants are extracted, including the possibility of conflict and the severity of conflict.

[0150] For adjacent traffic participants in the first right-of-way priority ranking result, the conflict risk assessment indicators between them are analyzed. If traffic participant C is located before traffic participant D in the first ranking result, but the conflict risk assessment indicators of the two show that the conflict risk of traffic participant C with other high-priority traffic participants is much higher than that of traffic participant D with other high-priority traffic participants, in order to reduce the overall conflict risk, the position of traffic participant D can be adjusted to before traffic participant C.

[0151] At the same time, for traffic participant combinations with higher conflict risks, if the priority of a certain traffic participant is too high and may lead to conflicts that are difficult to avoid, the priority of the traffic participant can be appropriately reduced; and for traffic participants with lower conflict risks and urgent tasks, the priority of the traffic participants can be appropriately increased.

[0152] Through the above adjustment, the right-of-way priority ranking result is more reasonable, both the importance of nodes and the conflict risk are considered, and the second right-of-way priority ranking result is obtained.

[0153] Step S136: constructing a cooperative path planning scheme based on the second right-of-way priority ranking result and the task coordination demand parameter, the cooperative path planning scheme including recommended driving paths, predicted passing times and path change permission conditions of each traffic participant.

[0154] In this embodiment, when constructing the cooperative path planning scheme based on the second right-of-way priority ranking result and the task coordination demand parameter, first, a recommended driving path is planned for each traffic participant according to the priority and the task target location of the traffic participant. Traffic participants with high priority have more selection rights in path selection and can preferentially select better and smoother paths while avoiding excessive coincidence with paths of other high-priority traffic participants.

[0155] When planning the recommended driving path, the path coincidence degree in the task coordination demand parameter is fully considered, and for traffic participant combinations with high path coincidence degrees, their driving order and time interval are reasonably arranged to reduce path conflicts. For example, traffic participants with high priority are allowed to pass through a certain coincident section first, and traffic participants with low priority are allowed to enter the section after a period of time.

[0156] The predicted passing time is calculated according to the total length of the recommended driving path, the average driving speed of the traffic participant and the waiting time that may be encountered in the path. For a section that needs to wait for other traffic participants to pass, the waiting time is reasonably estimated according to the predicted passing time and interval requirement of other traffic participants and is included in the predicted passing time.

[0157] The setting of the path change permission condition is based on the real-time conflict risk assessment and traffic flow changes, and clearly indicates which alternative paths the traffic participants can choose when the conflict risk of a certain road segment exceeds the set threshold, as well as the procedures and requirements for reporting to the central control system when changing paths.

[0158] Thus, the above contents are integrated to form a complete cooperative path planning scheme, ensuring that each traffic participant can drive in an orderly manner according to the planning, thereby improving the port operation efficiency.

[0159] Step S140: generating a set of cooperative control instructions based on the priority sequence of the right-of-way and the cooperative path planning scheme, the set of cooperative control instructions including speed coordination parameters, path adjustment sequences, and conflict avoidance operation rules.

[0160] Step S141: analyzing the priority sequence of the right-of-way to determine the position index of each traffic participant in the sequence.

[0161] In this embodiment, when analyzing the priority sequence of the right-of-way, each traffic participant is assigned a position index according to the arrangement order in the sequence. The position index starts from 1 and increases in order of priority from high to low, i.e., the traffic participant with the highest priority has a position index of 1, the next highest has a position index of 2, and so on.

[0162] The position index can intuitively reflect the relative position of each traffic participant in the priority sequence of the right-of-way. The smaller the position index, the higher the priority. In subsequent generation of speed coordination parameters and other control instructions, the position index will be an important reference, for example, traffic participants with smaller position indexes will have more favorable parameter settings when adjusting speed.

[0163] Step S142: planning a specific driving trajectory for each traffic participant based on the recommended driving path in the cooperative path planning scheme.

[0164] Step S1421: extracting the recommended driving path in the cooperative path planning scheme to determine the path segments available to each traffic participant and the corresponding usage time window.

[0165] In this embodiment, when extracting the recommended driving path in the cooperative path planning scheme, the unique identifier of each traffic participant is matched to obtain the corresponding recommended driving path for each traffic participant.

[0166] Then, the recommended driving path is divided into multiple continuous path segments, each of which is bounded by two adjacent landmark locations (such as intersections, yard entrance, specific coordinate points, etc.). For each path segment, according to the predicted passing time in the cooperative path planning scheme and the use of the path segment by other traffic participants, the time window in which each traffic participant can use the path segment, i.e., the earliest time allowed to enter the path segment and the latest time to leave, is determined.

[0167] For example, for path segment P1, the use time window of traffic participant A is T1 to T2, and the use time window of traffic participant B is T3 to T4, and T2 is less than T3, so as to avoid the simultaneous presence of both on the path segment.

[0168] Step S1422: Starting from the current location of the traffic participant and ending at the target work area, an initial path candidate set is constructed in combination with the available path segments and the use time window.

[0169] In this embodiment, when the initial path candidate set is constructed starting from the current location of the traffic participant and ending at the target work area, multiple complete paths from the starting point to the ending point are generated through different combination manners according to the available path segments.

[0170] When the path segments are combined, it is ensured that the path segments can be smoothly connected, and the use time window of the path segments contained in each complete path can be connected to each other, i.e., the latest leaving time of a previous path segment is not later than the earliest entering time of a next path segment.

[0171] All complete paths that meet the conditions are collected to form the initial path candidate set, and each candidate path contains the path segment sequence passed and the corresponding time window information.

[0172] Step S1423: The length of each path in the initial path candidate set is calculated, and paths with length values in a preset range are selected as path candidate schemes.

[0173] In this embodiment, when the length of each path in the initial path candidate set is calculated, the lengths of the path segments contained in the path are added to obtain the total length of each path.

[0174] Then, a preset range of path length is set, which is determined according to the straight-line distance from the starting point to the ending point and the road layout of the port work area. Paths with total length values in the preset range are selected as path candidate schemes, and unreasonable paths that are too long or too short are excluded.

[0175] For example, if the straight-line distance from the starting point to the ending point is L, the preset range can be 1.2L to 1.8L, and paths with total lengths in the range are retained as path candidate schemes.

[0176] Step S1424: Analyzing the conflict situation between the path candidate and the path candidates of other traffic participants, calculating the path conflict times and the conflict duration.

[0177] In this embodiment, when analyzing the conflict situation between the path candidate and the path candidates of other traffic participants, firstly, each path candidate of the current traffic participant is compared with each path candidate of all other traffic participants in the port operation area one by one. The comparison content includes the path segment involved in the path candidate, the use time window of each path segment, and the spatial position distribution of the path.

[0178] For two path candidates, if there is a common path segment and the use time windows of the two path candidates corresponding to the traffic participants on the common path segment overlap, it is determined that the two path candidates have a conflict on the path segment. The path segment involved in each conflict and the time period of the time window overlap are recorded.

[0179] The calculation method of the path conflict times is to count the total number of conflicts between the current path candidate and the path candidates of all other traffic participants. For example, the current path candidate A has 2 conflicts with the path candidate B1 of traffic participant B and 1 conflict with the path candidate C1 of traffic participant C, and the path conflict times of the path candidate A is 3.

[0180] The calculation of the conflict duration is to add up the time lengths of the time window overlap time periods of each conflict. For each common path segment with time window overlap, the time length of the overlap time period is the end time minus the start time of the overlapping part of the two time windows. All the time lengths are added up to obtain the total conflict duration of the path candidate.

[0181] By calculating the path conflict times and the conflict duration, the conflict situation between each path candidate and other path candidates can be effectively obtained.

[0182] Step S1425: The path candidate with path conflict times less than the set number threshold and conflict duration less than the set time threshold is determined as a high-quality path candidate.

[0183] In this embodiment, when determining the high-quality path candidate, firstly, the set number threshold of the path conflict times and the set time threshold of the conflict duration are set according to the traffic flow, the path complexity, and the requirement on the operation efficiency of the port operation area.

[0184] Then, the path conflict times of each path alternative are compared with a set number threshold, and the conflict duration is compared with a set time threshold. If the path conflict times of a path alternative are less than the set number threshold, and the conflict duration is less than the set time threshold, it means that the path alternative performs well in path coordination with other traffic participants, and the possibility and impact of conflict are relatively low. Therefore, it is determined as a high-quality path alternative.

[0185] For example, the set number threshold is 3 times, the set time threshold is 10 time units, the conflict times of a path alternative are 2 times, and the conflict duration is 8 time units. The path alternative is determined as a high-quality path alternative.

[0186] Step S1426: The high-quality path alternative is estimated for the passing time, and a high-quality path alternative with a passing time less than a set passing time and meeting the use time window requirement is selected as a recommended driving path to generate a specific driving trajectory.

[0187] In this embodiment, when the high-quality path alternative is estimated for the passing time, the driving time through each path segment is calculated according to the length of each path segment included in the high-quality path alternative and the estimated driving speed of the traffic participant on the different path segments. The driving time is calculated by dividing the path segment length by the estimated driving speed.

[0188] Then, the waiting time that may occur on each path segment is considered, including the waiting caused by the sequence of other traffic participants on the same path segment and the use time window, and the waiting caused by traffic control at the path segment entrance and other factors. The waiting time is added to the driving time of the corresponding path segment.

[0189] The driving time and the waiting time of all path segments are added to obtain the total passing time of the high-quality path alternative.

[0190] Then, the total passing time is compared with a set passing time, which is determined according to the completion time of the task requirement and the overall scheduling arrangement of the port operation. At the same time, it is checked whether the use time window of each path segment of the high-quality path alternative meets the time requirement in the coordinated path planning scheme, to ensure that there is no conflict with the use time window of the path of other traffic participants.

[0191] The high-quality path alternative with a total passing time less than the set passing time and meeting the use time window requirement is selected as the recommended driving path of the traffic participant. Then, based on the recommended driving path, a specific driving trajectory is generated according to the method described in step S142.

[0192] Step S143: According to the travel trajectory and the task completion time requirement in the task coordination demand parameter, the expected travel speed of the traffic participant is calculated, and the expected travel speed is the ratio of the travel trajectory length to the task completion time requirement.

[0193] In this embodiment, when calculating the expected travel speed of the traffic participant, the total length of the specific travel trajectory of the traffic participant is first obtained, and the total length of the travel trajectory is the sum of the lengths of all path segments and transition curves contained in the trajectory, which is obtained by analyzing and accumulating the travel trajectory data.

[0194] Then, the task completion time requirement in the task coordination demand parameter is extracted, and the time requirement is the time interval from the current time to the deadline when the task must be completed.

[0195] The calculation method of the expected travel speed is to divide the total length of the travel trajectory by the task completion time requirement. In this way, the expected travel speed calculated in this way can ensure that the traffic participant reaches the target work area according to the planned travel trajectory within the task completion time requirement.

[0196] For example, if the total length of the travel trajectory is S and the task completion time requirement is T, then the expected travel speed V is equal to S divided by T.

[0197] In the calculation process, it is necessary to ensure that the unit of the travel trajectory length matches the unit of the task completion time requirement, so as to ensure that the dimension of the expected travel speed is correct.

[0198] Step S144: Adjust the expected travel speed in combination with the position index in the priority sequence of the right-of-way, and determine the adjusted expected travel speed as the speed coordination parameter. The expected travel speed adjustment coefficient of the traffic participant with a higher position index in the priority sequence of the right-of-way is greater than that of the traffic participant with a lower position index.

[0199] In this embodiment, when adjusting the expected travel speed in combination with the position index in the priority sequence of the right-of-way, first, set the corresponding expected travel speed adjustment coefficient for each position index. The earlier the position index, the larger the corresponding adjustment coefficient; the later the position index, the smaller the adjustment coefficient. The size of the adjustment coefficient is determined according to factors such as the traffic flow, path conditions and task urgency of each traffic participant in the port work area.

[0200] Then, the expected driving speed of each traffic participant is multiplied by its corresponding adjustment coefficient to obtain an adjusted expected driving speed. For example, the adjustment coefficient of the traffic participant with position index 1 is K1, the adjustment coefficient of the traffic participant with position index 2 is K2, and K1 is greater than K2, the position index of the traffic participant A is 1, the expected driving speed of the traffic participant A is V1, and the adjusted expected driving speed of the traffic participant A is V1 multiplied by K1; the position index of the traffic participant B is 2, the expected driving speed of the traffic participant B is V2, and the adjusted expected driving speed of the traffic participant B is V2 multiplied by K2.

[0201] The adjusted expected driving speed comprehensively considers the priority of the traffic right of the traffic participant, so that the traffic participant with high priority can drive at a relatively high speed to improve the task completion efficiency, while avoiding unnecessary speed conflicts with other traffic participants. The adjusted expected driving speed is determined as the speed coordination parameter of the traffic participant.

[0202] Step S145: analyzing the predicted passing time and the path change permission condition in the cooperative path planning scheme to construct a path adjustment sequence, the path adjustment sequence including a path change opportunity, a lane selection suggestion, and a confluence point passing sequence.

[0203] In this embodiment, when analyzing the predicted passing time and the path change permission condition in the cooperative path planning scheme to construct the path adjustment sequence, first, the time points at which each traffic participant reaches the key nodes (such as intersections, confluence points, path segment start and end points, etc.) in the path are determined according to the predicted passing time. These time points are important references for determining the path change opportunity.

[0204] The determination of the path change opportunity needs to be combined with the path change permission condition. When a traffic participant reaches a certain key node, and it is judged according to the real-time traffic situation that the path needs to be changed to avoid conflicts or improve driving efficiency, and at the same time meets the change requirements (such as pre-warning time before change, distance requirements of surrounding traffic participants, etc.) specified in the path change permission condition, the time point corresponding to the key node is set as the path change opportunity.

[0205] The construction of the lane selection suggestion is based on the lane distribution of the path and the traffic flow situation. For a path segment with multiple lanes, a suitable lane is recommended for the current traffic participant according to the lane selection situation and driving speed of other traffic participants. For example, if the traffic flow on a certain lane is small and the driving speed is fast, and it meets the driving direction of the traffic participant, then the lane is taken as the lane selection suggestion.

[0206] The passing order of the confluence point is determined according to the priority sequence of the traffic participants and the time of the traffic participants expected to arrive at the confluence point. The traffic participants with high priority pass the confluence point first. If the time of multiple traffic participants expected to arrive at the confluence point is close, the passing order is determined according to the priority sequence, so as to ensure the traffic order of the confluence point and reduce congestion and conflicts.

[0207] The path change opportunity, the lane selection suggestion and the passing order of the confluence point are arranged according to time sequence and path node sequence to form a complete path adjustment sequence.

[0208] Step S146: Based on the conflict risk evaluation index in the dynamic association representation data, a conflict avoidance operation rule is constructed, which includes deceleration avoidance operation parameters, parking waiting time threshold, path detour offset and sound and light warning starting conditions.

[0209] Step S1461: The conflict risk evaluation index in the dynamic association representation data is analyzed to determine the conflict risk level. The conflict risk level is divided into different levels according to the numerical range of the conflict risk evaluation index.

[0210] In this embodiment, when the conflict risk evaluation index in the dynamic association representation data is analyzed to determine the conflict risk level, the quantitative values reflecting the conflict possibility and the conflict severity in the conflict risk evaluation index are first extracted. These values are usually obtained by comprehensively analyzing the position, speed, direction of travel, path coincidence degree and other factors of the traffic participants.

[0211] Then, according to the safety standards and historical conflict data of the port operation, the numerical range of multiple conflict risk evaluation indexes is preset, and each numerical range corresponds to a conflict risk level. For example, the numerical range of the conflict risk evaluation index is divided into four levels, and the range with the smallest value corresponds to the lowest conflict risk level, and the range with the largest value corresponds to the highest conflict risk level.

[0212] The quantitative values of the conflict risk evaluation index obtained by analysis are compared with the preset numerical range. If it falls into which numerical range, the conflict risk level of the traffic participant combination is determined as the level corresponding to the range. By determining the conflict risk level, the severity of the conflict can be more intuitively judged.

[0213] Step S1462: For different conflict risk levels, corresponding basic conflict avoidance operation rules are constructed.

[0214] In this embodiment, when the corresponding basic conflict avoidance operation rules are constructed for different conflict risk levels, a set of basic avoidance operation framework is set for each conflict risk level.

[0215] For the lowest conflict risk level, the basic conflict avoidance operation rule mainly focuses on maintaining the current driving state and enhancing observation, such as appropriately increasing vigilance, closely observing the motion state of the conflict object, and not needing to take active deceleration or steering operations.

[0216] For the lower conflict risk level, the basic conflict avoidance operation rule includes slight deceleration, adjustment of lane keeping safety distance, etc. The deceleration amplitude is small, and the lane adjustment range is limited, so as not to affect the overall driving efficiency.

[0217] For the higher conflict risk level, the basic conflict avoidance operation rule requires obvious deceleration operation, and small-range path detouring if necessary. The deceleration amplitude and detouring offset amount need to ensure that the conflict risk can be effectively reduced, and the pre-warning device is started to remind the surrounding traffic participants.

[0218] For the highest conflict risk level, the basic conflict avoidance operation rule includes emergency deceleration, immediate parking and waiting, or large-amplitude path detouring, etc. to ensure that the conflict can be completely avoided, and strong sound and light warning signals are sent to remind all relevant traffic participants.

[0219] Step S1463: Extract the real-time running data and operation state data of the traffic participants, and obtain the current driving speed, acceleration, and braking performance parameters of the vehicle.

[0220] In this embodiment, when extracting the real-time running data and operation state data of the traffic participants, for an automatic driving vehicle, the real-time driving speed, acceleration (including longitudinal and lateral acceleration), response time of the braking system, maximum braking deceleration, and other braking performance parameters are obtained through the vehicle-mounted sensors and control system; for a manually driven device, the real-time driving speed, acceleration generated by the driver's operation, stroke of the brake pedal, and maximum braking force of the braking system are obtained through the data acquisition device installed on the device.

[0221] Step S1464: Calculate the operation margin required for conflict avoidance by combining the current driving state parameters of the vehicle and the basic conflict avoidance operation rule, wherein the operation margin includes time margin and space margin.

[0222] In this embodiment, when calculating the operation margin in combination with the current driving state parameters of the vehicle and the basic conflict avoidance operation rules, the time margin is calculated as follows: according to the current driving speed of the traffic participant, the distance from the conflict object, and the deceleration or stopping requirement specified in the basic conflict avoidance operation rules, the shortest time to avoid the conflict after taking the avoidance operation is calculated, and then the time required to take the avoidance operation is subtracted to obtain the time margin. For example, if the shortest time to avoid the conflict is T1 and the time required to take the avoidance operation is T2, the time margin is T1 minus T2.

[0223] The space margin is calculated as follows: according to the driving speed of the traffic participant, the steering performance, and the path bypassing requirement specified in the basic conflict avoidance operation rules, the minimum safety distance between the traffic participant and the conflict object after taking the avoidance operation is calculated, and then the actual distance between the two is subtracted to obtain the space margin. If the calculated minimum safety distance is S1 and the current actual distance is S2, the space margin is S1 minus S2.

[0224] The time margin and the space margin together constitute the operation margin required for conflict avoidance. The larger the operation margin, the greater the buffer space for taking the avoidance operation, and the higher the likelihood of successful avoidance.

[0225] Step S1465: adjusting the specific parameters of the basic conflict avoidance operation rules according to the comparison result of the operation margin and the preset threshold value, wherein the deceleration avoidance operation rule is adopted when the time margin is greater than the preset threshold value, and the stopping and waiting operation rule is adopted when the time margin is less than or equal to the preset threshold value.

[0226] In this embodiment, when adjusting the specific parameters of the basic conflict avoidance operation rules according to the comparison result of the operation margin and the preset threshold value, the preset threshold value of the time margin and the preset threshold value of the space margin are first set. These preset threshold values are determined according to the type of the traffic participant, the driving speed, and the environmental characteristics of the port operation area.

[0227] When the time margin is greater than the preset threshold value, it means that there is enough time to avoid the conflict through deceleration operation, and the deceleration avoidance operation rule is adopted at this time. The deceleration avoidance operation parameters are adjusted, such as increasing the absolute value of the deceleration acceleration, so that the traffic participant can reduce the speed in a shorter distance and ensure a safe distance from the conflict object. At the same time, according to the size of the space margin, the lane selection is appropriately adjusted. If the space margin is large, a small amount of lane offset can be selected to assist in deceleration avoidance.

[0228] When the time margin is less than or equal to the preset threshold, it is indicated that the conflict cannot be effectively avoided only by the deceleration operation, and the parking and waiting operation rule is adopted at this time. A parking and waiting time threshold is set, and the parking and waiting time threshold is determined according to the time required for the conflict object to pass through the conflict region, to ensure that the parking and waiting time is sufficient for the conflict object to pass safely. At the same time, the specific position of parking is determined according to the space margin, and it is necessary to ensure that the parking position does not affect the normal driving of other road users, and there is enough space to restart driving later.

[0229] For the comparison result of the space margin, if the space margin is less than the preset threshold, it is indicated that the current path may not meet the detour requirement in the basic conflict avoidance operation rule, and the path detour offset needs to be further increased, or it needs to be re-evaluated whether the parking and waiting operation needs to be adopted.

[0230] Through the above adjustment, the conflict avoidance operation rule is more in line with the actual conflict situation, and the effectiveness of the avoidance operation is improved.

[0231] Step S1466: Analyzing the traffic right priority sequence of the traffic participants involved in the conflict, assigning a set number of traffic participants in the rear position in the traffic right priority sequence to the main conflict avoidance responsibility, and obtaining the conflict avoidance operation rule.

[0232] In this embodiment, when analyzing the traffic right priority sequence of the traffic participants involved in the conflict to assign the main conflict avoidance responsibility, first, all the traffic participants involved in the conflict are determined, and their position indexes in the traffic right priority sequence are checked.

[0233] According to the rule of the traffic right priority sequence, the traffic participants in the rear position have lower priority. A number, such as 2 or 3, is set, and the traffic participants in the rear position in the traffic right priority sequence are determined as the objects that need to bear the main conflict avoidance responsibility.

[0234] For these traffic participants bearing the main conflict avoidance responsibility, more stringent avoidance requirements are specified in their conflict avoidance operation rules, such as earlier start of avoidance operation, larger deceleration amplitude, longer parking and waiting time, or larger path detour offset, etc. The traffic participants in the front position bear the secondary conflict avoidance responsibility, and their avoidance operation rules are relatively relaxed, mainly to maintain the driving state and cooperate with the operation of the main avoidance responsibility party.

[0235] For example, in a conflict involving 3 traffic participants, their position indexes are 2, 5, and 7 respectively, and the set number is 2, then the traffic participants with position indexes 5 and 7 bear the main conflict avoidance responsibility, and more stringent parameters are set in their conflict avoidance operation rules.

[0236] Through the above responsibility distribution, the conflict avoidance operation is more targeted, fully considers the priority of the right-of-way, and ensures the order and efficiency of the overall traffic.

[0237] Step S147: Fuse the speed coordination parameter, the path adjustment sequence, and the conflict avoidance operation rule to generate a local coordination control instruction, and aggregate the local coordination control instructions of all traffic participants to generate a coordination control instruction set.

[0238] In this embodiment, when generating the local coordination control instruction by fusing the speed coordination parameter, the path adjustment sequence, and the conflict avoidance operation rule, first, the three parts of content are arranged in a unified format. The speed coordination parameter is expressed in the form of a clear speed value and a speed change interval; the path adjustment sequence is arranged in the order of time and path nodes; and the conflict avoidance operation rule is classified according to different conflict scenarios and corresponding operation steps.

[0239] Then, an association mapping relationship is established between the speed coordination parameter and the key nodes in the path adjustment sequence. For example, for a node that needs to turn in the path adjustment sequence, the corresponding speed coordination parameter is set to a speed value suitable for turning, to ensure that the traffic participant has a reasonable speed when turning and to ensure driving safety. At the same time, the conflict avoidance operation rule is associated with the high-risk area on the path, and when the traffic participant is about to enter the high-risk area, the corresponding conflict avoidance operation rule is activated in priority to provide clear operation guidance.

[0240] Then, the three parts of content are logically integrated to form a coherent local coordination control instruction. In the instruction, the speed coordination parameter of the traffic participant in the current stage is first specified, then the time sequence and node requirements of the path adjustment sequence are specified to explain when the path adjustment is performed and the specific way of adjustment, and finally the conflict avoidance operation rule to be followed in different scenarios is attached. For example, the local coordination control instruction first specifies that the traffic participant maintains a certain speed for the next 10 minutes, then explains that the traffic participant turns at a certain coordinate point, the speed during turning needs to be reduced to a certain range, and points out that if another traffic participant is encountered during turning, the operation should be performed according to the conflict avoidance rule.

[0241] During the integration process, it is necessary to ensure that there is no logical contradiction between the parts of content, for example, the turning operation time point required in the path adjustment sequence and the speed change time point set in the speed coordination parameter need to be matched with each other, and the operation requirements in the conflict avoidance operation rule cannot conflict with the provisions in the speed coordination parameter and the path adjustment sequence.

[0242] After the local cooperative control instruction generation of a single traffic participant is completed, the local cooperative control instructions of all traffic participants are aggregated. During aggregation, the local cooperative control instructions of each traffic participant are classified and arranged according to the unique identifier of the traffic participant, and each local cooperative control instruction of a traffic participant is taken as an independent entry, which contains the identifier information of the traffic participant and the corresponding instruction content. At the same time, consistency check is performed on the aggregated instruction set to ensure that the instructions of different traffic participants do not conflict with each other, for example, the instructions of two traffic participants in the same time period and in the same area do not contradict each other.

[0243] Finally, all the local cooperative control instructions that have passed the consistency check are integrated into a unified cooperative control instruction set, which is stored and transmitted in a standardized data format so that the control systems of various traffic participants can accurately parse and execute the relevant instructions.

[0244] Step S150: Distribute the cooperative control instruction set to the corresponding traffic participant control system to execute dynamic cooperative driving control in the port mixed driving scenario.

[0245] In this embodiment, when the cooperative control instruction set is distributed to the corresponding traffic participant control system, the connection with each traffic participant control system is first established through the communication network inside the port. The communication network uses high-reliability wireless communication technology to ensure the real-time and stability of instruction transmission.

[0246] Then, according to the traffic participant identifier corresponding to each local cooperative control instruction in the cooperative control instruction set, the instruction is accurately routed to the corresponding traffic participant control system. During transmission, the instruction is encrypted to prevent tampering or leakage of the instruction and to ensure the security of the instruction.

[0247] After receiving the local cooperative control instruction, the traffic participant control system parses the instruction, extracts key information such as speed coordination parameters, path adjustment sequence, and conflict avoidance operation rules, and converts these information into control signals executable by itself. For example, the control system of an autonomous vehicle will convert the speed coordination parameters into control signals for the accelerator and brake, convert the path adjustment sequence into control signals for the steering system, and perform the corresponding operations according to the instruction requirements.

[0248] During the execution, the traffic participant control system monitors the running state of itself and the changes of the surrounding environment in real time, and transmits the execution and feedback information back to the port central data processing center through the communication network. The central data processing center analyzes the feedback information, and if it finds that a certain traffic participant does not execute according to the instruction or an abnormal situation occurs, it can generate a collaborative control instruction according to the real-time dynamic interaction data set and distribute it to the traffic participant control system in time for dynamic adjustment, so as to ensure that the traffic participants in the port mixed driving scene can always keep collaborative driving, and improve the efficiency and safety of the port operation.

[0249] Figure 2 A schematic diagram of exemplary hardware and software components of the collaborative control system 100 based on the port automatic driving mixed driving scene that can implement the idea of the present application provided by some embodiments of the present application is shown. For example, the processor 120 can be used in the collaborative control system 100 based on the port automatic driving mixed driving scene, and is used to execute the functions in the present application.

[0250] For example, the collaborative control system 100 based on the port automatic driving mixed driving scene can include a network port 110 connected to a network, one or more processors 120 for executing program instructions, a communication bus 130, and different forms of storage media 140, such as a disk, a ROM, or a RAM, or any combination thereof. Exemplarily, the collaborative control system 100 based on the port automatic driving mixed driving scene can also include program instructions stored in a ROM, a RAM, or other types of non-transitory storage media, or any combination thereof. The methods of the present application can be implemented according to these program instructions. The collaborative control system 100 based on the port automatic driving mixed driving scene also includes an I / O interface 150 between the computer and other input / output devices.

[0251] In addition, the present application also provides a readable storage medium, wherein computer executable instructions are pre-set in the readable storage medium, and when the processor executes the computer executable instructions, the collaborative control method based on the port automatic driving mixed driving scene is realized.

[0252] It should be noted that, in order to simplify the description of the present application and to help understand one or more embodiments of the present application, in the foregoing description of the embodiments of the present application, various features are sometimes combined into one embodiment, drawing or description thereof.

Claims

1. A cooperative control method for a mixed-traffic scenario of autonomous driving in a port, characterized in that, The method includes: Acquire a dynamic interactive data set within the port operation area, the dynamic interactive data set including real-time operation data of autonomous vehicles, operation status data of manually driven equipment, port operation task scheduling information, and behavioral intention data of traffic participants; The dynamic interactive data set is processed by extracting associated features to generate dynamic associated representation data of the port mixed traffic scenario; Based on the dynamic correlation representation data, collaborative decision analysis is performed to determine the right-of-way priority sequence and collaborative path planning scheme for each traffic participant in the port operation area. The right-of-way priority sequence is generated by sorting through the node importance index, and the collaborative path planning scheme includes recommended driving routes, estimated transit time, and route change permission conditions. A set of collaborative control instructions is generated based on the right-of-way priority sequence and collaborative path planning scheme. The set of cooperative control instructions is distributed to the corresponding traffic participant control system to execute dynamic cooperative driving control in port mixed traffic scenarios. The step of performing collaborative decision analysis based on the dynamic correlation representation data to determine the right-of-way priority sequence and collaborative path planning scheme for each traffic participant within the port operation area includes: The association characteristics of traffic participants in the dynamic association representation data are analyzed to determine the association influence range of each traffic participant. The association influence range includes other traffic participants with major potential interaction relationships with the traffic participant. A traffic participant collaboration network is constructed based on the aforementioned scope of influence. Nodes in the traffic participant collaboration network are traffic participants, and the edges between nodes represent the relationships between the traffic participants. The weight of each edge is a correlation value. Analyze the task collaboration requirement parameters in the dynamic association representation data, extract the task completion time requirement and path overlap, and identify the traffic participants whose task completion time requirement is less than a first set value and whose path overlap is greater than a second set value as the key objects for collaborative decision-making. The importance of nodes in the traffic participant cooperative network is assessed, and the priority ranking of traffic participants' right-of-way is initially determined based on the node importance assessment results, resulting in the first priority ranking result of right-of-way. Based on the conflict risk assessment indicators in the dynamic association representation data, the first right-of-way priority ranking result is adjusted to obtain the second right-of-way priority ranking result. Based on the priority ranking result of the second right-of-way and the task coordination requirement parameters, a collaborative path planning scheme is constructed. The collaborative path planning scheme includes the recommended driving routes, estimated passage times, and route change permission conditions for each traffic participant.

2. The cooperative control method based on a port autonomous driving mixed-traffic scenario according to claim 1, characterized in that, The step of performing correlation feature extraction processing on the dynamic interactive data set to generate dynamic correlation representation data for the port mixed traffic scenario includes: The port operation task scheduling information in the dynamic interactive data set is analyzed, and the task attribute parameters of each traffic participant are extracted. The task attribute parameters include the urgency of the task, the characteristics of the cargo type, and the correlation of the operation area. A task association matrix is ​​constructed based on the task attribute parameters. The task urgency level is positively correlated with the urgency weight value, and the work area association value is positively correlated with the association value between traffic participants. Extract traffic participant behavior intention data from the dynamic interaction data set, identify the driving intention category of each traffic participant, and the driving intention category includes straight passage, turning operation, stopping operation and emergency avoidance; Based on the driving intention category, potential interaction relationships between traffic participants are determined, including cross-traffic, merging, following, and parallel driving. The traffic participant relationship features are generated by combining the task association matrix and potential interaction relationships. These traffic participant relationship features are used to describe the degree of interaction between different traffic participants during task execution. Analyze the real-time operation data of autonomous vehicles and the operation status data of manual driving equipment in the dynamic interactive data set, and extract task collaboration requirement parameters, which include task completion time requirements, path overlap and resource sharing requirements. A conflict risk assessment model is constructed based on the relationship characteristics of traffic participants and the parameters of task coordination requirements. The conflict risk assessment model is used to calculate the conflict risk assessment index between each traffic participant. The conflict risk assessment index is used to characterize the possibility and severity of conflict between traffic participants during driving. By integrating the relationship characteristics of traffic participants, task collaboration requirements parameters, and conflict risk assessment indicators, dynamic relationship representation data is generated.

3. The cooperative control method based on port autonomous driving mixed traffic scenario according to claim 2, characterized in that, The construction of the task association matrix based on the task attribute parameters includes: The task urgency level in the task attribute parameters is divided into different levels, and a corresponding urgency weight value is assigned to each level. The urgency weight values ​​corresponding to the task urgency levels increase in order of level. The cargo type features in the task attribute parameters are classified to determine the cargo type weight coefficient; The work area correlation in the task attribute parameters is calculated. The work area correlation is determined by the normalized distance between the current work area and the target work area of ​​the traffic participant and the work process dependency. The comprehensive weight of task attributes is constructed based on the urgency weight value, the cargo type weight coefficient, and the correlation of the work area. The comprehensive weight of task attributes is the weighted sum of the urgency weight value, the cargo type weight coefficient, and the correlation of the work area. Using traffic participants within the port operation area as the matrix rows and columns, the product of the combined weights of the task attributes of any two traffic participants is used as the matrix element value to construct an initial task association matrix. The initial task association matrix is ​​normalized, and then adjusted according to port operation rules and historical collaboration data. The correlation values ​​between traffic participants with direct operational process associations are increased, while the correlation values ​​between traffic participants without operational associations are decreased, resulting in the final task association matrix.

4. The cooperative control method based on port autonomous driving mixed traffic scenario according to claim 2, characterized in that, Determining the potential interaction relationships between traffic participants based on the driving intention category includes: Construct a driving intention interaction rule base, which contains rules for judging potential interaction relationships corresponding to different combinations of driving intention categories; For any two traffic participants within the port operation area, obtain their combination of driving intention categories; The driving intent category combination is matched with the driving intent interaction rules in the driving intent interaction rule base to determine the preliminary judgment result of the corresponding potential interaction relationship; Extract the real-time location data and movement direction data of traffic participants from the dynamic interaction data set, and combine them with the preliminary judgment results of the potential interaction relationship to verify the authenticity of the potential interaction relationship; If the real-time location data and movement direction data of two traffic participants indicate that they will appear in the same area within a preset time, then the potential interaction relationship is confirmed. The interaction intensity is assessed by using the ratio of distance to relative speed between traffic participants to obtain a numerical value for the interaction intensity of potential interactions that have been established. Potential interactions with an interaction intensity value greater than a preset threshold are marked as primary potential interactions, while interactions with an interaction intensity value less than or equal to the preset threshold are marked as secondary potential interactions. Record the primary and secondary potential interactions among all traffic participants within the port operation area, and establish a list of potential interactions.

5. The cooperative control method for mixed-traffic scenarios of autonomous driving in ports according to claim 1, characterized in that, The construction of a traffic participant collaborative network based on the associated influence range includes: Each traffic participant within the port operation area is treated as an independent node, and a unique identifier is assigned to each node. The identifier includes traffic participant type information and task number information. For each traffic participant node, find other traffic participant nodes with major potential interaction relationships within its associated influence range; Establish connection edges between a traffic participant node and other traffic participant nodes within its associated influence range, wherein the direction of the connection edge indicates the direction of the interaction; The weight value of the connecting edge is determined based on the correlation value between the traffic participants; Add attribute labels to the connection edges in the traffic participant collaborative network, the attribute labels including the interaction relationship type, interaction strength and interaction duration; A topology analysis is performed on the traffic participant collaborative network to identify key nodes and key paths. The key nodes are traffic participant nodes that have connection edges with multiple other nodes, and the key paths are sequences of connection edges that connect multiple key nodes. After spatial constraint processing of the traffic participant collaborative network based on the physical layout of the port operation area, the constructed traffic participant collaborative network is visualized to generate the traffic participant collaborative network.

6. The cooperative control method based on a port autonomous driving mixed-traffic scenario according to claim 1, characterized in that, The step of generating a set of cooperative control instructions based on the right-of-way priority sequence and cooperative path planning scheme includes: The right-of-way priority sequence is analyzed to determine the position index of each traffic participant in the sequence; Based on the recommended driving route in the collaborative path planning scheme, a specific driving trajectory is planned for each traffic participant; Based on the driving trajectory and the task completion time requirement in the task coordination requirement parameters, the expected driving speed of the traffic participants is calculated, where the expected driving speed is the ratio of the driving trajectory length to the task completion time requirement. Based on the position index in the right-of-way priority sequence, the expected driving speed is adjusted, and the adjusted expected driving speed is determined as the speed coordination parameter. The expected driving speed adjustment coefficient of the traffic participant with the earlier position index is greater than that of the traffic participant with the later position index. Analyze the estimated passage time and route change permission conditions in the collaborative route planning scheme, and construct a route adjustment sequence, which includes the timing of route changes, lane selection suggestions, and the order of passing through merging points; Based on the conflict risk assessment indicators in the dynamic correlation representation data, conflict avoidance operation rules are constructed. The conflict avoidance operation rules include deceleration and avoidance operation parameters, parking waiting time threshold, path detour offset and sound and light warning activation conditions. By integrating the speed coordination parameters, the path adjustment sequence, and the conflict avoidance operation rules, a local coordination control command is generated, and the local coordination control commands of all traffic participants are aggregated to generate a coordination control command set.

7. The cooperative control method based on a port autonomous driving mixed-traffic scenario according to claim 6, characterized in that, The step of planning a specific travel trajectory for each traffic participant based on the recommended travel route in the collaborative path planning scheme includes: Extract recommended travel routes from the collaborative path planning scheme, and determine the path segments available to each traffic participant and their corresponding time windows; Starting from the current location of traffic participants and ending at the target work area, an initial set of candidate routes is constructed by combining available route segments and usage time windows. The length of each path in the initial path candidate set is calculated, and paths with length values ​​within a preset range are selected as path candidates. Analyze the conflicts between alternative routes and alternative routes of other traffic participants, and calculate the number of route conflicts and the duration of the conflicts; Path alternatives with fewer than the set resignation threshold in terms of the number of path conflicts and shorter than the set time threshold in terms of conflict duration are identified as high-quality path alternatives. The travel time of the selected high-quality route alternatives is estimated, and the route alternatives with a travel time less than the set travel time and meeting the usage time window requirements are selected as recommended driving routes to generate specific driving trajectories.

8. The cooperative control method based on a port autonomous driving mixed-traffic scenario according to claim 6, characterized in that, The conflict avoidance operational rules are constructed based on the conflict risk assessment indicators in the dynamically correlated representation data, including: Analyze the conflict risk assessment indicators in the dynamic correlation representation data to determine the conflict risk level. The conflict risk level is divided into different levels according to the numerical range of the conflict risk assessment indicators. Develop corresponding basic conflict avoidance operation rules for different conflict risk levels; Extract real-time operational data and status data of traffic participants to obtain the vehicle's current speed, acceleration, and braking performance parameters; Based on the vehicle's current driving status parameters and basic conflict avoidance operation rules, the required operational margin for conflict avoidance is calculated, which includes time margin and space margin. Based on the comparison between the operating margin and the preset threshold, the specific parameters of the basic conflict avoidance operation rule are adjusted. When the time margin is greater than the preset threshold, a deceleration avoidance operation rule is adopted, and when the time margin is less than or equal to the preset threshold, a stop and wait operation rule is adopted. Analyze the right-of-way priority sequence of traffic participants involved in the conflict, and assign the main conflict avoidance responsibility to a set number of traffic participants who are at the end of the right-of-way priority sequence, thus obtaining the conflict avoidance operation rules.

9. A cooperative control system for a port autonomous driving mixed-traffic scenario, characterized in that, The system includes a processor and a memory, the memory being connected to the processor. The memory is used to store programs, instructions, or code, and the processor is used to execute the programs, instructions, or code in the memory to implement the cooperative control method based on a port autonomous driving mixed traffic scenario as described in any one of claims 1-8.

Images