A method and system for dynamic control of emergency lanes in highway accident scenarios

Abstract

This invention relates to the field of intelligent transportation technology, and provides a method and system for dynamic control of emergency lanes in highway accident scenarios. The method includes: acquiring accident reporting information; using a deep learning network model to predict the accident handling time; based on the accident handling time, using a traffic wave model to predict the queue length and dissipation time under the impact of the accident; using the queue length under the impact of the accident as the total length of the controlled road segment and discretizing it into segments; using the dissipation time under the impact of the accident as the control duration and intervalizing it; constructing a binary control strategy combining time windows and spatial segments; constructing an objective function with the goal of minimizing the weighted average of total travel time and total collision exposure time; solving the objective function using a Kriging surrogate model; and obtaining the dynamic control strategy for the emergency lane based on the binary control strategy. This improves the response speed of accident handling.

Description

Technical Field

[0001] This invention relates to the field of intelligent transportation technology, and in particular to a method and system for dynamic control of emergency lanes in highway accident scenarios. Background Technology

[0002] The statements in this section are merely background information related to the present invention and do not necessarily constitute prior art.

[0003] As the core arteries of modern transportation networks, highways directly impact regional economic vitality and social order through their traffic efficiency and safety. However, with the continuous growth of traffic flow, localized congestion caused by traffic accidents is becoming increasingly prominent. After an accident, traditional response methods rely on manual experience-based judgment and decentralized information feedback, often leading to delayed responses that cause congestion to spread and even trigger secondary accidents. Achieving accurate prediction of accident impacts, dynamic traffic diversion, and proactive management has become a key challenge for the intelligent operation of highways.

[0004] Meanwhile, emergency lane opening and control technology is a crucial technology in intelligent highway traffic management. Opening emergency lanes at appropriate times, especially after an accident, can effectively alleviate highway congestion and improve highway operational safety to some extent. Different emergency lane opening strategies will lead to different application effects in different scenarios. After an accident, the impact on the highway varies depending on the severity of the accident, with differences in indicators such as lane congestion range, congestion duration, average lane speed, average lane travel time, and lane occupancy. The differences in congestion length and duration under different accident levels have a significant impact on both highway efficiency and safety. Summary of the Invention

[0005] To address the technical problems mentioned above, this invention provides a dynamic management method and system for emergency lanes in highway accident scenarios. This invention integrates real-time accident parameters with historical travel patterns, and uses machine learning and traffic wave models to dynamically predict accident impacts. Based on the obtained spatiotemporal impact range, the continuous emergency lane opening problem is transformed into a binary decision variable of "time window × spatial segment," forming a scalable discretized spatiotemporal strategy library. Minimizing both "total travel time" and "total collision exposure time" are simultaneously incorporated into the objective function, achieving a balance between efficiency and safety. Finally, for the first time, the Kriging surrogate model is applied to highway emergency lane management scenarios. Leveraging its accurate fitting ability for high-dimensional, nonlinear traffic states, the accuracy of management strategy evaluation is significantly improved. Coupled with a genetic algorithm with strong global search capabilities, the Kriging surrogate model reduces the number of simulation calls and escapes local optima, solving complex discrete-continuous hybrid optimization problems.

[0006] To achieve the above objectives, the present invention adopts the following technical solution:

[0007] The first aspect of the present invention provides a method for dynamic control of emergency lanes in highway accident scenarios.

[0008] A method for dynamic control of emergency lanes in highway accident scenarios includes:

[0009] Accident reporting information is obtained, and a deep learning network model is used to predict the accident handling time. Based on the accident handling time, a traffic wave model is used to predict the queue length and dissipation time under the impact of the accident.

[0010] The total length of the controlled road segment is determined by the queue length under the impact of the accident, and the segment is discretized. The control duration is determined by the dissipation time under the impact of the accident, and the segment is intervalized. A binary control strategy combining time window and spatial segment is constructed. The objective function is constructed with the goal of minimizing the weighted average of total travel time and total collision exposure time.

[0011] The objective function is solved using the Kriging surrogate model, and the dynamic control strategy for the emergency lane is obtained based on the binary control strategy.

[0012] Furthermore, the method for obtaining accident reporting information and predicting accident response time using a deep learning network model includes:

[0013] The system obtains the route name, number of lanes, traffic flow, accident type, reporting source, number of vehicles involved, vehicle type, number of lanes occupied, whether a non-emergency lane was occupied, and the time of occurrence. Using an improved TabNet network model, the system obtains the accident handling time.

[0014] The improved TabNet network model includes: introducing a hierarchical multi-step attention masking mechanism in the attention converter, first filtering important groups from among feature groups, and then selecting key features in fine-grained manner within each group; and introducing a random perturbation mechanism during the training process of the TabNet network model to perturb the generated feature mask.

[0015] Furthermore, the total length of the controlled road segment is determined by the queue length affected by the accident, and this segment is discretized; the control duration is determined by the dissipation time affected by the accident, and this duration is intervalized; the method includes:

[0016] Spatially, the total length of the controlled road segment is discretized into segments of roughly equal length. The length of each segment is The length is within a certain range;

[0017] Control duration in terms of time Interval processing is performed, with a fixed time interval defined as the time interval. Within the same time interval, the open and closed states of different emergency lane segments remain fixed.

[0018] set up To characterize the first A 0-1 variable representing the open / closed status of the emergency lane in a spatial segment. When it is 1, it means the first... Emergency vehicle lanes in the space are open; A value of 0 indicates that the emergency lane is closed;

[0019] In the same time interval The mathematical form of the control strategy for all spatial segments is: ;

[0020] Control duration By time interval Divided into The segment refers to the duration. The control strategy for emergency lanes can be mathematically represented as follows: ;

[0021] The control strategy is mathematically a two-dimensional... A 0-1 variable matrix describing the entire control period Emergency lane control section The open and closed states of each spatial segment at different time intervals.

[0022] Furthermore, based on the accident handling time, a traffic wave model is used to predict the queue length and dissipation time under the influence of the accident; the method includes: calculating the rally wave velocity and the initiation wave velocity respectively, and calculating the queue length and dissipation time under the influence of the accident according to the relationship between the rally wave and the initiation wave.

[0023] Furthermore, the objective function is expressed by the following formula:

[0024]

[0025]

[0026]

[0027] in, Indicates the first The first cycle The first time step The vehicle's TTS (Traffic Time Tolerance) value; Indicates the first The first cycle Vehicle TET value at each time step; , These represent the weights of efficiency and safety indicators, respectively. Indicates the number of vehicles currently connected to the network; A 0 / 1 variable indicating whether the vehicle is in the simulation area; Indicates the simulation interval time; This represents the total simulation time; Indicates the first The first time step The car value; This represents the threshold for TTC.

[0028] Furthermore, in the process of solving the problem using the Kriging surrogate model, an online update mechanism is adopted. For candidate solutions with large prediction errors or excellent target values, a portion of samples are dynamically selected for real simulation evaluation and included in the training set to update the Kriging surrogate model. Through iterative updates, the dynamic evolution of the Kriging surrogate model from global coarse modeling to local high-precision fitting is realized.

[0029] A second aspect of the present invention provides a dynamic control system for emergency lanes in highway accident scenarios.

[0030] A dynamic control system for emergency lanes in highway accident scenarios includes:

[0031] The accident spatiotemporal impact prediction module is configured to: acquire accident reporting information, use a deep learning network model to predict the accident handling time; and based on the accident handling time, use a traffic wave model to predict the queue length and dissipation time under the impact of the accident.

[0032] The model building module for road segment discretization is configured as follows: the total length of the controlled road segment is taken as the queue length under the impact of the accident, and the segment is discretized; the control duration is taken as the dissipation time under the impact of the accident, and the interval is set; a binary control strategy combining time window and spatial segment is constructed; and an objective function is constructed with the goal of minimizing the weighted average of total travel time and total collision exposure time.

[0033] The solution module is configured to: solve the objective function using the Kriging surrogate model, and obtain the dynamic control strategy for the emergency lane based on the binary control strategy.

[0034] A third aspect of the present invention provides a computer device comprising:

[0035] A processor, adapted to execute computer programs;

[0036] A computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps in the dynamic control method for emergency lanes in highway accident scenarios as described in the first aspect above.

[0037] A fourth aspect of the present invention provides a computer-readable storage medium storing a computer program adapted to be loaded by a processor and to execute the steps of the dynamic control method for emergency lanes in a highway accident scenario as described in the first aspect above.

[0038] The fifth aspect of the present invention provides a computer program product or computer program.

[0039] This invention provides a computer program product or computer program including computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the steps in the dynamic control method for emergency lanes in a highway accident scenario as described in the first aspect above.

[0040] Compared with the prior art, the beneficial effects of the present invention are:

[0041] 1. Accident Spatiotemporal Impact Prediction Based on Handling Time Prediction and Traffic Wave Theory: Based on the received highway accident reports, a deep learning network model integrating hierarchical attention mechanism and mask perturbation feature selection training method is proposed. This model is used to predict the accident handling time, and combined with the road network traffic conditions at the time of the accident, the traffic wave model is used to predict the spatiotemporal impact range of the accident, thereby realizing the priority division of accident response.

[0042] 2. Construction of an Emergency Lane Opening Optimization Model Based on Road Segment Discretization: Combining the predicted spatiotemporal impact range of accidents, this paper for the first time transforms the continuous emergency lane opening problem into a binary decision variable of "time window × spatial segment," forming an scalable discretized spatiotemporal strategy library, which is convenient for genetic algorithm chromosome encoding. Considering both highway operation efficiency and safety, the objective function of the emergency lane opening optimization model is constructed.

[0043] 3. Dynamic Opening Solution Algorithm for Emergency Lanes Based on Kriging-GA: This is the first time that the Kriging model has been applied to the emergency lane management scenario on highways. During the solution process, the Kriging model is used to represent the implicit relationship between the input and output of the simulation model, leveraging its accurate fitting ability for high-dimensional and nonlinear traffic states to improve the accuracy of management strategy evaluation. A genetic algorithm with strong global search capabilities is coupled with the Kriging surrogate model, and a dynamic update mechanism is introduced to form a three-layer framework of "surrogate-evolution-iterative optimization." This reduces the number of simulation calls and allows the model to escape local optima, solving complex discrete-continuous hybrid optimization problems. Attached Figure Description

[0044] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an improper limitation of the invention.

[0045] Figure 1 This is a flowchart illustrating the dynamic control method for emergency lanes in highway accident scenarios, as shown in an embodiment of the present invention.

[0046] Figure 2 This is a technical roadmap of the dynamic control method for emergency lanes in highway accident scenarios shown in the embodiments of the present invention;

[0047] Figure 3 This is a schematic diagram of an attention converter according to an embodiment of the present invention;

[0048] Figure 4 This is a schematic diagram of a traffic wave model shown in an embodiment of the present invention;

[0049] Figure 5 This is a structural diagram of the emergency lane dynamic control system in a highway accident scenario, as shown in an embodiment of the present invention.

[0050] Figure 6 This is a structural diagram of a computer device shown in an embodiment of the present invention. Detailed Implementation

[0051] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0052] It should be noted that the following detailed description is illustrative and intended to provide further explanation of the invention. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0053] It should be noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the scope of exemplary embodiments according to the invention. As used herein, the singular form is intended to include the plural form as well, unless the context clearly indicates otherwise. Furthermore, it should be understood that when the terms "comprising" and / or "including" are used in this specification, they indicate the presence of features, steps, operations, devices, components, and / or combinations thereof.

[0054] This invention addresses the congestion problem caused by highway traffic accidents, aiming to solve three major technical bottlenecks in traditional emergency management: data fragmentation, delayed response, and crude strategies. By integrating historical travel patterns with real-time accident parameters, it achieves accurate prediction of the accident's impact range and dynamic optimization of emergency lane opening strategies, forming a dynamic management method and system for emergency lanes in highway accident scenarios.

[0055] Figure 1 This is a flowchart illustrating a dynamic control method for emergency lanes in highway accident scenarios, as shown in an embodiment of the present invention; see reference. Figure 1 The method includes:

[0056] Accident reporting information is obtained, and a deep learning network model is used to predict the accident handling time. Based on the accident handling time, a traffic wave model is used to predict the queue length and dissipation time under the impact of the accident.

[0057] The total length of the controlled road segment is determined by the queue length under the impact of the accident, and the segment is discretized. The control duration is determined by the dissipation time under the impact of the accident, and the segment is intervalized. A binary control strategy combining time window and spatial segment is constructed. The objective function is constructed with the goal of minimizing the weighted average of total travel time and total collision exposure time.

[0058] The objective function is solved using the Kriging surrogate model, and the dynamic control strategy for the emergency lane is obtained based on the binary control strategy.

[0059] This invention focuses on the deep application of the ETC gantry system, breaking through the limitations of a single data dimension. It constructs a dynamic road network impact assessment model through the fusion of all-element data (integrating structured data such as ETC route flow, historical OD selection ratios, and real-time accident parameters), and combines machine learning prediction models and traffic wave models to predict the scale of accident development. It also establishes a spatiotemporal model of emergency lane opening strategies, abstracting the actual opening strategy into a mathematical model problem. Combining highway operating efficiency and safety, it constructs an objective function for the emergency lane opening optimization model and selects a suitable solution algorithm. Finally, it introduces a response surface methodology to address the problems of long solution time, complex optimization process, and unsuitability for practical engineering applications in the emergency lane opening optimization model.

[0060] Figure 2This is a technical roadmap illustrating the dynamic control method for emergency lanes in highway accident scenarios according to an embodiment of the present invention; according to the technical roadmap, the dynamic control method for emergency lanes in highway accident scenarios includes:

[0061] Step 1: Prediction of the spatiotemporal impact of the accident.

[0062] (1) Access to accident reporting information

[0063] Access to detailed accident information proactively discovered or passively reported through various methods such as video surveillance, manual surveillance, video AI, and telephone reporting, including the accident location, mileage marker, direction, accident type, number of vehicles involved, and lane occupancy status.

[0064] (2) Spatiotemporal impact prediction of accidents based on deep learning and traffic wave model

[0065] Based on accident reporting information, the scope and duration of the accident's impact are predicted by combining the road network and the background and predicted traffic flow of the accident section.

[0066] 1) Predict accident handling time based on machine learning model. Input fields include route name, number of lanes, traffic flow, accident type, reporting source, number of vehicles involved, type of vehicles involved, number of lanes occupied, whether non-emergency lanes are occupied, and time of occurrence.

[0067] ① First, analyze the relevant influencing factors, including accident attribute factors, road attribute factors, environmental attribute factors, and traffic attribute factors; second, take the accident type, rescue method, number of lanes occupied, month of accident, date of accident, hour of accident, whether trucks or freight trucks were involved, road management company of the accident section, whether it was daytime, traffic flow saturation, road speed limit, and number of lanes of the road section as inputs, and the accident handling time as output.

[0068] ②Then, train the improved TabNet network model.

[0069] TabNet's core design philosophy is to simulate the decision manifold of decision trees through a neural network structure. Traditional decision trees achieve data classification or regression through feature partitioning, while TabNet uses Mask layers and Attentive Transformers to achieve sparse feature selection, and then uses Feature Transformers to process the selected features. This design not only preserves the interpretability of decision trees but also achieves more efficient weight feature allocation through sparsity regularization. TabNet's model structure is similar to an additive model, composed of multiple basic decision trees, each corresponding to a feature selection and processing step. This structure allows TabNet to perform feature selection like a decision tree while possessing the powerful representational capabilities of a neural network.

[0070] To further improve the performance of the TabNet model in terms of feature selection accuracy and generalization ability, improvements were made in two dimensions: model structure and training strategy. Specifically, the following two key innovations were included:

[0071] First: Hierarchical Multi-step Attention Mask Mechanism: To address the limitation of the mask generation process in the original TabNet relying on a single layer of sparse attention, this embodiment introduces a two-level attention mechanism of "between-group and within-group" based on the Attentive Transformer module. First, important groups are selected from between feature groups, and then key features are selected at a fine-grained level within the groups, thereby constructing a more structured, stable, and semantically constrained sparse mask, improving the expressive power and generalization ability of feature selection.

[0072] The hierarchical multi-step attention masking mechanism is mainly reflected in the improvements of the attention converter, such as... Figure 3 As shown.

[0073] The Attention Transformer is a key component of TabNet for feature selection. This module outputs an attention mask based on the feature representation of the previous layer. It introduces a hierarchical multi-step attention masking mechanism, injecting structural priors into feature selection by introducing two levels of attention: "group-wise" and "intra-group," guiding the model to form a more stable and interpretable mask sequence.

[0074] Input features Pre-divided into G feature groups: In the first layer of attention (Attention1), weights are generated for each group:

[0075]

[0076] in, For between groups, Networks for generating attention, This represents the global features of the transformation of the original input X.

[0077] Then, in the second attention layer (Attention2), for the selected group g, local attention is used to perform sparse selection of features within the group:

[0078]

[0079] in, For group weights, Networks for generating attention, This is the feature of the g-th group.

[0080] The final attention mask M is:

[0081]

[0082] To prevent reuse, feature usage is introduced:

[0083]

[0084] In the formula, i Let γ be the number of decision steps, and γ be the relaxation parameter. M [ j ] is the first j The mask is obtained step by step. To improve the sparsity of the mask, a sparse regularization term is introduced into the loss function, and the value of each step is included in the overall loss:

[0085]

[0086] In the formula, ε is passed into the model as a parameter, and it is generally a very small value. For decision-making steps, B For batch size, D For the total number of features, In the first i In step-by-step decision-making, samples b The j The mask value of each feature, Used to constrain the sparsity of attention masks and prevent them from becoming too uniform.

[0087] Second: Mask Perturbation Training Strategy: To mitigate the risk of overfitting due to excessive reliance on certain feature masks during training, this embodiment introduces a random perturbation mechanism during the training phase. This mechanism applies perturbations (such as mask discarding, perturbation noise, or sequence perturbation) to the generated feature masks, forcing the model to maintain prediction robustness under various mask combinations, effectively enhancing the model's generalization performance and anti-interference ability.

[0088] The mask perturbation training mechanism aims to break the deterministic structure of the mask generation process and improve the robustness of the model under various feature combinations by introducing multiple forms of mask perturbations. The specific method is as follows:

[0089] Before each decision step, a random perturbation is introduced into the original mask, forcing the model to adapt to diverse "feature attention patterns" and improving its generalization ability. The perturbation method is mask dropout, and the calculation formula is as follows:

[0090]

[0091] In the formula, To generate the selection mask vector for the features in step i, This is the mask after Dropout perturbation. p To preserve the probability, Bernoulli ( p Let be a random vector following a Bernoulli distribution, where each element has a probability of p Retain, with probability 1- p Set to zero.

[0092] In each training iteration, a random perturbation strategy is applied to the currently generated data to generate a perturbed mask, which is used for subsequent feature selection operations.

[0093]

[0094] in, X [ i The feature subset input to the decision module in step i. X This is the input feature vector.

[0095] 2) Using the accident handling time as input to the traffic wave model, the queue length and dissipation time under the impact of the accident are predicted to obtain the predicted spatiotemporal impact range of the accident. The traffic wave model is as follows: Figure 4 As shown.

[0096] First, the speed of the gathered wave. for:

[0097]

[0098] In the formula, , These are the traffic flow density and flow rate of the road under normal conditions. , ; , The density and traffic capacity at the accident location under accident conditions. , When the accident completely blocks the road, .

[0099] After specifying the formula using the Green Shields linear relationship model, the traffic characteristic parameters upstream of the accident site are obtained as follows:

[0100]

[0101] In the formula, The actual traffic capacity of the road under normal conditions ( ); Traffic congestion density under accident conditions ( ).

[0102] make , It is a dimensionless parameter between 0 and 1, reflecting the continuous deterioration of traffic conditions from free flow to congestion. .

[0103] The wave velocity of the condensed wave is:

[0104]

[0105] In the formula, The normal speed of vehicles on the road ( ).

[0106] Secondly, the starting wave velocity is:

[0107]

[0108] In the formula, The initial traffic flow when traffic begins to resume; .

[0109] Finally, At that moment, the initiating wave catches up with the rallying wave, and the two waves travel the same distance, that is:

[0110]

[0111] In the formula, For the time required for accident response, This refers to the duration of the accident. for:

[0112]

[0113] This is the maximum queue length. for:

[0114]

[0115] Step 2: Constructing an optimization model for emergency lane opening based on road segment discretization.

[0116] (1) Discretization of controlled road sections

[0117] Based on the above prediction of the accident impact range, the accident control strategy consists of control time and control distance, which describes the starting point and length of the entire emergency lane that need to be opened within a certain period of time.

[0118] Implementing emergency lane control strategies on highways can improve traffic conditions on congested or accident-prone sections.

[0119] Assuming the duration of control is This value is taken as the duration of the accident in step one. The total length of the road section requiring control is: This value is taken as the maximum queue length from step one. .

[0120] Spatially, the emergency lane control section is segmented and discretized, dividing the total road segment into spatial segments of roughly equal length. The length of each segment is The length is within a certain range.

[0121] In terms of time, the duration of control Interval processing is performed, with a fixed time interval defined as the time interval. Within the same time interval, the open and closed states of different emergency lane segments remain fixed.

[0122] set up To characterize the first The 0-1 variables represent the open / closed status of the emergency lane in each spatial segment. When it is 1, it means the first... Emergency vehicle lanes in the space are open; A value of 0 indicates that the emergency lane is closed.

[0123] In the same time interval The mathematical form of the control strategy for all spatial segments is: .

[0124] Control duration By time interval Divided into The segment refers to the duration. The control strategy for emergency lanes can be mathematically represented as follows: .

[0125] The control strategy is mathematically a two-dimensional... A 0-1 variable matrix describing the entire control period Emergency lane control section The open and closed states of each spatial segment at different time intervals.

[0126] (2) Optimization goals for opening emergency lanes in order to ensure safety and efficiency

[0127] Considering that some control strategies are unreasonable, time and space constraints are added to limit these strategies. This is because each control strategy is implemented... Depending on the different impacts on highway traffic conditions, the optimal control strategy that maximizes highway operating efficiency and safety is sought under constraints. This is an optimization problem that urgently needs to be solved.

[0128] The optimization of target selection takes into account both efficiency and security, seeking the most efficient opening strategy while ensuring security.

[0129] The optimization objectives consider both travel efficiency and driving safety, seeking the most efficient opening strategy while ensuring safety. Total Travel Spend (TTS) and Total Time Exposed to Collision (TET) are selected as the travel efficiency objective and the driving safety objective, respectively, and the two objectives are weighted and combined into a total objective. Due to strategy With the goal Without explicit mathematical analytical relationships, this problem is considered a "black box" problem. Introducing the Kriging surrogate model from response surface methodology can approximate the input strategy. With output target The relationships between the strategies are as follows: Each strategy corresponds to a solution. Considering the versatility and wide search range of the Genetic Algorithm (GA), it is chosen as the optimization algorithm. During the search for the optimal solution using the optimization algorithm, the Kriging surrogate model is used to accelerate the optimization process.

[0130] The total travel time of all vehicles in the road network is chosen as the first objective function. The road utilization rate is highest and the output flow per unit time is maximized when the total travel time of all vehicles is minimized. The following formula expresses the total travel time (TTS):

[0131]

[0132]

[0133] In the formula, Indicates the first The first cycle The first time step The vehicle's TTS (Traffic Time Tolerance) value; Indicates the number of vehicles currently connected to the network; A 0 / 1 variable indicating whether the vehicle is in the simulation area; Indicates the simulation step time; This represents the total simulation time.

[0134] From a safety perspective, the Time Exposed Time to Collision (TET) of all vehicles in the road network is selected as the second objective function. TET represents the total time spent in a safety emergency, and is calculated using a threshold below the TTC. The total time is used to characterize the collision. Time to Collision (TTC) refers to the time between two vehicles traveling on the same route at their respective speeds until a collision occurs. Referring to relevant literature, 1.5 seconds is taken as the TTC threshold. When the TTC is less than 1.5 seconds, the vehicles are considered to be in a dangerous emergency situation and the TET time is included. The TET calculation formula is as follows:

[0135]

[0136]

[0137] In the formula, Indicates the first The first cycle The first time step Vehicle TET value; Indicates the number of vehicles currently connected to the network; A 0 / 1 variable indicating whether a dangerous or emergency situation is in progress; Indicates the simulation interval time; This represents the total simulation time; Indicates the first The first time step The car value; This represents the threshold for TTC.

[0138] Taking into account both safety and efficiency indicators, a weighted approach is used to combine the two objectives of safety and efficiency for optimization. By normalizing the safety and efficiency indicators and then performing a weighted sum, the overall objective function is obtained. The expression is as follows:

[0139]

[0140] In the formula, , These represent the weights of efficiency and safety indicators, respectively.

[0141] Step 3: Algorithm for dynamic opening of emergency lanes based on improved Kriging-GA.

[0142] To accelerate the optimization model's solution speed, the Kriging meta-model is used as an auxiliary model. By using the Kriging meta-model to represent the implicit relationship between the simulation model's input and output, and combining it with the search algorithm, the optimization speed is accelerated.

[0143] The basic assumptions of the Kriging metamodel are as follows:

[0144]

[0145] In the formula, These are observed values, calculated from global and local estimates. It is a regression function vector. It is a vector of regression coefficients. This represents the global estimate. It is a function with a mean of zero and a variance of . The stochastic process function. Its covariance is as follows:

[0146]

[0147] In the formula, It's about variables. stochastic process function, and These are sample points. Function Take a Gaussian random process, the expression is as follows:

[0148]

[0149] In the formula, , and They are , and The One component. This represents the dimension of the design space.

[0150] The principle of the Kriging metamodel is to address unknown points in the experimental data. At this point, an unbiased estimate is given, and the basic expression is as follows:

[0151]

[0152] In the formula, Indicates in The estimated value of the point; This is the regression coefficient vector, estimated using the generalized least squares estimate. Each component is represented by the following formula:

[0153]

[0154]

[0155]

[0156] In the formula, Representing unknown points With training points ( The correlation vector between ) Let be a correlation matrix with dimension . In the formula This refers to the column vector of the training data. In the formula... It is by A calculated matrix.

[0157] Unknown parameter vector To encode traffic control strategies as candidate solutions in the solution space, namely, which road segments' emergency lanes should be opened and their distances, a Kriging meta-model is introduced as a surrogate model for simulation due to the high computational cost of simulation-based strategy evaluation. This model efficiently approximates the mapping relationship between strategies and their effects. The prediction accuracy of the Kriging model depends on the setting of its internal parameter θ. Therefore, a genetic algorithm is used to globally optimize θ, aiming to maximize the likelihood of the Kriging model to the simulation data, thereby ensuring the reliability of the surrogate model. Within this framework, the genetic algorithm is responsible for optimizing the Kriging model parameters, while the high-precision Kriging model is responsible for quickly evaluating the merits of the control strategies, ultimately driving the optimization process to discover a globally better strategy.

[0158] To address the issue of insufficient local accuracy caused by the traditional Kriging model's "static training-fixed prediction" approach, this embodiment designs an online update mechanism. During the optimization process, for candidate solutions with large prediction errors or excellent target value performance, a subset of samples are dynamically selected for real-world simulation evaluation and incorporated into the training set to update the Kriging model. Through iterative updates, the model dynamically evolves from "global coarse modeling" to "local high-precision fitting," effectively improving the surrogate model's ability to represent key search regions and accelerating the optimization convergence speed.

[0159] The core of online updates is selecting new sample points, adding them to the existing sample set, and retraining the Kriging model. A sampling criterion based on prediction uncertainty is introduced, such as Expected Improvement (EI):

[0160]

[0161]

[0162] In the formula, Let x be the mean of the Kriging model's predictions at the beginning of x. To predict the standard deviation, This is the predicted value corresponding to the current optimal solution. and Let Z be the distribution function and probability density of a normal distribution, and Z be the standardization improvement measure, which is used to measure the potential improvement of a candidate point relative to the current optimal solution.

[0163] During the model update phase, let the initial training set be... Choose a new point Construct a new training set:

[0164]

[0165] Then, the Kriging model is trained based on the new training set.

[0166] This invention achieves dynamic prediction of accident impact by integrating real-time accident parameters with historical travel patterns, based on machine learning and traffic wave models. Based on the obtained spatiotemporal impact range, the continuous emergency lane opening problem is transformed into a binary decision variable of "time window × spatial segment," forming a scalable discretized spatiotemporal strategy library. Minimizing both "total travel time" and "total collision exposure time" are simultaneously incorporated into the objective function, achieving a balance between efficiency and safety. Finally, for the first time, the Kriging surrogate model is applied to highway emergency lane management scenarios. Leveraging its accurate fitting capability for high-dimensional, nonlinear traffic states, the accuracy of management strategy evaluation is significantly improved. Coupled with a genetic algorithm possessing strong global search capabilities, the Kriging surrogate model reduces the number of simulation calls and escapes local optima, solving complex discrete-continuous hybrid optimization problems.

[0167] The above combination Figure 1 , Figure 2 The method for dynamic control of emergency lanes in highway accident scenarios provided by the embodiments of the present invention has been described in detail. Next, the dynamic control system for emergency lanes in highway accident scenarios provided by the embodiments of the present invention will be described in conjunction with the accompanying drawings.

[0168] Figure 5 This is a schematic diagram of the structure of a dynamic control system for emergency lanes in a highway accident scenario, as shown in an embodiment of the present invention. Figure 5 The system described in this invention includes:

[0169] The accident spatiotemporal impact prediction module is configured to: acquire accident reporting information, use a deep learning network model to predict the accident handling time; and based on the accident handling time, use a traffic wave model to predict the queue length and dissipation time under the impact of the accident.

[0170] The model building module for road segment discretization is configured as follows: the total length of the controlled road segment is taken as the queue length under the impact of the accident, and the segment is discretized; the control duration is taken as the dissipation time under the impact of the accident, and the interval is set; a binary control strategy combining time window and spatial segment is constructed; and an objective function is constructed with the goal of minimizing the weighted average of total travel time and total collision exposure time.

[0171] The solution module is configured to: solve the objective function using the Kriging surrogate model, and obtain the dynamic control strategy for the emergency lane based on the binary control strategy.

[0172] In some embodiments, obtaining accident reporting information and using a deep learning network model to predict accident handling time includes:

[0173] The system obtains the route name, number of lanes, traffic flow, accident type, reporting source, number of vehicles involved, vehicle type, number of lanes occupied, whether a non-emergency lane was occupied, and the time of occurrence. Using an improved TabNet network model, the system obtains the accident handling time.

[0174] The improved TabNet network model includes: introducing a hierarchical multi-step attention masking mechanism in the attention converter, first filtering important groups from among feature groups, and then selecting key features in fine-grained manner within each group; and introducing a random perturbation mechanism during the training process of the TabNet network model to perturb the generated feature mask.

[0175] In some embodiments, the total length of the controlled road segment is taken as the queue length affected by the accident, and then segmented and discretized; the control duration is taken as the dissipation time affected by the accident, and then intervalized; the method includes:

[0176] Spatially, the total length of the controlled road segment is discretized into segments of roughly equal length. The length of each segment is The length is within a certain range;

[0177] Control duration in terms of time Interval processing is performed, with a fixed time interval defined as the time interval. Within the same time interval, the open and closed states of different emergency lane segments remain fixed.

[0178] set up To characterize the first A 0-1 variable representing the open / closed status of the emergency lane in a spatial segment. When it is 1, it means the first... Emergency vehicle lanes in the space are open; A value of 0 indicates that the emergency lane is closed;

[0179] In the same time interval The mathematical form of the control strategy for all spatial segments is: ;

[0180] Control duration By time interval Divided into The segment refers to the duration. The control strategy for emergency lanes can be mathematically represented as follows: ;

[0181] The control strategy is mathematically a two-dimensional... A 0-1 variable matrix describing the entire control period Emergency lane control section The open and closed states of each spatial segment at different time intervals.

[0182] In some embodiments, the step of predicting the queue length and dissipation time under the influence of an accident using a traffic wave model based on the accident handling time includes: calculating the rally wave velocity and the initiation wave velocity respectively, and calculating the queue length and dissipation time under the influence of the accident based on the relationship between the rally wave and the initiation wave.

[0183] In some embodiments, the objective function is expressed by the following formula:

[0184]

[0185]

[0186]

[0187] in, Indicates the first The first cycle The first time step The vehicle's TTS (Traffic Time Tolerance) value; It means that in the t-th period, the first... Vehicle TET value at each time step; , These represent the weights of efficiency and safety indicators, respectively. Indicates the number of vehicles currently connected to the network; A 0 / 1 variable indicating whether the vehicle is in the simulation area; Indicates the simulation interval time; This represents the total simulation time; Indicates the first The first time step The car The time to TTC is the time required for a vehicle to collide with an obstacle in front of it. This represents the threshold for TTC.

[0188] In some embodiments, during the solution process using the Kriging surrogate model, an online update mechanism is employed. For candidate solutions with large prediction errors or excellent target values, a subset of samples are dynamically selected for real simulation evaluation and incorporated into the training set to update the Kriging surrogate model. Through iterative updates, the dynamic evolution of the Kriging surrogate model from global coarse modeling to local high-precision fitting is achieved.

[0189] According to embodiments of the present invention, the dynamic control system for emergency lanes in highway accident scenarios can correspond to the execution of the methods described in the embodiments of the present invention, and the above-mentioned and other operations and / or functions of each module of the dynamic control system for emergency lanes in highway accident scenarios are respectively for implementing Figure 1 , Figure 2 For the sake of brevity, the corresponding processes of each method in the code will not be elaborated here.

[0190] See Figure 6 The diagram shows the structure of a computer device, which includes a processor, a communication interface, and a computer-readable storage medium. The processor, communication interface, and computer-readable storage medium are connected via a bus or other means. The communication interface is used to receive and send data. The computer-readable storage medium can be stored in the computer device's memory. The computer-readable storage medium stores computer programs, including program instructions, and the processor executes the program instructions stored in the computer-readable storage medium. The processor (or CPU, Central Processing Unit) is the computing and control core of the computer device, suitable for implementing one or more instructions, specifically suitable for loading and executing one or more instructions to achieve the corresponding steps in the embodiment of the dynamic management method for emergency lanes in highway accident scenarios.

[0191] This embodiment provides a computer-readable storage medium (Memory), which is a memory device in a computer device used to store programs and data. It is understood that the computer-readable storage medium here can include both the built-in storage medium in the computer device and extended storage media supported by the computer device. The computer-readable storage medium provides storage space that stores the processing system of the computer device.

[0192] Furthermore, this storage space also contains one or more instructions suitable for loading and execution by the processor. These instructions can be one or more computer programs (including program code). It should be noted that the computer-readable storage medium here can be high-speed RAM memory or non-volatile memory, such as at least one disk storage device; optionally, it can also be at least one computer-readable storage medium located remotely from the aforementioned processor.

[0193] In one embodiment, the computer-readable storage medium stores one or more instructions; the processor loads and executes one or more instructions stored in the computer-readable storage medium to implement the corresponding steps in the above-described embodiment of the dynamic control method for emergency lanes in highway accident scenarios.

[0194] This embodiment provides a computer program product or computer program, which includes computer instructions stored in a computer-readable storage medium. The processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the corresponding steps in the above embodiment of the dynamic control method for emergency lanes in highway accident scenarios.

[0195] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of hardware embodiments, software embodiments, or embodiments combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage and optical storage) containing computer-usable program code.

[0196] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0197] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0198] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0199] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. The storage medium can be a magnetic disk, optical disk, read-only memory (ROM), or random access memory (RAM), etc.

[0200] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. An emergency lane dynamic management and control method in a highway accident scene, characterized in that, The method comprises the following steps: obtaining accident report information, using a deep learning network model to predict accident disposal time; based on the accident disposal time, using the traffic wave model to predict the queue length and dissipation time under the influence of the accident; taking the queue length under the influence of the accident as the total length of the control section, and performing segmented discretization; taking the dissipation time under the influence of the accident as the control time length, and performing intervalization; a binary control strategy combining time window and space section is constructed; taking the weighted minimum of total travel time and total collision exposure time as the target, a target function is constructed; using the Kriging surrogate model to solve the target function, and obtaining the dynamic control strategy of the emergency lane according to the binary control strategy; taking the queue length under the influence of the accident as the total length of the control section, and performing segmented discretization; taking the dissipation time under the influence of the accident as the control time length, and performing intervalization; the method comprises: In space, the total length of the controlled road section is segmented and discretized, and the total length of the controlled road section is divided into spatial segments with the same length The length of each segment is , and the length is within a certain range. Controlling the time length Intervalization, taking a fixed time as a time interval The opening and closing states of different emergency lane space segments are fixed and unchanged within the same time interval. Let be a 0-1 variable representing the open-close status of the i-th space segment emergency lane, be 1 if the i-th space segment emergency lane is open, be 0 if the i-th space segment emergency lane is closed.​​ In the same time interval The mathematical representation of the control policy for all spatial segments is given by: ; Controlling the time length By time interval Divided into Segments, the time length Emergency lane control strategy in mathematical form is expressed as: ; The control strategy is a two-dimensional 0-1 variable matrix in mathematical form, which describes the opening and closing states of each spatial segment of the emergency lane control section at different time intervals within the entire control duration. ​​​ the method comprises the following steps: the method comprises the following steps: wherein, represents the TTS value of the vehicle at the tth cycle and the th time step; represents the TET value of the vehicle at the tth cycle and the th time step; represents the TTC value of the vehicle at the tth cycle and the th time step; represents the TTC value of the vehicle at the tth cycle and the th time step; represents the TTC value of the vehicle at the tth cycle and the th time step; represents the TTC value of the vehicle at the tth cycle and the th time step; , represent the weights of efficiency and safety indicators, respectively; represents the number of vehicles in the network; is a 0 / 1 variable indicating whether the vehicle is in the simulation area or not; represents the simulation interval time; represents the total simulation time; represents the TTC value of the vehicle at the tth cycle and the th time step; represents the TTC value of the vehicle at the tth cycle and the th time step; represents the TTC value of the vehicle at the tth cycle and the th time step; represents the TTC value of the vehicle at the tth cycle and the th time step; represents the threshold value of TTC.

2. The method of claim 1, wherein, obtaining the route name, lane number, traffic volume, accident type, report source, number of vehicles involved, vehicle type involved, number of occupied lanes, whether to occupy non-emergency lane and occurrence time, using an improved TabNet network model to obtain the accident disposal time; wherein, the improved TabNet network model comprises: introducing a hierarchical multi-step attention mask mechanism in the attention transformer, first screening important groups from the feature groups, and then selecting key features in the group in a fine-grained manner; in the training process of the TabNet network model, a random disturbance mechanism is introduced to disturb the generated feature mask. In the solving process of the Kriging surrogate model, an online updating mechanism is used, and for candidate solution points with large prediction error or excellent target value, part of the samples are dynamically selected for real simulation evaluation, and are included in the training set to update the Kriging surrogate model. Through cyclic iterative updating, the dynamic evolution of the Kriging surrogate model from global rough modeling to local high-precision fitting is realized. 3.The method of claim 1, wherein, The method comprises the following steps:

4. An emergency lane dynamic management and control system in a highway accident scene, characterized in that, The accident space-time influence prediction module is configured to: obtain accident report information, and use a deep learning network model to predict accident disposal time; based on the accident disposal time, using the traffic wave model to predict the queue length and dissipation time under the influence of the accident; the model construction module of section discretization is configured to: taking the queue length under the influence of the accident as the total length of the control section, and performing segmented discretization; taking the dissipation time under the influence of the accident as the control time length, and performing intervalization; a binary control strategy combining time window and space section is constructed; taking the weighted minimum of total travel time and total collision exposure time as the target, a target function is constructed; the solving module is configured to: using the Kriging surrogate model to solve the target function, and obtaining the dynamic control strategy of the emergency lane according to the binary control strategy; ​ Taking the queue length under the influence of the accident as the total length of the managed section, and performing segmentation and discretization; Taking the dissipation time under the influence of the accident as the management time length, and performing intervalization; the method comprises: In space, the total length of the controlled road section is segmented and discretized, and the total length of the controlled road section is divided into spatial segments with the same length The length of each segment is , and the length is within a certain range. Controlling the time length Intervalization, taking a fixed time as a time interval The opening and closing states of different emergency lane space segments are fixed and unchanged within the same time interval; Let be a 0-1 variable representing the open-close status of the i-th spatial segment emergency lane, be 1 if the i-th spatial segment emergency lane is open, be 0 if the i-th spatial segment emergency lane is closed.​​ In the same time interval , the control policy of all spatial segments is mathematically represented as: ; Controlling the time length According to time intervals Divided into Segments, the time length The control strategy of the emergency lane in the time length ; The control strategy is a two-dimensional 0-1 variable matrix in mathematical form, which describes the open and close states of each spatial segment of the emergency lane control section in different time intervals within the entire control duration. ​​​ Based on the accident handling time, the traffic wave model is used to predict the queue length and dissipation time under the influence of the accident; the method comprises: respectively calculating the gathering wave speed and the starting wave speed, and calculating the queue length and dissipation time under the influence of the accident according to the relationship between the gathering wave and the starting wave; The target function is expressed by the following formula: wherein, represents the TTS value of the vehicle at the tth cycle and the th time step; represents the TET value of the vehicle at the tth cycle and the th time step; represent the weights of the efficiency and safety indicators, respectively; represents the number of vehicles in the network; is a 0 / 1 variable indicating whether the vehicle is in the simulation area or not; represents the simulation interval time; represents the total simulation time; represents the TTC value of the vehicle at the tth cycle and the th time step; represents the threshold value of the TTC.​​​​​​​​ 5.A computer device, characterized in that, a processor adapted to execute a computer program; A computer readable storage medium having a computer program stored therein, wherein the computer program is executed by the processor to implement the steps of the emergency lane dynamic management and control method under the expressway accident scene according to any one of claims 1-3.

6. A computer-readable storage medium, characterized in that, The computer readable storage medium stores a computer program, and the computer program is adapted to be loaded and executed by the processor to implement the steps of the emergency lane dynamic management and control method under the expressway accident scene according to any one of claims 1-3.

7. A computer program product, characterised in that, The computer program product comprises a computer program, and the computer program is executed by the processor to implement the steps of the emergency lane dynamic management and control method under the expressway accident scene according to any one of claims 1-3.

Images