A smart emergency lane opening system based on traffic flow prediction and its implementation method

Abstract

This invention discloses an intelligent emergency lane opening system based on traffic flow prediction and its implementation method. The system includes a data acquisition module, a data processing module, a central control module, a power supply module, and a display module. The method includes collecting and processing traffic flow over a period of time to construct a spatiotemporal feature matrix of traffic flow; constructing a CNN-GRU-Attention prediction model; optimizing the parameters of the CNN-GRU-Attention prediction model, including the weight matrix and bias vector, using the improved Big Cane Mouse algorithm IGCRA; obtaining traffic flow prediction results based on the optimized model; and determining whether to open the emergency lane based on the obtained traffic flow prediction results. This invention can accurately predict traffic flow, promptly identify and open emergency lanes before congestion occurs, provide additional passage space for vehicles, avoid excessive traffic concentration, thereby effectively alleviating traffic congestion, reducing vehicle waiting time, and improving overall traffic efficiency.

Description

Technical Field

[0001] This invention relates to the fields of traffic flow prediction and algorithm optimization, and in particular to an intelligent emergency lane opening system based on traffic flow prediction and its implementation method. Background Technology

[0002] With the acceleration of urbanization and the continuous increase in the number of motor vehicles, traffic congestion has become increasingly severe. Emergency lanes, as a key resource for ensuring smooth traffic flow and responding to emergencies, require increasingly scientific and effective management. However, traditional emergency lane management models have numerous drawbacks, greatly limiting the full realization of their potential.

[0003] Traditional traffic management models rely heavily on manual experience or fixed-time opening strategies, with a primary problem being poor dynamic adaptability. Urban traffic flow is constantly changing, with significant differences in traffic volume across different road sections and time periods. However, traditional models lack a real-time, comprehensive traffic flow monitoring system, making it impossible to adjust the use of emergency lanes in a timely manner based on traffic flow changes. This often results in the waste or over-occupancy of emergency lane resources, reducing traffic efficiency. Furthermore, insufficient accuracy in traffic flow prediction is another major challenge of traditional models. Without advanced prediction models, it is difficult to accurately predict short-term fluctuations in traffic flow and sudden congestion, preventing traffic management departments from planning and deploying emergency lane resources in advance. This often leads to delayed responses to sudden traffic events, exacerbating traffic congestion.

[0004] While some existing systems attempt to incorporate sensors to monitor traffic flow, they have failed to deeply integrate the spatiotemporal feature analysis of traffic data with deep learning prediction techniques. This results in inaccurate judgments regarding the timing of emergency lane opening, poor timeliness, and an inability to meet the practical needs of modern traffic management. Against this backdrop, the development of an intelligent emergency lane opening system based on traffic flow prediction is urgently needed. Summary of the Invention

[0005] Purpose of the invention: The purpose of this invention is to provide an intelligent emergency lane opening system based on traffic flow prediction and its implementation method.

[0006] Technical Solution: The intelligent emergency lane opening system based on traffic flow prediction described in this invention includes a data acquisition module, a data processing module, a central control module, a power supply module, and a display module. The data acquisition module detects traffic flow, inputs the detected traffic flow data into the data processing module, and predicts the traffic flow for a future period. The prediction results are then input into the central control module. The central control module determines whether the emergency lane needs to be opened based on the prediction results and displays the determination results through the display module.

[0007] Furthermore, the power supply module includes a solar array panel, a wind turbine generator, and a battery. The solar array panel is used to realize photovoltaic power generation, the wind turbine generator is used to realize wind power generation, and the battery is used to realize complementarity and mutual backup between different types of power sources.

[0008] The implementation method of the intelligent emergency lane opening system based on traffic flow prediction described in this invention includes the following steps:

[0009] Step 1: Collect and process traffic flow data over a period of time to construct a spatiotemporal feature matrix of traffic flow;

[0010] Step 2: Based on the traffic flow data collected in Step 1, construct a CNN-GRU-Attention prediction model;

[0011] Step 3: Based on the model obtained in Step 2, optimize the parameters of the CNN-GRU-Attention prediction model, including the weight matrix and bias vector, using the improved Big Cane Mouse algorithm IGCRA.

[0012] Step 4: Obtain traffic flow prediction results based on the optimized model from Step 3;

[0013] Step 5: Based on the traffic flow prediction results obtained in Step 4, determine whether it is necessary to open the emergency lane.

[0014] Furthermore, in step 1, the traffic flow data includes temporal and spatial features, and the spatiotemporal feature matrix is ​​as follows:

[0015]

[0016] in, The spatiotemporal feature matrix of traffic flow is composed of the feature vectors of each node in space. constitute; The number of nodes; For the first Nodes at each time point The time characteristics.

[0017] Furthermore, in step 2, the CNN-GRU-Attention prediction model is established as follows:

[0018] Step 2.1: Extract the temporal and spatial features of traffic flow data using the convolutional layers in a Convolutional Neural Network (CNN). Its output... It can be represented as:

[0019]

[0020] in, and They represent the first The weights and biases of each convolutional layer; * indicates a convolution operation; It is an activation function, usually a calibrated linear unit;

[0021] Step 2.2: The GRU neural network unit consists of two parts: an update gate and a reset gate. Its operating mechanism is as follows:

[0022]

[0023]

[0024]

[0025]

[0026] in, Traffic flow information input at the current moment; This is the hidden state from the previous moment; The hidden state to be passed on to the next moment; This is the candidate hidden state; , These are the reset door and the update door, respectively. It is the sigmoid activation function; As an activation function, it assigns the data to ; and These are the weights and bias matrices for each control gate, respectively;

[0027] Step 2.3: Introduce an attention mechanism based on the CNN-GRU model to redistribute weights and generate feature sequences, as shown in the following expression:

[0028]

[0029]

[0030] in, The number of samples; Let be the attention distribution function; For attention scoring functions, a scaled dot product model is often used; This is a weighted sum that uses an information selection mechanism to aggregate the input information.

[0031] Furthermore, in step 3, the weight matrix of the CNN-GRU-Attention model is optimized using the improved Big Cane Mouse algorithm. , , and bias constant , , The process is as follows:

[0032] Step 3.1: Initialize the population and parameters. The initial positions of the large cane rats in the search space are randomly assigned, and their population matrix is ​​as follows:

[0033]

[0034] in, It is a population of large sugarcane rats; For the first The first proposed by the large cane rat The values ​​of the problem variables; The number of samples; It is the number of decision variables;

[0035] Step 3.2: Substitute the randomly generated positions of the large cane rats into the objective function for evaluation. Each large cane rat represents a candidate solution to the optimization problem. The objective function formula is as follows:

[0036]

[0037] in, The number of samples; The actual value; This is a predicted value;

[0038] Step 3.3: Update the population members. Determine the new positions of the remaining mice in the search space based on the position of the dominant male mouse, i.e., update the optimal weight matrix. , , and bias constant , , The position formula is as follows:

[0039]

[0040]

[0041]

[0042] in, For the first The new location of the giant cane rat. For its first The value of the dimension; This is the value of the large cane rat's current position; The dominant male mouse position; The location of the randomly selected female mouse; This is the latest objective function value; This is the current objective function value; The objective function value for the dominant male rat; Random numbers defined within the problem space boundary; To simulate the effects of abundant food sources; This represents the current iteration number; This represents the maximum number of iterations. To simulate the coefficient of reduced food sources; The coefficient that encourages large cane rats to move to other abundant food sources within the breeding area; for Random numbers between;

[0043] Step 3.4: Update the search space by developing areas with abundant food sources based on the location of the female mouse, i.e., update the optimal weight matrix. , , and bias constant , , The position formula is as follows:

[0044] ;

[0045] Step 3.5: Determine whether the termination condition is met. If the termination condition is not met, return to step 3.3; if the constraint condition is met, proceed to step 3.6.

[0046] Step 3.6: Output the optimal weight matrix , , and bias constant , , .

[0047] Furthermore, in step 3.4, an adaptive weighting factor is introduced to improve the position of the newly generated particles, thereby enhancing the local optimization capability of IGCRA. The calculation formula is as follows:

[0048]

[0049] Where ω is the adaptive weighting factor.

[0050] Furthermore, in step 5, the decision-making model for determining whether the emergency lane needs to be opened is as follows:

[0051]

[0052] in, For the location of Traffic congestion index corresponding to the time period; For the location of Road time occupancy rate at any given moment; For the location of Traffic flow at any given moment; For the location of The time, location, and speed.

[0053] Furthermore, in step 5, the judgment model for determining whether the emergency lane needs to be opened is as follows:

[0054]

[0055] in, For the location of The traffic congestion level assessment value at any given time, where 1 represents smooth traffic and 2 represents congested traffic. At the location of The assessment value of the traffic condition level at any given time; The set number of tests.

[0056] Beneficial effects: Compared with the prior art, the present invention has the following significant advantages: The present invention utilizes CNN-GRU-

[0057] The Attention prediction model effectively processes time-series data and captures the dynamic characteristics of traffic flow changes over time, thereby improving the accuracy of traffic flow prediction. Compared to using only a single prediction method, this invention uses a combined prediction model, resulting in more accurate and reliable predictions. The prediction model is optimized using the GCRA optimization algorithm, and errors are further reduced by precisely controlling the model parameters. In addition, the position of newly generated particles is improved by introducing adaptive weight factors, which enhances the local optimization capability of the IGCRA algorithm. It can accurately predict traffic flow, promptly identify and open emergency lanes before congestion occurs, provide additional passage space for vehicles, avoid excessive concentration of traffic flow, and thus effectively alleviate traffic congestion, reduce vehicle waiting time, and improve overall traffic efficiency. Attached Figure Description

[0058] Figure 1 A framework diagram of an intelligent emergency lane opening system based on traffic flow prediction;

[0059] Figure 2 Flowchart for improving the Big Cane Mouse algorithm. Detailed Implementation

[0060] The technical solution of the present invention will be further described below with reference to the accompanying drawings.

[0061] like Figure 1As shown, the intelligent emergency lane opening system based on traffic flow prediction of the present invention includes a data acquisition module, a data processing module, a central control module, a power supply module, and a display screen module.

[0062] The data acquisition module includes a loop detector, which is used to detect traffic flow, that is, the number of vehicles passing through a certain section of road per unit time.

[0063] The data processing module includes a prediction model and an algorithm. The prediction model predicts traffic flow over a future period based on traffic flow data measured by loop detectors; the algorithm is used to optimize the model parameters.

[0064] The central control module determines whether the emergency lane needs to be opened based on the forecast results;

[0065] The power supply module includes a solar array panel, a wind turbine, and a battery. The solar array panel is used to generate photovoltaic power, the wind turbine is used to generate wind power, and the battery is used to achieve complementarity and mutual backup between different types of power sources. The display module shows the judgment results of the central control module, providing traffic police and drivers with information on the opening status of the emergency lane.

[0066] The implementation method of the intelligent emergency lane opening system based on traffic flow prediction described in this invention includes the following steps: Step 1: Collect and process traffic flow over a period of time, and construct a spatiotemporal feature matrix of traffic flow;

[0067] Traffic flow data includes temporal and spatial features, and the spatiotemporal feature matrix is ​​as follows:

[0068]

[0069] in, The spatiotemporal feature matrix of traffic flow is composed of the feature vectors of each node in space. constitute; The number of nodes; For the first Nodes at each time point The time characteristics.

[0070] Step 2: Based on the traffic flow data collected in Step 1, construct a CNN-GRU-Attention prediction model;

[0071] like Figure 2 As shown, the process of establishing the CNN-GRU-Attention prediction model is as follows:

[0072] Step 2.1: Extract the temporal and spatial features of traffic flow data using the convolutional layers in a Convolutional Neural Network (CNN). Its output... It can be represented as:

[0073]

[0074] in, and They represent the first The weights and biases of each convolutional layer; * indicates a convolution operation; It is an activation function, usually a calibrated linear unit;

[0075] Step 2.2: The GRU neural network unit consists of two parts: an update gate and a reset gate. Its operating mechanism is as follows:

[0076]

[0077]

[0078]

[0079]

[0080] in, Traffic flow information input at the current moment; This is the hidden state from the previous moment; The hidden state to be passed on to the next moment; This is the candidate hidden state; , These are the reset door and the update door, respectively. It is the sigmoid activation function; As an activation function, it assigns the data to ; and These are the weights and bias matrices for each control gate, respectively;

[0081] Step 2.3: Introduce an attention mechanism based on the CNN-GRU model to redistribute weights and generate feature sequences, as shown in the following expression:

[0082]

[0083]

[0084] in, For a specific moment; Let be the attention distribution function; For attention scoring functions, a scaled dot product model is often used; This is a weighted sum that uses an information selection mechanism to aggregate the input information.

[0085] Step 3: Based on the model obtained in Step 2, optimize the parameters of the CNN-GRU-Attention prediction model, including the weight matrix and bias vector, using the improved Big Cane Mouse algorithm IGCRA.

[0086] Optimizing the weight matrix of the CNN-GRU-Attention model using an improved Big Cane Mouse algorithm , , and bias constant , , The process is as follows:

[0087] Step 3.1: Initialize the population and parameters. The initial positions of the large cane rats in the search space are randomly assigned, and their population matrix is ​​as follows:

[0088]

[0089] in, It is a population of large sugarcane rats; For the first The first proposed by the large cane rat The values ​​of the problem variables; The total number of zebra population members; It is the number of decision variables;

[0090] Step 3.2: Substitute the randomly generated positions of the large cane rats into the objective function for evaluation. Each large cane rat represents a candidate solution to the optimization problem. The objective function formula is as follows:

[0091]

[0092] in, The number of samples; The actual value; This is a predicted value;

[0093] Step 3.3: Update the population members. Determine the new positions of the remaining mice in the search space based on the position of the dominant male mouse, i.e., update the optimal weight matrix. , , and bias constant , , The position formula is as follows:

[0094]

[0095]

[0096]

[0097] in, For the first The new location of the giant cane rat. For its first The value of the dimension; This is the value of the large cane rat's current position; The dominant male mouse position; The location of the randomly selected female mouse; This is the latest objective function value; This is the current objective function value; The objective function value for the dominant male rat; Random numbers defined within the problem space boundary; To simulate the effects of abundant food sources; This represents the current iteration number; This represents the maximum number of iterations. To simulate the coefficient of reduced food sources; The coefficient that encourages large cane rats to move to other abundant food sources within the breeding area; for Random numbers between;

[0098] Step 3.4: Update the search space by developing areas with abundant food sources based on the location of the female mouse, i.e., update the optimal weight matrix. , , and bias constant , , The position formula is as follows:

[0099] ;

[0100] Step 3.5: Determine whether the termination condition is met. If the termination condition is not met, return to step 3.3; if the constraint condition is met, proceed to step 3.6.

[0101] Step 3.6: Output the optimal weight matrix , , and bias constant , , .

[0102] The improvement in step 3.4 involves introducing an adaptive weighting factor to improve the position of the newly generated particles, thereby enhancing the local optimization capability of IGCRA. The calculation formula is as follows:

[0103]

[0104] Where ω is the adaptive weighting factor.

[0105] Step 4: Obtain traffic flow prediction results based on the optimized model from Step 3;

[0106] Step 5: Based on the traffic flow prediction results obtained in Step 4, determine whether it is necessary to open the emergency lane.

[0107] The decision-making model for determining whether to open the emergency lane is as follows:

[0108]

[0109] in, For the location of Traffic congestion index corresponding to the time period; For the location of Road time occupancy rate at any given moment; For the location of Traffic flow at any given moment; For the location of The time, location, and speed.

[0110] The model for determining whether an emergency lane needs to be opened is as follows:

[0111]

[0112] in, For the location of The traffic congestion level assessment value at any given time, where 1 represents smooth traffic and 2 represents congested traffic. At the location of The assessment value of the traffic condition level at any given time; The set number of tests.

Claims

1. A method for implementing an intelligent emergency lane opening system based on traffic flow prediction, characterized in that, Includes the following steps: Step 1: Collect and process traffic flow data over a period of time to construct a spatiotemporal feature matrix of traffic flow; Step 2: Based on the traffic flow data collected in Step 1, construct a CNN-GRU-Attention prediction model; Step 3: Based on the model obtained in Step 2, optimize the CNN-GRU algorithm using the improved Big Cane Mouse algorithm IGCRA. Attention predicts model parameters, including the weight matrix and bias vector; Step 4: Obtain traffic flow prediction results based on the optimized model from Step 3; Step 5: Based on the traffic flow prediction results obtained in Step 4, determine whether it is necessary to open the emergency lane; In step 2, the CNN-GRU-Attention prediction model is established as follows: Step 2.1: Extract the temporal and spatial features of traffic flow data using the convolutional layers in a Convolutional Neural Network (CNN). Its output... Represented as: ； in, and They represent the first The weights and biases of each convolutional layer; * indicates a convolution operation; It is an activation function, usually a calibrated linear unit; Step 2.2: The GRU neural network unit consists of two parts: an update gate and a reset gate. Its operating mechanism is as follows: ； ； ； ； in, Traffic flow information input at the current moment; This is the hidden state from the previous moment; The hidden state to be passed on to the next moment; This is the candidate hidden state; , These are the reset door and the update door, respectively. It is the sigmoid activation function; As an activation function, it assigns the data to ; and These are the weights and bias matrices for each control gate; Step 2.3: Introduce an attention mechanism based on the CNN-GRU model to redistribute weights and generate feature sequences, as shown in the following expression: ； ； in, The number of samples; Let the attention distribution function be used. For attention scoring functions, a scaled dot product model is often used; A weighted sum that aggregates input information using an information selection mechanism; In step 3, the weight matrix of the CNN-GRU-Attention model is optimized using the improved Big Cane Mouse algorithm. , , and bias constant , , The process is as follows: Step 3.1: Initialize the population and parameters. The initial positions of the large cane rats in the search space are randomly assigned, and their population matrix is ​​as follows: ； in, It is a population of large sugarcane rats; For the first The first proposed by the large cane rat The values ​​of the problem variables; The number of samples; It is the number of decision variables; Step 3.2: Substitute the randomly generated positions of the large cane rats into the objective function for evaluation. Each large cane rat represents a candidate solution to the optimization problem. The objective function formula is as follows: ； in, The number of samples; The actual value; This is a predicted value; Step 3.3: Update the population members. Determine the new positions of the remaining mice in the search space based on the position of the dominant male mouse, i.e., update the optimal weight matrix. , , and bias constant , , The position formula is as follows: ； ； ； in, For the first The new location of the giant cane rat. For its first The value of the dimension; This is the value of the large cane rat's current position; The dominant male mouse position; The location of the randomly selected female mouse; This is the latest objective function value; This is the current objective function value; The objective function value for the dominant male rat; Random numbers defined within the problem space boundary; To simulate the effects of abundant food sources; This represents the current iteration number; This represents the maximum number of iterations. To simulate the coefficient of reduced food sources; The coefficient that encourages large cane rats to move to other abundant food sources within the breeding area; for Random numbers between; Step 3.4: Update the search space by developing areas with abundant food sources based on the location of the female mouse, i.e., update the optimal weight matrix. , , and bias constant , , The position formula is as follows: ； Step 3.5: Determine whether the termination condition is met. If the termination condition is not met, return to step 3.3; if the constraint condition is met, proceed to step 3.

6. Step 3.6: Output the optimal weight matrix , , and bias constant , , ; In step 3.4, an adaptive weighting factor is introduced to improve the position of the newly generated particles, thereby enhancing the local optimization capability of IGCRA. The calculation formula is as follows: ； Where ω is the adaptive weighting factor.

2. The implementation method of the intelligent emergency lane opening system based on traffic flow prediction according to claim 1, characterized in that, In step 1, the traffic flow data includes temporal and spatial features, and the spatiotemporal feature matrix is ​​as follows: ； in, The spatiotemporal feature matrix of traffic flow is composed of the feature vectors of each node in space. constitute; The number of nodes; For the first Nodes at each time point The time characteristics.

3. The implementation method of the intelligent emergency lane opening system based on traffic flow prediction according to claim 1, characterized in that, In step 5, the decision-making model for determining whether the emergency lane needs to be opened is as follows: ； in, For the location of Traffic congestion index corresponding to the time period; For the location of Road time occupancy rate at any given moment; For the location of Traffic flow at any given moment; For the location of The time, location, and speed.

4. The implementation method of the intelligent emergency lane opening system based on traffic flow prediction according to claim 1, characterized in that, In step 5, the judgment model for determining whether the emergency lane needs to be opened is as follows: ； in, For the location of The traffic congestion level assessment value at any given time, where 1 represents smooth traffic and 2 represents congested traffic. At the location of The assessment value of the traffic condition level at any given time; The set number of tests.

5. An intelligent emergency lane opening system based on traffic flow prediction, applied to the method described in claims 1-4, characterized in that, It includes a data acquisition module, a data processing module, a central control module, a power supply module, and a display module. The data acquisition module detects traffic flow, inputs the detected traffic flow data into the data processing module, and predicts the traffic flow for a future period. The prediction results are then input into the central control module, which determines whether the emergency lane needs to be opened based on the prediction results and displays the results on the display module.

6. The intelligent emergency lane opening system based on traffic flow prediction according to claim 5, characterized in that, The power supply module includes a solar array panel, a wind turbine, and a battery. The solar array panel is used to realize photovoltaic power generation, the wind turbine is used to realize wind power generation, and the battery is used to realize complementarity and mutual backup between different types of power sources.

Images