Expressway quasi-all-weather traffic control strategy generation method based on heterogeneous traffic flow simulation training

Abstract

The invention discloses an expressway quasi-all-weather traffic control strategy generation method based on heterogeneous traffic flow simulation training, and relates to the technical field of expressway traffic control. Splashing water drops are subjected to feature calibration to obtain the splashing height; simulating brake distances required by vehicles running at different spattering heights and at different vehicle speeds to obtain a brake distance table; then, vehicles running on the highway in rainy days in real time are monitored, the real-time splashing height and the real-time speed are matched with the braking distance table, and the most suitable limiting speed on the highway under the current condition can be obtained; therefore, the driving speed limit on the highway in rainy days is more reasonable, and the driving safety can be ensured more flexibly; the method is simple, effective, easy and practical.

Description

Technical Field

[0001] The invention belongs to the technical field of highway traffic control, and in particular is a method for generating a quasi-all-weather traffic control strategy for a highway based on heterogeneous traffic flow simulation training. Background Art

[0002] Currently, for the traffic control of highways, many control strategies or simulation models are disclosed in the prior art; generally, they are aimed at how to control the speed of vehicles during driving, and the timing of vehicles changing lanes or merging into traffic; of course, they also include some other contents;

[0003] However, for vehicles traveling on highways on rainy days, visibility and braking distance of vehicles will be affected. Visibility will be strongly blocked by water droplets splashed by the vehicle in front at a higher speed, and the braking distance of vehicles will also increase due to the slippery road surface caused by rain, thereby increasing the risk of driving. Therefore, it is necessary to adaptively modify the speed limit of vehicles on rainy days to protect the safe driving of vehicles and avoid difficulties in highway travel due to vehicle collisions or other accidents. Summary of the invention

[0004] The present invention aims to solve at least one of the technical problems existing in the prior art;

[0005] To this end, the present invention proposes a method for generating a quasi-all-weather traffic control strategy for a highway based on heterogeneous traffic flow simulation training, and the method specifically comprises the following steps:

[0006] Build a recognition model to identify the landmark features of vehicles driving on the highway in rainy days through a neural network. The landmark features are the scene of water droplets splashing when the vehicle is driving in rainy days.

[0007] The splashing water droplets are characterized to obtain the splashing height, and then the braking distance required for vehicles traveling at different speeds at different splashing heights is simulated to obtain a braking distance table, which includes splashing points obtained by splitting the splashing height, braking distances corresponding to different driving speeds, and marking braking distances exceeding a preset value as overspeeding distances;

[0008] Vehicles traveling on rainy highways in real time are monitored, and the real-time splash height and real-time speed are matched with the braking distance table. If the braking distance obtained at this time is the speeding distance, the driving speed corresponding to the maximum braking distance at the current splash height will be obtained, marked as the speed limit, and then the speed limit will be updated according to the change of splash height at the same real-time speed.

[0009] Furthermore, the recognition model is a neural network recognition model.

[0010] Furthermore, the recognition model uses a convolutional neural network to recognize image features.

[0011] Furthermore, the specific method of obtaining the braking distance table is as follows:

[0012] Firstly, the maximum splash height on the expressway in the past three years is divided into several splash points, and the largest splash point corresponds to the maximum splash height;

[0013] Test the braking distance at different speeds at each splash point, starting with the first splash point;

[0014] Then set the driving speed, starting from the base value, and obtain the braking distance at the driving speed;

[0015] Then, the driving speed points are continuously increased with the set speed increment until the driving speed points are increased to the speed upper limit value, which is the value preset by the administrator; each time a set speed increment is added, the system distance under the currently selected splash point is collected for the increased driving speed point;

[0016] Then, the next splash point is obtained, and the same processing is repeated for each splash point to obtain the braking distance of each splash point at different driving speed points;

[0017] The braking distance exceeding X1 is marked as the overspeed distance; a braking distance table is obtained, including the braking distances of all splash points corresponding to different driving speeds; X1 is a preset value.

[0018] Furthermore, the splash point is determined as follows:

[0019] The data of past highways are monitored to obtain the maximum splash height on the highway in the past three years. Then, starting from zero, the set height value is accumulated at each interval. Each time a set height value is accumulated, a splash point is marked until the splash point reaches the maximum splash height, and several splash points are obtained; each splash point corresponds to a different splash height.

[0020] Furthermore, the specific method for determining the speed limit is:

[0021] Analyze and identify the real-time video of the highway to obtain the splash height of water droplets splashed by the vehicle, and at the same time obtain the real-time speed of the corresponding vehicle;

[0022] The splash height and real-time speed at this time are matched with the braking distance table to obtain the braking distance at this time. When the braking distance at this time is the speeding distance, the driving speed corresponding to the maximum braking distance under the current splash height will be obtained and marked as the speed limit.

[0023] Furthermore, the speed limit is updated as follows:

[0024] When the real-time speed remains unchanged and the splash height does not undergo an integrated change, no processing is performed; the integrated change here means that the absolute value of the splash height change does not exceed X2, where X2 is a preset value;

[0025] It is necessary to re-acquire the real-time speed of the vehicles on the highway, and then repeat the steps of determining the speed limit to update the speed limit.

[0026] Furthermore, after the speed limit is determined, it is also necessary to control the speed limit of vehicles on the expressway according to the determined speed limit. The specific method of control is to adjust the real-time speed limit of the expressway in real time.

[0027] Compared with the prior art, the present invention has the following beneficial effects:

[0028] The present invention can calibrate the characteristics of splashed water droplets to obtain the splash height, and then simulate the braking distance required for vehicles traveling at different speeds at different splash heights to obtain a braking distance table; then, by monitoring vehicles traveling on the highway in rainy days in real time, the real-time splash height and real-time speed are matched with the braking distance table to obtain the most suitable speed limit on the highway under the current situation;

[0029] Therefore, the speed limit for driving on the highway in rainy days is more reasonable, so that driving safety can be guaranteed more flexibly; the present invention is simple, effective, and easy to use. DETAILED DESCRIPTION

[0030] The technical solution of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0031] The present application provides a method for generating a quasi-all-weather traffic control strategy for a highway based on heterogeneous traffic flow simulation training, the method specifically comprising the following steps:

[0032] Step 1: Measure the highest point of the water droplets splashed by the vehicle, and the corresponding splash height is marked as Fi, i=1,...,n; the splashed water droplet image is marked as the target object, and the target object is identified by the recognition model. The specific identification method is:

[0033] The target object is identified by using a neural network recognition model, which is an existing technology and may use CNN;

[0034] Collect images of rain splashed by cars on the highway;

[0035] Then the collected images are preprocessed, including denoising, contrast enhancement, brightness adjustment, etc., to improve image quality and reduce the complexity of subsequent processing;

[0036] Use convolutional neural network (CNN) to automatically extract image features; CNN can learn useful feature representations from images through multi-layer convolution and pooling operations;

[0037] Then choose a suitable neural network architecture, such as ResNet, VGG, Inception, etc. These architectures perform well in image classification tasks;

[0038] Train the model using a labeled dataset; the labeled data should include labels for splashing rain so that the model can learn to distinguish this phenomenon;

[0039] During the training process, cross-validation, regularization and other techniques can be used to prevent overfitting and optimize the performance of the model.

[0040] Then, a test set with known results is collected to verify the accuracy of the recognition model. If the accuracy exceeds the preset ratio B1, it means that the recognition model is available. Otherwise, the training data is collected again for retraining until the recognition model is available to obtain the recognition model. This is a prior art, so it will not be described in detail here.

[0041] Step 2: Perform braking simulation. The specific simulation method is as follows:

[0042] S1: First, the data of the past expressways are monitored to obtain the maximum splash height on the expressways in the past three years, and then starting from zero, the set height value is accumulated at each interval, and each accumulated set height value is marked as a splash point until the splash point reaches the maximum splash height, and a number of splash points are obtained; the splash points correspond to different splash heights;

[0043] S2: Test the braking distance at different speeds at each splash point, and first select the first splash point; the test here can be carried out under the same road conditions on a closed highway;

[0044] Then set the driving speed, starting from the basic value. The basic value is generally set to 60km / h. Of course, the administrator can also set it to other values ​​to obtain the braking distance under the driving speed. Here, multiple simulations are used and then the average value is obtained;

[0045] Then, the driving speed points are continuously increased with the set speed increment until the driving speed points are increased to the speed upper limit value, which is the value preset by the administrator, and can generally be set to 120km / h, and can of course be adjusted to 132km / h according to the situation; each time a set speed increment is added, the system distance under the currently selected splash point is collected for the increased driving speed point;

[0046] Then, the next splash point is obtained, and the above step S2 is repeated to perform the same processing on each splash point to obtain the braking distance of each splash point at different driving speed points;

[0047] The braking distance exceeding X1 is marked as the overspeed distance; a braking distance table is obtained, including the braking distances of all splash points corresponding to different driving speeds; X1 is a preset value;

[0048] Step 3: Analyze and identify the real-time video of the highway to obtain the splash height of the water droplets splashed by the vehicle, and at the same time obtain the real-time speed of the corresponding vehicle;

[0049] The splash height and real-time speed at this time are matched with the braking distance table to obtain the braking distance at this time. When the braking distance at this time is the speeding distance, the driving speed corresponding to the maximum braking distance under the current splash height will be obtained and marked as the speed limit;

[0050] If the real-time speed remains unchanged and the splash height does not undergo an integrated change, the integrated change here means that the absolute value of the splash height change does not exceed X2, where X2 is a preset value, and no processing is performed at this time;

[0051] Step 4: If it exceeds X2, it is necessary to re-acquire the real-time speed of the vehicle on the highway, and then repeat step 3 to determine the new speed limit;

[0052] Step 5: Control the speed limit of vehicles on the highway based on the determined speed limit, including synchronously updating the current vehicle speed limit with the help of various navigation software, or informing the car owner of the latest highway speed limit through other advertising screens, and informing them that speeding vehicles will be captured by cameras at the same time.

[0053] The above embodiments are only used to illustrate the technical method of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical method of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical method of the present invention.

Claims

1. A method for generating quasi-all-weather traffic control strategies for highways based on heterogeneous traffic flow simulation training, characterized in that: The method specifically comprises the following steps: Build a recognition model to identify the landmark features of vehicles driving on the highway in rainy days through a neural network. The landmark features are the scene of water droplets splashing when the vehicle is driving in rainy days. The splashing water droplets are characterized to obtain the splashing height, and then the braking distance required for vehicles traveling at different speeds at different splashing heights is simulated to obtain a braking distance table, which includes splashing points obtained by splitting the splashing height, braking distances corresponding to different driving speeds, and marking braking distances exceeding a preset value as overspeeding distances; Vehicles traveling on rainy highways in real time are monitored, and the real-time splash height and real-time speed are matched with the braking distance table. If the braking distance obtained at this time is the speeding distance, the driving speed corresponding to the maximum braking distance at the current splash height will be obtained, marked as the speed limit, and then the speed limit will be updated according to the change of splash height at the same real-time speed.

2. The method for generating a quasi-all-weather traffic control strategy for a highway based on heterogeneous traffic flow simulation training according to claim 1 is characterized in that: The recognition model is a neural network recognition model.

3. The method for generating a quasi-all-weather traffic control strategy for a highway based on heterogeneous traffic flow simulation training according to claim 1 is characterized in that: The recognition model uses convolutional neural network to recognize image features.

4. The method for generating a quasi-all-weather traffic control strategy for a highway based on heterogeneous traffic flow simulation training according to claim 1 is characterized in that: The specific method of obtaining the braking distance table is: Firstly, the maximum splash height on the expressway in the past three years is divided into several splash points, and the largest splash point corresponds to the maximum splash height; Test the braking distance at different speeds at each splash point, starting with the first splash point; Then set the driving speed, starting from the base value, and obtain the braking distance at the driving speed; Then, the driving speed points are continuously increased with the set speed increment until the driving speed points are increased to the speed upper limit value, which is the value preset by the administrator; each time a set speed increment is added, the system distance under the currently selected splash point is collected for the increased driving speed point; Then, the next splash point is obtained, and the same processing is repeated for each splash point to obtain the braking distance of each splash point at different driving speed points; The braking distance exceeding X1 is marked as the overspeed distance; a braking distance table is obtained, including the braking distances of all splash points corresponding to different driving speeds; X1 is a preset value.

5. The method for generating a quasi-all-weather traffic control strategy for a highway based on heterogeneous traffic flow simulation training according to claim 1 is characterized in that: The splash point is determined by: The data of past highways are monitored to obtain the maximum splash height on the highway in the past three years. Then, starting from zero, the set height value is accumulated at each interval. Each time a set height value is accumulated, a splash point is marked until the splash point reaches the maximum splash height, and several splash points are obtained; each splash point corresponds to a different splash height.

6. The method for generating a quasi-all-weather highway traffic control strategy based on heterogeneous traffic flow simulation training according to claim 1 is characterized in that: The specific method for determining the speed limit is: Analyze and identify the real-time video of the highway to obtain the splash height of water droplets splashed by the vehicle, and at the same time obtain the real-time speed of the corresponding vehicle; The splash height and real-time speed at this time are matched with the braking distance table to obtain the braking distance at this time. When the braking distance at this time is the speeding distance, the driving speed corresponding to the maximum braking distance under the current splash height will be obtained and marked as the speed limit.

7. The method for generating a quasi-all-weather highway traffic control strategy based on heterogeneous traffic flow simulation training according to claim 6 is characterized in that: The way to update the speed limit is: When the real-time speed remains unchanged and the splash height does not undergo an integrated change, no processing is performed; the integrated change here means that the absolute value of the splash height change does not exceed X2, where X2 is a preset value; It is necessary to re-acquire the real-time speed of the vehicles on the highway, and then repeat the steps of determining the speed limit to update the speed limit.

8. The method for generating a quasi-all-weather highway traffic control strategy based on heterogeneous traffic flow simulation training according to claim 1 is characterized in that: After determining the speed limit, it is also necessary to control the speed limit of vehicles on the highway based on the determined speed limit. The specific method of control is to adjust the real-time speed limit of the highway in real time.