Anti-collision method and system

Abstract

The invention provides an anti-collision method and system, and relates to the technical field of underground engineering machinery, and the method comprises the steps: obtaining a multi-view visual image of an engineering vehicle in a preset range of a driving direction; performing feature extraction on the multi-view visual image, and determining obstacles existing in a preset range; detecting an obstacle in the multi-view visual image to obtain motion information of the obstacle relative to the engineering vehicle; according to the motion information of the obstacle and driving information of the engineering vehicle, obtaining a braking strategy of the engineering vehicle; and assisting the engineering vehicle in braking according to the braking strategy. According to the method, braking assistance is carried out on the engineering vehicle based on the multi-view visual image, it is ensured that the vehicle can respond to potential collision threats timely and accurately, the braking effect of the engineering machinery vehicle is improved, collision accidents are reduced, and the safety of engineering operation is guaranteed.

Description

Technical Field

[0001] The present invention relates to the technical field of underground engineering machinery, and in particular to an anti-collision method and system. Background Art

[0002] In explosion-proof environments such as underground, due to weak light and narrow roads, it is difficult for drivers to perceive obstacles in the direction of travel when driving engineering vehicles underground, which can easily cause unexpected injuries or losses. Therefore, in order to prevent collisions, monocular cameras are usually used to perceive the front of the engineering machinery.

[0003] In related technologies, due to insufficient light and heavy dust in explosion-proof environments such as underground, monocular cameras are prone to errors in their perception of obstacles, and are unable to provide timely reminders and take corresponding braking measures, which in turn leads to braking actions when there are no obstacles or no braking actions when there are obstacles, posing a great threat to the safety of equipment and personnel. Summary of the invention

[0004] The problem solved by the invention is how to improve the braking effect of engineering machinery.

[0005] In order to solve the above problems, the present invention provides an anti-collision method and system.

[0006] In a first aspect, the present invention provides an anti-collision method, comprising:

[0007] Acquire multi-eye visual images of the engineering vehicle within a preset range of the driving direction;

[0008] Extracting features from the multi-view image to determine obstacles within the preset range;

[0009] Detecting the obstacle in the multi-view image to obtain movement information of the obstacle relative to the engineering vehicle;

[0010] Obtaining a braking strategy for the engineering vehicle based on the movement information of the obstacle and the driving information of the engineering vehicle;

[0011] The engineering vehicle is assisted in braking according to the braking strategy.

[0012] Optionally, the multi-eye visual image includes a plurality of monocular visual images, and extracting features from the multi-eye visual image to determine obstacles existing within the preset range includes:

[0013] Input each of the monocular visual images into a preset target detection network;

[0014] Extracting local features of the monocular visual image through multiple convolutional layers of the preset target detection network in sequence;

[0015] By performing feature fusion on the local features of each convolutional layer, a multi-scale feature map is obtained;

[0016] Predicting based on the multi-scale feature map to determine all object labels in the monocular visual image;

[0017] The obstacle existing within the preset range is determined according to the object label.

[0018] Optionally, detecting the obstacle in the multi-view image to obtain movement information of the obstacle relative to the engineering vehicle includes:

[0019] Performing a disparity analysis on the obstacle in each monocular vision image of the multi-eye vision image to determine the distance between the obstacle and the engineering vehicle;

[0020] Determine the moving speed of the obstacle based on the change of the distance within a preset time period;

[0021] The distance between the obstacle and the engineering vehicle and the moving speed are used as the motion information of the obstacle.

[0022] Optionally, performing a disparity analysis on the obstacle in each monocular visual image of the multi-eye visual image to determine the distance between the obstacle and the engineering vehicle includes:

[0023] Matching the obstacle in each monocular vision image in the multi-eye vision image to obtain the same feature point of the same obstacle in each monocular vision image;

[0024] Determine, according to the position of the feature point in each monocular vision image, a lateral distance difference between the same feature points in each monocular vision image;

[0025] The depth of the feature point in the three-dimensional space is obtained by calculating the lateral distance difference between the same feature points in each monocular vision image in combination with a preset baseline length;

[0026] The depth of the feature point is used as the distance between the obstacle and the engineering vehicle.

[0027] Optionally, determining the moving speed of the obstacle in combination with the change of the distance within a preset time period includes:

[0028] Determining the relative speed between the obstacle and the engineering vehicle according to the change value of the distance between the obstacle and the engineering vehicle within the preset time period;

[0029] The moving speed of the obstacle is determined according to the relative speed combined with the speed of the engineering vehicle in the preset time period.

[0030] Optionally, obtaining a braking strategy of the engineering vehicle based on the movement information of the obstacle and the driving information of the engineering vehicle includes:

[0031] Predicting the collision risk level of the engineering vehicle within a preset future time period according to the distance between the obstacle and the engineering vehicle and the moving speed;

[0032] The braking strategy of the engineering vehicle under the predicted collision risk level is determined through graded braking strategies.

[0033] Optionally, the collision risk level includes multiple levels;

[0034] The predicting of the collision risk level of the engineering vehicle within a preset future time period according to the distance between the obstacle and the engineering vehicle and the moving speed includes:

[0035] Determining a braking distance and a braking time of the engineering vehicle according to the current speed of the engineering vehicle;

[0036] Determine a safe distance range between the engineering vehicle and the obstacle and a dangerous distance range corresponding to each level according to the braking distance, the braking time and the moving speed of the obstacle;

[0037] When the distance between the obstacle and the engineering vehicle is within the safety distance, the level is determined to be level 0;

[0038] When the distance between the obstacle and the engineering vehicle is within the dangerous distance range, the level corresponding to the dangerous distance range is used as the collision risk level of the engineering vehicle.

[0039] Optionally, determining the braking strategy of the engineering vehicle at the predicted collision risk level by using a graded braking strategy includes:

[0040] Determining a maximum speed limit and an expected speed of the engineering vehicle based on the predicted collision risk level of the engineering vehicle and in combination with the current speed of the engineering vehicle;

[0041] Determining the braking mode, braking acceleration, braking opening and braking time of the engineering vehicle according to the current vehicle speed and the expected vehicle speed;

[0042] The maximum speed limit, the braking acceleration, the braking method, the braking opening and the braking time are used as the braking strategy.

[0043] Optionally, assisting the engineering vehicle in braking according to the braking strategy includes:

[0044] Adjusting the speed limit of the engineering vehicle to the maximum speed limit;

[0045] When the current speed of the engineering vehicle is greater than the maximum speed limit, the engineering vehicle is assisted in braking according to the braking acceleration, the braking opening and the braking time in accordance with the braking method.

[0046] In a second aspect, the present invention provides an anti-collision system, characterized in that it includes:

[0047] A multi-eye perception unit, used to obtain multi-eye visual images of the engineering vehicle within a preset range of the driving direction;

[0048] A feature extraction unit, used to extract features from the multi-view image to determine obstacles existing within the preset range;

[0049] An image analysis unit, used to detect the obstacle in the multi-view image and obtain movement information of the obstacle relative to the engineering vehicle;

[0050] A strategy generating unit, configured to obtain a braking strategy for the engineering vehicle based on the movement information of the obstacle and the driving information of the engineering vehicle;

[0051] The auxiliary braking unit is used to assist the engineering vehicle in braking according to the braking strategy.

[0052] The anti-collision method and system of the present invention can simultaneously monitor the front environment from multiple perspectives by acquiring multi-eye visual images of the engineering vehicle within a preset range of the driving direction, thereby increasing the breadth and depth of perception and improving the accuracy and reliability of obstacle detection. Next, feature extraction is performed on the multi-eye visual image to determine the obstacles existing within the preset range, and the obstacles are accurately identified and classified through image processing technology to provide key information for auxiliary braking. Then, the obstacles in the image are detected, the motion information of the obstacles relative to the engineering vehicle is obtained, the dynamic characteristics of the obstacles are evaluated, and the potential collision risk is predicted. Subsequently, the braking strategy of the engineering vehicle is formulated in combination with the motion information of the obstacle and the driving information of the engineering vehicle, and the vehicle state and environmental information are integrated to determine the most appropriate braking measures. Finally, the engineering vehicle is assisted in braking according to the formulated braking strategy, realizing closed-loop control from perception to execution, ensuring that the vehicle can respond to potential collision threats in a timely and accurate manner, improving the braking effect of engineering machinery vehicles, reducing the occurrence of collision accidents, and ensuring the safety of engineering operations. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 This is a flow chart of an anti-collision method in one embodiment of the present invention;

[0054] Figure 2 A schematic diagram of the structural relationship between a binocular camera and an engineering vehicle in another embodiment of the present invention;

[0055] Figure 3 A schematic diagram of a control flow of an anti-collision system in another embodiment of the present invention;

[0056] Figure 4 A schematic diagram of the distance calculation between an obstacle and an engineering vehicle in another embodiment of the present invention;

[0057] Figure 5 A schematic diagram of the classification of collision risk levels between obstacles and engineering vehicles in another embodiment of the present invention;

[0058] Figure 6 FIG. 4 is a schematic structural diagram of an anti-collision system in another embodiment of the present invention. DETAILED DESCRIPTION

[0059] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific embodiments of the present invention are described in detail below in conjunction with the accompanying drawings. Although certain embodiments of the present invention are shown in the accompanying drawings, it should be understood that the present invention can be implemented in various forms and should not be interpreted as being limited to the embodiments described herein. On the contrary, these embodiments are provided to provide a more thorough and complete understanding of the present invention. It should be understood that the drawings and embodiments of the present invention are only for exemplary purposes and are not intended to limit the scope of protection of the present invention.

[0060] It should be understood that the various steps described in the method embodiments of the present invention may be performed in different orders and / or in parallel. In addition, the method embodiments may include additional steps and / or omit the steps shown. The scope of the present invention is not limited in this respect.

[0061] The term "including" and its variations used in this document are open inclusions, that is, "including but not limited to"; the term "based on" means "based at least in part on"; the term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one other embodiment"; the term "some embodiments" means "at least some embodiments"; the term "optionally" means "optional embodiments". The relevant definitions of other terms will be given in the following description. It should be noted that the concepts of "first", "second", etc. mentioned in the present invention are only used to distinguish different devices, modules or units, and are not used to limit the order or interdependence of the functions performed by these devices, modules or units.

[0062] It should be noted that the modifications of "one" and "plurality" mentioned in the present invention are illustrative rather than restrictive, and those skilled in the art should understand that, unless otherwise clearly indicated in the context, it should be understood as "one or more".

[0063] The names of the messages or information exchanged between multiple devices in the embodiments of the present invention are only used for illustrative purposes, and are not used to limit the scope of these messages or information.

[0064] In view of the problems existing in the above-mentioned related technologies, this embodiment provides an anti-collision method and system.

[0065] Combination Figure 1 As shown, an anti-collision method provided by an embodiment of the present invention includes:

[0066] Acquire multi-eye visual images of the engineering vehicle within a preset range in the driving direction.

[0067] Specifically, a multi-camera system is used to capture multi-view visual images of the environment in front of the engineering vehicle. By acquiring images from different angles through a multi-camera, it can provide more comprehensive field of view coverage and depth perception than a single camera. This multi-view data acquisition method enhances the system's ability to understand the environment, especially in complex underground environments, and can more accurately identify and locate obstacles. Figure 2 As shown, in a preferred embodiment of the present invention, the multi-eye camera adopts a binocular camera, and the binocular camera is set on the front outer side of the engineering vehicle. The advantage of binocular ranging is high accuracy. It directly uses the principle of disparity map to measure distance directly, without maintaining a sample database, and the sorted ranging accuracy is high.

[0068] Feature extraction is performed on the multi-view visual image to determine obstacles existing within the preset range.

[0069] Specifically, after acquiring multi-view images, computer vision technology is used to extract features from the images to identify and determine obstacles within a preset range. Deep learning algorithms, such as convolutional neural networks, are usually required to analyze image content, extract key features, and classify and identify obstacles. The accuracy of feature extraction directly affects the reliability of obstacle detection.

[0070] The obstacle in the multi-view image is detected to obtain movement information of the obstacle relative to the engineering vehicle.

[0071] Specifically, after successfully identifying obstacles, the system needs to further analyze the motion state of these obstacles relative to the engineering vehicle, including calculating the speed, direction and possible trajectory of the obstacles. By analyzing the motion information of obstacles, the system can predict future positions and behaviors, which is crucial for assessing collision risks and formulating corresponding braking strategies.

[0072] The braking strategy of the engineering vehicle is obtained according to the movement information of the obstacle and the driving information of the engineering vehicle.

[0073] Specifically, by combining the movement information of obstacles and the current driving status of the engineering vehicle, such as the current vehicle speed, the system can calculate the optimal braking strategy to avoid collision and determine what braking measures to take in specific situations, including early warning, deceleration, emergency braking, etc., to ensure the safety of vehicles and personnel.

[0074] The engineering vehicle is assisted in braking according to the braking strategy.

[0075] Specifically, according to the braking strategy calculated in the above steps, the engineering vehicle is assisted to perform actual braking operations and the braking system of the vehicle, such as hydraulic braking or electronic braking system, is assisted to achieve a predetermined deceleration or parking. In a preferred embodiment of the present invention, the system sends the braking strategy to the braking system of the vehicle via the CAN bus for execution. Figure 3As shown, in a preferred embodiment of the present invention, a 24V DC power supply is provided to the auxiliary driving controller through the vehicle-mounted power distribution system to ensure its normal operation. The auxiliary driving controller is connected to the switch signal conversion isolation module and the intrinsically safe power supply isolation module through the digital input interface. These modules are responsible for receiving the signal from the multi-eye visual intelligent sensor and converting it into a format suitable for processing by the vehicle-mounted controller. The multi-eye visual intelligent sensor is connected to the Ethernet conversion isolation module through Ethernet, which converts the signal into a format suitable for the vehicle network and communicates with the auxiliary driving controller through Ethernet. The auxiliary driving controller performs real-time analysis and processing based on the data provided by the multi-eye visual intelligent sensor, such as the distance, speed and type of the obstacle, to assess the risk of collision. If a potential collision is detected, the auxiliary driving controller will communicate with the vehicle-mounted controller through the CAN bus and send a braking command. After receiving the command, the vehicle-mounted controller will execute the corresponding braking strategy, such as adjusting the vehicle speed, starting emergency braking, etc., to avoid or mitigate the collision. In addition, the auxiliary driving controller can also control the function switch and the override switch through the digital output interface to achieve auxiliary control of other functions of the vehicle.

[0076] The anti-collision method and system of the present invention can simultaneously monitor the front environment from multiple perspectives by acquiring multi-eye visual images of the engineering vehicle within a preset range of the driving direction, thereby increasing the breadth and depth of perception and improving the accuracy and reliability of obstacle detection. Next, feature extraction is performed on the multi-eye visual image to determine the obstacles existing within the preset range, and the obstacles are accurately identified and classified through image processing technology to provide key information for auxiliary braking. Then, the obstacles in the image are detected, the motion information of the obstacles relative to the engineering vehicle is obtained, the dynamic characteristics of the obstacles are evaluated, and the potential collision risk is predicted. Subsequently, the braking strategy of the engineering vehicle is formulated in combination with the motion information of the obstacle and the driving information of the engineering vehicle, and the vehicle state and environmental information are integrated to determine the most appropriate braking measures. Finally, the engineering vehicle is assisted in braking according to the formulated braking strategy, realizing closed-loop control from perception to execution, ensuring that the vehicle can respond to potential collision threats in a timely and accurate manner, improving the braking effect of engineering machinery vehicles, reducing the occurrence of collision accidents, and ensuring the safety of engineering operations.

[0077] Optionally, the multi-eye visual image includes a plurality of monocular visual images, and extracting features from the multi-eye visual image to determine obstacles existing within the preset range includes:

[0078] Input each of the monocular visual images into a preset target detection network;

[0079] Extracting local features of the monocular visual image through multiple convolutional layers of the preset target detection network in sequence;

[0080] By performing feature fusion on the local features of each convolutional layer, a multi-scale feature map is obtained;

[0081] Predicting based on the multi-scale feature map to determine all object labels in the monocular visual image;

[0082] The obstacle existing within the preset range is determined according to the object label.

[0083] Specifically, the multi-eye visual image is composed of multiple monocular visual images, each of which is input into a preset target detection network. Through the multiple convolutional layers of the network, the system can extract local features of each monocular visual image, which reflect the basic patterns and edge information in the image. Subsequently, the system forms a multi-scale feature map by fusing the local features extracted by each convolutional layer, and can integrate features at different levels through a feature pyramid network to provide a more comprehensive object representation. Based on these multi-scale feature maps, predictions are made to determine all object labels in each monocular visual image, so that obstacles such as vehicles, pedestrians, and roads in the image can be identified. Finally, based on these object labels, the system determines the obstacles within a preset range. In a preferred embodiment of the present invention, the preset target detection network can be YoloV8, which is used for target detection, and the semantic segmentation effect is improved through a global convolutional network to achieve the classification and recognition of large-scale obstacles. Before application, YoloV8 needs to be trained. Generally, visual sensors are used to collect a large amount of video to establish video and image libraries under various driving environments and conditions. Then, based on the above video and image libraries, the end-to-end convolutional network is extended to semantic segmentation. Finally, the neural network is optimized to improve the scope of obstacle classification and recognition, thereby realizing the recognition of a wide range of obstacles.

[0084] In the embodiments of the present invention, the object detection network accurately identifies and locates obstacles from multiple perspectives, enabling it to maintain high-accuracy detection performance in various environmental conditions, such as low light or dusty environments. In addition, by fusing features at different levels, the system can simultaneously identify obstacles at close range and far range, expanding the detection range.

[0085] Optionally, detecting the obstacle in the multi-view image to obtain movement information of the obstacle relative to the engineering vehicle includes:

[0086] Performing a disparity analysis on the obstacle in each monocular vision image of the multi-eye vision image to determine the distance between the obstacle and the engineering vehicle;

[0087] Determine the moving speed of the obstacle based on the change of the distance within a preset time period;

[0088] The distance between the obstacle and the engineering vehicle and the moving speed are used as the motion information of the obstacle.

[0089] Specifically, firstly, a disparity analysis is performed on each monocular visual image in the multi-eye visual image, which is achieved by comparing the feature points of the same obstacle in the images captured by different cameras. In a preferred embodiment of the present invention, the multi-eye visual image is collected by a binocular camera, so there are two monocular visual images. The obstacles extracted from the monocular visual images of the left and right cameras are used to calculate the disparity between the images, thereby inferring the depth information of the obstacle, that is, the distance between the obstacle and the engineering vehicle. Then, the system determines the moving speed of the obstacle in combination with the distance change within a preset time period. Finally, the distance and moving speed between the obstacle and the engineering vehicle are integrated to form the motion information of the obstacle, which is convenient for hierarchical intelligent active warning brake control, which requires an accurate grasp of the motion state of the obstacle.

[0090] In the embodiment of the present invention, by detecting obstacles in the multi-view image, the system can accurately obtain the three-dimensional motion information of the obstacle relative to the engineering vehicle, including the distance and speed, so as to update the future position and trajectory of the obstacle in real time. In this way, the anti-collision system can more effectively assess the possibility of collision and respond in time, such as issuing a warning or performing automatic braking, thereby improving the safety of the engineering vehicle and its ability to avoid accidents.

[0091] Optionally, performing a disparity analysis on the obstacle in each monocular visual image of the multi-eye visual image to determine the distance between the obstacle and the engineering vehicle includes:

[0092] Matching the obstacle in each monocular vision image in the multi-eye vision image to obtain the same feature point of the same obstacle in each monocular vision image;

[0093] Determine, according to the position of the feature point in each monocular vision image, a lateral distance difference between the same feature points in each monocular vision image;

[0094] The depth of the feature point in the three-dimensional space is obtained by calculating the lateral distance difference between the same feature points in each monocular vision image in combination with a preset baseline length;

[0095] The depth of the feature point is used as the distance between the obstacle and the engineering vehicle.

[0096] Specifically, obstacle feature points are first matched for each monocular vision image in the multi-eye vision image. Image processing technology is generally used to identify and match the same feature points of the same obstacle in different images, which is a key step in stereo vision. Then, by measuring the position of the same feature point in different images, the lateral distance difference between the feature points is determined according to the position of the same feature point in each monocular vision image. Then, according to the lateral distance difference and the preset baseline length, that is, the distance between each lens of the multi-eye camera, the principle of similar triangles is used to calculate and obtain the depth information of the feature point in three-dimensional space, that is, the distance between the obstacle and the engineering vehicle. In a preferred embodiment of the present invention, AAnet can be used for stereo matching. AAnet is a deep learning model for stereo matching tasks, which realizes fast and accurate matching of feature points. First, AAnet receives each monocular vision image as input, and then extracts the features of the image through a convolutional neural network. Then, an adaptive aggregation module is used to integrate these features to better capture disparity information. Finally, AAnet uses a discrete belief propagation module to infer the depth map, thereby realizing the stereo matching task. It is worth noting that AAnet consists of two modules, including the same-scale aggregation module and the cross-scale aggregation module. AANet can be used to replace the 3D convolution based on the matching cost volume, speeding up the inference speed while maintaining a high accuracy.

[0097] In another embodiment of the present invention, the multi-eye camera is a binocular camera, so there are images from the left and right cameras. Feature points are extracted from the images of the left and right cameras for matching to calculate the disparity between the images. When the corresponding feature points in one image are searched for in another image, their pixel offset, i.e., disparity, can be calculated. The farther the object is, the smaller the disparity is. Conversely, the larger the disparity is. Then, based on the relationship between disparity and distance, the distance between the obstacle and the engineering vehicle can be determined. Figure 4 As shown in the figure, Tx is the preset baseline length, that is, the distance between the two cameras, f is the focal length of the camera, Xl and Xr are the lateral positions of the target in the imaging planes of the left and right cameras, respectively, P is the feature point, (X, Y, Z) is the lateral, longitudinal, and depth coordinates of the feature point in three-dimensional space, the lateral distance difference is (Xl-Xr), and Z is the distance between the obstacle and the engineering vehicle. Therefore, based on the triangle formula, the distance between the obstacle and the engineering vehicle can be obtained:

[0098] Z=f×T x / (Xl-Xr).

[0099] In an embodiment of the present invention, through precise parallax analysis, the system can obtain the exact position of the obstacle in three-dimensional space, which provides richer spatial information than the traditional monocular vision system, enabling the anti-collision system to more accurately predict the trajectory of the obstacle and possible collision points, thereby more effectively performing early warning and braking, and improving the safety of engineering vehicles.

[0100] Optionally, determining the moving speed of the obstacle in combination with the change of the distance within a preset time period includes:

[0101] Determining the relative speed between the obstacle and the engineering vehicle according to the change value of the distance between the obstacle and the engineering vehicle within the preset time period;

[0102] The moving speed of the obstacle is determined according to the relative speed combined with the speed of the engineering vehicle in the preset time period.

[0103] Specifically, the relative speed between the obstacle and the engineering vehicle is first determined by monitoring the change in the distance between the obstacle and the engineering vehicle within a preset time period. The speed change of the obstacle relative to the engineering vehicle is calculated by tracking the position of the obstacle in continuous frames and analyzing the time series of these position data, so as to achieve real-time update of the obstacle depth information. This information can be used to monitor the movement trend and speed of the obstacle. Then, combined with the speed of the engineering vehicle within the preset time period, the system can determine the absolute moving speed of the obstacle. In a preferred embodiment of the present invention, the speed of the engineering vehicle within the preset time period can be obtained through the control system of the engineering vehicle, and this data can be combined with the relative speed of the obstacle to more accurately predict the actual movement state of the obstacle.

[0104] In an embodiment of the present invention, by determining the moving speed of an obstacle, the system is able to monitor the movement state of the obstacle, including its speed and direction, in real time and accurately. This precise grasp of the movement of the obstacle allows the system to predict potential collisions and respond in a timely manner, such as avoiding or mitigating collisions through early warning or automatic braking.

[0105] Optionally, obtaining a braking strategy of the engineering vehicle based on the movement information of the obstacle and the driving information of the engineering vehicle includes:

[0106] Predicting the collision risk level of the engineering vehicle within a preset future time period according to the distance between the obstacle and the engineering vehicle and the moving speed;

[0107] The braking strategy of the engineering vehicle under the predicted collision risk level is determined through graded braking strategies.

[0108] Specifically, the distance and moving speed information between the obstacle and the engineering vehicle is first used to analyze the current speed and direction of the obstacle and the driving state of the engineering vehicle to predict the collision risk level of the engineering vehicle in the preset future time period, so as to realize hierarchical intelligent active warning braking control. Then, the system determines the braking strategy of the engineering vehicle under different collision risk levels through a graded braking strategy. In a preferred embodiment of the present invention, the collision risk levels include 0, 1, 2, 3 and 4, which correspond to no risk, warning, speed limit, braking and emergency braking, respectively, and different levels correspond to different response measures. This hierarchical strategy can flexibly adjust the braking measures according to the actual risk level, ensuring that the most appropriate actions can be taken in different situations to avoid or mitigate collisions.

[0109] In an embodiment of the present invention, by predicting the risk of collision and taking graded braking measures in advance, the system can intervene in time before a collision occurs to reduce or avoid the occurrence of accidents, so that the system can dynamically adjust the braking strategy according to real-time obstacle and vehicle status information, thereby significantly improving the adaptability and effectiveness of the anti-collision system.

[0110] Optionally, the collision risk level includes multiple levels;

[0111] The predicting of the collision risk level of the engineering vehicle within a preset future time period according to the distance between the obstacle and the engineering vehicle and the moving speed includes:

[0112] Determining a braking distance and a braking time of the engineering vehicle according to the current speed of the engineering vehicle;

[0113] Determine a safe distance range between the engineering vehicle and the obstacle and a dangerous distance range corresponding to each level according to the braking distance, the braking time and the moving speed of the obstacle;

[0114] When the distance between the obstacle and the engineering vehicle is within the safety distance, the level is determined to be level 0;

[0115] When the distance between the obstacle and the engineering vehicle is within the dangerous distance range, the level corresponding to the dangerous distance range is used as the collision risk level of the engineering vehicle.

[0116] Specifically, the braking distance and braking time are calculated according to the current speed of the engineering vehicle, based on the kinematic formula in physics, taking into account the dynamic characteristics of the vehicle and the performance of the braking system. Then, combined with the moving speed of the obstacle, the safe distance range between the engineering vehicle and the obstacle, as well as the dangerous distance range corresponding to each collision risk level are determined. This process involves an in-depth understanding and precise calculation of the dynamic behavior of the vehicle to ensure that the collision risk can be accurately assessed at different speeds and obstacle movement states. When the distance between the obstacle and the engineering vehicle is within the safe distance, the system determines the collision risk level to be 0, indicating no collision risk; when the distance enters the dangerous distance range, the system will determine the collision risk level of the engineering vehicle according to the level corresponding to the range, and adopt the corresponding braking strategy.

[0117] In a preferred embodiment of the present invention, the collision risk level is divided into level 0, level 1, level 2, level 3 and level 4, corresponding to no risk, warning, speed limit, braking and emergency braking respectively. According to the braking distance, braking time and the moving speed of the obstacle, the safe distance range between the engineering vehicle and the obstacle and the dangerous distance range corresponding to each level are determined, wherein the safe distance range is the distance between the engineering vehicle and the obstacle greater than or equal to 70m, level 1 is the distance between the engineering vehicle and the obstacle greater than or equal to 50m and less than 70m, level 2 is the distance between the engineering vehicle and the obstacle greater than or equal to 30m and less than 50m, level 3 is the distance between the engineering vehicle and the obstacle greater than or equal to 10m and less than 30m, and level 4 is the distance between the engineering vehicle and the obstacle less than 10m.

[0118] In an embodiment of the present invention, by calculating the braking distance and time, the system can predict the braking behavior of the vehicle at different speeds, and then determine the safe and dangerous distance ranges from obstacles. This hierarchical risk assessment method enables the system to flexibly adjust the braking strategy based on real-time distance and speed information, so that the most appropriate action can be taken in different situations to avoid or mitigate collisions. This method not only improves the safety of engineering vehicles, but also enhances the adaptability and response speed of the anti-collision system. Especially in complex underground environments, this intelligent braking strategy based on real-time data analysis is crucial to ensuring the safety of operators and equipment, and also improves the efficiency and reliability of engineering operations.

[0119] Optionally, determining the braking strategy of the engineering vehicle at the predicted collision risk level by using a graded braking strategy includes:

[0120] Determining a maximum speed limit and an expected speed of the engineering vehicle based on the predicted collision risk level of the engineering vehicle and in combination with the current speed of the engineering vehicle;

[0121] Determining the braking mode, braking acceleration, braking opening and braking time of the engineering vehicle according to the current vehicle speed and the expected vehicle speed;

[0122] The maximum speed limit, the braking acceleration, the braking method, the braking opening and the braking time are used as the braking strategy.

[0123] Specifically, the maximum speed limit and the expected speed of the vehicle are determined according to the collision risk level and the current speed of the engineering vehicle. This decision-making process takes into account the safety performance and road conditions of the vehicle to ensure that the vehicle can travel at an appropriate speed under different risk levels. Then, the system determines the braking method of the vehicle, such as electric braking or mechanical braking, braking acceleration, braking opening and braking time, according to the current speed and the expected speed. These parameters are based on the precise calculation of vehicle dynamics and braking system characteristics to achieve smooth and effective deceleration or parking. Finally, the system integrates these parameters as a braking strategy and sends them to the vehicle's braking system for execution through the CAN bus. In a preferred embodiment of the present invention, the generation and execution of the braking strategy can be carried out based on the human brain regulation mechanism through the driver's braking behavior model, wherein an intelligent control system is divided into three levels: organizational level, coordination level and execution level, and the three levels are connected in a hierarchical form to form an intelligent control system, and the driver's braking behavior model is divided into three parts: sensory center model, behavioral center model and execution system, and the functions of the cerebellum, brainstem and spinal cord are simulated in a three-layer hierarchical structure. Among them, the cerebellum model is used as a comparator to realize the movement regulation function. A simplified model of the behavioral center is simulated by a neural network based on the cerebellum model to achieve real-time control, coordination, balance and rapid response of the controlled object. In order to simulate the control mechanism of different control levels of the human brain model, control strategies such as open-loop and closed-loop control, variable sampling cycle and control cycle, and variable control method are adopted under different control conditions to establish task coordination and communication among multiple centers, which can realize intelligent interaction among sensors, controllers, and actuators, and arbitrate conflicting control input and output signals. Its specific responsibilities are to analyze the target motion trajectory given by the upper layer, and then control the algorithm and the corresponding parameters according to the current state of the vehicle to obtain the braking strategy, pass the braking strategy to the execution layer, and feedback the execution status of the execution layer to the upper organizational layer, so as to realize the transmission and execution of the braking strategy.

[0124] Combination Figure 5As shown, in a preferred embodiment of the present invention, the collision risk levels 1, 2, 3 and 4 correspond to warning, speed limit, braking and emergency braking, respectively. According to the braking distance, braking time and the moving speed of the obstacle, the safe distance range between the engineering vehicle and the obstacle and the dangerous distance range corresponding to each level are determined, wherein Level 1 means that the distance between the engineering vehicle and the obstacle is greater than or equal to 50m and less than 70m, Level 2 means that the distance between the engineering vehicle and the obstacle is greater than or equal to 30m and less than 50m, Level 3 means that the distance between the engineering vehicle and the obstacle is greater than or equal to 10m and less than 30m, and Level 4 means that the distance between the engineering vehicle and the obstacle is less than 10m. Each level corresponds to a different braking strategy:

[0125] When the distance is greater than 50m, the system will first respond to the Level 1 warning. If the vehicle speed is greater than 30km / h, the Level 1 active braking function will be activated, and the vehicle's maximum speed limit will be adjusted to 30km / h. When the maximum speed limit is exceeded, the electric brake will be used to automatically control the vehicle to brake below 30km / h.

[0126] When the distance is shortened to less than 50m and more than 30m, the system will first respond to the Level 2 warning. If the vehicle speed is greater than 20km / h, the Level 2 active braking function will be activated, and the vehicle's maximum speed limit will be adjusted to 20km / h. When the maximum speed limit is exceeded, the electric braking method will be used to automatically control the vehicle to brake to less than 20km / h.

[0127] When S is shortened to less than 30m and more than 10m, firstly, a Level 3 warning response is performed. If the vehicle speed is greater than 10km / h, the Level 2 active braking function is activated, and the maximum speed limit of the vehicle is adjusted to 10km / h. When the maximum speed limit is exceeded, the electric braking method is used to automatically control the vehicle to brake to less than 10km / h. The braking acceleration and brake opening are calculated and sent in real time according to the braking time and braking distance.

[0128] When the system detects that S is shortened to less than 10m, it will first respond to the Level 4 warning. If the vehicle speed is greater than 2km / h, emergency braking will be performed and the hazard lights will be turned on to avoid rear-end collision.

[0129] In an embodiment of the present invention, the graded braking strategy not only improves the safety and effectiveness of braking, especially in complex underground environments, but also reduces vehicle damage and passenger discomfort caused by emergency braking. By precisely controlling the braking force and timing, the system can avoid collisions while maintaining smooth operation of the vehicle and improving ride comfort.

[0130] Optionally, assisting the engineering vehicle in braking according to the braking strategy includes:

[0131] Adjusting the speed limit of the engineering vehicle to the maximum speed limit;

[0132] When the current speed of the engineering vehicle is greater than the maximum speed limit, the engineering vehicle is assisted in braking according to the braking acceleration, the braking opening and the braking time in accordance with the braking method.

[0133] Specifically, first, the speed limit of the engineering vehicle is adjusted to the maximum speed limit according to the previously determined braking strategy, and the speed of the vehicle is immediately reduced to reduce the energy and damage during a potential collision. In a preferred embodiment of the present invention, the maximum speed limit of the vehicle is adjusted to 30km / h, 20km / h or 10km / h according to the situation. Subsequently, if the current speed of the engineering vehicle exceeds this maximum speed limit, the system will start the auxiliary braking program to assist the vehicle in braking according to the predetermined braking method, braking acceleration, braking opening and braking time. During this process, the system uses electric braking to automatically control the vehicle to brake below the speed limit to ensure that the vehicle travels within a safe speed range.

[0134] In the embodiment of the present invention, by adjusting the vehicle speed to the safe speed limit in real time and assisting braking when necessary, this strategy can significantly improve the safety of the vehicle and reduce accidents caused by speeding or untimely driver response.

[0135] Combination Figure 6 As shown, the present invention also provides an anti-collision system, comprising:

[0136] A multi-eye perception unit, used to obtain multi-eye visual images of the engineering vehicle within a preset range of the driving direction;

[0137] A feature extraction unit, used to extract features from the multi-view image to determine obstacles existing within the preset range;

[0138] An image analysis unit, used to detect the obstacle in the multi-view image and obtain movement information of the obstacle relative to the engineering vehicle;

[0139] A strategy generating unit, configured to obtain a braking strategy for the engineering vehicle based on the movement information of the obstacle and the driving information of the engineering vehicle;

[0140] The auxiliary braking unit is used to assist the engineering vehicle in braking according to the braking strategy.

[0141] The anti-collision system of the present invention can simultaneously monitor the front environment from multiple perspectives by acquiring multi-eye visual images of the engineering vehicle within a preset range of the driving direction, thereby increasing the breadth and depth of perception and improving the accuracy and reliability of obstacle detection. Next, feature extraction is performed on the multi-eye visual image to determine the obstacles existing within the preset range, and the obstacles are accurately identified and classified through image processing technology to provide key information for auxiliary braking. Then, the obstacles in the image are detected, the movement information of the obstacles relative to the engineering vehicle is obtained, the dynamic characteristics of the obstacles are evaluated, and the potential collision risk is predicted. Subsequently, the braking strategy of the engineering vehicle is formulated in combination with the movement information of the obstacle and the driving information of the engineering vehicle, and the vehicle state and environmental information are integrated to determine the most appropriate braking measures. Finally, the engineering vehicle is assisted in braking according to the formulated braking strategy, realizing closed-loop control from perception to execution, ensuring that the vehicle can respond to potential collision threats in a timely and accurate manner, improving the braking effect of engineering machinery vehicles, reducing the occurrence of collision accidents, and ensuring the safety of engineering operations.

[0142] A person skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a computer program, and the program can be stored in a computer-readable storage medium, and when the program is executed, it can include the processes of the embodiments of the above-mentioned methods. The storage medium can be a disk, an optical disk, a read-only memory (ROM) or a random access memory (RAM), etc.

[0143] Although the present invention is disclosed as above, the protection scope of the present invention is not limited thereto. Those skilled in the art may make various changes and modifications without departing from the spirit and scope of the present invention, and these changes and modifications will fall within the protection scope of the present invention.

Claims

1. A collision prevention method, characterized in that: include: Acquire multi-eye visual images of the engineering vehicle within a preset range of the driving direction; Extracting features from the multi-view image to determine obstacles within the preset range; Detecting the obstacle in the multi-view image to obtain movement information of the obstacle relative to the engineering vehicle; Obtaining a braking strategy for the engineering vehicle based on the movement information of the obstacle and the driving information of the engineering vehicle; The engineering vehicle is assisted in braking according to the braking strategy.

2. The anti-collision method according to claim 1, characterized in that: The multi-eye visual image includes a plurality of monocular visual images, and the feature extraction of the multi-eye visual image to determine obstacles existing within the preset range includes: Input each of the monocular visual images into a preset target detection network; Extracting local features of the monocular visual image through multiple convolutional layers of the preset target detection network in sequence; By performing feature fusion on the local features of each convolutional layer, a multi-scale feature map is obtained; Predicting based on the multi-scale feature map to determine all object labels in the monocular visual image; The obstacle existing within the preset range is determined according to the object label.

3. The anti-collision method according to claim 2, characterized in that: The detecting the obstacle in the multi-view image to obtain the motion information of the obstacle relative to the engineering vehicle includes: Performing a disparity analysis on the obstacle in each monocular vision image of the multi-eye vision image to determine the distance between the obstacle and the engineering vehicle; Determine the moving speed of the obstacle based on the change of the distance within a preset time period; The distance between the obstacle and the engineering vehicle and the moving speed are used as the motion information of the obstacle.

4. The anti-collision method according to claim 3, characterized in that: The performing disparity analysis on the obstacle in each monocular visual image of the multi-eye visual image to determine the distance between the obstacle and the engineering vehicle comprises: Matching the obstacle in each monocular vision image in the multi-eye vision image to obtain the same feature point of the same obstacle in each monocular vision image; Determine, according to the position of the feature point in each monocular vision image, a lateral distance difference between the same feature points in each monocular vision image; The depth of the feature point in the three-dimensional space is obtained by calculating the lateral distance difference between the same feature points in each monocular vision image in combination with a preset baseline length; The depth of the feature point is used as the distance between the obstacle and the engineering vehicle.

5. The anti-collision method according to claim 3, characterized in that: The determining the moving speed of the obstacle in combination with the change of the distance within a preset time period includes: Determining the relative speed between the obstacle and the engineering vehicle according to the change value of the distance between the obstacle and the engineering vehicle within the preset time period; The moving speed of the obstacle is determined according to the relative speed combined with the speed of the engineering vehicle in the preset time period.

6. The anti-collision method according to claim 3, characterized in that: The step of obtaining a braking strategy for the engineering vehicle based on the movement information of the obstacle and the driving information of the engineering vehicle includes: Predicting the collision risk level of the engineering vehicle within a preset future time period according to the distance between the obstacle and the engineering vehicle and the moving speed; The braking strategy of the engineering vehicle under the predicted collision risk level is determined through graded braking strategies.

7. The anti-collision method according to claim 6, characterized in that: The collision risk level includes multiple levels; The predicting of the collision risk level of the engineering vehicle within a preset future time period according to the distance between the obstacle and the engineering vehicle and the moving speed includes: Determining a braking distance and a braking time of the engineering vehicle according to the current speed of the engineering vehicle; Determine a safe distance range between the engineering vehicle and the obstacle and a dangerous distance range corresponding to each level according to the braking distance, the braking time and the moving speed of the obstacle; When the distance between the obstacle and the engineering vehicle is within the safety distance, the level is determined to be level 0; When the distance between the obstacle and the engineering vehicle is within the dangerous distance range, the level corresponding to the dangerous distance range is used as the collision risk level of the engineering vehicle.

8. The anti-collision method according to claim 6, characterized in that: The step of determining the braking strategy of the engineering vehicle at the predicted collision risk level through a graded braking strategy includes: Determining a maximum speed limit and an expected speed of the engineering vehicle based on the predicted collision risk level of the engineering vehicle and in combination with the current speed of the engineering vehicle; Determining the braking mode, braking acceleration, braking opening and braking time of the engineering vehicle according to the current vehicle speed and the expected vehicle speed; The maximum speed limit, the braking acceleration, the braking method, the braking opening and the braking time are used as the braking strategy.

9. The anti-collision method according to claim 8, characterized in that: Assisting the engineering vehicle to brake according to the braking strategy includes: Adjusting the speed limit of the engineering vehicle to the maximum speed limit; When the current speed of the engineering vehicle is greater than the maximum speed limit, the engineering vehicle is assisted in braking according to the braking acceleration, the braking opening and the braking time in accordance with the braking method.

10. An anti-collision system, characterized in that: include: A multi-eye perception unit, used to obtain multi-eye visual images of the engineering vehicle within a preset range of the driving direction; A feature extraction unit, used to extract features from the multi-view image to determine obstacles existing within the preset range; An image analysis unit, used to detect the obstacle in the multi-view image and obtain movement information of the obstacle relative to the engineering vehicle; A strategy generating unit, configured to obtain a braking strategy for the engineering vehicle based on the movement information of the obstacle and the driving information of the engineering vehicle; The auxiliary braking unit is used to assist the engineering vehicle in braking according to the braking strategy.

Images