Automatic coupling method, automatic uncoupling method and electronic device for a rail vehicle

Abstract

The application provides an automatic coupling method, an automatic uncoupling method and an electronic device for a rail vehicle. The application precisely controls the automatic coupling process of the rail vehicle by combining laser radar and three-dimensional point cloud data, which has significant beneficial effects. First, the laser radar continuously collects three-dimensional point cloud data, which can monitor the position of the target coupler in real time, and is not affected by environmental factors such as light, rain and snow, overcoming the identification deficiency of the traditional vision system in complex weather. Second, by calculating the distance deviation data between the couplers and adjusting the driving speed according to the preset speed control strategy, it ensures that the first rail vehicle can accurately approach the target coupler at an appropriate speed, reducing the risk of collision or misplacement. In addition, after judging that the preset coupling distance is reached, the coupling closing operation is automatically executed, ensuring the automation and efficiency of the coupling process, reducing the need for manual intervention, and improving the operation safety and reliability.

Description

Technical Field

[0001] This invention relates to the technical field of automatic control of rail vehicles, specifically to an automatic coupling method, an automatic uncoupling method, and electronic equipment for rail vehicles. Background Technology

[0002] In the current rail transit and industrial transportation sectors, coupling operations between vehicles mostly rely on manual labor or semi-automatic equipment. Traditional coupling and uncoupling operations are not only inefficient but also pose significant safety hazards, especially in high-temperature, high-noise, dusty, or toxic environments, where manual operation is highly prone to misoperation and personal injury.

[0003] In related technologies, there is a technical problem that the coupler positioning is not accurate enough during the automatic coupling process between two rail vehicles. Summary of the Invention

[0004] The purpose of this invention is to overcome the above-mentioned technical deficiencies and provide an automatic coupling method, an automatic uncoupling method, and an electronic device for rail vehicles, so as to solve the technical problem that the coupler positioning is not accurate enough during the automatic coupling process of two rail vehicles in related technologies.

[0005] To achieve the above-mentioned technical objectives, the present invention adopts the following technical solution:

[0006] In a first aspect, the present invention provides an automatic coupling method for rail vehicles, comprising:

[0007] The first rail vehicle is controlled to travel at a preset first speed, and during the travel, three-dimensional point cloud data in front of the first rail vehicle is continuously collected by a preset lidar; wherein, the front of the first rail vehicle refers to the front of the main body coupler of the first rail vehicle; the lidar is installed on the first rail vehicle and is on the same side as the main body coupler.

[0008] Based on the three-dimensional point cloud data, it is determined whether a target coupler exists; wherein, the target coupler is configured on one side of the main coupler of the second rail vehicle relative to the first rail vehicle.

[0009] If a target coupler is determined to exist, the docking deviation data between the target coupler and the main coupler is calculated; wherein, the docking deviation data includes at least the docking distance deviation data;

[0010] Based on the distance deviation data and the preset speed control strategy, the travel speed of the first rail vehicle is controlled so that the first rail vehicle approaches the second rail vehicle at a speed that is not higher than and not equal to the first speed.

[0011] If the preset hook distance is reached, the hook will be closed.

[0012] Furthermore, the step of determining whether a target coupler exists based on the three-dimensional point cloud data includes:

[0013] For the three-dimensional point cloud data, feature extraction is performed to obtain feature data used to characterize the coupler features;

[0014] Based on the feature data used to characterize the coupler features, it is determined whether there is a target coupler in front of the main coupler.

[0015] Furthermore, the docking deviation data also includes: docking angle deviation data;

[0016] The step of calculating the docking deviation data between the target coupler and the main coupler when it is determined that a target coupler exists includes:

[0017] Based on the latest 3D point cloud data, target 3D point cloud data is extracted; wherein, the target 3D point cloud data is 3D point cloud data representing the area where the target coupler is located;

[0018] For the target 3D point cloud data, a preset surface fitting algorithm is executed to generate a fitted quadratic surface model;

[0019] Based on the quadratic surface model, calculate the coordinates of the center point of the target coupler and the direction of its normal vector.

[0020] Based on the pre-calibrated center point coordinates and normal vector direction of the main coupler and the center point coordinates and normal vector direction of the target coupler, the distance deviation data and docking angle deviation data are calculated.

[0021] Furthermore, the step of controlling the travel speed of the first rail vehicle based on the distance deviation data and the preset speed control strategy includes:

[0022] If a target coupler is detected, the first speed is switched to the second speed, and the first rail vehicle is controlled to travel at a constant speed at the second speed; wherein the second speed is lower than the first speed.

[0023] Determine whether the distance deviation data has reached the preset speed switching threshold;

[0024] When a preset speed switching threshold is reached, the second speed is switched to the third speed, and the first rail vehicle is controlled to move at the third speed; wherein the third speed is lower than the second speed.

[0025] Furthermore, based on the aforementioned distance deviation data and a preset speed control strategy, the travel speed of the first rail vehicle is controlled so that the first rail vehicle approaches the second rail vehicle at a speed not higher than and not equal to the first speed. And when the coupler is a James coupler, the method further includes:

[0026] Send a control signal to the power unit of the main coupler to open the main coupler so that the main coupler is in a ready-to-connect state.

[0027] The step of closing the hook when it is determined that a preset hook distance has been reached includes:

[0028] If it is determined that the preset hooking distance has been reached, the first rail vehicle is controlled to make contact with the second rail vehicle at a speed not exceeding the third speed.

[0029] After a touch is detected, three-dimensional point cloud data of the hook area is collected using a pre-set lidar.

[0030] For the three-dimensional point cloud data of the hook area, extract the point cloud voxel block of the locking pin area of ​​the James coupler;

[0031] Based on the extracted point cloud voxel blocks of the locking pin area and the preset locking pin placement model, an initial judgment on whether the hooking was successful is performed.

[0032] If the initial assessment indicates successful hooking, a pre-defined final judgment strategy is executed to determine whether the hooking was ultimately successful.

[0033] Furthermore, the step of performing an initial judgment on whether the hooking was successful based on the extracted point cloud voxel blocks of the locking pin region and the preset locking pin placement model includes:

[0034] The point cloud voxel blocks of the locking pin area are matched and compared with the pre-stored reference point cloud spatial distribution in the locking pin placement model to obtain the spatial alignment relationship between the two.

[0035] Based on the spatial alignment relationship, the point cloud clusters corresponding to the locking pin endpoints are extracted from the point cloud voxel blocks, and the center coordinates of the point cloud clusters are calculated.

[0036] Determine whether the center coordinates of the point cloud cluster fall within the three-dimensional tolerance range of the hook pin hole area calibrated in the locking pin positioning model, and determine whether the absolute value of the vertical coordinate deviation is within the preset vertical deviation range.

[0037] If all conditions are met, it is determined that the locking pin has been correctly positioned, and the hook is initially determined to be successfully hooked.

[0038] Furthermore, the step of executing a preset final judgment strategy to determine whether the hooking was successful after the initial judgment is successful includes:

[0039] The first rail vehicle is controlled to move in the opposite direction of travel by a preset distance, and during the reverse movement, the lidar continuously collects three-dimensional point cloud data of the locking pin area.

[0040] Based on the three-dimensional point cloud data collected during the reverse movement, the real-time point cloud clusters corresponding to the locking pin endpoints are extracted, and their real-time center coordinates are calculated.

[0041] Based on the continuous change in the real-time center coordinates, it is determined whether the position of the locking pin meets the preset stability conditions; wherein, the preset stability conditions include: the change in the real-time center coordinates of the locking pin endpoint in the horizontal direction does not exceed the horizontal deviation range; and the change in the vertical direction does not exceed the vertical deviation range.

[0042] If the real-time center coordinates meet the preset stability conditions, the hooking is ultimately determined to be successful.

[0043] Secondly, the present invention provides an automatic uncoupling method for rail vehicles, comprising:

[0044] In response to the received uncoupling command, a control signal is sent to the power unit of the main coupler of the first rail vehicle to drive the power unit to perform the coupler unlocking operation; wherein, the main coupler and the target coupler are in a hooking state; wherein, the first rail vehicle uses the above-mentioned automatic hooking method of rail vehicle to perform the hooking operation of the main coupler and the target coupler, so that the main coupler and the target coupler are in a hooking state.

[0045] During the coupler unlocking operation, a preset lidar is used to collect three-dimensional point cloud data of the coupler area and extract real-time point cloud voxel blocks of the locking pin area.

[0046] Matching and comparison are performed based on real-time point cloud voxel blocks and a preset lock pin not in place model to obtain the matching and comparison results; wherein, the lock pin not in place model includes the baseline point cloud distribution features after the lock pin is pulled out.

[0047] Based on the matching and comparison results, it is determined whether the locking pin has been successfully pulled out;

[0048] If the locking pin is successfully removed, the first track vehicle is controlled to move in the opposite direction until the lidar detects that the hook separation distance is greater than the preset safe distance threshold, and then the hook removal is confirmed to be completed.

[0049] Furthermore, the step of determining whether the locking pin has been successfully removed based on the matching comparison result includes:

[0050] If the point cloud voxel block of the locking pin area does not cover the predefined locking pin feature area in the locking pin not in place model, or if the center coordinate of the point cloud cluster corresponding to the locking pin endpoint deviates from the center of the hook hole by more than the preset unhooking tolerance range, then it is determined that the locking pin has been successfully pulled out.

[0051] If the center coordinates of the point cloud cluster corresponding to the locking pin endpoint are still within the preset unhooking tolerance range of the hook hole area, it is determined that the unhooking is not completed and an unlocking abnormal alarm signal is generated.

[0052] Thirdly, the present invention provides an electronic device, comprising: a memory, and one or more processors communicatively connected to the memory; the memory stores instructions executable by the one or more processors, the instructions being executed by the one or more processors to cause the one or more processors to implement the method described above.

[0053] Beneficial effects:

[0054] This invention combines lidar and 3D point cloud data to precisely control the automatic coupling process of rail vehicles, offering significant advantages. First, by continuously acquiring 3D point cloud data through lidar, the position of the target coupler can be monitored in real time, unaffected by environmental factors such as lighting, rain, or snow, overcoming the limitations of traditional vision systems in complex weather conditions. Second, by calculating the distance deviation between the couplers and adjusting the travel speed according to a preset speed control strategy, it ensures that the first rail vehicle can accurately approach the target coupler at an appropriate speed, reducing the risk of collisions or misalignment. Furthermore, after determining that the preset coupling distance has been reached, the coupling closing operation is automatically executed, ensuring the automation and efficiency of the coupling process, reducing the need for manual intervention, and improving operational safety and reliability. Attached Figure Description

[0055] Figure 1 This is a flowchart illustrating an automatic hooking method for a rail vehicle provided in an embodiment of the present invention;

[0056] Figure 2 This is a flowchart illustrating an automatic uncoupling method for a rail vehicle provided in an embodiment of the present invention.

[0057] Figure 3 This is a block diagram of an electronic device used in an embodiment of the present invention. Detailed Implementation

[0058] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort should fall within the scope of protection of the present application.

[0059] In the rail transit sector, automated coupler technology is used to improve transportation efficiency and reduce manual operations. However, existing automated coupler systems face numerous challenges, particularly regarding the accuracy of coupler positioning. Accurate coupler positioning is crucial for ensuring successful and safe coupler operations. Most existing technologies rely on purely visual solutions and laser ranging methods, but both have limitations, resulting in insufficient coupler positioning accuracy.

[0060] In some technical solutions, automatic coupling systems primarily rely on visual recognition technology and lidar ranging to detect and locate the coupling. Visual recognition technology acquires image data of the target coupling using cameras or other imaging devices and analyzes it through image processing algorithms. However, visual solutions are highly susceptible to environmental interference. For example, in rainy, foggy, or smoky environments, cameras cannot clearly capture images of the target coupling, leading to decreased image recognition accuracy and even the inability to identify the coupling's location. Worse still, cameras are prone to misjudgment or missed detection in low light or under conditions of significant light variation.

[0061] Furthermore, laser ranging methods rely on lidar systems to measure the distance to the target coupler. While laser ranging technology offers high reliability under certain environmental conditions, lidar typically only provides distance information and cannot accurately determine whether the vehicle and target coupler have been successfully coupled. A single laser ranging method cannot effectively determine whether the coupling operation has been completed. Therefore, although laser ranging can provide some distance information, its limitations make it difficult to meet the requirements of high-precision coupling operations.

[0062] Therefore, the application of vision and lidar technologies in automated hooking systems faces a series of challenges. First, vision systems are easily affected by the environment; blurry images, insufficient light, and smoke interference can all lead to recognition failures or reduced accuracy. Second, laser ranging methods cannot comprehensively acquire the geometric information of the coupler, only providing target distance data and lacking a comprehensive assessment of the coupler's shape and docking status. Consequently, these two technologies struggle to ensure accurate automated hooking operations under various complex environments, especially in extreme weather conditions.

[0063] In summary, in the relevant technologies, there is a technical problem of inaccurate coupler positioning during the automatic coupling process of two rail vehicles.

[0064] like Figure 1 As shown, this embodiment provides an automatic coupling method for rail vehicles, the method comprising:

[0065] Step S10: Control the first rail vehicle to travel at a preset first speed, and during the travel, continuously collect three-dimensional point cloud data in front of the first rail vehicle through a preset lidar; wherein, the front of the first rail vehicle refers to the front of the main body coupler of the first rail vehicle; the lidar is installed on the first rail vehicle and is on the same side as the main body coupler.

[0066] In this embodiment, the entity executing the method can be a control device or control system of the first rail vehicle. Specifically, the control device can be an MCU (microcontroller unit), an FPGA (field-programmable gate array), an embedded controller, an industrial PC (IPC), a PLC (programmable logic controller), or a single-chip microcomputer, etc.

[0067] In this embodiment, the first rail vehicle is a rail vehicle that needs to be moved to perform coupling. It is understood that, in a specific scenario, a first rail vehicle and a second rail vehicle are parked on a certain track. The first rail vehicle has a main coupler and a lidar on one side. Correspondingly, the second rail vehicle has a target coupler on one side. For example, the main coupler and lidar are located at the rear of the first rail vehicle, and the target coupler is located at the front of the second rail vehicle. Therefore, the main coupler and the target coupler can be coupled together, thereby connecting the first rail vehicle and the second rail vehicle.

[0068] Specifically, the first rail vehicle can be a complete locomotive, located at the front of the train, and equipped with a power system and control system or control device, responsible for traction and control of the entire train's operation. The second rail vehicle can be a complete carriage, either a passenger car or a freight car. The first rail vehicle can also be one of multiple carriages, located at the front or middle of the train, and can be connected to other carriages. This carriage can also be equipped with a control system to coordinate coupling operations. The second rail vehicle can be another carriage in the same train, waiting to be connected to the first rail vehicle. The first rail vehicle can serve as a power unit, either an electric locomotive or a diesel locomotive, providing traction and control functions. The second rail vehicle can serve as a trailer unit, either a non-powered carriage or freight car, relying on the traction of the power unit.

[0069] Therefore, in this embodiment, "in front of the first rail vehicle" refers to the front of the main coupler of the first rail vehicle. That is, when the main coupler is located at the rear of the first rail vehicle, then "in front of the first rail vehicle" is also the front of the rear. Correspondingly, when the main coupler is located at the front of the first rail vehicle, then "in front of the first rail vehicle" is also the front of the front. In other words, "in front of the first rail vehicle" is the front of the first rail vehicle in its direction of travel, which is also the front of the main coupler (the second rail vehicle is parked at a certain distance in front of the main coupler).

[0070] In this embodiment, the first rail vehicle and the second rail vehicle can be rail vehicles used in urban rail transit scenarios, such as urban subways or light rail.

[0071] In this embodiment, the first rail vehicle and the second rail vehicle can be rail vehicles used in long-distance rail transport scenarios. For example, they can be trains.

[0072] In this embodiment, the first rail vehicle and the second rail vehicle can be rail vehicles in an industrial rail system scenario. For example, they can be dedicated locomotives in industrial parks or mining areas.

[0073] In this embodiment, the main coupler can be a Jameson coupler, and correspondingly, the target coupler is a Jameson coupler that can be adapted to the main coupler. The two couplers can be coupled or uncoupled automatically or manually.

[0074] In this embodiment, the lidar can be installed on the top of one side of the first rail vehicle to acquire three-dimensional point cloud data of the area in front of that side. One or more lidars can be installed.

[0075] In this embodiment, the lidar can be a mechanically rotating lidar.

[0076] In this embodiment, the lidar can be a solid-state lidar.

[0077] In this embodiment, the lidar can be a multi-beam lidar.

[0078] In this embodiment, the lidar can be a single-beam lidar.

[0079] In this embodiment, the lidar can be a long-range lidar.

[0080] It is understood that the lidar can communicate with the control device or control system of the first rail vehicle. Specifically, the lidar can be fixed to the roof or a rigid structure at the front of the first rail vehicle. The lidar's optical axis is tilted downwards to ensure that the scanning range covers the area where the coupler is located. The three-dimensional point cloud data collected by the lidar can be transmitted to the control device or control system of the first rail vehicle via Ethernet (UDP protocol) or a dedicated LVDS interface. Control signals from the control device or control system of the first rail vehicle can be sent to the lidar or other modules (e.g., the power module of the rail vehicle) via a CAN bus.

[0081] It is also understandable that the lidar can be pre-calibrated.

[0082] In this embodiment, the first speed is a pre-set speed, which can be the fastest travel speed during the automated hooking process of the first rail vehicle. It is understood that the speed of the first rail vehicle can vary during the automated hooking process. For example, the speed of the first rail vehicle can be controlled according to a strategy of uniform approach-deceleration upon contact-precise hooking. Alternatively, the speed of the first rail vehicle can be controlled based on the relative distance between the first and second rail vehicles. For example, the closer the distance, the lower the speed of the first rail crane (e.g., dynamically adjusted using a PID control algorithm).

[0083] In this embodiment, the first speed can be 0.4 m / s, 0.5 m / s, or 0.6 m / s, etc. Specifically, the first speed is the maximum initial travel speed preset by the rail vehicle in the automated coupling process, used to quickly approach the target vehicle in the initial stage. The role of the first speed is, on the one hand, to shorten the approach time as much as possible within a safe range; on the other hand, to provide sufficient distance buffer for the subsequent precise alignment and deceleration stages; and on the other hand, it can serve as a benchmark parameter for speed graded control, supporting speed adjustment logic under different scenarios. For example, in a possible and specific embodiment, during the uniform approach stage, the vehicle can travel at a uniform speed of the first speed (e.g., 0.5 m / s) to quickly reduce the distance to the target vehicle; during the contact deceleration stage, when the docking deviation (relative distance) reaches a preset threshold, the vehicle can switch to the second speed (e.g., 0.3 m / s) to enter the precise alignment mode, in which the speed can be finely adjusted in real time according to the docking deviation (e.g., using a PID control algorithm); during the precise coupling stage, the vehicle can switch to the third speed when approaching contact to ensure low-speed contact and avoid mechanical impact.

[0084] Step S12: Based on the three-dimensional point cloud data, determine whether a target coupler exists; wherein, the target coupler is configured on one side of the main coupler of the second rail vehicle relative to the first rail vehicle.

[0085] In this embodiment, the target coupler is a coupler configured on the second rail vehicle, which needs to be connected to the main coupler. Specifically, the target coupler is fixedly installed at the end of the second rail vehicle, and its installation position is symmetrically arranged with the main coupler of the first rail vehicle. For example, if the main coupler of the first rail vehicle is located on the left side of the car body, the target coupler is correspondingly installed on the right side of the second rail vehicle; the two can meet the alignment conditions of coaxial alignment or within a preset angular deviation range in space to ensure the feasibility of mechanical engagement.

[0086] In this embodiment, the point cloud data collected by lidar can be used to reconstruct the three-dimensional geometric contour and spatial pose of the target coupler in real time.

[0087] In this embodiment, the step of determining whether a target coupler exists based on the three-dimensional point cloud data may include:

[0088] First, feature extraction is performed on the three-dimensional point cloud data to obtain feature data that characterizes the coupler features.

[0089] Specifically, the feature data used to characterize the coupler features may include:

[0090] 1. Hook surface features, such as a curvature distribution histogram, surface fitting parameters (quadratic surface coefficients), or principal curvature directions (maximum / minimum curvature), etc. It can be understood that these hook surface features can be used to describe the geometry of the hook's concave surface.

[0091] 2. Planar features of the hook tail, such as plane equation parameters, plane area, or plane normal vector direction, etc. It can be understood that these planar features of the hook tail can be used to describe the planar structure and spatial attitude of the hook tail.

[0092] 3. Spatial relationship characteristics, such as the distance between the hook head and the hook tail, the perpendicular distance between the center point of the hook head and the plane of the hook tail, or the angle between the normal vector of the hook head and the normal vector of the hook tail, etc.

[0093] 4. Statistical characteristics, such as point cloud density, reflection intensity distribution (to distinguish materials), or symmetry score (left-right symmetry of the hook), etc. These statistical characteristics can be used to eliminate interference from foreign objects (such as metal supports or track reflections).

[0094] In this embodiment, a curvature-based partitioning fitting algorithm can be used to extract the hook-head surface features.

[0095] In this embodiment, the RANSAC (Random Sample Consensus) algorithm can be used to extract the hook-tail plane features.

[0096] In this embodiment, a geometric relationship analysis algorithm can be used to extract spatial relationship features.

[0097] Then, based on the feature data used to characterize the coupler features, it is determined whether there is a target coupler in front of the main coupler.

[0098] In this embodiment, feature data characterizing coupler features can be matched with a preset feature template (which may store various coupler features characterizing the target coupler) to determine whether a target coupler exists. It is understood that if a preset similarity threshold is met, it is determined that a target coupler exists in front of the main coupler. Conversely, if the similarity threshold is not met, it is determined that no target coupler exists in front of the main coupler (the first and second rail vehicles are still far apart).

[0099] In this embodiment, feature data characterizing the coupler can be used to further extract the point cloud of the ROI region (which may be a complete point cloud segment where the coupler head is located) from the three-dimensional point cloud data. Then, the point cloud of the ROI region is used to match the point cloud of the coupler with a preset coupler point cloud template to determine whether the target coupler exists.

[0100] Step S14: If it is determined that a target coupler exists, calculate the docking deviation data between the target coupler and the main coupler; wherein the docking deviation data includes at least the docking distance deviation data.

[0101] In this embodiment, the docking deviation data may further include docking angle deviation data. Specifically, the docking deviation data can be represented as a set of quantitative indicators characterizing the comprehensive degree of deviation between the two couplers in spatial pose. The docking distance deviation data can be used to describe the straight-line distance between the center points of the two couplers in a specified coordinate system. The docking angle deviation data can be used to describe the spatial angular difference between the normal vector directions of the two couplers.

[0102] More specifically, the docking distance deviation data can be used to quantify the relative positional deviation of the center points of the two couplers in three-dimensional space, which directly affects the approach speed and path planning. The docking distance deviation data can be a three-dimensional Euclidean distance (fully representing spatial deviation), a two-dimensional projected distance (i.e., the Z-axis is fixed, requiring only horizontal alignment), or a one-dimensional distance (i.e., both the Z-axis and Y-axis are fixed, only the X-axis distance needs to be determined; understandably, with the track as the X-axis, the axis perpendicular to the ground as the Z-axis, and on the ground, the axis perpendicular to the X-axis as the Y-axis). The docking angle deviation data can be used to quantify the alignment error of the normal vectors of the two couplers, ensuring no lateral stress during mechanical engagement. Similarly, this docking angle deviation data can be the angle between the normal vectors in space (three-dimensional angle), the yaw angle of the projected plane (two-dimensional simplification), or a one-dimensional angle (one-dimensional simplification).

[0103] In one specific implementation, step S14 may include:

[0104] The step of calculating the docking deviation data between the target coupler and the main coupler when it is determined that a target coupler exists includes:

[0105] Step S142: Based on the latest 3D point cloud data, extract the target 3D point cloud data; wherein, the target 3D point cloud data is the 3D point cloud data representing the area where the target coupler is located;

[0106] Step S144: For the target 3D point cloud data, execute a preset surface fitting algorithm to generate a fitted quadratic surface model;

[0107] Step S146: Based on the quadratic surface model, calculate the coordinates of the center point and the direction of the normal vector of the target coupler;

[0108] Step S148: Based on the pre-calibrated center point coordinates and normal vector direction of the main coupler and the center point coordinates and normal vector direction of the target coupler, calculate the distance deviation data and docking angle deviation data.

[0109] Step S16: Based on the distance deviation data and the preset speed control strategy, control the travel speed of the first rail vehicle so that the first rail vehicle approaches the second rail vehicle at a speed that is not higher than and not equal to the first speed.

[0110] In this embodiment, the preset speed control strategy can be a speed-graded control strategy. For example, multiple distance intervals can be preset based on the distance deviation data, with different distance intervals corresponding to different speeds. Within a certain distance interval, the first rail vehicle is controlled to maintain the speed corresponding to that distance interval and perform uniform speed travel. The closer the first rail vehicle is to the second rail vehicle (i.e., the smaller the distance deviation data), the lower its speed.

[0111] In this embodiment, the preset speed control strategy can be a linearly decreasing speed control strategy. More specifically, the entire approach distance can be divided into multiple intervals (e.g., 0-0.3 meters, 0.3-1 meter, and over 1 meter). When the distance is >1 meter, the speed is constant at 0.4 m / s; when the distance is 0.3-1 meter, the speed decreases linearly from 0.4 m / s to 0.1 m / s; and when the distance is <0.3 meters, the speed further decreases linearly to 0.05 m / s. A gradual adjustment algorithm is used at the interval boundaries to avoid sudden speed changes. In a specific scenario example, the vehicle approaches from 2 meters away with an initial speed of 0.4 m / s; when the distance decreases to 0.8 meters, the speed decreases proportionally to 0.25 m / s; at 0.2 meters, the final fine-tuning stage begins, with the speed decreasing to 0.05 m / s to ensure no impact contact.

[0112] In this embodiment, the preset speed control strategy can be a speed control strategy that reduces the driving speed non-linearly.

[0113] In this embodiment, the preset speed control strategy can also be a PID feedback-based control strategy.

[0114] Specifically, distance deviation data can be used as input, and the speed adjustment amount can be calculated in real time by a proportional (P), integral (I), and derivative (D) controller.

[0115]

[0116] In the formula, This can be expressed as the adjusted speed. This can be represented as the real-time deviation between the target value and the actual value (distance deviation data). It can be expressed as the integral gain coefficient. This can be represented as an integral term, that is, the cumulative sum of errors from the initial time to the current time t. It can be expressed as the differential gain coefficient. It can be expressed as the rate of change of error over time, that is, the derivative of the error.

[0117] In this embodiment, the preset speed control strategy can also be a speed control strategy based on fuzzy logic control. For example, a fuzzy set of distance deviations (such as "far", "medium", "near") and fuzzy rules for speed output can be defined, and nonlinear mapping can be achieved through membership functions. More specifically, precise distance deviations (such as 1.5 meters, 0.6 meters) can be converted into fuzzy linguistic values ​​such as "far", "medium", and "near". For example, when the distance is >1 meter, the degree of belonging to "far" is 80%; when the distance is 0.3-1 meter, the degree of belonging to "medium" is 60%; and when the distance is <0.3 meters, the degree of belonging to "near" is 90%. Fuzzy rules can be defined, for example, if the distance is "far", then output "high speed"; if the distance is "medium" and the angle deviation is "small", then output "medium speed"; if the distance is "near" or the angle deviation is "large", then output "low speed". Finally, the fuzzy output is converted into precise speed values ​​(such as "high speed" corresponding to 0.4 m / s, and "low speed" corresponding to 0.05 m / s). In a specific scenario example, when the vehicle is 2 meters away from the target, the fuzzy system determines it to be "far" and approaches at a constant speed of 0.4 m / s; when the distance decreases to 0.5 meters, the system judges it to be "medium speed" based on the comprehensive angle deviation and adjusts the speed to 0.2 m / s; when the distance is 0.2 meters, "low speed" is triggered, and millimeter-level micro-motion is entered.

[0118] In this embodiment, the preset speed control strategy can also be a speed control strategy based on machine learning or deep learning. More specifically, historical operation data can be collected in advance, including distance deviation, angle deviation, speed command, and hooking results (success / failure). The historical operation data is then used to train the neural network, which (such as a fully connected network or LSTM) learns the mapping relationship between input (distance, angle) and output (speed). The model is optimized through supervised learning, with the goal of minimizing the difference between the predicted speed and the actual operating speed. In a specific scenario example, in a dusty environment, the model learns from historical data that "speed should be reduced in advance when dust concentration is high." When the current dust concentration exceeds the standard, even if the distance is 1 meter, the speed is actively reduced from 0.4 m / s to 0.3 m / s to avoid the risk of misjudgment.

[0119] In this embodiment, the preset speed control strategy can also be a speed control strategy based on reinforcement learning (RL). For example, a speed control strategy network can be trained using a deep reinforcement learning model, and the reward function can be designed to balance rapid approach with low impact.

[0120] Therefore, in this embodiment, the phrase "making the first rail vehicle approach the second rail vehicle at a speed not higher than and not equal to the first speed" can be expressed as follows: during the process of the first rail vehicle moving towards the second rail vehicle, any of the above-mentioned speed control strategies are adopted to reduce the first speed. For example, the speed can be reduced in a linear manner, or in a nonlinear manner, or it can be reduced by model prediction, thereby making the first rail vehicle approach the second rail vehicle at a speed not higher than and not equal to the first speed.

[0121] Step S18: If it is determined that the preset hook distance has been reached, perform hook closure.

[0122] In this embodiment, the preset hook distance can be a relatively short preset distance, such as 5 cm, 4 cm, or 6 cm, etc. It is understood that, on the one hand, the point cloud data collected by the lidar during vehicle movement has a certain amount of noise and delay, with a typical ranging error of about ±1 cm. Therefore, the preset distance is generally selected in the range of 4–6 cm, with 5 cm being the preferred value. This ensures sufficient safety margin while allowing enough response time before mechanical closure to avoid impact or misalignment. On the other hand, the insertion of the coupler head into the hook opening requires a certain buffer stroke; too short a stroke will result in insufficient time for the closing action, while too long a stroke increases the risk of secondary collisions. A distance of about 5 cm is sufficient in most James Bond coupler structures to ensure that the coupler head has basically entered the hook opening and overlaps with the locking pin alignment range.

[0123] This implementation method, by combining LiDAR and 3D point cloud data, precisely controls the automatic coupling process of rail vehicles, offering significant advantages. First, by continuously acquiring 3D point cloud data through LiDAR, the position of the target coupler can be monitored in real time, unaffected by environmental factors such as lighting, rain, or snow, overcoming the limitations of traditional vision systems in complex weather conditions. Second, by calculating the distance deviation between the couplers and adjusting the travel speed according to a preset speed control strategy, it ensures that the first rail vehicle can accurately approach the target coupler at an appropriate speed, reducing the risk of collisions or misalignment. Furthermore, after determining that the preset coupling distance has been reached, the coupling closing operation is automatically executed, ensuring the automation and efficiency of the coupling process, reducing the need for manual intervention, and improving operational safety and reliability. Therefore, this implementation method has high precision and adaptability in automated coupling in rail transit, improving operational efficiency and reducing safety hazards caused by misoperation.

[0124] In some implementations, the step of determining whether a target coupler exists based on the three-dimensional point cloud data includes:

[0125] Step S122: Perform feature extraction on the three-dimensional point cloud data to obtain feature data used to characterize the coupler features.

[0126] Specifically, the feature data used to characterize the coupler features may include:

[0127] 1. Hook surface features, such as a curvature distribution histogram, surface fitting parameters (quadratic surface coefficients), or principal curvature directions (maximum / minimum curvature), etc. It can be understood that these hook surface features can be used to describe the geometry of the hook's concave surface.

[0128] 2. Planar features of the hook tail, such as plane equation parameters, plane area, or plane normal vector direction, etc. It can be understood that these planar features of the hook tail can be used to describe the planar structure and spatial attitude of the hook tail.

[0129] 3. Spatial relationship characteristics, such as the distance between the hook head and the hook tail, the perpendicular distance between the center point of the hook head and the plane of the hook tail, or the angle between the normal vector of the hook head and the normal vector of the hook tail, etc.

[0130] 4. Statistical characteristics, such as point cloud density, reflection intensity distribution (to distinguish materials), or symmetry score (left-right symmetry of the hook), etc. These statistical characteristics can be used to eliminate interference from foreign objects (such as metal supports or track reflections).

[0131] In this embodiment, a curvature-based partitioning fitting algorithm can be used to extract the hook-head surface features.

[0132] In this embodiment, the RANSAC (Random Sample Consensus) algorithm can be used to extract the hook-tail plane features.

[0133] In this embodiment, a geometric relationship analysis algorithm can be used to extract spatial relationship features.

[0134] In some implementations, PCA analysis can be performed on the neighborhood of each point to calculate the point cloud normal vector and principal curvature direction. Based on this, local descriptors such as Fast Point Feature Histograms (FPFH) or Signature of Histograms of Orientations (SHOT) can be constructed to encode the surface shape and normal vector distribution, which can be used to distinguish the coupler surface from surrounding metal or debris.

[0135] In some implementations, end-to-end feature extraction algorithms based on deep learning can also be used to perform feature extraction and obtain feature data to characterize the features of the coupler. For example, networks such as PointNet / PointNet++, DGCNN, and PointCNN can be used to learn high-dimensional feature vectors directly from the original or downsampled point cloud input. The above-mentioned networks can be pre-trained through classification or segmentation tasks and fine-tuned in the hooking scenario to obtain deep features that can distinguish components such as the hook head, hook tail, and locking pin.

[0136] Step S124: Based on the feature data used to characterize the coupler features, determine whether there is a target coupler in front of the main coupler.

[0137] In this embodiment, feature data characterizing coupler features can be matched with a preset feature template (which may store various coupler features characterizing the target coupler) to determine whether a target coupler exists. It is understood that if a preset similarity threshold is met, it is determined that a target coupler exists in front of the main coupler. Conversely, if the similarity threshold is not met, it is determined that no target coupler exists in front of the main coupler (the first and second rail vehicles are still far apart).

[0138] Specifically, the preset feature template can store the following in advance:

[0139] Geometric parameters: coefficients of the hook head surface equation, equation of the hook tail plane, coordinates of the center point, and direction of the normal vector;

[0140] Statistical parameters: curvature distribution histogram, mean reflection intensity, symmetry score;

[0141] Spatial relationships: standard distance between hook head and hook tail, range of angle between normal vectors.

[0142] During matching, a coarse matching can be performed first, followed by a fine matching (ICP registration). More specifically, the cosine similarity between the real-time hook curvature histogram and the template can be calculated, and objects with excessively low cosine similarity can be directly excluded (using a pre-set similarity threshold, such as 85%). This can quickly exclude over 90% of non-target objects (such as metal supports beside the track). Then, the real-time hook point cloud is aligned with the template point cloud, the registration error (root mean square error < 5mm) is calculated, and the hook-tail distance and the angle between the normal vectors are verified to be within the standard range.

[0143] In this embodiment, feature data characterizing the coupler can be used to further extract the point cloud of the ROI region (which may be a complete point cloud segment where the coupler head is located) from the three-dimensional point cloud data. Then, the point cloud of the ROI region is used to match the point cloud of the coupler with a preset coupler point cloud template to determine whether the target coupler exists.

[0144] Specifically, firstly, a coupler point cloud template can be pre-constructed. For example, in a laboratory environment, a high-resolution LiDAR (such as a 128-line sensor) can be used to scan a standard coupler to obtain a noise-free point cloud. This can also include coupler models with different wear levels and installation angles (±5° tilt). Then, the coupler head region is located based on feature data. The curvature value of each point in the point cloud can be calculated first, and high-curvature regions (curvature > 0.05) can be selected to initially locate candidate coupler head regions. Then, Euclidean clustering (cluster radius can be 10 cm) is performed on the high-curvature points, retaining the largest cluster as the core area of ​​the coupler head. Next, the ROI region is delineated. This can be done by extending the core area of ​​the coupler head along the track extension direction (X-axis) by several centimeters (e.g., 20 cm), perpendicular to the track direction (Y-axis) by several centimeters (e.g., 30 cm), and vertically (Z-axis) by several centimeters (e.g., 20 cm), forming a three-dimensional ROI region. A pass-through filter can be further applied to precisely trim the ROI point cloud, excluding irrelevant objects (such as tracks and gravel). Further analysis can be conducted to identify outlier noise points and eliminate discrete noise points within the ROI. Finally, the ROI point cloud can be matched with templates. A coarse matching can be performed first. For example, the curvature distribution histogram of the ROI point cloud can be calculated, and cosine similarity can be calculated between it and the curvature histograms of all couplers in the template library, filtering candidate templates with a similarity > 0.8. Then, the distance between the hook head and hook tail is checked to see if it is within the standard range (500±10mm), eliminating templates that are clearly mismatched. After coarse matching, fine matching is performed. For example, the Iterative Closest Point (ICP) algorithm can be used to register the ROI point cloud with the candidate template point cloud, optimizing the rigid body transformation matrix (translation + rotation). After registration, the root mean square error (RMSE) between the two point clouds can be calculated, and the angle between the normal vectors of the two hook heads can be checked to ensure attitude alignment.

[0145] By extracting features from the 3D point cloud data in step S122 and determining the presence of the target coupler based on the extracted coupler feature data in step S124, this embodiment can quickly and accurately identify the object to be hooked in complex environments. Compared with coarse detection based directly on the original point cloud or 2D image, this method utilizes multi-dimensional geometric features such as the hook head surface, hook tail plane, and spatial relationships. This effectively filters environmental clutter (such as interference from track facilities and foreign object reflections) and reduces the risk of misjudgment caused by occlusion or partial point cloud loss. At the same time, by using feature template matching instead of simple threshold judgment, higher recognition accuracy and faster response speed are achieved, providing reliable input for subsequent deviation calculation and speed control, and significantly improving the stability and intelligence level of the system.

[0146] In some embodiments, the docking deviation data further includes: docking angle deviation data;

[0147] The step of calculating the docking deviation data between the target coupler and the main coupler when it is determined that a target coupler exists includes:

[0148] Step S142: Extract target three-dimensional point cloud data based on the latest three-dimensional point cloud data; wherein, the target three-dimensional point cloud data is three-dimensional point cloud data representing the area where the target coupler is located.

[0149] In this embodiment, the target 3D point cloud data can be extracted from the latest 3D point cloud data based on the feature data of the latest 3D point cloud data.

[0150] Specifically, firstly, for each point in the latest 3D point cloud data, a corresponding local feature descriptor (such as curvature, normal vector, FPFH / SHOT, etc.) is calculated. Then, the local feature descriptor of each point is matched with a preset feature template to obtain a similarity score. Next, high-scoring points that best match the feature template can be selected from the latest 3D point cloud data as "seed points." To prevent misselection, seed points can be required to cluster spatially—that is, only when a certain number of high-scoring points appear in a certain region will that region be used as the starting point for subsequent extraction. Then, a feature-based region growing algorithm is executed. For example, each seed point can be used as the center to search for neighboring points along its spatial neighborhood (e.g., radius r=5cm); for each neighboring point, if the similarity between its features and the feature template meets a threshold, it is also marked as a "coupler candidate point" and the growth continues outward; points that do not meet the conditions are not included, while preventing crossing into other objects. Next, all labeled candidate points are subjected to Euclidean clustering, dividing them into several clusters. Clusters that are significantly too large, too small, or too far off-center are removed based on their size, geometry (length, width, and height range), and distance from the cluster center to the expected coupler position. Finally, the cluster that best matches the coupler characteristics (which could be the one with the most points or the highest average similarity) is selected from the remaining clusters. The set of all points corresponding to this cluster is output as the "target 3D point cloud data" for the next step of surface fitting (S144) and deviation calculation (S148).

[0151] In this embodiment, the target 3D point cloud data can also be extracted from the latest 3D point cloud data based on prior knowledge.

[0152] Specifically, this prior knowledge can be a constraint on spatial scope, such as:

[0153] X-axis (track direction): 0.5m to 5m in front of the main body coupler (avoid scanning irrelevant areas too far away);

[0154] Y-axis (perpendicular to the track direction): ±0.5m (covering the lateral swing range of the coupler);

[0155] Z-axis (height direction): 0.2m to 0.8m on the track surface (coupler installation height range).

[0156] Pass-through filtering can be performed on point cloud data that meets spatial constraints to retain point clouds within the ROI.

[0157] Step S144: For the target 3D point cloud data, execute a preset surface fitting algorithm to generate a fitted quadratic surface model.

[0158] Specifically, step S144 may further include:

[0159] First, before executing the preset surface fitting algorithm, a statistical outlier removal (SOR) algorithm can be run again to remove noise points. Reflection intensity correction and data normalization can also be performed (for example, the point cloud coordinate system can be converted to a local coordinate system with the target coupler center as the origin to eliminate the influence of global position offset).

[0160] Then, the pre-defined quadratic surface equation is solved using least squares to obtain the fitted quadratic surface model.

[0161] More specifically, the preset quadratic surface equation can be:

[0162]

[0163] In the formula, , as well as These can be parameters that control the curvature and direction of a surface. and It can be a parameter that controls the tilt of the surface. This is a translation term used to adjust the reference height of the surface. This is understandable. - These are the model parameters to be determined.

[0164] The following error function can be pre-constructed to solve for the parameters to be solved in the above quadratic surface equation.

[0165]

[0166] In the formula, As a weighting coefficient, points in high curvature regions can be assigned higher weights. This represents the number of point clouds.

[0167] Using this error function, the sum of squared vertical distances from each data point to the surface can be defined as the optimization objective.

[0168] In this embodiment, the error function can be solved by converting it into a system of linear equations. For example:

[0169]

[0170] In the formula, Represented as a design matrix, with each row corresponding to a single point. , It is a diagonal weight matrix. For the parameters to be determined, , .

[0171] The optimal parameters can be obtained by solving the linear equation system using QR decomposition or SVD. .

[0172] Step S146: Based on the quadratic surface model, calculate the coordinates of the center point and the direction of the normal vector of the target coupler.

[0173] Specifically, the method of locating the vertex of a surface can be used to calculate the coordinates of the center point of the target coupler. Specifically, utilizing the inherent "vertex" property of a quadratic surface, the overall shape of the surface model is scanned to automatically find the point where the surface's inclination is exactly zero in all directions—that is, the position where the surface is most "flat" and most "convex." This position is then identified as the geometric center of the coupler.

[0174] Alternatively, the centroid projection method of the point cloud can be used to calculate the coordinates of the center point of the target coupler. Specifically, first calculate the overall average position (centroid) of the target point cloud in step S142, then project it onto the surface along the "shortest path" of the fitted surface to find the surface position closest to the centroid. This projected point is both closest to the original point cloud center and falls on the surface, and is regarded as the center of the hook.

[0175] More specifically, firstly, the target point cloud extracted in step S142 can be processed. The centroid is obtained by taking the arithmetic mean:

[0176]

[0177] Then, the center of mass Projecting onto the fitted surface can be done using minimum distance projection:

[0178]

[0179] In the formula, At the center of mass The gradient vector for position calculation, It is an implicit function value.

[0180] Next, the projection point can be obtained using Newton's iteration method. As an approximate "center point coordinate" of the hook head.

[0181] Finally, at the projection point The gradient is calculated and normalized to obtain the direction of the surface normal vector.

[0182] Step S148: Based on the pre-calibrated center point coordinates and normal vector direction of the main coupler and the center point coordinates and normal vector direction of the target coupler, calculate the distance deviation data and docking angle deviation data.

[0183] In this embodiment, the center point of the main coupler and the center point of the target coupler can be directly regarded as two endpoints in space, and the shortest straight-line distance between them can be measured. For example, it can be the three-dimensional Euclidean distance between the center points of the two couplers, which directly represents the spatial position deviation.

[0184] The docking angle deviation data can be the spatial angle between the normal vectors of the two couplers, directly representing the attitude alignment deviation. For example, the dot product of the two unit vectors can be calculated, and the inverse cosine function of the result can be taken to obtain the radian value, which can then be converted into an angle value.

[0185] By introducing quadratic surface fitting and normal vector calculation in steps S142–S148, this implementation not only obtains the three-dimensional distance deviation between the target coupler and the main coupler, but also accurately quantifies the angular deviation between their normal vectors, thereby achieving dual alignment control of attitude and position. Compared with methods that rely solely on distance information, angular deviation data can avoid lateral stress or engagement jamming caused by slight tilting of the coupler head; simultaneously, the smooth model obtained through quadratic surface fitting can suppress point cloud noise interference and improve the stability of normal vector estimation. Overall, when dynamically planning vehicle travel paths and speed classifications, this scheme can adjust the heading and fine-tune the wheelset steering in real time based on the angular deviation, achieving centimeter-level and degree-level precise alignment, significantly reducing the risk of coupler impact or misalignment, and improving the coupler success rate and system reliability.

[0186] In some embodiments, the step of controlling the speed of the first rail vehicle based on the distance deviation data and a preset speed control strategy includes:

[0187] Step S162: If it is determined that there is a target coupler, switch the first speed to the second speed and control the first rail vehicle to travel at a constant speed at the second speed; wherein the second speed is lower than the first speed.

[0188] In this embodiment, the first speed is a pre-set speed, which can be the fastest travel speed of the first rail vehicle during the automated hook-up process. It is understood that the speed of the first rail vehicle can vary during the automated hook-up process. For example, the speed of the first rail vehicle can be controlled according to a strategy of uniform approach-deceleration upon contact-precise hook-up. Alternatively, the speed of the first rail vehicle can be controlled based on the relative distance between the first and second rail vehicles. For example, the closer the distance, the lower the speed of the first rail vehicle (e.g., dynamically adjusted using a PID control algorithm). In this embodiment, the first speed can be 0.4 m / s, 0.5 m / s, or 0.6 m / s, etc.

[0189] In this embodiment, the second speed can be 0.2 m / s or 0.1 m / s, etc. It is understood that, on the one hand, compared to the first speed, the second speed significantly reduces the vehicle's kinetic energy, allowing the first rail vehicle to maintain a more stable and uniform speed when initially approaching the target coupler, reducing overshoot or jitter caused by inertia and ensuring the quality of point cloud acquisition and processing. On the other hand, although the speed is reduced, it is still sufficient to allow the vehicle to approach the target in a relatively short time, providing enough distance buffer for the next stage of fine-tuning at even lower speeds; simultaneously, in the transition range of a few meters to tens of centimeters from the target coupler, the second speed can quickly respond to control commands and achieve precise attitude correction. Moreover, at this speed level, both lidar point cloud sampling and algorithm processing can maintain a high frame rate and stability, avoiding point cloud blurring or omissions caused by excessively fast movement, providing a reliable data foundation for subsequent precise alignment.

[0190] Step S164: Determine whether the distance deviation data has reached the preset speed switching threshold.

[0191] In this embodiment, the preset speed switching threshold can be 0.3m, 0.5m, or 0.2m, etc.

[0192] In this embodiment, the speed switching threshold is a pre-set critical distance value used to trigger phased adjustments to the vehicle speed. When the real-time distance deviation between the couplers reaches this threshold, the driving speed will be automatically reduced to improve docking accuracy and avoid collision risks.

[0193] Step S166: When a preset speed switching threshold is reached, the second speed is switched to the third speed, and the first rail vehicle is controlled to move at the third speed; wherein the third speed is lower than the second speed.

[0194] In this embodiment, the third speed can be 0.05 m / s, 0.06 m / s, or 0.04 m / s, etc.

[0195] In this embodiment, the third speed is the ultra-low speed adopted by the first rail vehicle in the final contact phase, which can be millimeter-level movement, used to achieve precise alignment and eliminate inertial impact.

[0196] In this embodiment, switching the second speed to the third speed and controlling the first rail vehicle to travel at the third speed can mean controlling the first rail vehicle to travel at a constant speed using the third speed.

[0197] In this embodiment, switching from the second speed to the third speed and controlling the first rail vehicle to travel at the third speed can also be done by continuously reducing the travel speed based on the third speed (for example, by fine-tuning the speed through a PID control algorithm).

[0198] In this implementation, the two-stage speed switching method first switches from a higher speed to a medium speed after detecting the target coupler, smoothly approaching the target; when the distance deviation reaches a preset threshold, it then switches to a low speed stage to ensure a smooth contact during the final docking stage. This graded deceleration strategy, starting fast and then slowing down, shortens the overall approach time, improves operational efficiency, and provides sufficient buffering for precise alignment and closing actions, significantly reducing impact and misalignment risks, thus ensuring high precision and a high success rate for coupler docking.

[0199] In some embodiments, based on the distance deviation data and a preset speed control strategy, the travel speed of the first rail vehicle is controlled so that the first rail vehicle approaches the second rail vehicle at a speed not higher than and not equal to the first speed. Furthermore, when the coupler is a James coupler, the method further includes:

[0200] Send a control signal to the power unit of the main coupler to open the main coupler so that the main coupler is in a ready-to-connect state.

[0201] In this embodiment, the power unit of the main coupler can be an electric actuator. For example, it can be driven by a DC motor or a servo motor, converting rotary motion into linear stroke through reduction gears, worm gears, or ball screws to push and pull the coupler. It can also be a hydraulic cylinder system. For example, it can consist of a hydraulic pump, valves, and hydraulic cylinders, using high-pressure oil to drive a piston and generate high thrust. It can also be an electro-hydraulic servo actuator, etc.

[0202] In this embodiment, opening the main body coupler can be achieved by using the aforementioned power device to push the locking pin assembly, causing the locking pin to disengage (lift) from the pin hole, thus releasing the fixed constraint on the hook jaw. At this time, the hook jaw can move freely in subsequent steps. After the locking pin is disengaged, the power device continues to act on the rotating arm (pry bar or connecting rod) of the hook jaw, rotating the hook jaw from the closed position counterclockwise (or clockwise, depending on the mechanism design) to the predetermined "opening" angle.

[0203] The step of closing the hook when it is determined that a preset hook distance has been reached includes:

[0204] Step S182: If it is determined that the preset hook distance has been reached, control the first rail vehicle to contact the second rail vehicle at a speed not exceeding the third speed.

[0205] In this embodiment, controlling the first rail vehicle to contact the second rail vehicle at a speed not exceeding the third speed can be either directly controlling the first rail vehicle to travel at the third speed, or starting with the third speed and continuously decreasing the speed (the closer the distance, the smaller the speed) to control the first rail vehicle to contact the second rail vehicle.

[0206] Step S184: After a touch is detected, three-dimensional point cloud data of the hook area is collected using a preset lidar.

[0207] In this embodiment, the detection of a touch can be achieved by using a preset force sensor to determine whether a touch has occurred. For example, a force sensor can be preset on one side of the coupler body. Alternatively, dynamic analysis can be performed using a preset inertial measurement unit (IMU).

[0208] Step S186: Extract the point cloud voxel block of the locking pin area of ​​the James coupling from the three-dimensional point cloud data of the hook area.

[0209] In a specific implementation plan, the process may include: after the hook is closed, preprocessing the high-density point cloud of the hook area acquired by the lidar to remove environmental noise and unify the resolution. Specifically, the statistical outlier removal (SOR) algorithm is first used to remove outliers with abnormal density, and then radius filtering is used to remove isolated noise points with few surrounding points. Next, voxel mesh filtering is applied to the remaining point cloud to unify the resolution to small cubes with a side length of approximately 2 mm per voxel.

[0210] After preprocessing, and combining the geometric calibration information of the Jan coupler, a candidate region for extracting the locking pin is cropped in the three-dimensional coordinate system at a certain distance (within 50–80 mm) offset below the center of the hook head and within a horizontal radius of approximately 20 mm. Within this region, a set of high-curvature points (points with curvature exceeding approximately 0.05 corresponding to the cylindrical locking pin surface) is quickly identified through curvature analysis. Then, the RANSAC cylindrical fitting algorithm is used, with preset locking pin radius (25–40 mm) and height (80–120 mm) as constraints, and the optimal cylindrical model is determined after multiple iterations. Finally, the cluster of voxels that best matches this cylindrical model is considered the locking pin region, and it is reorganized with a voxel size of 1 cm³, retaining the representative point with the highest reflection intensity in each voxel, outputting an approximately 20×20×20 voxel block that fully covers the geometric information of the locking pin.

[0211] Step S188: Based on the extracted point cloud voxel blocks of the locking pin area and the preset locking pin placement model, perform an initial judgment on whether the hooking was successful.

[0212] In this embodiment, reference voxel models of three locking pin placement states—"fully placed," "partially placed," and "not placed"—can be pre-stored in a template library. Each model is generated from a high-precision point cloud and includes an FPFH (Fast Point Feature Histogram) descriptor to characterize local geometric features.

[0213] For the real-time locking voxel block obtained in step S186, its FPFH features can be extracted first, and a nearest neighbor search can be performed in the template library to calculate the average cosine similarity of all feature vectors. When the similarity is greater than or equal to the similarity threshold (which can be 0.85), it indicates that the real-time voxel and a certain template are highly consistent in geometric distribution, and can be preliminarily determined to be in the corresponding state.

[0214] In one specific implementation, to enhance accuracy, high-scoring candidate voxel blocks can undergo another ICP (Iterative Closest Point) registration. Rigid body transformation is used to ensure close alignment between the real-time voxel and the template, and the root mean square error after registration is checked to be ≤2mm. If both the ICP error and FPFH matching meet the standards, further multi-level verification is performed on geometric and optical characteristics such as the angle between the locking pin axis and the pin hole axis (must be less than 0.5°), the distance between the bottom of the locking pin and the bottom surface of the pin hole (must be less than 1mm), and the average reflection intensity of the locking pin area (must be greater than 80). If all conditions are met, "Initial hooking successful" is output; otherwise, anomaly handling is triggered, such as retrying the docking or triggering an alarm to prompt manual intervention.

[0215] Step S1810: If the initial judgment is that the hooking is successful, execute the preset final judgment strategy to finally determine whether the hooking is successful.

[0216] This implementation method automatically activates the main body coupler to enter a ready-to-hook state during the approach phase, and makes smooth contact at an extremely low speed upon touch. Then, a high-density 3D point cloud scan of the hooking area is performed using LiDAR to accurately extract the locking pin voxel, completing the initial locking pin placement judgment. Finally, a reverse verification or other final judgment strategy is used for final confirmation. This forms a closed-loop control system from opening the hook, micro-speed docking, real-time point cloud detection, and dual verification. This not only ensures smooth and shock-free hooking action but also detects and corrects locking pin misplacement or foreign object interference in real time, significantly improving the hooking success rate and system reliability, minimizing mechanical damage and safety risks, and achieving truly unmanned, all-weather automatic hooking.

[0217] In some implementations, the step of performing an initial determination of whether the hooking was successful based on the extracted point cloud voxel blocks of the locking pin region and a preset locking pin placement model includes:

[0218] Step S1882: Match and compare the point cloud voxel block of the locking pin region with the pre-stored reference point cloud spatial distribution in the locking pin placement model to obtain the spatial alignment relationship between the two.

[0219] In this implementation, firstly, a coarse registration method (such as based on voxel centers or known hook poses) is used to roughly align the two sets of voxels. Then, a fine registration algorithm (iterative nearest point ICP) is run, continuously adjusting translation and rotation parameters in 3D space to achieve optimal overlap between the real-time voxel block and the model voxel block in overall shape. Finally, a set of rigid body transformation matrices is output, which is the "spatial alignment relationship," accurately describing how the real-time point cloud voxels are mapped to the model coordinate system.

[0220] Step S1884: Based on the spatial alignment relationship, extract the point cloud clusters corresponding to the locking pin endpoints from the point cloud voxel blocks, and calculate the center coordinates of the point cloud clusters.

[0221] In this embodiment, firstly, a small cluster of points representing the end of the locking pin (endpoint cluster) can be identified and separated in the transformed voxel, because this cluster is spatially closest to the end position of the locking pin in the template. Then, cluster analysis is performed on this cluster to remove surrounding noise points, retaining only the densest endpoint cluster. Finally, the average position of all points in this cluster is calculated, which is the "center coordinate of the locking pin endpoint," used for subsequent tolerance determination.

[0222] Step S1886: Determine whether the center coordinates of the point cloud cluster fall within the three-dimensional tolerance range of the hook pin hole area calibrated in the locking pin positioning model, and determine whether the absolute value of the vertical coordinate deviation is within the preset vertical deviation range.

[0223] In this embodiment, determining whether the center coordinates of the point cloud cluster fall within the three-dimensional tolerance range of the hook pin hole area calibrated in the locking pin positioning model can be done by checking the deviation distance of the endpoint center relative to the center of the template hole in the hole plane (horizontal XY plane), requiring it to be within a preset horizontal tolerance radius (e.g., ±5mm). Determining whether the absolute value of the vertical coordinate deviation is within a preset vertical deviation range can be done by comparing the height difference of the endpoint centers in the vertical direction (Z-axis), requiring it to fall within the template hole depth tolerance range (e.g., ±3mm).

[0224] Step S1888: If all conditions are met, determine that the locking pin has been correctly positioned and initially determine that the hook has been successfully hooked.

[0225] In this embodiment, by precisely registering the real-time extracted locking pin voxel block with a pre-stored positioning model, and then accurately locating the locking pin endpoint from the aligned data and calculating its spatial center, it is possible to determine whether the locking pin has truly fallen into the pin hole with high-resolution three-dimensional tolerance. This method avoids misjudgments caused by simply relying on overall shape matching. By performing dual tolerance checks on the locking pin endpoint in both the horizontal and vertical directions, it achieves a "fine-grained" initial determination of the hook status, significantly improving the accuracy and reliability of locking pin positioning detection, thereby significantly reducing repetitive operations and safety risks caused by misjudgments.

[0226] In some implementations, the step of executing a preset final judgment strategy to determine whether the hooking was successful after the initial determination is that the hooking was successful includes:

[0227] Step S18102: Control the first rail vehicle to move a preset distance in the opposite direction along the travel direction, and continuously collect three-dimensional point cloud data of the locking pin area through the lidar during the reverse movement.

[0228] In this embodiment, after initially determining that the locking pin has engaged, the control device can automatically issue a slight backward control command, causing the first rail vehicle to move a small, preset distance (from a few centimeters to a fraction of a meter) in the opposite direction of its original travel direction. During this process, the onboard LiDAR continuously collects point cloud data of the locking pin area, ensuring that multiple frames of continuous 3D point cloud data can be acquired regardless of the slightest movement of the vehicle, for subsequent stability analysis. Both the moving distance and speed are strictly controlled to prevent secondary disengagement or data ambiguity caused by excessive speed or large displacement.

[0229] Step S18104: Based on the three-dimensional point cloud data collected during the reverse movement, extract the real-time point cloud cluster corresponding to the locking pin endpoint and calculate its real-time center coordinates.

[0230] In this embodiment, at each moment of the vehicle's reverse micro-movement, the newly acquired point cloud of the locking pin region can be precisely segmented, and the point cloud cluster representing the locking pin endpoints can be extracted. It is understood that this point cloud cluster has already been pre-located within the initially extracted ROI range; therefore, it is only necessary to quickly filter the locking pin endpoint clusters in the latest data using clustering or feature thresholding, and calculate the average position of all points within that cluster—that is, the "real-time center coordinates" at that moment. In this way, a series of spatial positions of the locking pin endpoints updated as the vehicle moves backward can be obtained.

[0231] Step S18106: Based on the continuous change of the real-time center coordinates, determine whether the position of the locking pin meets the preset stability conditions; wherein, the preset stability conditions include: the change of the real-time center coordinates of the locking pin endpoint in the horizontal direction does not exceed the horizontal deviation range; and the change of the vertical direction displacement does not exceed the vertical deviation range.

[0232] In this embodiment, the real-time center coordinates of the locking pin endpoints calculated in multiple time frames (e.g., 5–1 consecutive frames) can be compared pairwise to statistically determine their maximum displacements in the horizontal direction (along the X / Y axes in the track plane) and the vertical direction (Z axis). If none of these maximum displacement values ​​exceed the preset horizontal / vertical deviation thresholds (e.g., horizontal ≤ 1 mm, vertical ≤ 0.5 mm), it is considered that the locking pin endpoints maintain extremely high stability during the reverse micro-motion process—meaning that the locking pin has been firmly engaged in the pin hole without any loosening or retraction.

[0233] Step S18108: If the real-time center coordinates meet the preset stability conditions, the hooking is finally determined to be successful.

[0234] In this embodiment, the control device only confirms successful hooking if and only if the locking pin endpoint continuously meets the aforementioned horizontal and vertical stability conditions throughout the entire reverse micro-motion verification process. At this point, the control device generates a "hooking successful" signal to notify the dispatcher and allows the vehicle to release the hook brake and enter train formation operation. If the stability conditions are not met, an abnormal handling process is triggered, such as an alarm prompt or a re-attempt to hook the vehicle. This multi-frame timing and micro-motion verification final judgment mechanism effectively eliminates false locking pin phenomena caused by vibration, impact, or external forces, greatly improving the reliability and safety of automatic hooking.

[0235] After initially determining "successful hooking," this final judgment strategy dynamically tracks the positional changes of the lock pin's endpoint center by slightly reversing the vehicle and continuously collecting point cloud data of the lock pin area. Only when the displacement fluctuations in both the horizontal and vertical directions remain within a preset small tolerance range is the hooking finally confirmed. This multi-frame temporal stability verification can promptly capture any risks of lock pin retraction or loosening caused by vibration, inertia, or external forces, avoiding false "successful hooking" determinations and significantly improving the reliability and safety of the entire automatic hooking process.

[0236] like Figure 2 As shown, this embodiment provides an automatic uncoupling method for rail vehicles, including:

[0237] Step S20: In response to the received uncoupling command, a control signal is sent to the power unit of the main coupler of the first rail vehicle to drive the power unit to perform the coupler unlocking operation; wherein, the main coupler and the target coupler are in a hooked state; wherein, the first rail vehicle uses an automatic hooking method for rail vehicles as described in the above embodiments to perform the hooking operation between the main coupler and the target coupler, so that the main coupler and the target coupler are in a hooked state.

[0238] Step S22: During the coupler unlocking operation, the three-dimensional point cloud data of the coupler area is collected by the preset lidar, and the real-time point cloud voxel block of the locking pin area is extracted.

[0239] Step S24: Perform a matching comparison based on the real-time point cloud voxel block and the preset lock pin not in place model to obtain the matching comparison result; wherein, the lock pin not in place model includes the baseline point cloud distribution characteristics after the lock pin is pulled out.

[0240] In this embodiment, the pin-not-placed model can be a reference model of the point cloud and geometric features corresponding to the state where the pin is lifted but has not yet fallen into the pin hole after the unlocking operation is performed. Specifically, the model can record the typical vertical offset of the pin relative to the hook reference plane when it is in the released position, such as the absolute height range after rising several centimeters from the depth of the hole in the normal placement position. After the pin is pushed out of the pin hole, its rod and end cap are not obscured by the hook hole wall, so the complete cylindrical surface shape and clear end face can be presented in the point cloud. The model stores the radius, height, and end face shape of this exposed cylinder in voxel or mesh form. Unlike when the pin end is almost at the center of the hole in the normal placement position, the pin-not-placed model can preset the horizontal and vertical offset vectors of the pin after unlocking, helping the control device to quickly locate the pin position area after being lifted. To support fast matching, the model also includes local geometric descriptors (such as FPFH histograms or SHOT features) of the exposed surface of the lock pin and the side of the cylinder, which can accurately reflect the normal vector distribution and curvature pattern of the exposed lock pin.

[0241] Step S26: Based on the matching comparison results, determine whether the locking pin has been successfully pulled out.

[0242] Step S28: If the locking pin is successfully pulled out, control the first track vehicle to move in the opposite direction until the laser radar detects that the hook separation distance is greater than the preset safe distance threshold, and then confirm that the hook removal is completed.

[0243] In this embodiment, while activating the unlocking mechanism, a laser radar is used to scan the hook area in real time and extract the locking pin voxel. By accurately matching it with the "not in place" model, the pin removal status is determined, effectively avoiding mechanical jamming caused by incomplete unlocking. If it is confirmed that the locking pin has been removed, the vehicle can be safely reversed until a sufficient separation distance is reached before the process ends. The entire process requires no manual intervention, ensuring the reliability and accuracy of the hook removal action, and greatly improving operational safety and system automation through early warning and safe distance threshold mechanisms.

[0244] In some implementations, the step of determining whether the locking pin has been successfully removed based on the matching comparison result includes:

[0245] Step S262: If the point cloud voxel block of the locking pin area does not cover the predefined locking pin feature area in the locking pin not in place model, or if the center coordinate of the point cloud cluster corresponding to the locking pin endpoint deviates from the center of the hook hole by more than the preset unhooking tolerance range, then it is determined that the locking pin has been successfully pulled out.

[0246] In this embodiment, the real-time extracted lock pin region voxel blocks can be mapped to the feature region framework defined in the lock pin not in place model. This feature region can correspond to the spatial position that the lock pin end may occupy after unlocking. If almost no real-time voxels are detected within this model feature region (i.e., the voxel block does not cover the region), it indicates that the lock pin has completely exited its position.

[0247] In this embodiment, the point cloud cluster of the locking pin endpoint can also be identified from real-time voxels, and the geometric center coordinates of the cluster can be calculated. This center is then compared with the center of the hook hole. If the deviation in either the horizontal or vertical direction exceeds a pre-set unhooking tolerance threshold (e.g., ±10mm), it indicates that the locking pin end has moved out of the hole boundary.

[0248] Step S264: If the center coordinates of the point cloud cluster corresponding to the locking pin endpoint are still within the preset unhooking tolerance range of the hook hole area, it is determined that the unhooking is not completed and an unlocking abnormal alarm signal is generated.

[0249] According to an embodiment of the present invention, an electronic device is provided; please refer to... Figure 3The electronic device in this embodiment may include one or more of the following components: a processor, a network interface, memory, non-volatile memory, and one or more application programs, wherein the one or more application programs may be stored in non-volatile memory and configured to be executed by one or more processors, and the one or more programs are configured to perform the methods as described in the foregoing method embodiments.

[0250] According to embodiments of the present invention, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a computer, causes the computer to perform the method described in any of the above embodiments.

[0251] According to embodiments of the present invention, a computer program product comprising instructions is also provided, which, when executed by a computer, cause the computer to perform a method in any of the above embodiments.

[0252] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0253] Optionally, specific examples in this embodiment can refer to the examples described in the above embodiments, and will not be repeated here.

[0254] The sequence numbers of the embodiments in this application are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0255] In the above embodiments of this application, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0256] The above description is only a preferred embodiment of this application. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of this application, and these improvements and modifications should also be considered within the scope of protection of this application.

Claims

1. An automatic coupling method for rail vehicles, characterized in that, include: The first rail vehicle is controlled to travel at a preset first speed, and during the travel, three-dimensional point cloud data in front of the first rail vehicle is continuously collected by a preset lidar; wherein, the front of the first rail vehicle refers to the front of the main body coupler of the first rail vehicle; the lidar is installed on the first rail vehicle and is on the same side as the main body coupler. Based on the three-dimensional point cloud data, it is determined whether a target coupler exists; wherein, the target coupler is configured on one side of the main coupler of the second rail vehicle relative to the first rail vehicle. If a target coupler is determined to exist, the docking deviation data between the target coupler and the main coupler is calculated; wherein, the docking deviation data includes at least the docking distance deviation data; Based on the distance deviation data and the preset speed control strategy, the travel speed of the first rail vehicle is controlled so that the first rail vehicle approaches the second rail vehicle at a speed that is not higher than and not equal to the first speed. If the preset hook distance is reached, the hook will be closed. The step of controlling the travel speed of the first rail vehicle based on the distance deviation data and the preset speed control strategy includes: If a target coupler is detected, the first speed is switched to the second speed, and the first rail vehicle is controlled to travel at a constant speed at the second speed; wherein the second speed is lower than the first speed. Determine whether the distance deviation data has reached the preset speed switching threshold; When a preset speed switching threshold is reached, the second speed is switched to the third speed, and the first rail vehicle is controlled to move at the third speed; wherein the third speed is lower than the second speed; Based on the aforementioned distance deviation data and a preset speed control strategy, the travel speed of the first rail vehicle is controlled so that the first rail vehicle approaches the second rail vehicle at a speed not higher than and not equal to the first speed. Furthermore, when the coupler is a James coupler, the method further includes: Send a control signal to the power unit of the main coupler to open the main coupler so that the main coupler is in a ready-to-connect state. The step of closing the hook when it is determined that a preset hook distance has been reached includes: If it is determined that the preset hooking distance has been reached, the first rail vehicle is controlled to make contact with the second rail vehicle at a speed not exceeding the third speed. After a touch is detected, three-dimensional point cloud data of the hook area is collected using a pre-set lidar. For the three-dimensional point cloud data of the hook area, extract the point cloud voxel block of the locking pin area of ​​the James coupler; Based on the extracted point cloud voxel blocks of the locking pin area and the preset locking pin placement model, an initial judgment on whether the hooking was successful is performed. If the initial assessment indicates successful hooking, a pre-defined final judgment strategy is executed to determine whether the hooking was ultimately successful.

2. The method according to claim 1, characterized in that, The step of determining whether a target coupler exists based on the three-dimensional point cloud data includes: For the three-dimensional point cloud data, feature extraction is performed to obtain feature data used to characterize the coupler features; Based on the feature data used to characterize the coupler features, it is determined whether there is a target coupler in front of the main coupler.

3. The method according to claim 1, characterized in that, The docking deviation data also includes: docking angle deviation data; The step of calculating the docking deviation data between the target coupler and the main coupler when it is determined that a target coupler exists includes: Based on the latest 3D point cloud data, target 3D point cloud data is extracted; wherein, the target 3D point cloud data is 3D point cloud data representing the area where the target coupler is located; For the target 3D point cloud data, a preset surface fitting algorithm is executed to generate a fitted quadratic surface model; Based on the quadratic surface model, calculate the coordinates of the center point of the target coupler and the direction of its normal vector. Based on the pre-calibrated center point coordinates and normal vector direction of the main coupler and the center point coordinates and normal vector direction of the target coupler, the distance deviation data and docking angle deviation data are calculated.

4. The method according to claim 1, characterized in that, The step of performing an initial judgment on whether the hooking was successful, based on the extracted point cloud voxel blocks of the locking pin region and a preset locking pin placement model, includes: The point cloud voxel blocks of the locking pin area are matched and compared with the pre-stored reference point cloud spatial distribution in the locking pin placement model to obtain the spatial alignment relationship between the two. Based on the spatial alignment relationship, the point cloud clusters corresponding to the locking pin endpoints are extracted from the point cloud voxel blocks, and the center coordinates of the point cloud clusters are calculated. Determine whether the center coordinates of the point cloud cluster fall within the three-dimensional tolerance range of the hook pin hole area calibrated in the locking pin positioning model, and determine whether the absolute value of the vertical coordinate deviation is within the preset vertical deviation range. If all conditions are met, it is determined that the locking pin has been correctly positioned, and the hook is initially determined to be successfully hooked.

5. The method according to claim 1, characterized in that, The step of executing a preset final judgment strategy to determine whether the hooking was successful after the initial judgment is successful includes: The first rail vehicle is controlled to move in the opposite direction of travel by a preset distance, and during the reverse movement, the lidar continuously collects three-dimensional point cloud data of the locking pin area. Based on the three-dimensional point cloud data collected during the reverse movement, the real-time point cloud clusters corresponding to the locking pin endpoints are extracted, and their real-time center coordinates are calculated. Based on the continuous change in the real-time center coordinates, it is determined whether the position of the locking pin meets the preset stability conditions; wherein, the preset stability conditions include: the change in the real-time center coordinates of the locking pin endpoint in the horizontal direction does not exceed the horizontal deviation range; and the change in the vertical direction does not exceed the vertical deviation range. If the real-time center coordinates meet the preset stability conditions, the hooking is ultimately determined to be successful.

6. An automatic uncoupling method for rail vehicles, characterized in that, include: In response to the received uncoupling command, a control signal is sent to the power unit of the main coupler of the first rail vehicle to drive the power unit to perform the coupler unlocking operation; wherein the main coupler and the target coupler are in a hooked state; wherein the first rail vehicle uses an automatic hooking method for rail vehicles according to any one of claims 1-5 to perform the hooking operation between the main coupler and the target coupler, so that the main coupler and the target coupler are in a hooked state. During the coupler unlocking operation, a preset lidar is used to collect three-dimensional point cloud data of the coupler area and extract real-time point cloud voxel blocks of the locking pin area. Matching and comparison are performed based on real-time point cloud voxel blocks and a preset lock pin not in place model to obtain the matching and comparison results; wherein, the lock pin not in place model includes the baseline point cloud distribution features after the lock pin is pulled out. Based on the matching and comparison results, it is determined whether the locking pin has been successfully pulled out; If the locking pin is successfully removed, the first track vehicle is controlled to move in the opposite direction until the lidar detects that the hook separation distance is greater than the preset safe distance threshold, and then the hook removal is confirmed to be completed.

7. The method according to claim 6, characterized in that, The step of determining whether the locking pin has been successfully removed based on the matching comparison result includes: If the point cloud voxel block of the locking pin area does not cover the predefined locking pin feature area in the locking pin not in place model, or if the center coordinate of the point cloud cluster corresponding to the locking pin endpoint deviates from the center of the hook hole by more than the preset unhooking tolerance range, then it is determined that the locking pin has been successfully pulled out. If the center coordinates of the point cloud cluster corresponding to the locking pin endpoint are still within the preset unhooking tolerance range of the hook hole area, it is determined that the unhooking is not completed and an unlocking abnormal alarm signal is generated.

8. An electronic device, characterized in that, include: A memory, and one or more processors communicatively connected to the memory; The memory stores instructions that can be executed by the one or more processors to cause the one or more processors to implement the method as described in any one of claims 1 to 5.

Images