A highly redundant autonomous alignment method for unmanned container trucks under a port quay crane

Abstract

The embodiment of the present application provides a highly redundant autonomous alignment method for unmanned container trucks under a port quay crane. The method is applied to the on-board terminal of an unmanned vehicle. The unmanned vehicle accepts the target location task and autonomously moves to the vicinity of the target location; it determines whether the point cloud information output by the first laser radar can be received normally; if all are abnormal, the vehicle stops, and the quay crane alignment system intervenes in the alignment, and the parking guidance value is calculated to guide the unmanned container truck to align; if all are normal, the operator chooses whether the quay crane alignment system needs to intervene in the alignment; when the quay crane alignment system needs to intervene in the alignment, the first laser radar is selected to identify the relative position of the quay crane lifting equipment and the unmanned vehicle to calculate the parking distance, and the parking is completed directly under the guidance of the parking distance. The present invention can be effectively combined with the quay crane alignment system, adapt to the situation when the laser hardware of the unmanned vehicle fails, and ensure the parking alignment accuracy under various problems.

Description

Technical Field

[0001] The present application relates to the field of unmanned driving technology, and in particular to a highly redundant autonomous alignment method for unmanned container trucks under a port quay crane. Background Art

[0002] Nowadays, many ports have introduced unmanned container trucks to carry out operations in the process of transformation to automation, which has greatly improved port operation efficiency while saving labor costs.

[0003] In the prior art, the quay crane parking method for unmanned vehicles often relies on the laser hardware equipment on the unmanned container truck to output point cloud data, and based on the point cloud data, it issues instructions on specific moving distances and positions to the hoisting equipment operating on the shore. The accuracy of its position directly affects the efficiency and safety of ship loading and unloading.

[0004] Therefore, the method of relying solely on the single conventional point cloud data output by the laser equipment for alignment often cannot meet the complex conditions of unmanned ports. Once the laser hardware on the unmanned container truck cannot output point cloud data or the output point cloud data is abnormal or the weather is too bad to ensure accuracy through point cloud data alone, the unmanned container truck can only be forced to stop and wait for maintenance or further instructions. The overall task execution speed will also decrease, so its adaptability and scalability are poor.

[0005] Therefore, how to improve the accuracy and efficiency of quay crane parking detection for unmanned vehicles in unmanned ports has become a technical problem that needs to be solved urgently. Summary of the invention

[0006] The embodiments of the present application provide a highly redundant autonomous positioning method for an unmanned container truck under a port quay crane, which can be effectively combined with a quay crane positioning system to adapt to the parking control effect when the laser hardware of the unmanned vehicle fails.

[0007] Other features and advantages of the present application will become apparent from the following detailed description, or may be learned in part by the practice of the present application.

[0008] A highly redundant autonomous alignment method for unmanned container trucks under a port quay bridge is applied to an on-board terminal on an unmanned vehicle, wherein the on-board terminal is communicatively connected to a first laser radar; the first laser radar is arranged on the top of the unmanned vehicle; the on-board terminal performs parking control on the unmanned vehicle according to point cloud data sent by the first laser radar;

[0009] The method comprises the following steps:

[0010] S1: The unmanned vehicle accepts the target location mission and autonomously moves to the vicinity of the target location;

[0011] S2: The on-board terminal determines whether the laser hardware on the unmanned vehicle is working properly and confirms whether the point cloud information output by the first laser radar can be received normally; if both are not normal, the vehicle stops, the quay crane alignment system intervenes in the alignment, and the parking guidance value is calculated to guide the unmanned container truck to align;

[0012] If all are normal, the operator will choose whether the quay crane alignment system needs to intervene in the alignment; if the quay crane alignment system needs to intervene in the alignment, the operator will choose to use the first laser radar to identify the relative position between the quay crane crane and the unmanned vehicle to calculate the parking distance, and directly complete the parking under the guidance of the parking distance;

[0013] After parking, the point cloud data of the unmanned vehicle is obtained through the laser hardware installed on the quay crane, and the current parking position of the unmanned vehicle is calculated by the quay crane alignment system. If the parking error threshold is met, the alignment is terminated. If the parking error threshold is not met, the parking guidance value calculated by the quay crane alignment system is used to guide the unmanned vehicle to align, and the parking distance under the first laser radar scan is calculated based on the alignment position and calibrated as the parking reference distance;

[0014] When it is selected that the quay crane alignment system does not need to intervene in the alignment, the first laser radar is directly selected to identify the relative position between the quay crane crane and the unmanned vehicle to obtain the parking distance, and parking is directly completed under the guidance of the parking distance;

[0015] S3: After reporting that the vehicle has completed autonomous alignment and the bridge equipment has confirmed that it is correct, loading / unloading operations are carried out.

[0016] As a further improvement, in step S2, when the first laser radar is selected to identify the relative position between the quay crane and the unmanned vehicle to obtain the parking distance, the following steps are specifically included:

[0017] When the unmanned vehicle arrives at the target location and the quay crane equipment is in place, the crane's hoist can be used to directly load or unload containers at the target location; the target location is located in the target area close to the quay crane equipment when the unmanned vehicle is moving towards the quay crane equipment. When the unmanned vehicle autonomously drives to the target area, the unmanned vehicle enters the alignment and parking process. During the alignment and parking process, the unmanned vehicle can identify the front beam of the quay crane equipment in real time, and the unmanned vehicle moves in a straight line during the entire alignment and parking process;

[0018] The random sampling consistency method and the prior normal vector are used to perform plane segmentation, and the bottom surface point cloud set of the front beam of the quay crane equipment identified by the first laser radar is determined. The distance to the unmanned vehicle is calculated according to the coordinate information of the bottom surface point cloud set of the front beam to obtain the parking reference distance.

[0019] As a further improvement, when the first laser radar is selected to identify the relative position between the quay crane and the unmanned vehicle to obtain the parking distance, the specific steps include:

[0020] Establish a coordinate system, in which the traveling direction of the unmanned vehicle during the parking process is set as the X-axis, the direction perpendicular to the X-axis and parallel to the ground is set as the Y-axis, and the direction perpendicular to the X-axis and perpendicular to the ground is set as the Z-axis;

[0021] The first straight-pass filter is set to filter the point cloud set. The range value on the X-axis is determined by the difference between the position information of the quay crane crane after it is in place through the target location task and the current position of the unmanned vehicle. The range value on the Y-axis is determined according to the experience value of the width of multiple lanes under the quay crane. The range value on the Z-axis is determined according to the experience value of the front beam height.

[0022] The plane is segmented by using the random sampling consistency method and the prior normal vector. For the point cloud data collected by the first laser radar, the invalid points including the laser points of the camera on the quay bridge and the side of the bridge are eliminated, and then the first point cloud set is obtained by the range value of the first straight-pass filter.

[0023] The point cloud data of the sling when the sling is flush with the front beam is eliminated by Euclidean clustering, and the point cloud data of the sling in the first point cloud set is further eliminated to obtain the point cloud set of the bottom surface of the front beam;

[0024] Project the point cloud set on the bottom surface of the front beam into two-dimensional data, and obtain the X value of the two-dimensional data. At this time, the minimum value comes from the point on the bottom of the front beam closest to the first laser device. Through this point, traverse all points to find the row of points on the bottom of the front beam closest to the unmanned vehicle;

[0025] The set of points found is fitted with a least squares fitting algorithm to obtain a straight line representing the bottom of the front beam closest to the unmanned vehicle, that is, the bottom edge of the rear side of the front beam, and the distance value B1 between the unmanned vehicle and the straight line is obtained, and the parking distance is calculated through B1.

[0026] As a further improvement, the method further comprises the following steps:

[0027] A second straight-through filter is set to filter the point cloud set. The range of the second straight-through filter is determined according to B1, the height difference between the top surface and the bottom surface of the front beam on the Z axis, and the width difference on the Y axis. The point cloud set of the rear side of the front beam is obtained. After plane segmentation, all X values ​​in the point cloud set are averaged to obtain B2.

[0028] The third straight-through filter is set to filter the point cloud set. It is known that the distance between the rear side surface of the front beam and the front side surface of the rear beam is B3. According to B1, B3, and the height difference between the upper top surface and the lower bottom surface of the rear beam on the Z axis and the width difference on the Y axis, the range of the third straight-through filter is determined to obtain the point cloud set of the front side surface of the rear beam, and the average value of all X values ​​in the point cloud set is calculated to obtain B4;

[0029] The stopping distance B is (B1+B2+B3+B4) / 3.

[0030] As a further improvement, when the unmanned vehicle autonomously drives to the target area, the unmanned vehicle enters the alignment and parking process. During the alignment and parking process, the unmanned vehicle can identify the front beam of the quay crane in real time. When the unmanned vehicle moves to the threshold range directly below the front beam, it is determined whether the first laser radar outputs point cloud data normally;

[0031] The judgment method is as follows: after obtaining the point cloud set of the bottom surface of the front beam, the point cloud set of the bottom surface of the front beam is further classified by the second Euclidean clustering to form multiple side-by-side clusters, and the judgment is made by the number of clusters or the difference in the number of laser point clouds between any two different clusters;

[0032] If it is judged to be normal, the point cloud data of the current first laser radar is output; if it is judged to be abnormal, the point cloud data of the first laser radar is processed abnormally and re-output.

[0033] As a further improvement, the threshold of the second Euclidean clustering is set to 0.3m, and the minimum number of each cluster is four; when the number of clusters is less than 12, or the difference in the number of laser point clouds between any two different clusters is greater than 15, it is judged as an abnormal situation.

[0034] As a further improvement, if an abnormality is determined, the point cloud data of the first laser radar is processed for abnormality and re-output, specifically including:

[0035] The X values ​​of all points in the point cloud set of the bottom surface of the front beam after the first Euclidean clustering are averaged to obtain Q, and the fourth direct filter is set to filter the point cloud set. According to the width W of the front beam, the reserved margin distance E, and the distance R from the rear side of the front beam to the front side of the rear beam, the range value of the second direct filter on the X axis is determined to be (QWRE, Q-R+E), the range value on the Y axis is determined according to the empirical value of the width of multiple lanes under the quay crane, and the range value on the Z axis is determined according to the empirical value of the height of the rear beam. The point cloud set of the front side of the rear beam is obtained through the range value of the fourth direct filter; the average value of all X values ​​in the point cloud set of the front side of the rear beam is obtained to obtain B5, and the parking distance B=R+B5.

[0036] As a further improvement, when the quay crane positioning system intervenes in the positioning, after the unmanned vehicle stops, the quay crane positioning system waits for 1s to 3s, and finally collects multiple quay crane guidance parking distance information and executes the average value of its parking distance.

[0037] As a further improvement, when all are abnormal, the container truck will stop, and the quay crane alignment system will intervene in the alignment, and the parking guidance value will be calculated to guide the unmanned container truck to align, including:

[0038] When the quay crane equipment is in place, the unmanned vehicle will autonomously drive to the quay crane position / yard bay target area, and brake and stop when the unmanned vehicle is within 2m of the destination point issued by the TMC system;

[0039] The quay crane alignment system outputs the quay crane guidance parking distance information for more than 7 times after the unmanned vehicle stops for 1.5 seconds, and judges the validity of the parking guidance value output by the quay crane alignment system. If it is valid, the average value of the parking distance of the unmanned container truck guidance alignment controller is used for the first execution until the vehicle stops;

[0040] The quay crane alignment system again outputs the average value of the quay crane guidance parking distance information of more than 7 times after the unmanned vehicle stops for 1.5s, and determines whether the unmanned vehicle is within the parking threshold range after the first execution of its parking distance average value. If it is within the range, it reports to the TMC horizontal transportation system that the container truck alignment is completed. Otherwise, the unmanned container truck guidance alignment controller is used to execute the average value output again until the unmanned vehicle is within the parking threshold range and then reports the alignment completion status to complete the container truck autonomous alignment process.

[0041] As a further improvement, the vehicle terminal is further connected to the two second laser radars for communication, and the two second laser radars are respectively arranged on both sides of the unmanned vehicle; the vehicle-mounted terminal performs parking control on the unmanned vehicle according to the point cloud data sent by the second laser radar, characterized in that the target parking distance is obtained by subtracting the box offset distance from the parking distance, specifically including:

[0042] Establish a coordinate system, in which the traveling direction of the unmanned vehicle during the parking process is set as the X-axis, the direction perpendicular to the X-axis and parallel to the ground is set as the Y-axis, and the direction perpendicular to the X-axis and perpendicular to the ground is set as the Z-axis;

[0043] A fifth direct filter is set to filter the point cloud set, and a range value of the fifth direct filter is determined according to different types of box delivery tasks in the target location task;

[0044] For the point cloud data collected by the second laser radar, after removing invalid points, a point cloud set to be processed is obtained by using the range value of the fifth pass filter;

[0045] Using a random sampling consistency method and a priori normal vector to perform plane segmentation, determine a point cloud set of the front side of the vehicle container near the front of the vehicle identified by the second laser radar;

[0046] Take the average of the X values ​​of the point cloud set to get the box offset distance S;

[0047] The target stop distance is further compensated for the test calibration value to obtain the target precise stop distance, which specifically includes:

[0048] When the quay crane crane equipment at the target location is in place and the task is to deliver containers; in the case of test calibration, when the parking of the unmanned vehicle is completed, the test calibration value is the relative distance between the container on the vehicle and the quay crane crane equipment;

[0049] When the quay crane hoisting equipment at the target location is in place and the task is to collect containers; in the test calibration case, when the parking of the unmanned vehicle is completed, the test calibration value is the relative distance between the unmanned vehicle and the quay crane hoisting equipment.

[0050] In the present application, the overall method is designed with high redundancy and can select the optimal alignment method under different data conditions. On the one hand, when the point cloud data of the first laser radar may be erroneous due to bad weather or when the unmanned container truck performs the task of the quay crane for the first time and needs to determine a correction reference value through the initial test alignment process, the laser hardware of the unmanned container truck is combined with the laser hardware of the quay crane crane to perform alignment. In this case, although the laser hardware of the quay crane crane needs to be stopped for identification first, which will affect the alignment speed, it can ensure higher alignment accuracy. On the other hand, when the external environment When it is normal or the vehicle-mounted terminal stores the correction reference value obtained from the initial test alignment, the unmanned container truck’s laser hardware is used for real-time recognition, and the alignment is completed directly in one go without stopping the vehicle, which can effectively improve the overall alignment efficiency. At the same time, the feedback of the correction reference value obtained during the initial test can further ensure the accuracy requirements under the premise of meeting the alignment efficiency. In addition, when the laser hardware of the unmanned container truck cannot work normally, the laser hardware of the quay crane lifting equipment is used for alignment to ensure that the entire alignment process can be carried out smoothly. Compared with the existing quay crane alignment system, the overall solution has stronger adaptability and scalability.

[0051] It should be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] The drawings herein are incorporated into the specification and constitute a part of the specification, showing embodiments consistent with the present application, and together with the specification, are used to explain the principles of the present application. Obviously, the drawings described below are only some embodiments of the present application, and for ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work. In the drawings:

[0053] Figure 1 A schematic flow chart of a method for highly redundant autonomous alignment of unmanned container trucks under a port quay crane according to an embodiment of the present application is shown;

[0054] Figure 2 A schematic diagram showing the situation in which the tractor head of the unmanned vehicle moves between the front beam and the rear beam according to an embodiment of the present application;

[0055] Figure 3 The figure shows the load status of the unmanned vehicle under different box delivery tasks according to the embodiment of the present application;

[0056] Figure 4 A schematic diagram of the structure of a computer system suitable for implementing an electronic device of an embodiment of the present application is shown. DETAILED DESCRIPTION

[0057] Example embodiments will now be described more fully with reference to the accompanying drawings. However, example embodiments can be implemented in a variety of forms and should not be construed as limited to the examples set forth herein; rather, these embodiments are provided so that this application will be more comprehensive and complete and fully convey the concept of the example embodiments to those skilled in the art.

[0058] In addition, described feature, structure or characteristic can be combined in one or more embodiments in any suitable manner. In the following description, many specific details are provided to provide a full understanding of the embodiments of the present application. However, those skilled in the art will appreciate that the technical scheme of the present application can be put into practice without one or more of the specific details, or other methods, components, devices, steps, etc. can be adopted. In other cases, known methods, devices, realizations or operations are not shown or described in detail to avoid blurring the various aspects of the application.

[0059] The block diagrams shown in the accompanying drawings are merely functional entities and do not necessarily correspond to physically independent entities. That is, these functional entities may be implemented in software form, or in one or more hardware modules or integrated circuits, or in different networks and / or processor devices and / or microcontroller devices.

[0060] The flowcharts shown in the accompanying drawings are only exemplary and do not necessarily include all the contents and operations / steps, nor must they be executed in the order described. For example, some operations / steps can be decomposed, and some operations / steps can be combined or partially combined, so the actual execution order may change according to actual conditions.

[0061] Figure 1 A flow chart of a highly redundant autonomous positioning method for an unmanned container truck under a port quay crane according to an embodiment of the present application is shown. The method can be applied to an on-board terminal on an unmanned vehicle, wherein the on-board terminal is communicatively connected to a first laser radar; the first laser radar is disposed on the top of the unmanned vehicle; the on-board terminal performs parking control on the unmanned vehicle based on point cloud data sent by the first laser radar.

[0062] like Figure 2 As shown, the parking standard in this case is to meet the operational requirements. When the quay crane equipment is in place, the unmanned vehicle can rely on the position of a certain part of the quay crane lifting equipment to complete the parking process, and determine its specific parking position according to the lane and related parameters of each quay crane, so that the quay crane lifting equipment and the unmanned vehicle can complete the loading and receiving of containers only through the lifting equipment after parking (before the quay crane lifting equipment is in place, it is also parked according to the position of the container to be placed on the ship or the position of the container to be received on the ship, and it also moves to the above position during the lifting process; therefore, the unmanned vehicle in the following text can use the relevant position of a certain part of the quay crane lifting equipment as an intermediate reference to achieve alignment with the target position).

[0063] The coordinate origin in this case is the center of the rear axle of the tractor head of the unmanned vehicle; for the sake of production operation safety, the vehicle needs to be straightened in advance when the quay crane is in alignment, so the alignment process or 20m before alignment is a straight line condition, and no nonlinear irregular route will appear; the prior experience in the article refers to the data obtained by actual vehicle testing, and corresponding operations can be performed by discovering the commonalities of the data.

[0064] Please refer to Figure 1 , the method comprises the following steps:

[0065] S1: The unmanned vehicle accepts the target location mission and autonomously moves to the vicinity of the target location;

[0066] The task includes the destination location / container type (40 feet or 20 feet) / container collection or delivery task / container location (front, middle, and back). Due to the error of the quay crane when moving, it is not very accurate, and the vehicle GPS positioning itself has a certain error, so the destination location and the current unmanned vehicle position cannot be directly used to stop and complete the loading and unloading of containers, but it can be used to determine the approximate location of the vehicle after arriving at the destination to assist in selecting laser point data;

[0067] S2: The vehicle terminal determines whether the laser hardware on the unmanned vehicle is working properly and confirms whether the point cloud information output by the first laser radar can be received normally. It can also further confirm whether the following test calibration value and correction reference value are normally imported as correction parameters. The reason why the point cloud information output by the first laser radar is used for alignment first is that the main feature point of the quay crane equipment that is easier to identify is the top beam. The point cloud data of the top beam is used as the calculation source, which has a higher accuracy rate.

[0068] If all are abnormal, the vehicle will stop and the quay crane alignment system will intervene in the alignment. First, the unmanned vehicle will brake and stop within 2m of the target location. After it stops, it will stably receive the parking guidance value issued by the quay crane alignment system, that is, the laser device on the quay crane will emit a laser to identify the unmanned container truck and output the corresponding point cloud data. Then, the parking guidance value will be initialized to false. After that, it will be confirmed in turn whether the quay crane is in place, whether the container truck is in the effective area of ​​the quay crane, and whether the vehicle position is within the execution threshold (the execution threshold range represents the real-time distance range between the unmanned vehicle and the target location with the center of the rear axle of the tractor head as the origin, and -1.5~ Within 10m, since the first laser radar can identify the quay crane and output point cloud data when it is about 20m away from the quay crane, the vehicle will slow down and enter the alignment process within 10m; therefore, in order to ensure the operating efficiency, the vehicle does not slow down when the distance is 10m-20m; and greater than -1.5m is because there is no relevant perception sensor installed behind the vehicle, so there is a certain risk of long-distance reversing). If all are satisfied, the parking guidance value is initialized to true, and then the parking guidance value is calculated through the point cloud data to guide the unmanned container truck to align, and it is determined whether the distance between the parking position and the target location is within the completion threshold of 0.05m. If not, it returns to the step of initialization to false again. If yes, the alignment is completed; since this step is directly aligned through the quay crane alignment system and does not use the laser radar, there is no need to correct the parameters to speed up the overall alignment efficiency;

[0069] If all are normal, the operator will choose whether the quay crane alignment system needs to intervene in the alignment; when encountering bad weather or when the unmanned container truck performs the task of the quay crane for the first time, a correction reference value needs to be determined through the initial test alignment process, and the quay crane alignment system needs to intervene in the alignment, the parking distance is initialized to false at first, and then the parameters of the parking distance are corrected (the first alignment can only be corrected by testing the calibration value. After the correction reference value is obtained in the first alignment, the parameters can be corrected by the correction reference value before each subsequent alignment). After that, it is confirmed in turn whether the quay crane is in place, whether the container truck is in the effective area of ​​the quay crane, and whether the vehicle position is within the execution threshold. If all are met, the parking distance is initialized to true, and the first laser radar is selected to identify the relative position of the quay crane lifting equipment and the unmanned vehicle to calculate the parking distance, and the parking is completed directly under the guidance of the parking distance;

[0070] After parking, the point cloud data of the unmanned vehicle, i.e., the unmanned container truck, is obtained through the laser hardware installed on the quay crane, and the current parking position of the unmanned vehicle is calculated through the quay crane alignment system. If the parking error threshold is met, the alignment is terminated. If the parking error threshold is not met, the unmanned vehicle is guided to align through the parking guidance value calculated by the quay crane alignment system, and the correction reference value is calculated through the difference between the parking guidance value and the parking distance for subsequent selection to use the first laser radar to identify the relative position of the quay crane lifting equipment and the unmanned vehicle. After the parking distance is calculated, the correction reference value is directly added or subtracted to obtain the optimized parking distance for execution;

[0071] When it is selected that the quay crane alignment system does not need to intervene in the alignment, the first laser radar is directly selected to identify the relative position of the quay crane lifting equipment and the unmanned vehicle to obtain the parking distance. Initially, the parking distance is initialized to false, and then the parameters of the parking distance are corrected. After that, it is confirmed in turn whether the quay crane is in place, whether the container truck is located in the effective area of ​​the quay crane, and whether the vehicle position is within the execution threshold. If all of them are met, the parking distance is initialized to true, and parking is completed directly under the guidance of the parking distance, and it is determined whether the distance between the parking position and the target location is within the completion threshold of 0.05m. If not, it returns to the step of initialization to false again. If yes, the alignment is completed. In addition, if the correction reference value has been obtained through the intervention of the quay crane alignment system in the initial test alignment process, the correction reference value can be directly added or subtracted from the parking distance after the parking distance is calculated to obtain the optimized parking distance for execution.

[0072] S3: After reporting that the vehicle has completed autonomous alignment and the bridge equipment has confirmed that it is correct, loading / unloading operations are carried out.

[0073] The overall method is designed with high redundancy and can select the optimal alignment method under different data conditions. On the one hand, when the point cloud data of the first laser radar may be incorrect due to bad weather or when the unmanned container truck performs the task of the quay crane for the first time and needs to determine a correction reference value through the initial test alignment process, the laser hardware of the unmanned container truck is used in combination with the laser hardware of the quay crane crane to perform alignment. In this case, although the laser hardware of the quay crane crane needs to be stopped for identification first, which will affect the alignment speed, it can ensure higher alignment accuracy. On the other hand, when the external environment is normal or the on-board terminal stores the correction reference value obtained from the initial test alignment, the unmanned container truck is selected to perform alignment. The laser hardware of the container truck can perform real-time recognition and complete the alignment directly in one go without stopping the vehicle, which can effectively improve the overall alignment efficiency. At the same time, the feedback of the correction reference value obtained during the initial test can further ensure the accuracy requirements under the premise of meeting the alignment efficiency. In addition, when the laser hardware of the unmanned container truck cannot work normally, the laser hardware of the quay crane lifting equipment is used for alignment to ensure that the entire alignment process can be carried out smoothly. Compared with the existing quay crane alignment system, the overall solution can effectively improve the flexibility of unmanned vehicle yard crane operations, reduce the vehicle site parking time, improve the efficiency of unmanned vehicles in loading and unloading containers on the yard crane and their adaptability in actual operations, and ensure the parking control accuracy.

[0074] Please refer to Figure 2 In step S2, when the first laser radar is selected to identify the relative position between the quay crane and the unmanned vehicle to obtain the parking distance, the method specifically includes:

[0075] When the unmanned vehicle arrives at the target location and the quay crane is in place, the crane's hoist can be used to directly load or unload boxes at the target location; the target location is located in the target area close to the quay crane when the unmanned vehicle is moving towards the quay crane. When the unmanned vehicle autonomously drives to the target area, the unmanned vehicle enters the alignment and parking process. During the alignment and parking process, the unmanned vehicle can identify the front beam of the quay crane in real time through the first laser radar and output the parking distance through real-time processing of point cloud data, and the unmanned vehicle moves in a straight line during the entire alignment and parking process; since the distance between the front beam and the rear beam of the quay crane is relatively close, only about five meters, when the vehicle is about to arrive at the target location, its first laser radar is approximately located in the area directly below the front beam. At this time, if the crane's hoist is flush with the front beam, the rear beam will be partially blocked. In order to ensure recognition accuracy and reduce the amount of calculation, when there is no abnormality in the point cloud data output by the first laser radar, its first laser radar is controlled to only recognize the front beam.

[0076] The random sampling consistency method and the prior normal vector are used to perform plane segmentation, and the bottom surface point cloud set of the front beam of the quay crane equipment identified by the first laser radar is determined. The distance to the unmanned vehicle is calculated according to the coordinate information of the bottom surface point cloud set of the front beam to obtain the parking reference distance.

[0077] By limiting the target location and target area, it can be confirmed that the unmanned vehicle can always identify the point cloud set of the front beam of the quay crane equipment in real time and calculate the output relative distance during the positioning and parking process. When the point cloud data output by the first laser radar is abnormal, the point cloud sets of multiple surfaces can also be identified, and more data can be obtained, so that there are more calculation methods that can adapt to different port conditions, making the overall positioning method more stable; at the same time, ensuring linear movement during positioning can also make the subsequent range value screening of the direct-pass filter more accurate and simple.

[0078] For further information, please refer to Figure 2 When the first laser radar is selected to identify the relative position between the quay crane and the unmanned vehicle to obtain the parking distance, it specifically includes:

[0079] Establish a coordinate system, in which the traveling direction of the unmanned vehicle during the parking process is set as the X-axis, the direction perpendicular to the X-axis and parallel to the ground is set as the Y-axis, and the direction perpendicular to the X-axis and perpendicular to the ground is set as the Z-axis;

[0080] A first straight-pass filter is set to filter the point cloud set. The range value on the X-axis is determined by the difference between the position information of the quay crane hoisting equipment after the target location task is obtained and the current position of the unmanned vehicle. The range value on the Y-axis is determined according to the empirical value of the width of multiple lanes under the quay crane, and the range value on the Z-axis is determined according to the empirical value of the front beam height. In this embodiment, there are generally 6 lanes under the quay crane. The Y range value is set to plus or minus 8m based on the center line of the lane and the empirical value. The height of the beam is about 12.5m, and the Z range value is set to 12.0 to 13.5m based on the ground.

[0081] The plane is segmented by using the random sampling consistency method and the prior normal vector. For the point cloud data collected by the first laser radar, the invalid points including the laser points of the camera on the quay bridge, the sides of the bridge and infinity are eliminated, and then the first point cloud set is obtained by the range value of the first straight-pass filter.

[0082] The point cloud data of the sling when the sling is flush with the front beam is eliminated through Euclidean clustering, and the point cloud data of the sling in the first point cloud set is further eliminated to obtain the point cloud set of the bottom surface of the front beam; the value threshold of the Euclidean clustering is 0.7m. Through empirical values, it can be judged that the minimum distance between the point cloud set of the sling and the point cloud set of the front beam is greater than 0.7m, so the two point cloud sets are separated by this threshold and the point cloud set of the bottom surface of the front beam is retained.

[0083] Project the point cloud of the bottom surface of the front beam into two-dimensional data, and obtain the X value of the two-dimensional data. At this time, the minimum value comes from the point at the bottom of the front beam closest to the first laser device. Through this point, all points are traversed to find the row of points at the bottom of the front beam closest to the unmanned vehicle. According to the data obtained from the actual vehicle test, different beams have the same length. The rule for traversal is to find other X values ​​that differ from the X value by plus or minus 4mm (the X value is based on the laser accuracy and fewer possible interference items. Generally, the error is small, and it is within plus or minus 3mm under normal circumstances. When setting, the range is expanded to plus or minus 4mm). Points outside this distance are points on other surfaces of the front beam, so they do not need to be considered. The Y value needs to be separated by a certain distance. The distance is related to the laser model and can be obtained by experience. The distance between two adjacent Y values ​​is about 40cm. Finally, through multiple corresponding X values ​​and Y values, the row of points closest to the vehicle is found.

[0084] The set of points found is fitted with the least squares fitting algorithm to obtain the straight line representing the bottom of the front beam closest to the unmanned vehicle, that is, the bottom edge of the rear side of the front beam, and the distance value B1 between the unmanned vehicle and the straight line is obtained. The parking distance is calculated by B1. The straight line fitting least squares fitting algorithm is an integrated algorithm in the opencv library. It only needs to import the required fitting points, so it is not described in detail. After fitting into a straight line according to the above method, the straight line can be understood as a straight line parallel to the front beam. Because the unmanned vehicle will be straightened in advance before alignment, the straight line is also parallel to the unmanned vehicle, so from Figure 2 It can be seen that this value is the shortest distance between the straight line and the center of the rear axle of the traction head of the unmanned vehicle; when the vehicle is aligned, in order to simplify the calculation process or when the first laser radar is in the area directly below the front beam and cannot normally obtain the point cloud information of the front beam, the B1 value can be used as the parking distance for alignment throughout the process.

[0085] In order to further improve the alignment accuracy, the stop reference distance is further obtained by performing calculations through B1, including the following steps:

[0086] A second straight-through filter is set to filter the point cloud set. The range of the second straight-through filter is determined according to B1, the height difference between the top surface and the bottom surface of the front beam on the Z axis, and the width difference on the Y axis. The point cloud set of the rear side of the front beam is obtained. After plane segmentation, all X values ​​in the point cloud set are averaged to obtain B2.

[0087] The third straight-through filter is set to filter the point cloud set. It is known that the distance between the rear side surface of the front beam and the front side surface of the rear beam is B3. According to B1, B3, and the height difference between the upper top surface and the lower bottom surface of the rear beam on the Z axis and the width difference on the Y axis, the range of the third straight-through filter is determined to obtain the point cloud set of the front side surface of the rear beam, and the average value of all X values ​​in the point cloud set is calculated to obtain B4;

[0088] The stopping distance B is (B1+B2+B3+B4) / 3.

[0089] When the unmanned vehicle enters the target area, it is a process of receiving point cloud data in real time and outputting the target parking distance in real time. When the vehicle is close to the quay crane and when the vehicle is under the quay crane, the point cloud data it receives will be different. Among them, B1 is calculated by the rear edge line of the bottom of the front beam, which is the most stable compared with other parts of the yard crane. On the one hand, when the alignment is close to completion, the front of the vehicle has passed the position of the hoist, and the rear beam will be disturbed by the hoist when identifying the rear beam, so the front beam is more stable than the rear beam. At the same time, after the vehicle continues to move forward for alignment, the front of the vehicle may go out of the front beam. At this time, the laser equipment cannot identify the rear side of the front beam. In summary, the rear edge line of the bottom of the front beam is the most stable and can be identified under normal circumstances during the entire alignment process. Therefore, the detection process is based on the B1 obtained by the rear edge line of the bottom of the front beam. In order to ensure the stability and adaptability of the entire detection, the following other detection distance items are introduced. In special cases, when B2 and B4 cannot be obtained, the alignment and parking can be performed only based on the B1 value.

[0090] When the quay crane is in place and the vehicle approaches the quay crane, please refer to Figure 4 In this case, after the unmanned container truck receives the task from the TMC horizontal transportation system, it plans the global driving trajectory and enters the task execution state. At the same time, the quay crane equipment receives the task from the ITC dispatching command center to move to the designated quay crane position; the unmanned container truck drives autonomously to the task quay crane position until the first lidar sensor can detect the point cloud information of the quay crane equipment.

[0091] When the unmanned truck continues to move forward and the tractor head of the vehicle is between the front beam and the rear beam, please refer to Figure 2 In this case, the laser recognition bridge in this article uses the rear bottom of the front beam, the rear side of the front beam, and the front side of the rear beam to calculate B2 and B4 (because the coordinate origin is the center of the rear axle of the tractor head, B4 is a negative value at this time). At this time, B3 is the distance from the rear side of the front beam to the front side of the rear beam; the B value calculated by combining multiple surfaces is more stable, which is better than directly using B1 as the B value for alignment. Of course, the final parking alignment of the vehicle can also be completed only by B1.

[0092] In the embodiment, there may be a situation where the output of the first laser radar is abnormal. Before the unmanned vehicle is about to arrive at the target location, the first laser radar on the unmanned truck head is located just below the bottom of the front beam at a distance. The first laser radar is directly installed on the roof and the shape of the laser radar is a horizontally placed cylinder. Due to long-term dust accumulation and rain erosion, a layer of gelatinous attachment will form on the top of the laser radar. The gelatinous attachment will block and affect the formation of point cloud data directly above the internal laser sensor; therefore, when the unmanned vehicle autonomously drives to the target area, the unmanned vehicle enters the alignment and parking process. During the alignment and parking process, the unmanned vehicle can identify the front beam of the quay crane lifting equipment in real time. When the unmanned vehicle moves to the threshold range directly below the front beam (when the first laser radar moves in the X-axis direction to within the range of plus or minus 3.5 meters from the front side of the front beam), the on-board terminal determines in real time whether the first laser radar outputs point cloud data normally;

[0093] The judgment method is as follows: after obtaining the point cloud set of the bottom surface of the front beam, the point cloud set of the bottom surface of the front beam is further classified by the second Euclidean clustering to form multiple side-by-side clusters, and the judgment is made by the number of clusters or the difference in the number of laser point clouds between any two different clusters; the normally output point cloud set of the bottom surface of the front beam will form 16 rows of side-by-side and almost equal-length clusters after the second Euclidean clustering, and the number of laser point clouds in each cluster is between 40 and 50. However, due to the occlusion effect of its gel-like attachment on the internal laser head in different situations (possibly continuous and intermittent occlusion, possible partial and whole block occlusion, possible large-area occlusion), After the second Euclidean clustering, the point cloud data output under abnormal circumstances are either between 12 and 16 rows, but the difference in the number of laser point clouds between each row is larger than 15 (which may be caused by continuous intermittent occlusion and local block occlusion), or the number of rows is less than 12 rows, but the difference in the number of laser point clouds between each row is not large (which may be caused by large-area occlusion). Of course, in theory, it is also possible that there are less than 12 rows, but the difference in the number of laser point clouds between each row is larger than 15. If this occurs in the actual point cloud detection situation, the judgment logic of the difference in the number of laser point clouds in each cluster combined with the number of rows can be added to the following judgment method.

[0094] If it is judged to be normal, the point cloud data of the current first laser radar is output; if it is judged to be abnormal, the point cloud data of the first laser radar is processed abnormally and re-output.

[0095] Furthermore, the threshold of the second Euclidean clustering is set to 0.3m, and the minimum number of each cluster is four. By setting this threshold, the influence of other interference objects installed on the front beam (such as the camera on the quay crane, which generally has 3 or less laser point clouds after clustering) or noise can be eliminated; when the number of clusters is less than 12, or the difference in the number of laser point clouds between any two different clusters is greater than 15, it is judged as an abnormal situation.

[0096] By adding this judgment method to the alignment process, the abnormal working condition of the laser radar can be effectively identified by combining Euclidean clustering, and the abnormal output of the first laser radar when it is below the bottom of the front beam can be quickly identified. The real-time recognition results are fed back to the vehicle-mounted terminal for selecting the alignment distance calculation method, thereby ensuring the accuracy of the parking distance in the last stage based on the first laser radar.

[0097] Furthermore, if an abnormality is determined, the point cloud data of the first laser radar is processed for abnormality and re-output, specifically including:

[0098] The X values ​​of all points of the point cloud set on the bottom surface of the front beam after the first Euclidean clustering are averaged to obtain Q (at this time, due to the different point cloud sets under different abnormal conditions, the obtained Q values ​​may all fall within the range from the bottom line of the front side surface of the front beam to the bottom line of the rear side surface of the front beam), and the fourth straight-through filter is set to filter the point cloud set. According to the width W of the front beam (W is 2.2m in this embodiment), the reserved margin distance E (to ensure that the range value can completely cover the front side surface of the rear beam, generally a threshold of 1m is reserved at the front and rear), and the distance R from the rear side of the front beam to the front side of the rear beam (R is 17.5m in this embodiment), the second The range value of the pass filter on the X-axis is (QWRE, Q-R+E). In this embodiment, the range value is (Q-20.7, Q-16.5). The range value on the Y-axis is determined according to the empirical value of the width of multiple lanes under the quay crane (same as above, plus or minus 8m), and the range value on the Z-axis is determined according to the empirical value of the rear beam height (same as above, 12-13.5m). The point cloud set of the front side of the rear beam is obtained through the range value of the fourth pass filter; after plane segmentation and removal of noise, the average value of all X values ​​in the point cloud set of the front side of the rear beam is taken to obtain B5 (B5 is a negative value at this time), and the parking distance B=R+B5.

[0099] Through this calculation method, the point cloud set of the bottom of the front beam with abnormal output can be effectively used to obtain the point cloud set of the front side of the rear beam. When the first laser radar is located between the front beam and the rear beam, the front side of the rear beam located at the rear can be effectively identified, and the point cloud set of the front side of the rear beam is used to calculate the parking distance, thereby avoiding the problem of inaccurate final parking distance due to abnormal output of the point cloud set at the bottom of the front beam, and realizing a high-redundancy alignment method that can cope with various unexpected situations.

[0100] In this embodiment, when the quay crane alignment system intervenes in the alignment, after the unmanned vehicle stops, the quay crane alignment system waits for 1s to 3s, and finally collects multiple quay crane guidance parking distance information and executes the average value of its parking distance. Because the alignment process of the vehicle terminal based on the first laser radar has a certain deviation during the movement of the vehicle, the quay crane alignment system will not be guided until the quay crane alignment system is stable after the vehicle stops for 1.5S. Even if the vehicle is stationary, the quay crane alignment system still has deviations, but the deviations are small. In order to reduce the impact of this aspect, the guidance values ​​issued by the quay crane alignment system are collected 7 times and then averaged as the final alignment distance.

[0101] In this embodiment, when the laser hardware fails to work and the vehicle-mounted terminal fails to receive the normal point cloud data of the first laser radar, that is, both are abnormal, the vehicle stops, and only the quay crane alignment system intervenes in the alignment, and the parking guidance value is calculated to guide the unmanned container truck to align, which specifically includes:

[0102] When the quay crane hoisting equipment is in place, the unmanned vehicle autonomously drives to the quay crane position / yard bay target area, and brakes to stop when the unmanned vehicle is located within 2m of the destination point issued by the TMC system; because the quay crane guidance cannot be output throughout the field, the effective range is approximately plus or minus 5m. In actual experiments, the value within 2m is the most accurate, so the first stop should be controlled within 2m of the target location as much as possible;

[0103] The quay crane alignment system outputs the quay crane guidance parking distance information for more than 7 times after the unmanned vehicle stops for 1.5 seconds, and judges the validity of the parking guidance value output by the quay crane alignment system; (Judgment principle 1: The guidance issued by the quay crane alignment system is generally continuous, and starts to be received 1.5 seconds after parking. If the guidance value is not received 7 times within 5 seconds, the judgment is invalid, and it is necessary to restart the timing reception until it meets the judgment principle; Judgment principle 2: The guidance value issued by the quay crane alignment system is generally stable. If the maximum and minimum values ​​of the guidance values ​​received 7 times within 5 seconds exceed 0.5m, the judgment is invalid, and it is necessary to restart the timing reception until it meets the judgment principle);

[0104] If it is effective, the unmanned container truck will be used to guide the alignment controller to execute the average value of its parking distance for the first time until it stops; (the quay crane alignment system may also have the problem of inaccurate output due to the position offset of the laser sensor on the quay crane, but since the overall movement of the quay crane is stable and the working environment and recognition range are relatively stable, the probability of offset is not high and the bridge equipment will be regularly calibrated and maintained. The quay crane corresponding system requires the unmanned container truck to stop before it can be guided. Although it takes a few seconds longer, its accuracy will be more accurate);

[0105] The quay crane alignment system again outputs the average value of the quay crane guidance parking distance information of more than 7 times after the unmanned vehicle stops for 1.5s, and determines whether the unmanned vehicle is within the parking threshold range after the first execution of the average value of its parking distance (in this embodiment, the parking threshold range is within 0.05m of the target location). If it is within the range, it reports to the TMC horizontal transportation system that the container truck alignment is completed. Otherwise, the unmanned container truck guidance alignment controller is used to execute the average value output again until the unmanned vehicle is within the parking threshold range and then reports the alignment completion status to complete the container truck autonomous alignment process.

[0106] After at least one alignment of the quay crane corresponding system, a high parking accuracy can be ensured. Since it cannot be guaranteed that the unmanned vehicle actually executes the guidance value after the first execution of the guidance value (the execution amount of the vehicle is calculated based on the vehicle speed × time, and the vehicle speed is measured by the vehicle speed sensor. The measurement value cannot be 100% accurate, so it will cause the execution value and the guidance value to deviate), so a second or even multiple verifications are performed after the vehicle stops. For example, the first time the guidance value of 1m is received, but the actual vehicle only executes 0.9m, there is still a deviation of 0.1m, so it receives guidance again, and the second time it receives the guidance value of 0.1m, the vehicle executes again, but the actual vehicle only executes 0.05m, there is still a deviation of 0.05m, but it meets the parking threshold range requirements, then the alignment is reported to be completed, and the subsequent cabinet hanging or unloading work is carried out.

[0107] In this embodiment, please refer to Figure 2 , the vehicle terminal is further connected to the two second laser radars for communication, and the two second laser radars are respectively arranged on both sides of the unmanned vehicle; the vehicle terminal performs parking control on the unmanned vehicle according to the point cloud data sent by the second laser radar, and obtains the target parking distance by subtracting the box offset distance from the parking distance, specifically including:

[0108] Establish a coordinate system, in which the traveling direction of the unmanned vehicle during the parking process is set as the X-axis, the direction perpendicular to the X-axis and parallel to the ground is set as the Y-axis, and the direction perpendicular to the X-axis and perpendicular to the ground is set as the Z-axis;

[0109] A fifth direct filter is set to filter the point cloud set, and the range value of the fifth direct filter is determined according to different types of box delivery tasks in the target location task; for the point cloud data collected by the second laser radar, after removing invalid points, the point cloud set to be processed is obtained through the range value of the fifth direct filter; Figure 3As shown, the types of containers in port operations mainly include 20-foot containers and 40-foot containers. Therefore, the unmanned vehicle may have the following situations: a 20-foot container is placed in the center, front or rear end, two 20-foot containers are placed in the front and back, and a 40-foot container is placed in the center. The quay crane is a single-lift container, that is, one container is lifted at a time. Therefore, there are the following situations during alignment: 20 feet in front, 20 feet behind, 20 feet in the middle, and 40 feet in the middle. The range value of the direct-pass filter can be obtained based on the prior experience of testing different types.

[0110] Using a random sampling consistency method and a priori normal vector to perform plane segmentation, determine a point cloud set of the front side of the vehicle container near the front of the vehicle identified by the second laser radar;

[0111] Take the average of the X values ​​of the point cloud set to get the box offset distance S;

[0112] The box offset distance S is used to calculate the offset and distance of the box under different tasks, so as to further control the target stopping distance and ultimately achieve accurate grasping of the box.

[0113] In addition, when the laser equipment on the unmanned vehicle is offset during use or the laser hardware on the quay crane used to align the unmanned truck or container is offset, when the above scheme is used for identification, the output point cloud information is biased, which will cause the final parking position to be inaccurate. Therefore, it is necessary to observe the error situation and perform the final alignment compensation manually after completing the parking under the test condition, so as to obtain the error value, that is, the test calibration value; based on this, in the subsequent alignment parking, the error value is compensated in advance; that is, after further compensating the test calibration value by the target parking distance, the target precise parking distance is obtained; in general, when the current test calibration value cannot meet the operation requirements, the calibration value is re-determined, and the alignment compensation specifically includes the following steps:

[0114] When the quay crane crane equipment at the target location is in place and the task is to deliver containers; in the case of test calibration, when the parking of the unmanned vehicle is completed, the test calibration value is the relative distance between the container on the vehicle and the quay crane crane equipment;

[0115] When the quay crane hoisting equipment at the target location is in place and the task is to collect containers; in the test calibration case, when the parking of the unmanned vehicle is completed, the test calibration value is the relative distance between the unmanned vehicle and the quay crane hoisting equipment.

[0116] Through this compensation method, the overall parking method can be made more adaptable, and can effectively solve the errors caused by the offset of the laser equipment of the unmanned vehicle or the positioning equipment of the quay crane equipment, further improving the overall alignment accuracy; at the same time, the test calibration value can also be used as a correction parameter and imported into the vehicle terminal before the alignment begins.

[0117] In this embodiment, the terminal may be provided with an unmanned truck guidance and alignment controller to ensure that the unmanned truck can autonomously control the truck electric brake actuator to complete the guidance and self-execution process when the unmanned truck is loaded or unloaded. According to the determined target parking distance, when the target parking distance is greater than or equal to the ideal braking distance, the terminal can autonomously control the speed of the current vehicle to achieve parking. When the target parking distance is less than the ideal braking distance, the current vehicle can be emergency braked to prevent the vehicle from exceeding the target parking position.

[0118] It should be understood that the ideal braking distance of a vehicle is not the same when traveling at different speeds. The ideal braking distance is related to the vehicle's load and driving speed. Technical personnel in this field can measure and obtain the ideal braking distance of the vehicle under different loads and driving speeds based on the actual vehicle, so that they can directly call it for subsequent comparison during actual use.

[0119] In one example, the ideal braking distance S of the driverless vehicle under different conditions can be determined according to the following formula: BRK :

[0120] S BRK =P1V 2 +P2V+P3M+P4

[0121] Wherein, V is the travel speed, M is the mass of the container, P1, P2, P3 and P4 are predetermined coefficients, which can be set by those skilled in the art based on prior experience.

[0122] It should be noted that the above-mentioned unmanned container truck guidance alignment controller can not only be used for the first parking of unmanned vehicles, but if there is a second parking position (that is, the first parking position is an invalid parking position), the vehicle parking controller can also realize the above-mentioned function to prevent the vehicle from exceeding the target parking position due to untimely braking, thereby ensuring the parking control effect.

[0123] In this embodiment, a computer-readable medium stores a computer program, and when the computer program is executed by a processor, a highly redundant autonomous alignment method for unmanned container trucks under a port quay crane is implemented as described in the above embodiment.

[0124] Figure 4 A schematic diagram of the structure of a computer system suitable for implementing an electronic device of an embodiment of the present application is shown.

[0125] Specifically, an electronic device includes: one or more processors; a storage device for storing one or more programs, and when the one or more programs are executed by the one or more processors, the one or more processors implement the above-mentioned high-redundancy autonomous alignment method for unmanned container trucks under a port quay crane.

[0126] It should be noted that Figure 4 The computer system of the electronic device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0127] like Figure 4 As shown, the computer system includes a central processing unit (CPU) 401, which can perform various appropriate actions and processes according to the program stored in the read-only memory (ROM) 402 or the program loaded from the storage part 408 to the random access memory (RAM) 403, such as executing the method described in the above embodiment. In the RAM 403, various programs and data required for system operation are also stored. The CPU 401, the ROM 402 and the RAM 403 are connected to each other through the bus 404. The input / output (I / O) interface 405 is also connected to the bus 404.

[0128] The following components are connected to the I / O interface 405: an input section 406 including a keyboard, a mouse, etc.; an output section 407 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker, etc.; a storage section 408 including a hard disk, etc.; and a communication section 409 including a network interface card such as a LAN (Local Area Network) card, a modem, etc. The communication section 409 performs communication processing via a network such as the Internet. A drive 410 is also connected to the I / O interface 405 as needed. A removable medium 411, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 410 as needed so that a computer program read therefrom is installed into the storage section 408 as needed.

[0129] In particular, according to an embodiment of the present application, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present application includes a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a computer program for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through the communication part 409, and / or installed from a removable medium 411. When the computer program is executed by the central processing unit (CPU) 401, various functions defined in the system of the present application are executed.

[0130] It should be noted that the computer-readable medium shown in the embodiment of the present application may be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM), a flash memory, an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present application, a computer-readable storage medium may be any tangible medium containing or storing a program, which may be used by or in combination with an instruction execution system, device or device. In the present application, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, which carries a computer-readable computer program. Such propagated data signals may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. Computer-readable signal media may also be any computer-readable medium other than computer-readable storage media, which may send, propagate, or transmit programs for use by or in conjunction with an instruction execution system, apparatus, or device. The computer program contained on the computer-readable medium may be transmitted using any appropriate medium, including but not limited to: wireless, wired, etc., or any suitable combination of the above.

[0131] The flowchart and block diagram in the accompanying drawings illustrate the possible architecture, functions and operations of the system, method and computer program product according to various embodiments of the present application. Wherein, each box in the flowchart or block diagram can represent a module, a program segment, or a part of the code, and the above-mentioned module, program segment, or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram or flowchart, and the combination of boxes in the block diagram or flowchart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0132] The units involved in the embodiments described in this application may be implemented by software or hardware, and the units described may also be set in a processor. The names of these units do not, in some cases, constitute limitations on the units themselves.

[0133] As another aspect, the present application further provides a computer program product or a computer program, the computer program product or the computer program including computer instructions, the computer instructions being stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device executes a highly redundant autonomous alignment method for unmanned container trucks under a port quay bridge provided in the above embodiment.

[0134] The computer readable medium may be included in the electronic device described in the above embodiment; or it may exist independently without being assembled into the electronic device. The above computer readable medium carries one or more programs, and when the above one or more programs are executed by an electronic device, the electronic device implements the method described in the above embodiment.

[0135] It should be noted that, although several modules or units of the equipment for action execution are mentioned in the above detailed description, this division is not mandatory. In fact, according to the embodiments of the present application, the features and functions of two or more modules or units described above can be embodied in one module or unit. On the contrary, the features and functions of one module or unit described above can be further divided into being embodied by multiple modules or units.

[0136] Through the description of the above implementation methods, it is easy for those skilled in the art to understand that the example implementation methods described here can be implemented by software, or by combining software with necessary hardware. Therefore, the technical solution according to the implementation method of the present application can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.) or on a network, including several instructions to enable a computing device (which can be a personal computer, a server, a touch terminal, or a network device, etc.) to execute the method according to the implementation method of the present application.

[0137] Those skilled in the art will readily appreciate other embodiments of the present application after considering the specification and practicing the embodiments disclosed herein. The present application is intended to cover any variations, uses or adaptations of the present application, which follow the general principles of the present application and include common knowledge or customary technical means in the art that are not disclosed in the present application.

[0138] It should be understood that the present application is not limited to the precise structures that have been described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof. The scope of the present application is limited only by the appended claims.

Claims

1. A highly redundant autonomous alignment method for unmanned container trucks under a port quay crane, characterized in that: An on-board terminal applied to an unmanned vehicle, the on-board terminal being in communication connection with a first laser radar; the first laser radar being arranged on the top of the unmanned vehicle; the on-board terminal performing parking control on the unmanned vehicle according to the point cloud data sent by the first laser radar; The method comprises the following steps: S1: The unmanned vehicle accepts the target location mission and autonomously moves to the vicinity of the target location; S2: The on-board terminal determines whether the laser hardware on the unmanned vehicle is working properly and confirms whether the point cloud information output by the first laser radar can be received normally; if both are not normal, the vehicle stops, the quay crane alignment system intervenes in the alignment, and the parking guidance value is calculated to guide the unmanned container truck to align; If all are normal, the operator will choose whether the quay crane alignment system needs to intervene in the alignment; if the quay crane alignment system needs to intervene in the alignment, the operator will choose to use the first laser radar to identify the relative position between the quay crane crane and the unmanned vehicle to calculate the parking distance, and directly complete the parking under the guidance of the parking distance; After parking, the point cloud data of the unmanned vehicle is obtained through the laser hardware installed on the quay crane, and the current parking position of the unmanned vehicle is calculated by the quay crane alignment system. If the parking error threshold is met, the alignment is terminated. If the parking error threshold is not met, the parking guidance value calculated by the quay crane alignment system is used to guide the unmanned vehicle to align, and the parking distance under the first laser radar scan is calculated based on the alignment position and calibrated as the parking reference distance; When it is selected that the quay crane alignment system does not need to intervene in the alignment, the first laser radar is directly selected to identify the relative position between the quay crane crane and the unmanned vehicle to obtain the parking distance, and parking is directly completed under the guidance of the parking distance; S3: After reporting that the vehicle has completed autonomous alignment and the bridge equipment has confirmed that it is correct, loading / unloading operations are carried out; In step S2, when the first laser radar is selected to identify the relative position between the quay crane and the unmanned vehicle to obtain the parking distance, the following steps are specifically performed: When the unmanned vehicle arrives at the target location and the quay crane equipment is in place, the crane's hoist can be used to directly load or unload containers at the target location; the target location is located in the target area close to the quay crane equipment when the unmanned vehicle is moving towards the quay crane equipment. When the unmanned vehicle autonomously drives to the target area, the unmanned vehicle enters the alignment and parking process. During the alignment and parking process, the unmanned vehicle can identify the front beam of the quay crane equipment in real time, and the unmanned vehicle moves in a straight line during the entire alignment and parking process; Using the random sampling consistency method and the prior normal vector to perform plane segmentation, determine the bottom surface point cloud set of the front beam of the quay crane equipment identified by the first laser radar, calculate the distance to the unmanned vehicle according to the coordinate information of the bottom surface point cloud set of the front beam, and obtain the parking reference distance; When the first laser radar is selected to identify the relative position between the quay crane and the unmanned vehicle to obtain the parking distance, it specifically includes: Establish a coordinate system, in which the traveling direction of the unmanned vehicle during the parking process is set as the X-axis, the direction perpendicular to the X-axis and parallel to the ground is set as the Y-axis, and the direction perpendicular to the X-axis and perpendicular to the ground is set as the Z-axis; The first straight-pass filter is set to filter the point cloud set. The range value on the X-axis is determined by the difference between the position information of the quay crane crane after it is in place through the target location task and the current position of the unmanned vehicle. The range value on the Y-axis is determined according to the experience value of the width of multiple lanes under the quay crane. The range value on the Z-axis is determined according to the experience value of the front beam height. The plane is segmented by using the random sampling consistency method and the prior normal vector. For the point cloud data collected by the first laser radar, the invalid points including the laser points of the camera on the quay bridge and the side of the bridge are eliminated, and then the first point cloud set is obtained by the range value of the first straight-pass filter. The point cloud data of the sling when the sling is flush with the front beam is eliminated by Euclidean clustering, and the point cloud data of the sling in the first point cloud set is further eliminated to obtain the point cloud set of the bottom surface of the front beam; Project the point cloud set on the bottom surface of the front beam into two-dimensional data, and obtain the X value of the two-dimensional data. At this time, the minimum value comes from the point on the bottom of the front beam closest to the first laser device. Through this point, traverse all points to find the row of points on the bottom of the front beam closest to the unmanned vehicle; The set of points found is fitted with a least squares fitting algorithm to obtain a straight line representing the bottom of the front beam closest to the unmanned vehicle, that is, the bottom edge of the rear side of the front beam, and the distance value B1 between the unmanned vehicle and the straight line is obtained, and the parking distance is calculated by B1; When the unmanned vehicle autonomously drives to the target area, the unmanned vehicle enters the alignment and parking process. During the alignment and parking process, the unmanned vehicle can identify the front beam of the quay crane in real time. When the unmanned vehicle moves to the threshold range directly below the front beam, it determines whether the first laser radar outputs point cloud data normally; The judgment method is as follows: after obtaining the point cloud set of the bottom surface of the front beam, the point cloud set of the bottom surface of the front beam is further classified by the second Euclidean clustering to form multiple side-by-side clusters, and the judgment is made by the number of clusters or the difference in the number of laser point clouds between any two different clusters; If the judgment is normal, the point cloud data of the current first laser radar is output; if the judgment is abnormal, the point cloud data of the first laser radar is processed abnormally and re-output, and the re-output specifically includes: The X values ​​of all points of the point cloud set on the bottom surface of the front beam after the first Euclidean clustering are averaged to obtain Q, and the fourth direct filter is set to filter the point cloud set. According to the width W of the front beam, the reserved margin distance E, and the distance R from the rear side of the front beam to the front side of the rear beam, the range value of the second direct filter on the X axis is determined to be (QWRE, Q-R+E), the range value on the Y axis is determined according to the empirical value of the width of multiple lanes under the quay crane, and the range value on the Z axis is determined according to the empirical value of the height of the rear beam. The point cloud set of the front side of the rear beam is obtained through the range value of the fourth direct filter; the average value of all X values ​​in the point cloud set on the front side of the rear beam is obtained to obtain B5, and the parking distance B=R+B5.

2. The method according to claim 1, characterized in that Further comprising the steps of: A second straight-through filter is set to filter the point cloud set. The range of the second straight-through filter is determined according to B1, the height difference between the top surface and the bottom surface of the front beam on the Z axis, and the width difference on the Y axis. The point cloud set of the rear side of the front beam is obtained. After plane segmentation, all X values ​​in the point cloud set are averaged to obtain B2. The third straight-through filter is set to filter the point cloud set. It is known that the distance between the rear side surface of the front beam and the front side surface of the rear beam is B3. According to B1, B3, and the height difference between the upper top surface and the lower bottom surface of the rear beam on the Z axis and the width difference on the Y axis, the range of the third straight-through filter is determined to obtain the point cloud set of the front side surface of the rear beam, and the average value of all X values ​​in the point cloud set is calculated to obtain B4; The stopping distance B is (B1+B2+B3+B4) / 3.

3. The method according to claim 1, characterized in that: The threshold of the second Euclidean clustering is set to 0.3m, and the minimum number of each cluster is four; when the number of clusters is less than 12, or the difference in the number of laser point clouds between any two different clusters is greater than 15, it is judged as an abnormal situation.

4. The method according to claim 1, characterized in that: When the quay crane positioning system intervenes in the positioning, after the unmanned vehicle stops, the quay crane positioning system waits for 1s to 3s, and finally collects multiple quay crane guidance parking distance information and executes the average value of its parking distance.

5. The method according to claim 1, characterized in that If all are abnormal, the truck will stop, and the quay crane alignment system will intervene in the alignment, and the parking guidance value will be calculated to guide the unmanned container truck to align, including: When the quay crane equipment is in place, the unmanned vehicle will autonomously drive to the quay crane position / yard bay target area, and brake and stop when the unmanned vehicle is within 2m of the destination point issued by the TMC system; The quay crane alignment system outputs the quay crane guidance parking distance information for more than 7 times after the unmanned vehicle stops for 1.5 seconds, and judges the validity of the parking guidance value output by the quay crane alignment system. If it is valid, the average value of the parking distance of the unmanned container truck guidance alignment controller is used for the first execution until the vehicle stops; The quay crane alignment system again outputs the average value of the quay crane guidance parking distance information for more than 7 times after the unmanned vehicle stops for 1.5s, and determines whether the unmanned vehicle is within the parking threshold range after the first execution of the average value of its parking distance. If it is within the range, it reports to the TMC horizontal transportation system that the container truck alignment is completed. Otherwise, the unmanned container truck guidance alignment controller is used to execute the average value output again until the unmanned vehicle is within the parking threshold range and then reports the alignment completion status to complete the container truck autonomous alignment process.

6. The method according to claim 1, wherein the vehicle-mounted terminal is further connected to two second laser radars for communication, and the two second laser radars are respectively arranged on both sides of the unmanned vehicle; the vehicle-mounted terminal performs parking control on the unmanned vehicle according to the point cloud data sent by the second laser radars, characterized in that: The target parking distance is obtained by calculating the difference between the box offset distance and the parking distance, including: Establish a coordinate system, in which the traveling direction of the unmanned vehicle during the parking process is set as the X-axis, the direction perpendicular to the X-axis and parallel to the ground is set as the Y-axis, and the direction perpendicular to the X-axis and perpendicular to the ground is set as the Z-axis; A fifth direct filter is set to filter the point cloud set, and a range value of the fifth direct filter is determined according to different types of box delivery tasks in the target location task; For the point cloud data collected by the second laser radar, after removing invalid points, a point cloud set to be processed is obtained by using the range value of the fifth pass filter; Using a random sampling consistency method and a priori normal vector to perform plane segmentation, determine a point cloud set of the front side of the vehicle container near the front of the vehicle identified by the second laser radar; Take the average of the X values ​​of the point cloud set to get the box offset distance S; The target stop distance is further compensated for the test calibration value to obtain the target precise stop distance, which specifically includes: When the quay crane crane equipment at the target location is in place and the task is to deliver containers; in the case of test calibration, when the parking of the unmanned vehicle is completed, the test calibration value is the relative distance between the container on the vehicle and the quay crane crane equipment; When the quay crane hoisting equipment at the target location is in place and the task is to collect containers; in the test calibration case, when the parking of the unmanned vehicle is completed, the test calibration value is the relative distance between the unmanned vehicle and the quay crane hoisting equipment.

Images