Heavy truck automatic warehouse accurate alignment guiding method and system and vehicle

Abstract

The invention discloses a heavy truck automatic warehouse accurate alignment guiding method and system and a vehicle, and the method is used for the alignment of an intelligent heavy truck and an automatic warehouse, when the intelligent heavy truck receives a loading and unloading alignment task, the point cloud data of an alignment scene is obtained through a laser radar, the point cloud of a target area is stably extracted through point cloud preprocessing, and the point cloud is extracted; calculating the position and orientation of an alignment end point, guiding the intelligent heavy truck to park through data post-processing, and outputting the alignment end point and keeping updating; site transformation is not needed, the operation cost is effectively reduced, parking guidance of the intelligent heavy truck in an alignment scene is effectively output through laser radar detection, the parking operation efficiency is improved, and the alignment time loss is reduced; through an adaptive point cloud preprocessing method, stable extraction of alignment target point clouds of the intelligent heavy truck under different distances and postures is completed, the adjustment distance and direction of the trailer of the intelligent heavy truck are obtained after processing, the adjustment distance and direction are converted into an accurate alignment end point, and the automation level is improved.

Description

Technical Field

[0001] The present invention belongs to the field of intelligent driving technology and relates to the measurement and positioning of laser radars, and specifically to a method, system and vehicle for precise alignment guidance of heavy trucks in an automated warehouse. Background Art

[0002] In industrial parks, warehouse material transfer is a crucial part of daily production. The automation rate and continuous operation duration of transfers have a decisive impact on the overall production efficiency of the park. During the material loading and unloading process in automated warehouses, transfer vehicles require precise alignment and guidance to accurately reach material loading and unloading points, thereby enabling material transfer between warehouses.

[0003] In automated material transfer yards, heavy trucks must begin docking in the operating lane during material loading and unloading. They must accurately maneuver the vehicle to the docking equipment and precisely lock the trailer's rear cargo box with the warehouse's docking equipment before automated loading and unloading can begin. Furthermore, upon parking, the relative position and posture error between the heavy truck's trailer and the warehouse door must be kept within a minimal range to ensure precise locking with the docking equipment. Inadequate docking accuracy can lead to risks such as inefficient loading and unloading, vehicle collisions, and equipment damage.

[0004] In the prior art, publication number CN117935223A discloses a method and system for mechanical parking space detection using laser radar, which obtains entrance corner points and bottom corner points through side point clouds and bottom point clouds. It is necessary to obtain laser point clouds of the side and bottom structures of the parking space, which poses potential detection accuracy issues for parking spaces without side structural features. Publication number CN115410406A discloses a parking space detection method, device, parking system, and storage medium, which utilizes captured environmental images to identify whether a vehicle is in a familiar environment, captures a parking space image, and obtains a target parking space image corresponding to the target environmental image for image matching. However, when there is a large difference between the environmental image and the target environmental image or the image quality is poor, the accuracy of image matching may be affected, resulting in errors in parking space identification. In addition, this method relies on a pre-stored set of environmental images. If the images in the set are not comprehensive or are not updated in a timely manner, the accuracy and reliability of parking space detection may also be affected. Summary of the Invention

[0005] In response to the above problems, the main purpose of the present invention is to design a method, system and vehicle for precise positioning guidance of heavy-duty trucks in automated warehouses, which only uses a laser radar to achieve high-precision guidance of intelligent heavy-duty trucks, and solve the risks of low loading and unloading efficiency, difficult positioning, vehicle collisions, equipment damage, etc. caused by positioning accuracy problems when intelligent heavy-duty trucks precisely park and load and unload goods in warehouses.

[0006] In order to achieve the above purpose, the present invention adopts the following technical solutions:

[0007] A method for precise alignment guidance of heavy trucks in an automated warehouse, which is used for alignment of intelligent heavy trucks with the automated warehouse, comprises the following steps:

[0008] The intelligent heavy-duty truck receives the loading and unloading alignment task and uses LiDAR to obtain point cloud data of the alignment scene. The alignment scene includes the automated warehouse location, automated warehouse door, and ground lock. After the intelligent heavy-duty truck's trailer is aligned, it is locked with the ground lock and docked with the automated warehouse door.

[0009] Perform point cloud preprocessing on the point cloud data acquired by the LiDAR, and extract the aligned wall surface for detection and fitting as the target area point cloud;

[0010] Detection and fitting are performed based on the extracted point cloud of the target area, and the relative position of the smart heavy truck and the alignment endpoint is calculated, including the longitudinal distance, lateral distance, and heading angle. The alignment endpoint is the location of the origin of the trailer coordinate system when the smart heavy truck is aligned;

[0011] Based on the longitudinal distance, lateral distance and orientation angle of the intelligent heavy truck relative to the alignment end point, the position of the alignment end point in the origin coordinate system of the intelligent heavy truck trailer is obtained;

[0012] Based on the position of the alignment endpoint in the trailer's origin coordinate system, the intelligent heavy-duty truck is guided into the parking space and locked in place using a ground lock.

[0013] After the intelligent heavy truck parks at the storage location, it outputs the information of the positioning end point and keeps it updated.

[0014] As a further description of the present invention, one laser radar is provided and located in the middle of the rear of the trailer of the smart heavy truck, and is used to observe the positioning scene throughout the parking process of the smart heavy truck.

[0015] As a further description of the present invention, the process of point cloud preprocessing is:

[0016] In the full amount of point cloud data acquired by the LiDAR, the first frame selection is performed to retain the point cloud data within a fixed range from the center of the LiDAR, which is recorded as L1;

[0017] Calibrate the trailer's origin position when the trailer is parked and locked, and record it as the parking prior point position T prior ;

[0018] The relative position between the trailer origin and the parking point at that moment is calculated using the point cloud data obtained at the current moment.

[0019] Transform the point cloud data L1 from the trailer coordinate system to the parking point coordinate system to obtain the transformed point cloud data;

[0020] In the parking point coordinate system, the transformed point cloud data is selected for the second time, and the point cloud data within a fixed range is retained, which is recorded as L2;

[0021] The point cloud data L2 is output as the target area point cloud of the alignment wall.

[0022] As a further description of the present invention, point cloud detection and fitting includes the following steps:

[0023] Divide the point cloud of the target area into grids;

[0024] Calculate based on the divided grid to obtain the flatness and normal vector of each grid point cloud;

[0025] Based on the preset values ​​of flatness and normal vectors, as well as the required number of point clouds, grid filtering is performed to extract the ideal plane point cloud L3 that can represent the wall surface;

[0026] Project the extracted ideal plane point cloud L3 onto the ground;

[0027] Perform two-dimensional straight line fitting on the projected point cloud to obtain the straight line equation of the wall surface: Ax+By+C=0;

[0028] Among them, A, B, and C are the coefficients and constant terms of the straight line equation, and x and y are the coordinate variables of the point on the ground after projection.

[0029] As a further description of the present invention, based on the fitted straight line equation, the relative position of the intelligent heavy truck and the alignment end point is calculated, including the longitudinal distance, the lateral distance, and the heading angle;

[0030] Among them, the longitudinal distance is the longitudinal distance x between the origin of the trailer coordinate system and the coordinate system of the end point. trailer , the expression is:

[0031]

[0032] x trailer =dd fix ;

[0033] Among them, d is the longitudinal distance from the center of the radar point cloud to the corresponding wall, d fix It is the fixed longitudinal distance from the center of the radar point cloud to the alignment wall during alignment locking;

[0034] The heading angle is the relative heading angle θ between the trailer longitudinal direction and the normal direction of the opposing wall, and the expression is:

[0035]

[0036] Thus, the relative positions of the intelligent heavy truck and the alignment end point in the longitudinal and heading directions are obtained.

[0037] As a further description of the present invention, the lateral distance between the intelligent heavy truck and the alignment end point is obtained based on a reflective plate calibrated on the center axis of the storage location, and the reflective plate is located on the alignment wall, specifically including the following steps:

[0038] Obtain point cloud data of the alignment scene including the reflector through LiDAR;

[0039] Extracting point cloud data of the area containing the reflector through point cloud preprocessing;

[0040] Setting an intensity threshold and number of points for the reflector interior points, sorting the extracted regional point cloud data according to the intensity value of the point cloud, and extracting the reflector interior points from the sorted point cloud based on the settings;

[0041] The average value of the extracted inner point coordinates of the reflector is obtained to obtain the coordinates of the center of the reflector in the trailer coordinate system (x reflector ,y reflector );

[0042] Then, the lateral distance y between the origin of the trailer coordinate system and the coordinate system of the end point is trailer , the expression is:

[0043] y trailer =x reflector sinθ-y reflector ·cosθ;

[0044] Where θ is the relative heading angle between the trailer longitudinal direction and the normal direction of the opposing wall.

[0045] As a further description of the present invention, the process of obtaining the position of the positioning end point in the origin coordinate system of the intelligent heavy truck trailer includes the following steps:

[0046] Based on the trailer position (x trailer ,y trailer ) and the relative heading angle θ, using the transformation matrix between the coordinate system of the end point and the trailer coordinate system Perform coordinate transformation and transformation matrix expression:

[0047]

[0048] Among them, x trailer Represents the longitudinal distance between the origin of the trailer coordinate system and the end-point coordinate system, y trailer represents the lateral distance between the origin of the trailer coordinate system and the coordinate system of the alignment end point, and θ is the relative heading angle between the longitudinal direction of the trailer and the normal direction of the alignment wall;

[0049] Then, the alignment endpoint posture P in the trailer coordinate system is destination , the expression is:

[0050]

[0051] Among them, P initial It is a constant vector, indicating the position of the alignment end point in the alignment end point coordinate system.

[0052] As a further description of the present invention, the process of guiding the intelligent heavy truck to park in the parking space is to transform the alignment end point posture in the trailer coordinate system into the global coordinate system to obtain the latitude and longitude coordinates and heading angle of the alignment end point in the global coordinate system;

[0053] When the trailer origin reaches the alignment end point and the heading angle is consistent, the alignment lock is accurately completed.

[0054] A precise alignment guidance system for heavy trucks in an automated warehouse, comprising a task receiving and processing module, a data acquisition module, a point cloud preprocessing module, a detection and fitting calculation module, a relative posture calculation module, a parking guidance module, and an alignment information output and update module;

[0055] The task receiving and processing module is used to receive the loading and unloading alignment task issued by the upper system and trigger the alignment detection processing flow of the intelligent heavy truck when the task starts;

[0056] The data acquisition module is used to use the laser radar to collect point cloud data of the positioning scene in real time, including the storage locations, automated warehouse doors, and ground locks of the automated warehouse;

[0057] The point cloud preprocessing module is used to preprocess the point cloud data collected by the laser radar and extract the aligned wall point cloud for detection and fitting as the target area point cloud;

[0058] The detection and fitting calculation module is used to perform wall detection and fitting calculation based on the point cloud of the target area to obtain the relative position of the smart heavy truck and the alignment end point, including the longitudinal distance, lateral distance, and orientation angle;

[0059] The relative posture calculation module is used to calculate the posture of the alignment endpoint in the coordinate system of the intelligent heavy-duty truck trailer based on the relative posture of the intelligent heavy-duty truck and the alignment endpoint;

[0060] The parking guidance module is used to guide the intelligent heavy truck to park in a predetermined path and posture according to the results of the relative posture calculation module, and complete the positioning and locking through the ground lock;

[0061] The alignment information output and update module is used to output the alignment endpoint information after the intelligent heavy truck is parked in the storage space, and keep it updated in real time.

[0062] A vehicle comprises the above-mentioned guidance system and executes the above-mentioned guidance method.

[0063] Compared with the prior art, the technical effects of the present invention are:

[0064] The present invention provides a method, system and vehicle for precise alignment guidance of heavy trucks in automated warehouses. The method is used for the alignment of intelligent heavy trucks with automated warehouses. When the intelligent heavy trucks receive loading and unloading alignment tasks, point cloud data of the alignment scene is acquired through a laser radar, point clouds of target areas are stably extracted through point cloud preprocessing, and the position and orientation of the alignment points are calculated. The intelligent heavy trucks are guided to park through data post-processing, and the alignment endpoint is output and kept updated at the same time. The method does not require site reconstruction, effectively reduces operating costs, and effectively outputs parking guidance for intelligent heavy trucks in alignment scenarios through laser radar detection, thereby improving parking operation efficiency and reducing alignment time loss for single vehicles. The adaptive point cloud preprocessing method is used to complete the stable extraction of alignment target point clouds for intelligent heavy trucks at different distances and postures, and the adjustment distance and direction of the intelligent heavy truck trailer are obtained after processing, which are converted into precise alignment endpoints, thereby improving the level of automation. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] Figure 1 Schematic diagram of the overall method flow of the present invention;

[0066] Figure 2 This is a schematic diagram of the intelligent heavy truck alignment scenario;

[0067] Figure 3 Schematic diagram of the laser radar arrangement and its field of view angle of the present invention;

[0068] Figure 4 This is a comparison diagram of the wall point cloud fitting before and after the present invention;

[0069] Figure 5 A schematic diagram of obtaining the lateral offset of the alignment endpoint of the present invention;

[0070] Figure 6 Schematic diagram of generating the alignment endpoint of the present invention;

[0071] Figure 7 Schematic diagram of the precise alignment guidance data curve of the present invention. DETAILED DESCRIPTION

[0072] The present invention is described in detail below with reference to the accompanying drawings:

[0073] In one embodiment of the present invention, a method for accurately aligning and guiding heavy trucks in an automated warehouse is disclosed. Figure 1-7 As shown, this method is used for the alignment of intelligent heavy trucks and automated warehouses, where Figure 2The demonstration showcases an alignment scenario for intelligent heavy-duty trucks in automated material transfer yards. Specifically, during material loading and unloading, the intelligent heavy-duty truck must first perform an alignment parking operation in the operating lane, accurately driving the vehicle to the alignment equipment (ground lock). Only after the trailer's rear cargo box is precisely locked with the warehouse alignment equipment can automated material loading and unloading proceed. Upon parking, the relative position and posture error between the intelligent heavy-duty truck's trailer and the automated warehouse door must be kept within a small range to accurately lock with the alignment equipment. Insufficient alignment parking accuracy can lead to difficulties in alignment, low loading and unloading efficiency, vehicle collisions, and equipment damage.

[0074] For the above-mentioned alignment scenario, this embodiment uses only one laser radar to achieve high-precision guidance of the vehicle in the longitudinal, lateral, and heading directions, allowing the smart heavy truck to accurately park in the warehouse and lock with the ground lock, efficiently completing the loading and unloading work without modifying the warehouse. Specifically, this embodiment includes the following steps:

[0075] The intelligent heavy-duty truck receives the loading and unloading alignment task and uses LiDAR to obtain point cloud data of the alignment scene. The alignment scene includes the automated warehouse location, automated warehouse door, and ground lock. After the intelligent heavy-duty truck's trailer is aligned, it is locked with the ground lock and docked with the automated warehouse door.

[0076] The point cloud data acquired by the LiDAR is pre-processed to extract the alignment wall surface used for detection and fitting as the target area point cloud. The alignment wall surface refers to the wall surface behind the ground lock, that is, the wall surface below the cargo loading and unloading port.

[0077] Detection and fitting are performed based on the extracted point cloud of the target area, and the relative position of the smart heavy truck and the alignment endpoint is calculated, including the longitudinal distance, lateral distance, and heading angle. The alignment endpoint is the location of the origin of the trailer coordinate system (a fixed point selected for the trailer body) when the alignment of the smart heavy truck is completed;

[0078] Based on the longitudinal distance, lateral distance and orientation angle of the intelligent heavy truck relative to the alignment end point, the position of the alignment end point in the origin coordinate system of the intelligent heavy truck trailer is obtained;

[0079] Based on the position of the alignment endpoint in the trailer's origin coordinate system, the intelligent heavy-duty truck is guided into the parking space and locked in place using a ground lock.

[0080] After the intelligent heavy truck parks at the storage location, it outputs the information of the positioning end point and keeps it updated.

[0081] In this embodiment, in response to the needs of parking scenarios and actual loading and unloading, a laser radar is set up and located in the middle of the rear of the trailer of the intelligent heavy truck. It is used to observe the positioning scenario of the intelligent heavy truck during parking, providing an important point cloud basis for this embodiment, such as Figure 3As shown; specifically, the height and angle arrangement of the laser radar are set according to the actual needs of the positioning scene.

[0082] In this embodiment, the specific guidance method includes the following parts:

[0083] 1. Point cloud preprocessing

[0084] The point cloud preprocessing part mainly transforms and extracts the position and posture (collectively referred to as posture) of the lidar point cloud, so that the point cloud in the target area can be screened out when the trailer of the intelligent heavy truck is in different postures, providing stable point cloud input for subsequent processes, realizing noise removal and data simplification.

[0085] Specifically, the process of point cloud preprocessing is:

[0086] A first selection is performed on the full amount of point cloud data acquired by the LiDAR, retaining the point cloud data within a fixed range from the center of the LiDAR, denoted as L1. This step ensures that the point cloud of the target area remains within the selection range when the trailer enters the warehouse at the maximum heading angle, and removes a large amount of point cloud data from non-key areas.

[0087] Calibrate the trailer's origin position when the trailer is parked and locked, and record it as the parking prior point position T prior ;

[0088] The relative position between the trailer origin and the parking point at that moment is calculated using the point cloud data obtained at the current moment.

[0089] Transform the point cloud data L1 from the trailer coordinate system to the parking point coordinate system to obtain the transformed point cloud data;

[0090] In the parking point coordinate system, the transformed point cloud data is selected for the second time, and the point cloud data within a fixed range is retained, which is recorded as L2;

[0091] The point cloud data L2 is output as the target area point cloud of the alignment wall.

[0092] When parking, the intelligent heavy-duty truck adjusts the trailer's posture. To address the constant shifting relative posture of the trailer and the parking space, point cloud selection is decoupled from the trailer's posture, enabling stable extraction of the target area point cloud. By selecting the area twice, a stable target area point cloud can be obtained for any trailer posture, serving as the subsequent input point cloud.

[0093] 2. Detection and fitting of wall point clouds

[0094] In this example, point cloud preprocessing yields a target wall area point cloud L2 for detection and fitting. This point cloud data is small, facilitating rapid algorithm processing, but still contains various noise factors, such as small devices, equipment piping, and uneven areas on the wall, which directly affect the wall point cloud fitting accuracy. The point cloud detection and fitting process represents the wall surface as a straight line equation, thereby determining the vertical distance from the trailer's origin to the aligned wall surface and the trailer's heading angle relative to the parking location, providing longitudinal and directional guidance during parking.

[0095] Specifically, point cloud detection and fitting includes the following steps:

[0096] Divide the point cloud of the target area into grids;

[0097] Calculate based on the divided grid to obtain the flatness and normal vector of each grid point cloud;

[0098] Based on the preset values ​​of flatness and normal vector, as well as the number of point clouds required, grid filtering is performed to filter out grids with insufficient flatness, large differences between normal vectors and preset values, and too few points, and extract the ideal plane point cloud L3 that can represent the wall surface; Figure 4 The bold point cloud in the extraction rendering;

[0099] Project the extracted ideal plane point cloud L3 onto the ground;

[0100] Perform two-dimensional straight line fitting on the projected point cloud to obtain the straight line equation of the wall surface: Ax+By+C=0;

[0101] Among them, A, B, and C are the coefficients and constant terms of the straight line equation, and x and y are the coordinate variables of the point on the ground after projection.

[0102] The above plane extraction and two-dimensional point cloud fitting are performed on the point cloud L2. The fitting results are as follows: Figure 4 The horizontal straight line in the fitting effect diagram.

[0103] Based on the fitted straight line equation, the relative position between the intelligent heavy truck and the alignment end point is calculated, including the longitudinal distance, lateral distance, and heading angle;

[0104] Among them, the longitudinal distance is the longitudinal distance x between the origin of the trailer coordinate system and the coordinate system of the end point. trailer , the expression is:

[0105]

[0106] x trailer =dd fix ;

[0107] Among them, d is the longitudinal distance from the center of the radar point cloud to the corresponding wall, d fixThe fixed longitudinal distance from the center of the radar point cloud to the alignment wall during alignment locking is a fixed parameter.

[0108] The heading angle is the relative heading angle θ between the trailer longitudinal direction and the normal direction of the opposing wall, and the expression is:

[0109]

[0110] Thus, the relative positions of the intelligent heavy truck and the alignment end point in the longitudinal and heading directions are obtained.

[0111] 3. Horizontal distance detection at the end point of alignment

[0112] Smart heavy trucks consist of a tractor and trailer. Their non-rigid nature makes lateral control more difficult, requiring precise lateral distance guidance. Therefore, the relative position of the alignment endpoint also includes the lateral distance of the alignment endpoint. This embodiment processes a reflector calibrated on the centerline of the storage location to obtain the lateral distance between the trailer and the alignment endpoint, providing the data required for subsequent generation of the alignment endpoint. The reflector is located on the alignment wall. The process specifically includes the following steps:

[0113] Obtain point cloud data of the alignment scene including the reflector through LiDAR;

[0114] Extracting the point cloud data of the area containing the reflector through the above-mentioned point cloud preprocessing;

[0115] Setting an intensity threshold and number of points for the reflector interior points, sorting the extracted regional point cloud data according to the intensity value of the point cloud, and extracting the reflector interior points from the sorted point cloud based on the settings;

[0116] like Figure 5 As shown in the figure, by taking the average value of the extracted inner point coordinates of the reflector, the coordinates of the center of the reflector in the trailer coordinate system (x reflector ,y reflector );

[0117] By derivation, it can be obtained that the lateral distance y between the origin of the trailer coordinate system and the coordinate system of the end point is trailer , the expression is:

[0118] y trailer =x reflector sinθ-y reflector ·cosθ;

[0119] Where θ is the relative heading angle between the trailer longitudinal direction and the normal direction of the opposing wall.

[0120] In order to efficiently obtain the final alignment endpoint coordinates, the above data are all expressed in the alignment endpoint coordinate system.

[0121] 4. Guided data post-processing

[0122] Through the above content, the longitudinal distance, lateral distance, and angular position relationship of the trailer origin of the intelligent heavy-duty truck relative to the alignment end point are accurately obtained. The ultimate goal is to obtain the position of the alignment end point in the trailer origin coordinate system to accurately guide the heavy-duty truck to complete the entire process from parking to alignment locking.

[0123] Specifically, the process of obtaining the position of the end point in the origin coordinate system of the intelligent heavy truck trailer includes the following steps:

[0124] like Figure 6 As shown, based on the trailer position (x trailer ,y trailer ) and the trailer heading angle θ, using the transformation matrix between the coordinate system of the end point and the trailer coordinate system Perform coordinate transformation and transformation matrix expression:

[0125]

[0126] Among them, x trailer Represents the longitudinal distance between the origin of the trailer coordinate system and the end-point coordinate system, y trailer represents the lateral distance between the origin of the trailer coordinate system and the coordinate system of the alignment end point, and θ is the relative heading angle between the longitudinal direction of the trailer and the normal direction of the alignment wall;

[0127] Then, the alignment endpoint posture P in the trailer coordinate system is destination , the expression is:

[0128]

[0129] Among them, P initial It is a constant vector, indicating the position of the alignment end point in the alignment end point coordinate system.

[0130] Furthermore, in this embodiment, the process of guiding the intelligent heavy truck to park in the parking space is to convert the alignment terminal position P in the trailer coordinate system into the global trailer origin position after time synchronization. destination The coordinates are converted into the global coordinate system to obtain the latitude and longitude coordinates and heading angle of the alignment end point in the global coordinate system, thereby achieving precise guidance of the alignment end point. Finally, when the trailer origin reaches the alignment end point and the heading angle is adjusted to be consistent, the alignment lock is completed accurately and automatic loading and unloading operations are performed.

[0131] This embodiment discloses the technical solution of the present invention through the above content. The specific parking alignment technical effects are described as follows:

[0132] When manually driving an intelligent heavy truck to perform the alignment and locking task in an automated warehouse, it is necessary to repeatedly adjust the trailer back and forth to accurately lock it with the alignment device (ground lock). However, the alignment guidance method of this embodiment can realize the autonomous and accurate alignment of the intelligent heavy truck. Usually, the number of adjustments will not exceed two times, and the overall parking alignment time is less. Figure 7 As shown, it is a parking data curve of an intelligent heavy truck using the alignment guidance method of this embodiment. From the data curve, it can be found that accurate alignment can be completed after one adjustment.

[0133] Based on the above, the precise alignment guidance method of the present invention has the following advantages over the prior art:

[0134] 1. The method of the present invention uses laser radar and reflectors to detect the automatic loading and unloading of intelligent heavy-duty trucks. It accurately outputs real-time parking endpoints without relying on parking space structure, enabling high-precision parking guidance without requiring site modifications, effectively reducing operating costs and improving the automation level and efficiency of intelligent heavy-duty truck production operations.

[0135] 2. The method of the present invention uses an adaptive point cloud preprocessing method to complete the stable extraction of the alignment target point cloud of the intelligent heavy-duty truck at different distances and postures. After processing, the adjustment distance and direction of the intelligent heavy-duty truck trailer are obtained and converted into a precise alignment end point. It does not rely on pre-stored information, is less affected by the environment, and has good robustness.

[0136] In another embodiment of the present invention, a precise alignment guidance system for heavy trucks in an automated warehouse is disclosed, the system comprising a task receiving and processing module, a data acquisition module, a point cloud preprocessing module, a detection and fitting calculation module, a relative posture calculation module, a parking guidance module, and an alignment information output and update module;

[0137] The task receiving and processing module is used to receive the loading and unloading alignment task issued by the upper system and trigger the alignment detection processing flow of the intelligent heavy truck when the task starts;

[0138] The data acquisition module is used to use the laser radar to collect point cloud data of the positioning scene in real time, including the storage locations, automated warehouse doors, and ground locks of the automated warehouse;

[0139] The point cloud preprocessing module is used to preprocess the point cloud data collected by the laser radar and extract the aligned wall point cloud for detection and fitting as the target area point cloud;

[0140] The detection and fitting calculation module is used to perform wall detection and fitting calculation based on the point cloud of the target area to obtain the relative position of the smart heavy truck and the alignment end point, including the longitudinal distance, lateral distance, and orientation angle;

[0141] The relative posture calculation module is used to calculate the posture of the alignment endpoint in the coordinate system of the intelligent heavy-duty truck trailer based on the relative posture of the intelligent heavy-duty truck and the alignment endpoint;

[0142] The parking guidance module is used to guide the intelligent heavy truck to park in a predetermined path and posture according to the results of the relative posture calculation module, and complete the positioning and locking through the ground lock;

[0143] The alignment information output and update module is used to output the alignment endpoint information after the intelligent heavy truck is parked in the storage space, and keep it updated in real time.

[0144] In another embodiment of the present invention, a vehicle is disclosed. The vehicle includes the above-mentioned guidance system and executes the above-mentioned guidance method.

[0145] The above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the scope of the present invention. Other modifications or equivalent substitutions made to the technical solutions of the present invention by ordinary technicians in this field should be included in the scope of the claims of the present invention as long as they do not depart from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for precise positioning guidance of heavy trucks in an automated warehouse, characterized in that: This method is used to align intelligent heavy trucks with automated warehouses and includes the following steps: The intelligent heavy-duty truck receives the loading and unloading alignment task and uses LiDAR to obtain point cloud data of the alignment scene. The alignment scene includes the automated warehouse location, automated warehouse door, and ground lock. After the intelligent heavy-duty truck's trailer is aligned, it is locked with the ground lock and docked with the automated warehouse door. Perform point cloud preprocessing on the point cloud data acquired by the LiDAR, and extract the aligned wall surface for detection and fitting as the target area point cloud; Detection and fitting are performed based on the extracted point cloud of the target area, and the relative position of the smart heavy truck and the alignment endpoint is calculated, including the longitudinal distance, lateral distance, and heading angle. The alignment endpoint is the location of the origin of the trailer coordinate system when the smart heavy truck is aligned; Based on the longitudinal distance, lateral distance and orientation angle of the intelligent heavy truck relative to the alignment end point, the position of the alignment end point in the origin coordinate system of the intelligent heavy truck trailer is obtained; Based on the position of the alignment endpoint in the trailer's origin coordinate system, the intelligent heavy-duty truck is guided into the parking space and locked in place using a ground lock. After the intelligent heavy truck parks at the storage location, it outputs the information of the positioning end point and keeps it updated.

2. The method for precise alignment and guidance of heavy trucks in an automated warehouse according to claim 1, characterized in that: The laser radar is provided in one piece and is located in the middle of the rear of the trailer of the smart heavy truck, and is used for observing the positioning scene during the entire parking process of the smart heavy truck.

3. The method for precise alignment and guidance of heavy trucks in an automated warehouse according to claim 1, characterized in that: The process of point cloud preprocessing is: In the full amount of point cloud data acquired by the LiDAR, the first frame selection is performed to retain the point cloud data within a fixed range from the center of the LiDAR, which is recorded as L1; Calibrate the trailer's origin position when the trailer is parked and locked, and record it as the parking prior point position T prior ; Using the point cloud data obtained at the current moment, calculate the relative pose between the trailer origin and the parking prior point at that moment; Transform the point cloud data L1 from the trailer coordinate system to the parking point coordinate system to obtain the transformed point cloud data; In the parking point coordinate system, the transformed point cloud data is selected for the second time, and the point cloud data within a fixed range is retained, which is recorded as L2; The point cloud data L2 is output as the target area point cloud of the alignment wall.

4. The method for precise alignment and guidance of heavy trucks in an automated warehouse according to claim 1, characterized in that: Point cloud detection and fitting includes the following steps: Divide the point cloud of the target area into grids; Calculate based on the divided grid to obtain the flatness and normal vector of each grid point cloud; Based on the preset values ​​of flatness and normal vectors, as well as the required number of point clouds, grid filtering is performed to extract the ideal plane point cloud L3 that can represent the wall surface; Project the extracted ideal plane point cloud L3 onto the ground; Perform two-dimensional straight line fitting on the projected point cloud to obtain the straight line equation of the wall surface: Ax+By+C=0; Among them, A, B, and C are the coefficients and constant terms of the straight line equation, and x and y are the coordinate variables of the point on the ground after projection.

5. The method for precise alignment and guidance of heavy trucks in an automated warehouse according to claim 4 is characterized in that: Based on the fitted straight line equation, the relative position between the intelligent heavy truck and the alignment end point is calculated, including the longitudinal distance, lateral distance, and heading angle; Among them, the longitudinal distance is the longitudinal distance x between the origin of the trailer coordinate system and the coordinate system of the end point. trailer , the expression is: x trailer =d-d fix ; Where d is the longitudinal distance from the center of the radar point cloud to the corresponding wall, d fix It is the fixed longitudinal distance from the center of the radar point cloud to the alignment wall during alignment locking; The heading angle is the relative heading angle θ between the trailer longitudinal direction and the normal direction of the opposing wall, and the expression is: Thus, the relative positions of the intelligent heavy truck and the alignment end point in the longitudinal and heading directions are obtained.

6. The method for precise alignment and guidance of heavy trucks in an automated warehouse according to claim 5, characterized in that: The lateral distance between the intelligent heavy truck and the alignment end point is obtained by processing a reflective plate calibrated on the center axis of the storage location. The reflective plate is located on the alignment wall. The specific steps include: Obtain point cloud data of the alignment scene including the reflector through LiDAR; Extracting point cloud data of the area containing the reflector through point cloud preprocessing; Setting an intensity threshold and number of points for the reflector interior points, sorting the extracted regional point cloud data according to the intensity value of the point cloud, and extracting the reflector interior points from the sorted point cloud based on the settings; The average value of the extracted inner point coordinates of the reflector is obtained to obtain the coordinates of the center of the reflector in the trailer coordinate system (x reflector ,y reflector ); Then, the lateral distance y between the origin of the trailer coordinate system and the coordinate system of the end point is trailer , the expression is: y trailer =x reflector ·sinθ-y reflector ·cosθ; Where θ is the relative heading angle between the trailer longitudinal direction and the normal direction of the opposing wall.

7. The method for precise alignment and guidance of heavy trucks in an automated warehouse according to claim 1, characterized in that: The process of obtaining the position of the end point in the origin coordinate system of the intelligent heavy-duty truck trailer includes the following steps: Based on the trailer position (x trailer ,y trailer ) and the relative heading angle θ, using the transformation matrix between the coordinate system of the end point and the trailer coordinate system Perform coordinate transformation and transformation matrix expression: Among them, x trailer Represents the longitudinal distance between the origin of the trailer coordinate system and the end-point coordinate system, y trailer represents the lateral distance between the origin of the trailer coordinate system and the alignment endpoint coordinate system, θ is the relative heading angle between the trailer longitudinal direction and the alignment wall normal; then, the alignment endpoint pose P in the trailer coordinate system destination , the expression is: Among them, P initial It is a constant vector, indicating the position of the alignment end point in the alignment end point coordinate system.

8. The method for precise alignment and guidance of heavy trucks in an automated warehouse according to claim 7, characterized in that: The process of guiding the intelligent heavy truck to park in the parking space is to transform the alignment endpoint posture in the trailer coordinate system into the global coordinate system to obtain the latitude and longitude coordinates and heading angle of the alignment endpoint in the global coordinate system; When the trailer origin reaches the alignment end point and the heading angle is consistent, the alignment lock is accurately completed.

9. A precise alignment and guidance system for heavy trucks in an automated warehouse according to any one of claims 1 to 8, characterized in that: The system includes a task receiving and processing module, a data acquisition module, a point cloud preprocessing module, a detection and fitting calculation module, a relative posture calculation module, a parking guidance module, and an alignment information output and update module; The task receiving and processing module is used to receive the loading and unloading alignment task issued by the upper system and trigger the alignment detection processing flow of the intelligent heavy truck when the task starts; The data acquisition module is used to use the laser radar to collect point cloud data of the positioning scene in real time, including the storage locations, automated warehouse doors, and ground locks of the automated warehouse; The point cloud preprocessing module is used to preprocess the point cloud data collected by the laser radar and extract the aligned wall point cloud for detection and fitting as the target area point cloud; The detection and fitting calculation module is used to perform wall detection and fitting calculation based on the point cloud of the target area to obtain the relative position of the smart heavy truck and the alignment end point, including the longitudinal distance, lateral distance, and orientation angle; The relative posture calculation module is used to calculate the posture of the alignment endpoint in the coordinate system of the intelligent heavy-duty truck trailer based on the relative posture of the intelligent heavy-duty truck and the alignment endpoint; The parking guidance module is used to guide the intelligent heavy truck to park in a predetermined path and posture according to the results of the relative posture calculation module, and complete the positioning and locking through the ground lock; The alignment information output and update module is used to output the alignment endpoint information after the intelligent heavy truck is parked in the storage space, and keep it updated in real time.

10. A vehicle, characterized in that: The vehicle includes the guidance system of claim 9 and executes the guidance method of any one of claims 1 to 8.

Images