Vehicle alignment method and straddle carrier

Abstract

The invention discloses a vehicle alignment method and a straddle carrier, and relates to a method for accurately aligning a lifting appliance of the straddle carrier with a transport vehicle and / or a container on the transport vehicle, and the method comprises the following steps: a global navigation step: enabling the straddle carrier to move from a starting position to a first position according to task information; an automatic navigation step: enabling the straddle carrier to move from the first position to the second position, and obtaining the pose information of the transport vehicle by using the data of the sensor in the moving process; and an alignment adjustment step: comparing the obtained position and posture information with a pre-stored transport vehicle template to enable the straddle carrier to move from the second position to the third position. Therefore, accurate alignment of the straddle carrier, the transport vehicle and / or the container on the transport vehicle can be achieved.

Description

Technical Field

[0001] The present application relates to the fields of vehicle automation and data processing, and particularly to a vehicle alignment method and a straddle carrier. Background Art

[0002] As is well known, in scenarios such as ports and factories, straddle carriers are widely used for handling containers. It grabs a container, drives to a designated location, places the container at the designated position, and then drives away. Nowadays, with the increasing demand for automation, the demand for the automation of straddle carriers is also increasing day by day, and the container handling operation process of straddle carriers has been realized in the container handling in fixed areas. Specifically, when the straddle carrier grabs and places the container on the ground or directly above the container already placed on the ground, since the positions of these grabs and placements can be confirmed in advance through a global positioning system such as Beidou navigation, the straddle carrier can accurately reach these grab and place positions to perform the grabbing and placing of the container. Among them, the alignment accuracy requirement between the straddle carrier and the container is 3 - 5 cm.

[0003] However, when the straddle carrier grabs or places a container in a scenario outside the fixed area, the above-mentioned method based on global positioning often cannot be used. For example, when the straddle carrier loads and unloads containers on a manned or self-driving container truck (hereinafter also referred to as "truck"), because the position of the truck is not fixed, even if information such as a parking space is preset, the position deviation after parking is often more than dozens of centimeters, which cannot meet the requirements for loading and unloading containers. Specifically, at the beginning of the operation, the straddle carrier detects the parking position of the truck in real time and drives above the truck. During the process of the spreader on the upper part of the straddle carrier descending, the position of the locking hole on the top of the container is detected from top to bottom by a camera installed on the spreader, and the locking head of the spreader is aligned with the locking hole of the container to perform the box-grabbing operation. Then, the straddle carrier drives above another truck, and the position of the locking head on the trailer or the locking hole on the top of the container is detected by the camera installed on the spreader, and the locking hole below the container is aligned with the locking head of the trailer to perform the box-releasing operation.

[0004] However, during the above box-grabbing and box-releasing operation process, the alignment accuracy requirement is also 3 - 5 cm, which requires very high sensor detection. Especially during the box-releasing operation, since the sensor installed on the spreader will be blocked by the grabbed container, even if the spreader is lowered to a position close to the truck, the position of the locking head of the trailer below cannot be observed, resulting in the inability to accurately detect the position of the truck or the container thereon, and the inability to achieve precise alignment between the straddle carrier and the truck or the container thereon. Summary of the Invention

[0005] Technical Problems to be Solved by the Invention

[0006] This application is formed to solve the above technical problems, and its purpose is to provide a vehicle alignment method and a straddle carrier that can achieve precise alignment between the straddle carrier and the transport vehicle or the container thereon.

[0007] Technical solutions for solving the technical problems

[0008] This application provides a vehicle alignment method, which is a method for precisely aligning the spreader of the straddle carrier with the transport vehicle and / or the container thereon, including: a global navigation step of moving the straddle carrier from the starting position to the first position according to the task information; an automatic navigation step of moving the straddle carrier from the first position to the second position, and obtaining the pose information of the transport vehicle by using the data of the sensor during the movement; and an alignment adjustment step of comparing the obtained pose information with the pre-stored transport vehicle template, and moving the straddle carrier from the second position to the third position.

[0009] Preferably, the first position is any position at a specified distance from the area where the transport vehicle is located.

[0010] Preferably, the specified distance is 5 - 10m.

[0011] Preferably, the second position is a position where the spreader of the straddle carrier is located above the transport vehicle and / or the container thereon.

[0012] Preferably, obtaining the pose information of the transport vehicle by using the data of the sensor during the movement includes: a data acquisition step of enabling the sensor to collect the sensor data during the process of the straddle carrier moving from the first position to the second position; a data annotation step of annotating the transport vehicle in the sensor data; and a data processing step of processing the sensor data by using the trained neural network model and outputting the pose information of the transport vehicle.

[0013] Preferably, in the data annotation step, the annotation content includes the category and pose information of the transport vehicle.

[0014] Preferably, offline model training is performed before the data processing step, and in the data processing step, the trained model is used for online real-time processing, inputting the data of the sensor and outputting the pose information of the transport vehicle.

[0015] Preferably, offline template recording is performed before the alignment adjustment step, scanning the transport vehicle data in the state of the straddle carrier with a container to record the transport vehicle template, and in the alignment adjustment step, online template comparison is performed, comparing the transport vehicle template and the obtained pose information of the transport vehicle, and determining the third position in a manner that the pose information matches the transport vehicle template.

[0016] Preferably, in the alignment adjustment step, the third position is determined in such a way that the pose information is exactly matched with the transporter template.

[0017] Preferably, in the alignment adjustment step, the third position is determined in such a way that the pose difference between the pose information and the transporter template is within the movable range of the spreader.

[0018] Preferably, the pose difference is compensated by adjusting the spreader.

[0019] The present application further provides a straddle carrier, including: a main body portion that can move on the ground and has an accommodation space inside; a grasping portion that is slidably mounted on the main body portion relative to the main body portion within the accommodation space; a sensor unit that has a plurality of sensors; a driving unit that is used to drive the main body portion and the grasping portion; and a control unit that controls the sensor unit to collect data of the transporter and obtain the pose information of the transporter when the straddle carrier moves. Among them, in the sensor unit, at least one sensor is mounted on the main body portion in such a way that it can detect the transporter accommodated in the accommodation space. The control unit compares the pose information of the transporter with a pre-stored transporter template and controls the straddle carrier to move to a position where the pose information matches the transporter template.

[0020] Preferably, the main body portion includes: a plurality of legs; and a plurality of wheels mounted at the bottoms of the plurality of legs; the at least one sensor is mounted on the legs.

[0021] Preferably, the at least one sensor is a lidar.

[0022] Therefore, in the present application, sensors are installed on the legs of the straddle carrier, and through staged navigation, the neural network model and the template registration technology are respectively combined in each stage of navigation. The alignment accuracy can reach 3 - 5 cm, and a vehicle alignment method that can achieve precise alignment between the straddle carrier and the transporter and / or the container thereon is realized. Furthermore, the automated operation of the straddle carrier to pick up and place the container relative to the transporter at a non-fixed position is realized, so the application scenario is greatly expanded. Description of the Drawings

[0023] Figure 1 is a schematic structural diagram of the straddle carrier;

[0024] Figure 2 is a flowchart of vehicle alignment;

[0025] Figure 3 is a schematic diagram of a pre-recorded point cloud template.

[0026] Symbol Description:

[0027] 1 - Straddle carrier, 11 - Main body part, 12 - Grabbing part, 111 - Leg, 112 - Wheel, 4 - Sensor Detailed implementation manner

[0028] Hereinafter, the present application will be further described in conjunction with the following implementation manners. It should be understood that the following implementation manners are only used to illustrate the present application and do not limit the present application.

[0029] Figure 1 is a schematic structural diagram of a straddle carrier. The following will be combined with Figure 1 to illustrate the basic structure of the straddle carrier.

[0030] The straddle carrier 1 mainly includes: a main body part 11 that can move on the ground and has an accommodation space inside; a grabbing part 12 installed in the accommodation space of the main body part 11; a sensor unit; and a driving unit and a control unit (both not shown). The straddle carrier uses the control unit to make the sensor unit detect the surrounding environment, and makes the driving unit drive to a specified position to accommodate a transport vehicle (which may carry a container) in the accommodation space inside the main body part 11, and the grabbing part 12 located above the transport vehicle (which may carry a container) grabs or places the container.

[0031] In this implementation manner, the main body part 11 is in the shape of a gantry, and adopts a wide gauge design (also known as a wide-body structure, WS: Wide-Structure) and a low center of gravity design, thereby having sufficient lateral stability to prevent the vehicle from tilting or rolling over when carrying heavy containers. Specifically, the main body part 11 mainly includes: a plurality of legs 111 and a plurality of wheels 112 installed at the bottoms of the plurality of legs 111. Among them, the plurality of legs 111 are arranged in two parallel rows to form a portal frame structure, which is used to support the weight of the vehicle and provide stability, and an accommodation space for accommodating a transport vehicle (which may carry a container) is formed inside. The plurality of wheels 112 are arranged in two rows and correspondingly provided below the legs 111, and disperse and bear the weight of the whole vehicle while moving. In addition, as Figure 1 shown, the four legs 111 enclose a rectangular parallelepiped-shaped accommodation space, and the eight wheels 112 are correspondingly arranged at the bottoms of the legs 111. However, the number, configuration, etc. of the legs 111 and the wheels 112 are only examples, and the present application is not limited thereto.

[0032] The gripping part 12 is slidably mounted in the accommodation space relative to the main body part 11. More specifically, the gripping part 12 is mounted in the accommodation space between the legs 111 of the main body part 11 in a manner that can slide relative to the main body part 11 at least in the height direction (i.e., the specified track). It mainly includes a gripping device that can grip and lift a container, such as a spreader. A locking mechanism and a spreader sensor (both not shown) are provided on the spreader. The locking mechanism, such as a gripper, a hook, a lock head, etc., can be locked with the locked mechanism on the transport vehicle or the container. The spreader sensor, such as a camera, a lidar, etc., can take a top-down view or scan. When the spreader descends above the container for the operation of gripping the container, the position of the container is detected by the spreader sensor to achieve accurate alignment, and the firm connection between the spreader and the container is ensured through the locking mechanism to achieve safe handling.

[0033] In addition, the sensor unit of the straddle carrier 1 includes a plurality of sensors installed at various parts of the straddle carrier 1. More specifically, it further includes at least one sensor 4. The sensor 4 is mounted on the main body part 11 in a manner that can capture the accommodation space, and is used to detect the position and pose of the transport vehicle (which may carry a container) entering the accommodation space. It is preferably a lidar. The "position and pose information" described in this application refers to the position and orientation of an object in three-dimensional space. In addition, the installation position of the sensor 4 is not specifically limited as long as it can scan the transport vehicle (which may carry a container) in the accommodation space. In addition, as Figure 1 shown, the number of the sensors 4 is preferably two or more, and is preferably installed on the side of the legs 111 facing the accommodation space respectively. Further preferably, the sensors 4 are installed on the side of the legs 111 facing the accommodation space in a manner opposite to each other. More preferably, the sensor 4 is installed inside the accommodation space of the legs 111, and the height is equivalent to the height of the transport vehicle or its trailer, so as to scan as many features of the transport vehicle or its trailer as possible.

[0034] Figure 1The state shown is the empty container state where the straddle carrier has not grasped the container. Although the illustration of the container-carrying state where the straddle carrier grasps the container is omitted, those skilled in the art can understand it. Additionally, it should be clear that in this application, throughout the entire process of the straddle carrier's autonomous driving, sensors (such as lidar, cameras, ultrasonic sensors, etc.) and known algorithm support means are utilized to prevent the straddle carrier from colliding with other objects, which will not be elaborated further below. Moreover, the transport vehicle mentioned in this application can be, for example, a container truck or other vehicles dedicated to transporting containers. The container is usually of a cuboid structure and has locking holes or other locking mechanisms near all eight corners, but it is not limited specifically thereto as long as similar functions can be achieved. Further, it should be further clarified that when the transport vehicle is loaded with a container, the locking mechanism on the transport vehicle and the locked mechanism on the container are locked with each other to prevent accidents during transportation. Therefore, the relative position between the transport vehicle and the container can be regarded as unique and fixed. In other words, given the pose information of the known transport vehicle, the specific information of the container can be completely determined based on the known relative position relationship. Thus, in some cases in this article, the transport vehicle and the container it loads are described as a whole.

[0035] Figure 2 is a flowchart of the alignment of the straddle carrier and the transport vehicle. Hereinafter, in combination with Figure 2 The vehicle alignment method described in this application will be described in detail, including:

[0036] Global navigation step S1: According to the task information, move the straddle carrier from the starting position to the first position;

[0037] Automatic navigation step S2: Move the straddle carrier from the first position to the second position, and collect data using sensor 4 and obtain the pose information of the transport vehicle during the movement; and

[0038] Alignment adjustment step S3: Compare the obtained pose information of the transport vehicle with the pre-stored transport vehicle template, and move the straddle carrier from the second position to the third position.

[0039] Specifically, in the global navigation step S1, the straddle carrier moves from the starting position to the first position according to the global positioning navigation and the autonomous driving planned path. The first position is an arbitrary position that has a specified distance from the area where the transport vehicle is located. The specified distance can be, for example, 5 to 10 m or the like. In this embodiment, the straddle carrier moves from the starting position to an arbitrary position within 10 m around the area where the transport vehicle is located, that is, the first position. In addition, the global positioning navigation mentioned here is a technology that determines the precise position of an object on the earth based on a satellite or other technical systems and provides route navigation. For example, it can be GPS (Global Positioning System) or other similar systems, such as BDS (China Beidou), GLONASS (Russia), and GNSS (European Galileo), etc., without special limitation. However, as mentioned above, the positioning accuracy of the civilian version of the global navigation is usually within the range of 5 to 10 m. Therefore, this application sets the first position to overcome the problem that the global positioning accuracy cannot meet the alignment requirements.

[0040] In the automatic navigation step S2, the straddle carrier moves from the first position to the second position based on autonomous driving, etc. The second position is the position where the spreader of the straddle carrier is located above the transport vehicle and / or the container thereon. That is, the straddle carrier accommodates the transport vehicle into its own accommodation space in such a way that the transport vehicle is located below its spreader. However, even when the straddle carrier moves to the second position where the transport vehicle is located below the spreader, due to the accuracy requirement for the container handling operation being 3 to 5 cm, it is very difficult for the relative position between the straddle carrier and the transport vehicle or the container placed thereon to meet the grasping or placing requirements. Therefore, the pose information of the transport vehicle is required for further alignment adjustment.

[0041] Therefore, during the process of the straddle carrier moving from the first position to the second position, the sensor 4 collects a large amount of real-time data during this process to obtain the pose information of the transport vehicle. That is, in this embodiment, the sensor 4 is a lidar (LiDAR). Therefore, a large amount of point cloud data will be collected during the movement of the straddle carrier. The collected point cloud data is input into a pre-trained neural network model, and then the pose information of the transport vehicle is output. Specifically, it includes:

[0042] Data acquisition step S2-1: The sensor 4 collects a large amount of point cloud data of the lidar during the process of the straddle carrier moving from the first position to the second position. During the acquisition process, ensure that the data acquisition covers the information of the transport vehicle under different environmental conditions and at various angles and distances to improve the generalization ability of the model.

[0043] Data annotation step S2-2: Perform manual annotation and / or automatic annotation to accurately label the transport vehicle in each frame of the collected point cloud data. During the annotation process, the annotation content may include, for example, the category of the transport vehicle (such as empty load or full load, etc.), pose information (such as three-dimensional bounding box, attitude angle, etc.), and other annotation content can also be added according to specific requirements, actual scenarios, etc.

[0044] Data processing step S2-3: Use the trained neural network model to process the point cloud data collected by the sensor 4 in real time and output the pose information of the transport vehicle.

[0045] Furthermore, offline model training is performed before the data processing step, and online real-time processing is performed during the data processing step.

[0046] Among them, offline model training is usually completed in advance in an offline environment, and multiple rounds of training are performed using fixed computing resources until the model reaches a satisfactory detection accuracy, such as being able to accurately detect the transport vehicle. In addition, since the sensor 4 in this embodiment is a lidar, in this embodiment, lidar point cloud data will be used in advance, and deep learning techniques will be used to train the neural network model to detect the pose information of the transport vehicle. Specifically as follows:

[0047] First, select a deep learning model. As existing open-source deep learning models, for example, PointPillar, SECOND, VoxelNet, etc. are known. In this embodiment, for example, PointPillar is selected as the deep learning model. PointPillar is a network that efficiently processes three-dimensional point cloud data, which can divide the point cloud data into multiple pillar grid structures (pillars), and then extract features through deep learning networks such as convolutional neural networks.

[0048] Next, adopt the supervised learning method for training. Use the point cloud data (input) collected by the lidar (LiDAR) and the corresponding category and pose information of the transport vehicle (output), etc. to train the above deep learning model. In this embodiment, during the training phase, the above model will receive the point cloud data as input and use the manually annotated output (category and pose information of the transport vehicle) as the ground truth label for comparison, calculate the error between the prediction result of the model and the annotated data through a loss function, etc., and update the parameters through backpropagation, etc., gradually learning how to better map the input point cloud to the output category and pose information to improve its performance in the detection task.

[0049] In addition, during the above model training process, hyperparameters such as the learning rate and batch size can be adjusted as necessary for model tuning, or cross-validation can also be used to evaluate the performance of the model to ensure its consistency on the training data and validation data.

[0050] Online real-time processing usually involves deploying the trained neural network model in a straddle carrier or a corresponding system, receiving the real-time point cloud data stream from the lidar (sensor 4) online, processing it using the neural network model, and outputting the position information of the transport vehicle. The details are as follows:

[0051] First, the model can pre-process the received real-time point cloud data as needed, such as denoising and segmenting ROI (region of interest) to improve the detection efficiency of the model.

[0052] Next, the model outputs the position and orientation information of the transport vehicle based on the input point cloud data, such as the position and orientation in three-dimensional space (for example, expressed in Euler angles or quaternions).

[0053] Based on this, the output result can be used for subsequent tasks, such as autonomous driving path planning and cargo loading and unloading operations. In addition, although the above example illustrates the operation method when the sensor 4 is a laser radar, the sensor 4 can also be a visual detection device such as a camera, for example, through image analysis such as YOLO, the above similar method can be used for data collection and model training.

[0054] In the alignment adjustment step S3, the acquired posture information is compared with the pre-stored transport vehicle template, so that the straddle carrier is moved from the second position to the third position. The third position is a position where the acquired posture information of the transport vehicle matches the pre-stored transport vehicle template, and is also a position where the locking mechanism of the spreader of the straddle carrier can be aligned with the locked mechanism of the transport vehicle or container.

[0055] Specifically, offline template recording is performed before the alignment adjustment step S3, and online template comparison is performed in the alignment adjustment step S3.

[0056] Offline template recording is designed to generate an accurate transport vehicle template to assist the straddle carrier in identifying and locating the transport vehicle in the automated container grabbing and placing task, so as to accurately grab and place the container. As mentioned above, there are many situations when the straddle carrier grabs and places the container. Situation 1: The straddle carrier is empty and the transport vehicle is loaded with containers, and the straddle carrier performs a container grabbing operation relative to the transport vehicle; Situation 2: The straddle carrier is loaded with containers and the transport vehicle is empty, and the straddle carrier performs a container placing operation relative to the transport vehicle; Situation 3: The straddle carrier is loaded with containers and the transport vehicle is loaded with containers, and the straddle carrier performs a container placing (stacking) operation relative to the transport vehicle. It can be seen that in situations 2 and 3, since the straddle carrier is loaded with containers, the spreader sensor is blocked by the container and it is basically impossible to perform environmental detection. Therefore, template recording at this time is extremely important.

[0057] Template recording is usually also performed with the aid of sensor 4. In this embodiment, Figure 3As shown, when the sensor 4 uses a lidar, the transport vehicle template is a point cloud template, and the pose information obtained by real-time detection based on the sensor 4 is compared with the point cloud template pre-stored via the sensor 4. Specifically, the point cloud template is recorded as described below.

[0058] First, the operator drives the straddle carrier to move above the transport vehicle and zero the spreader. During this process, try to ensure the accurate relative position of the straddle carrier and the trailer to facilitate accurate subsequent data recording. Zeroing the spreader is to adjust the spreader of the straddle carrier to the standard initial position, thereby reducing variables during the recording process and ensuring the consistency of data recording and the reliability of the template.

[0059] Next, with the position of the spreader fixed, the operator adjusts the position of the straddle carrier until the locking mechanism of the spreader is accurately aligned with the locked mechanism of the transport vehicle or container below, and locks the two with each other.

[0060] Next, lift the spreader and the container vertically to a predetermined position, and record the scan data at this position through the sensor 4 (other sensors can also be used for auxiliary recording if necessary).

[0061] Next, through a preset range and filtering method, remove the redundant laser point clouds (such as the point clouds of the ground and surrounding objects), and only retain the point cloud data related to the transport vehicle as the point cloud template.

[0062] In addition, the point cloud template is preferably a three-dimensional point cloud template, but it can also be a two-dimensional point cloud template as long as the accuracy is sufficient. In addition, the sensor 4 is preferably converted to record the template from a top-down perspective. The pre-recorded point cloud template can be stored in the straddle carrier or related systems. When there are multiple types of transport vehicles, multiple point cloud templates can be recorded correspondingly.

[0063] Online template comparison refers to comparing the pre-recorded transport vehicle template with the pose information of the transport vehicle obtained from the real-time scan data based on the sensor 4, and moving the straddle carrier in a way that makes the two fully match. When the transport vehicle template is completely aligned with the actual pose of the transport vehicle, it means that the straddle carrier reaches the third position. At this time, the straddle carrier can perform the grasping and releasing of the container relative to the transport vehicle at the third position.

[0064] In this embodiment, for example, the ICP (Iterative Closest Point) algorithm can be used to achieve this. The ICP algorithm is a commonly used method for aligning two sets of point clouds in three-dimensional space. By continuously iterating to find the best registration, the finally output transformation matrix describes how the source point cloud (the data to be registered, that is, the pose of the transport vehicle in this embodiment) needs to be transformed to align with the target point cloud (the reference data, that is, the point cloud template in this embodiment) (that is, it contains the pose difference information).

[0065] Further, in this embodiment, in the alignment adjustment step S3, during the process of matching the acquired pose information of the transport vehicle with the pre-stored transport vehicle template, the third position can be determined in such a way that the pose information is exactly matched with the transport vehicle template. The straddle carrier moves from the second position to the third position, for example, by means of automatic adjustment such as slight forward movement, backward movement, left and right offset, and even turning, so as to align the actual pose of the transport vehicle with the template pose (exact match).

[0066] In addition, in the alignment adjustment step S3, during the process of matching the acquired pose information of the transport vehicle with the pre-stored transport vehicle template, the third position can also be determined in such a way that the pose difference between the pose information and the transport vehicle template is within the movable range of the spreader. Specifically, the spreader of the straddle carrier itself has a certain movable range (i.e., adjustment range), for example, it can linearly move a specified distance in the up and down, left and right, and front and back directions relative to the main body 11. Therefore, when the acquired pose difference is within the movable range of the spreader, instead of moving the straddle carrier, the spreader can be moved to compensate for the pose difference.

[0067] That is to say, if the pose difference is small, the third position can be determined as the position of exact match, and at this time, the straddle carrier is moved to eliminate the pose difference; or the third position can be determined as any position within the movable range of the spreader, and at this time, there is no need to move the straddle carrier, and the pose difference is eliminated by the movement of the spreader itself.

[0068] If the pose difference is large and cannot be covered only by the movement of the spreader itself, then the straddle carrier needs to be moved. The third position can be determined as the position of exact match, and at this time, the straddle carrier is moved to eliminate the pose difference; or the third position can be determined as any position within the movable range of the spreader, and at this time, the straddle carrier is moved to within the movable range of the spreader, and then the spreader is moved to compensate for the pose difference.

[0069] Thus, by combining spreader adjustment and vehicle movement, a larger operation range and operation precision are ensured, and by reasonably judging the deviation range, unnecessary vehicle movement is reduced, and the work efficiency is improved.

[0070] When the acquired pose information of the transport vehicle is aligned with the pre-stored transport vehicle template, that is, the straddle carrier is precisely aligned with the transport vehicle and / or the container. Then, during the container grasping operation, the locking mechanism of the spreader of the straddle carrier can be locked with the locked mechanism of the container on the transport vehicle (specifically, the locked mechanism located at the top). During the container releasing operation, the locked mechanism of the container grasped by the spreader of the straddle carrier (specifically, the locked mechanism located at the bottom) can be locked with the locking mechanism on the transport vehicle or neatly stacked with the container on the transport vehicle.

[0071] In addition, as described above, when the sensor 4 uses a sensor device such as a camera, template recording and graphic comparison (such as by means of image analysis, etc.) can also be performed in a similar manner to the above, so as to perform alignment adjustment.

[0072] In summary, in this application, a sensor is installed on the outrigger of the straddle carrier, and through staged navigation, a vehicle alignment method that can achieve precise alignment between the straddle carrier and the transport vehicle or the container thereon is realized, and then the automated operation of the straddle carrier to pick up and place the container relative to the transport vehicle at a non-fixed position is realized. Specifically, this application does not require additional sensors to be configured on the site, has good applicability and high practicality, and combines the neural network model and the template registration technology in each stage of navigation respectively. The alignment accuracy can reach 3-5 cm, which fully meets the requirements of the spreader to grasp and release the container. Moreover, due to the real-time detection of the pose of the transport vehicle, the application scenario is greatly expanded, and the target object is not limited to vehicles and can be extended to non-fixed loading and unloading platforms, etc.

[0073] In addition, any process or method description shown in the flowchart or described in other ways herein can be understood as representing a module, segment, or part of code including one or more executable instructions for implementing a specific logical function or process. The scope of the preferred embodiments of this application includes additional implementations, where the functions can be executed in a manner that is not shown or discussed, including in a substantially simultaneous manner or in the reverse order according to the involved functions, which should be understood by those skilled in the technical field to which the embodiments of this application belong.

[0074] In addition, those of ordinary skill in the art can realize that the units and algorithm steps of the examples described in conjunction with the embodiments disclosed herein can be implemented by electronic hardware, computer software, or a combination of the two. To clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described according to functions in the above description. Whether these functions are executed in a hardware or software manner depends on the specific application and design constraints of the technical solution. Professional technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of this application.

[0075] Finally, it should be understood that this application is not limited to the exact structure described above and shown in the drawings, and various modifications and changes can be made without departing from its scope. The scope of this application is only limited by the appended claims.

Claims

1. A vehicle alignment method, characterized in that: It is a method of accurately aligning the straddle carrier's spreader with the transport vehicle and / or the containers on it, including: a global navigation step, moving the straddle carrier from the starting position to the first position according to the task information; an automatic navigation step, moving the straddle carrier from the first position to the second position, and obtaining position information of the transport vehicle using data from the sensor during the movement; and The step of adjusting the position of the straddle carrier is to compare the acquired position information with a pre-stored transport vehicle template, so as to move the straddle carrier from the second position to the third position.

2. The vehicle alignment method according to claim 1, characterized in that: The first position is any position that is at a specified distance from the area where the transport vehicle is located.

3. The vehicle alignment method according to claim 1, characterized in that: The specified distance is 5 to 10 m.

4. The vehicle alignment method according to claim 1, characterized in that: The second position is a position where the spreader of the straddle carrier is located above the transport vehicle and / or a container thereon.

5. The vehicle alignment method according to claim 1, characterized in that: During the movement, the sensor data is used to obtain the position information of the transport vehicle, including: A data collection step, causing the sensor to collect sensor data during the process of the straddle carrier moving from the first position to the second position; a data labeling step, labeling the transport vehicle in the sensor data; and The data processing step uses the trained neural network model to process the sensor data and output the position information of the transport vehicle.

6. The vehicle alignment method according to claim 5, characterized in that: In the data labeling step, the labeling content at least includes the category and posture information of the transport vehicle.

7. The vehicle alignment method according to claim 5, characterized in that: Offline model training is performed before the data processing step, and in the data processing step, the trained model is used to perform online real-time processing, input the data of the sensor and output the position information of the transport vehicle.

8. The vehicle alignment method according to claim 1, characterized in that: Before the alignment adjustment step, an offline template recording is performed, and the transport vehicle data of the straddle carrier with boxes is scanned to record the transport vehicle template. In the alignment adjustment step, an online template comparison is performed to compare the transport vehicle template with the acquired position and posture information of the transport vehicle, so as to determine the third position in a manner that the position and posture information matches the transport vehicle template.

9. The vehicle alignment method according to claim 8, characterized in that: In the alignment adjustment step, the third position is determined in such a way that the posture information completely matches the transport vehicle template.

10. The vehicle alignment method according to claim 8, characterized in that: In the alignment adjustment step, the third position is determined in such a way that the posture difference between the posture information and the transport vehicle template is within the range of motion of the spreader.

11. The vehicle alignment method according to claim 10, characterized in that: The posture difference is compensated by adjusting the sling.

12. A straddle carrier, comprising: A main body, which can move on the ground and has a receiving space inside; a gripping portion, which is slidably mounted on the main body in the accommodation space relative to the main body; a sensor unit having a plurality of sensors; A driving unit, used for driving the main body and the grasping part; as well as a control unit, which controls the sensor unit to collect data of the transport vehicle and obtains position information of the transport vehicle when the straddle carrier moves. It is characterized in that In the sensor unit, at least one sensor is installed on the main body in a manner capable of detecting the transport vehicle accommodated in the accommodation space. The control unit compares the position and posture information of the transport vehicle with a pre-stored transport vehicle template, and controls the straddle carrier to move to a position where the position and posture information matches the transport vehicle template.

13. The straddle carrier according to claim 12, characterized in that: The main body includes: a plurality of legs; and a plurality of wheels mounted on the bottom of the plurality of legs; The at least one sensor is mounted to the leg.

14. The straddle carrier according to claim 12, characterized in that: The at least one sensor is a lidar.

Images