A method, system and product for parking control of a terminal automated guided vehicle intelligent unlocking station

Abstract

The present application relates to the technical field of automated control of guided vehicles, and provides a wharf automated guided vehicle intelligent unlocking station parking control method, system and product, comprising: in response to a first sensor detecting a feature of a first identification position of a container, controlling the automated guided vehicle to slow down; in response to detecting a feature of a second identification position, controlling the vehicle to further slow down and enable a second sensor; obtaining a first distance between the second identification position and a first target position and a second distance between the first identification position and a second target position at a low speed, and controlling braking to stop when the two distances satisfy a preset condition; after parking, obtaining the first distance and the second distance to determine an actual position offset, and sending the actual position offset to an unlocking mechanical arm control system to revise a grabbing coordinate. The present application realizes millimeter-level accurate parking of the automated guided vehicle, eliminates relative displacement errors between the vehicle body and the load, breaks through a data barrier between positioning and execution, and significantly improves the operation efficiency of the unlocking station.

Description

Technical Field

[0001] This application relates to a parking control method, system, and product for intelligent unlocking stations of automated guided vehicles (AGVs) at docks, used in the field of intelligent control of AGVs at docks. Background Technology

[0002] With the rapid development of automated container terminals worldwide, using Automated Guided Vehicles (AGVs or IGVs, etc.) in conjunction with automated quay cranes and rail-mounted gantry cranes for container transfer has become an important mode for improving the efficiency of container handling at terminals. In terminal operations, the installation and removal of container twistlocks (lock pins) has long relied on manual labor, which is not only high-risk and labor-intensive but also restricts the overall operational efficiency of the terminal. Therefore, the industry is gradually setting up intelligent unlocking stations in the landside interaction area of ​​terminals, deploying robots or robotic arms to automatically install and remove twistlocks on containers carried by AGVs, replacing manual operation.

[0003] In typical existing AGV parking control schemes, when an AGV enters a smart unlocking station, it primarily relies on global positioning methods such as onboard LiDAR, inertial navigation, or magnetic nails embedded in the ground for deceleration and stopping control. The system pre-calibrates a theoretical target position for stopping, and guides the AGV to gradually decelerate and eventually stop at this preset position through the onboard positioning system. Subsequently, the robotic arm of the unlocking station identifies and locates the container corner pieces or twist locks based on its own vision system, performing grasping and disassembly operations. The key to this control logic lies in the positioning and adjustment of the AGV's body position.

[0004] However, in practical applications, the above solution has revealed the following technical defects: Firstly, there is a clear contradiction between the required parking accuracy and the dynamic control capability of the AGV. The working range of the robotic arm imposes certain requirements on the final parking accuracy of the AGV. However, when the AGV is traveling on a wet, slippery, and slightly sloping road surface under heavy load, it is difficult for it to stop accurately at the preset target center in one go due to the influence of inertia and changes in wheel-ground forces. The actual parking position often exceeds the effective working tolerance range of the robotic arm.

[0005] Secondly, the existing control scheme is essentially for positioning the "vehicle" rather than the "container". Factors such as the swaying of the AGV during operation, the initial loading position of the container, the elastic deformation of the container on the trailer, and load offset will all cause a significant deviation between the actual spatial position of the container end face and corner pieces and the theoretical position of the AGV body. Relying solely on the global positioning of onboard sensors cannot effectively detect and reflect this deviation.

[0006] The accumulation of these deviations directly leads to frequent operational failures for the robotic arm due to its inability to identify corner components or out-of-tolerance twist lock positions. This necessitates multiple micro-adjustments by the AGV, and sometimes even manual intervention for reset and guidance. Such repeated interruptions severely reduce the single-station processing capacity and throughput of the unlocking station, and may cause queuing congestion for subsequent AGVs, resulting in delays in the terminal's logistics cycle. Summary of the Invention

[0007] The purpose of this application is to overcome the shortcomings of the prior art and provide a method, system and product for parking control of automated guided vehicles (AGVs) at intelligent unlocking stations in docks. For the parking process of AGVs at intelligent unlocking stations, it realizes precise position control and deviation perception between containers and robotic arm workstations, thereby improving the success rate of parking and the efficiency of automatic lock removal and installation operations.

[0008] Firstly, this application provides a docking control method for an intelligent unlocking station of an automated guided vehicle (AGV) at a wharf, the technical solution of which includes the following steps: In response to the first sensor detecting the positional characteristics of a first identification location on the container being transported by the automated guided vehicle, the automated guided vehicle is controlled to decelerate from a first speed to a second speed; In response to the first sensor detecting the positional characteristics of the second identification position on the container, the automated guided vehicle is controlled to decelerate from the second speed to the third speed, and the second sensor is enabled; wherein, the second identification position is located behind the first identification position in the direction of container movement, and the second sensor is located in front of the first sensor in the direction of container movement; At the third speed, the first distance between the second identification position of the container and the first target position collected by the first sensor, and the second distance between the first identification position of the container and the second target position collected by the second sensor are obtained. When the first distance and the second distance meet the preset conditions, the automated guided vehicle is controlled to brake and stop. After receiving the signal that the automated guided vehicle has come to a complete stop, the first distance and the second distance in the stopped state of the automated guided vehicle are obtained to determine the actual position offset with respect to the first target position and the second target position. The actual position offset is then sent to the unlocking robotic arm control system to revise the grasping coordinates.

[0009] By adopting the above technical solution, this application shifts the reference benchmark for parking control from the AGV body to the container itself. Dual sensors positioned at the front and rear directly detect the container's front and rear position characteristics, enabling direct perception of the container's actual position. Through a control link involving multi-stage deceleration and coordinated micro-motion alignment, the conflict between operational efficiency and parking accuracy is balanced. Gradual deceleration suppresses the inertial impact of the heavy-duty AGV, providing sufficient control and measurement margin for millimeter-level parking accuracy. Simultaneously, by introducing static deviation measurement and compensation after parking, and using this data as a pre-compensation parameter for the robotic arm's hand-eye calibration, the system's fault tolerance and adaptability are expanded, significantly improving the first-time success rate of automatic lock removal and installation.

[0010] Preferably, both the first sensor and the second sensor are lidar, the first identification position is the front face of the container, the second identification position is the rear face of the container, the first sensor and the second sensor acquire three-dimensional point clouds of the front face and the rear face of the container, and calibrate the positional features of the first identification position and the second identification position.

[0011] By adopting the above technical solution, specific sensor types and sensing data formats for achieving high-precision position perception are provided. LiDAR is used to acquire 3D point clouds of the front and rear faces, which are then used as the basis for determining the container's position. This eliminates the need for special structural modifications or feature structure settings and training for the container or automated guided vehicle (AGV). It effectively resists interference from partial occlusion, dirt, or deformation, providing a data foundation for subsequent position perception calculations and AGV control.

[0012] Preferably, the location feature is a trigger signal when the first or second identification position of the container enters a preset warning position within the range of the smart unlocking station.

[0013] By adopting the above technical solution, the identification location is specified as a virtual calibration line or spatial relationship preset within the unlocking station. This gives the system an objective and quantifiable spatial benchmark for judging when the AGV enters the work buffer zone and when it is about to reach the target. This ensures that when AGVs with different working conditions and different loads enter, the deceleration action can be accurately triggered at a certain and repeatable spatial position, improving the standard consistency of the control process and the certainty of system operation.

[0014] Preferably, in S200, after controlling the automated guided vehicle to decelerate from the first speed to the second speed, and before detecting the position feature of the second identified position, the following is further included: The first sensor acquires the three-dimensional point cloud trajectory of the front face of the container in real time to form the actual travel path; The actual travel path is compared with the preset path to obtain the deviation angle between the two, and the distance deviation between the travel path and the plane where the smart unlocking station is located is obtained. Based on the deviation angle and the distance deviation, the automated guided vehicle is controlled to steer, so that the automated guided vehicle moves along a preset path at a preset distance.

[0015] By adopting the above technical solution, dynamic control of the AGV's posture during travel is achieved. Since AGVs may still drift laterally or deviate in direction at low speeds due to road conditions or load, this solution can identify in advance that the AGV is not traveling straight towards the target workstation and proactively correct its direction. This ensures that before entering the final S300 alignment and parking stage, the AGV's body and container orientation are essentially parallel to the intelligent unlocking station workstation and at the correct lateral distance. By eliminating lateral and angular deviations, the complex multidimensional deviation problem during final parking is simplified to a distance deviation problem in the direction of travel, greatly reducing the difficulty of final coordinated alignment and further increasing the probability of accurate parking on the first attempt.

[0016] Preferably, the first target position and the second target position are the calibration positions of the unlocking robot arm in the point cloud space where the three-dimensional point cloud is located; the first distance is the distance from the center point of the rear end face of the container to the first target position, and the second distance is the distance from the center point of the front end face of the container to the second target position; the preset condition is that the first distance and the second distance are equal.

[0017] By adopting the above technical solution, the target position is directly defined as the calibration position point of the unlocking robotic arm, linking the definition of precise parking with the ideal working coordinate point of the robotic arm. This reduces the difficulty of subsequent identification of container corner fittings and twist locks by the robotic arm, improving the success rate and operational accuracy of lock removal and installation. Using the equality of the first and second distances as the parking condition essentially involves finding the geometric center point of the container relative to the two workstations before and after it. This is a simple, efficient, and computationally inefficient centering strategy that simultaneously ensures the convenience of the unlocking robotic arms before and after the operation while guaranteeing the real-time performance and robustness of the control system.

[0018] Preferably, there are at least three unlocking robotic arm stations, arranged at intervals along the direction of the intelligent unlocking station; Before the automated guided vehicle travels at the third speed to a braking stop, the following is also included: The length and model of the container are determined by the three-dimensional point cloud of the front and rear faces of the container. Based on the length and model of the container, select the unlocking robotic arm station with the corresponding spacing, and define the calibration point of the selected unlocking robotic arm station in the point cloud space as the first target position and the second target position.

[0019] By adopting the above solution, the berthing control system is able to handle the operational conditions of containers of different sizes. Automated terminals handle multiple container sizes simultaneously, such as 20-foot (double 20-foot), 40-foot, and 45-foot containers. This technical solution uses LiDAR point cloud data to identify the container length online and dynamically calls up the matching robotic arm station coordinate set. It can adapt to the container size and model for configuring and aligning the robotic arm station, without the need for manual presets or information on the container size and model through other systems. The entire unlocking station can achieve fully automatic, adaptive, and precise berthing of containers of multiple sizes, significantly improving the system's automation level and flexible operation capabilities.

[0020] Secondly, this application provides a docking control system for an intelligent unlocking station of an automated guided vehicle at a dock. The technical solution adopted includes an intelligent unlocking station, a sensor module, an unlocking robotic arm module, a spatial perception module, and a central controller. The sensor module includes a first sensor and a second sensor, which are respectively installed on the intelligent unlocking station. The first sensor is located behind the container in the direction of travel, and the second sensor is located in front of the container in the direction of travel. It is used to acquire the first identification position and the second identification position data on the container carried by the automated guided vehicle. The unlocking robotic arm module includes a plurality of unlocking robotic arms, and the intelligent unlocking station is provided with unlocking robotic arm workstations at intervals, the number of which corresponds to the number of unlocking robotic arms. The spatial perception module is used to construct and calculate the three-dimensional spatial position information of the container and its positional relationship with the smart unlocking station based on the data acquired by the first sensor and the second sensor. The sensor module is connected to the spatial perception module. The central controller is connected to both the spatial perception module and the unlocking robotic arm module, and is also communicatively connected to the external automated guided vehicle control system. The central controller is configured as follows: When the first sensor detects the positional characteristics of the first identification position on the container being transported by the automated guided vehicle, a second speed deceleration signal is sent to the external automated guided vehicle control system. When the first sensor detects the positional characteristics of the second identification position on the container being transported by the automated guided vehicle, a third speed deceleration signal is sent to the external automated guided vehicle control system, and the second sensor is enabled. The system acquires a first distance collected by a first sensor and a second distance collected by a second sensor. When the first distance and the second distance meet a preset condition, it sends a braking stop signal to the external automated guided vehicle control system. The system receives a stop signal from the external automated guided vehicle (AGV) control system, then sends detection commands to the first and second sensors to obtain the first and second distances when the AGV is stopped, in order to determine the actual position offsets of the first and second target positions. The actual position offsets are then sent to the unlocking robotic arm module to revise the grasping coordinates.

[0021] Preferably, the intelligent unlocking stations are arranged in two symmetrical rows, forming a travel channel for the automated guided vehicle between the two rows of intelligent unlocking stations, so as to simultaneously install and remove the locking pins at the four corners of the container. The first sensor and the second sensor are installed on one side of the intelligent unlocking station.

[0022] By adopting the above technical solution, a system architecture of dual-sided intelligent unlocking stations is provided. The two rows of intelligent unlocking stations are symmetrically arranged, allowing the AGV to pass through in a straight line so that the locking pins at the four corners can be disassembled and installed simultaneously by the robotic arms on both sides. Moreover, dual-sided operation can be achieved by deploying a sensor system on only one side of the intelligent unlocking station.

[0023] Preferably, a visual monitoring module and a manual control module are also included; The visualization monitoring module is connected to the spatial perception module and is used to display the point cloud spatial images of the automated guided vehicle, the container, and the smart unlocking station in real time. The manual control module is connected to the external automated guided vehicle control system through the central controller, and is used to receive input from monitoring personnel to manually take over the operation control of the automated guided vehicle.

[0024] By adopting the above technical solution, a human-machine collaborative monitoring and intervention mechanism was constructed. This allows monitoring personnel to observe the relative positional relationships between the AGV, containers, and unlocking station from a global perspective, and provides an effective intervention channel under abnormal operating conditions. When the system encounters sensor failure, extreme environmental interference, or sudden situations exceeding the boundaries of automatic processing logic, monitoring personnel can immediately take over the operation control of the AGV. This provides the system with fault tolerance, avoids the production risk of the entire unlocking station stopping due to automatic alignment failure, and significantly improves the system's practicality and reliability.

[0025] Thirdly, this application provides a computer program product, the technical solution of which includes a computer program or instructions, which, when executed by a processor, implement the steps of the above-mentioned dock automated guided vehicle intelligent unlocking station berthing control method.

[0026] In summary, this application includes at least one of the following beneficial technical effects: 1. This application uses dual sensors to directly detect the first and second position features on the container body, completely transferring the reference benchmark for parking control from the automated guided vehicle (AGV) body to the container itself. This avoids the perception blind spot of "vehicle inaccurate container" caused by factors such as vehicle body swaying, load offset, and trailer deformation, thus solving the parking accuracy bottleneck of the AGV and laying an accurate perception foundation for the subsequent automatic disassembly and assembly control of the unlocking robotic arm.

[0027] 2. This application establishes a three-level refined speed control strategy, from coarse deceleration and fine deceleration to dual-sensor collaborative micro-motion alignment, effectively balancing the contradiction between operational efficiency and parking accuracy. More importantly, a static deviation compensation stage is introduced after the automated guided vehicle (AGV) comes to a complete stop. This stage converts the actual positional offset of the AGV after it comes to a standstill into a calculable and compensable digital offset for the unlocking robotic arm control system. This provides a coordinate translation alignment reference for the unlocking robotic arm to identify the container corner fittings and torsion locks, forming a complete closed loop of "positioning-stopping-measurement-compensation." This scheme significantly reduces the stringent requirements for the absolute stopping accuracy of the AGV, greatly improves the success rate of automatic unlocking, and eliminates the efficiency losses caused by repeated fine-tuning and manual intervention.

[0028] 3. In the application of this application, on the one hand, the length and specifications of containers are identified online using 3D point cloud data of the container end face, and the corresponding robotic arm workstation coordinate group is dynamically selected, achieving precise positioning services for containers of different specifications such as 20-foot, 40-foot, and 45-foot, without the need for manual pre-setting or external system notification; on the other hand, the integration of the visualization monitoring module and the manual control module provides the system with a human-machine collaborative intervention channel under extreme abnormal operating conditions, enabling the system to have graceful degradation tolerance. These two designs significantly improve the adaptability and availability of the system in real and complex terminal environments from the dimensions of automation flexibility and operational reliability, respectively. Attached Figure Description

[0029] Figure 1 This is a flowchart illustrating a docking control method for an automated guided vehicle (AGV) at a smart unlocking station in an embodiment of this application. Figure 2 This is a schematic diagram of the spatial layout of a smart unlocking station in one embodiment of this application; Figure 3 This is a schematic diagram of the architecture of a dock automated guided vehicle intelligent unlocking station parking control system according to an embodiment of this application. Detailed Implementation

[0030] This specific embodiment is merely an explanation of this application and is not intended to limit it. After reading this specification, those skilled in the art can make modifications to this embodiment without contributing any inventive step, but such modifications are protected by patent law as long as they are within the scope of this application.

[0031] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application. It should be noted that in the optional embodiments of this application, the object information and other related data involved require the permission or consent of the object when the embodiments of this application are applied to specific products or technologies, and the collection, use, and processing of related data must comply with the relevant laws, regulations, and standards of the relevant countries and regions. That is to say, if the embodiments of this application involve data related to the object, it needs to be obtained with the authorization and consent of the object, the authorization and consent of the relevant departments, and in compliance with the relevant laws, regulations, and standards of the country and region. If personal information is involved in the embodiments, the acquisition of all personal information requires the consent of the individual. If sensitive information is involved, the separate consent of the information subject is required, and the embodiments also need to be implemented with the authorization and consent of the object.

[0032] The embodiments of this application will now be described in further detail with reference to the accompanying drawings. Example 1:

[0033] like Figure 1 As shown, this embodiment provides a method for controlling the parking of automated guided vehicles (AGVs) at intelligent unlocking stations in docks. This method, through the collaborative operation of two sensors, achieves precise parking and deviation compensation for AGVs carrying containers at intelligent unlocking stations, solving the problem of AGV position adjustments required due to insufficient parking accuracy causing robotic arm operation failures.

[0034] First, it should be noted that the berthing control system implementing this method is subordinate to the intelligent unlocking station control system. It interacts internally with the unlocking robotic arm control system under the intelligent unlocking station control system, and communicates with the AGV control system of the dock through the central controller of the intelligent unlocking station control system.

[0035] Specifically, this method includes the following steps.

[0036] In step S100, in response to the first sensor detecting the positional characteristics of the first identification position on the container being transported by the automated guided vehicle, the automated guided vehicle is controlled to decelerate from the first speed to the second speed.

[0037] In this embodiment, the AGV enters the working area of ​​the intelligent unlocking station at a relatively high speed. This initial speed is typically the AGV's full-load travel speed or medium-speed travel speed, and the specific speed is set according to the port's operational efficiency requirements and safety regulations. The AGV is not monitored by the intelligent unlocking station's control system before entering the working area of ​​the intelligent unlocking station.

[0038] The first sensor is deployed at the entrance area of ​​the intelligent unlocking station, its detection direction pointing towards the characteristic parts of the container carried by the AGV. When the container carried by the AGV moves with the vehicle body, and the first identification position on it enters the sensing range of the first sensor and triggers a specific position feature, the system determines that the AGV has entered the parking preparation area. At this time, the central controller of the intelligent unlocking station control system sends a command to the AGV control system, takes over the control of the AGV, and sends a deceleration command to reduce the vehicle speed to a second speed. This second speed is significantly lower than the first speed, for example, it can be a value in the range of 1m / s to 2m / s, aiming to reduce the vehicle's inertia, create conditions for subsequent precise alignment, and prevent the vehicle from overshooting the stop position due to excessive speed.

[0039] Regarding the location feature, it can specifically be a trigger signal indicating that the container's first or second identification position enters a preset warning position within the range of the smart unlocking station. For example, a virtual or physical warning line is preset on the ground or frame structure of the smart unlocking station. When the first sensor detects that the container's first identification position (such as the front face of the container) crosses this warning line, the trigger signal is generated. This triggering method is simple, reliable, and has a fast response speed, effectively identifying the timing of the container's entry. It should be understood that the specific coordinates of the preset warning position can be adjusted according to different container sizes or different unlocking station layouts; this embodiment does not impose any limitations on this.

[0040] In step S200, in response to the first sensor detecting the positional characteristics of the second identification position on the container, the automated guided vehicle is controlled to decelerate from the second speed to the third speed, and the second sensor is enabled; wherein, the second identification position is located behind the first identification position in the direction of container movement, and the second sensor is located in front of the first sensor in the direction of container movement.

[0041] As the AGV continues to move at the second speed, the first sensor continuously monitors the container's position. When the first sensor detects the second identification position on the container, the system determines that the AGV is about to reach the target working position. Since the second identification position is located behind the first identification position in the container's direction of movement (such as the rear end of the container), this detection process actually reflects the process of the container's overall structure entering the parking preparation area. At this time, the central controller sends a deceleration command again, reducing the vehicle speed to the third speed. This third speed is a crawling speed or micro-motion speed to ensure high-precision parking control at extremely low speeds.

[0042] Simultaneously, the system enables the second sensor, causing it to begin operation. Since the second sensor is located in front of the first sensor in the container's direction of movement, i.e., closer to the interior of the unlocking station, after the container enters the berthing preparation area, the second sensor can more accurately point to the container's first identification position in subsequent steps, precisely identifying the displacement of the first identification position and its distance from the target position. At this point, the first sensor, in subsequent steps, more accurately points to the container's second identification position, and the identification task shifts to monitoring the displacement of the second identification position.

[0043] The second sensor is deployed in front of the unlocking station. Initially, when the AGV enters, the container has not yet entered its effective detection range. Activating it prematurely would generate a large amount of empty data or false detection signals, increasing the system's unnecessary processing burden. Activating the second sensor only after the container has fully entered the berthing preparation area, and binding the sensor's activation time to the container's actual spatial position, avoids invalid detections and signal conflicts, ensuring that its output is always valid and usable measurement data.

[0044] In step S300, at the third speed, the first distance between the second identification position of the container and the first target position collected by the first sensor, and the second distance between the first identification position of the container and the second target position collected by the second sensor are obtained. When the first distance and the second distance meet the preset conditions, the automatic guided vehicle is controlled to brake and stop.

[0045] During the crawling speed phase, the first and second sensors work together to measure the container's position information in real time. Specifically, the first and second target positions are the unlocking robot arm positions pre-calibrated in the system. The first distance measured by the first sensor is the distance from the container's second identification position to the rear position of the unlocking robot arm; the second distance measured by the second sensor is the distance from the container's first identification position to the front position of the unlocking robot arm. The central controller monitors the values ​​of the first and second distances in real time. When both meet preset conditions, it indicates that the actual coordinates of the two identification positions of the container meet the working requirements of the unlocking robot arm's operating range, allowing the system to stop and begin the unlocking operation.

[0046] More specifically, the preset condition is that the first distance equals the second distance. The length of the container is a known value based on its model, and the positions of the two sets of unlocking robotic arms in the intelligent unlocking station can also be preset according to the size of the container. The distance between the two workstations matches the container length and the rated working range of the robotic arms. Under this premise, when the measured first distance and second distance are equal, it indicates that the container is exactly centered between the two sets of workstations, and the distance from each corresponding workstation is equal. It must be within the optimal working range of the unlocking robotic arms, which can simultaneously take into account the working range of the front and rear unlocking robotic arms and avoid any target working position from exceeding the flexible working space range of the unlocking robotic arms due to being too close to or too far from the workstation. This equidistant judgment logic is simple and efficient, and can make parking decisions in real time without complex calculations. It is the preferred solution for achieving fast and accurate positioning control in engineering practice. The parking control method based on dual-sensor distance criteria directly uses the characteristic position of the container itself as a reference, eliminating the relative displacement error that may exist between the AGV body and the container, thereby achieving millimeter-level parking accuracy.

[0047] Step S400: After obtaining the signal that the automated guided vehicle has completely stopped, obtain the first distance and the second distance in the stopped state of the automated guided vehicle to determine the actual position offset of the first target position and the second target position, and send the actual position offset to the unlocking robotic arm control system to revise the grasping coordinates.

[0048] The AGV's complete stop signal is provided by the AGV control system based on the AGV's motion status. After receiving the AGV's complete stop signal, the central controller records the first and second distances obtained by the first and second sensors in the current state.

[0049] When the AGV comes to a complete stop, due to factors such as inertia, braking delay, or slippery road surface, the actual stopping position of the container may deviate slightly from the ideal target position. At this time, the system reads the first and second distances measured by the first and second sensors, and calculates the actual position offset of the container relative to the first and second target positions. The system sends this actual position offset to the unlocking robotic arm control system via a data interface. The unlocking robotic arm control system automatically corrects the robotic arm's grasping coordinates based on this offset.

[0050] It should be noted that the aforementioned positional offset is essentially a macroscopic correction of the deviation between the robotic arm's original reference coordinate system and the actual container pose, rather than a microscopic positioning of the twist lock or corner piece gripping point. Specifically, before performing operations, the robotic arm's control system has a preset original reference coordinate system based on the intelligent unlocking station. Without this offset correction, the robotic arm will assume the container is parked at its theoretically calibrated position and activate its built-in machine vision system to search for and identify corner pieces and twist locks. However, due to the deviation between the actual stopping position and the theoretical position of the container, the robotic arm's visual search window may completely or partially deviate from the actual spatial range of the corner pieces, leading to recognition failures or requiring a significant expansion of the search area, consuming additional time or even causing operation timeouts. This solution sends the aforementioned offset as a pre-compensation parameter to the robotic arm control system, its essential function being to drive the robotic arm's original reference coordinate system to translate entirely to the correct position matching the actual container pose. After this correction, the robotic arm's vision recognition system no longer needs to blindly search the global space. Instead, it performs precise recognition within a defined area that highly overlaps with the actual position of the corner piece. This provides the robotic arm with prior area guidance, accurately directing its vision system's working range to the target neighborhood. The final precise positioning and grasping of the corner piece and twist lock are still completed by the robotic arm's built-in machine vision system. This collaborative approach essentially reduces the search uncertainty and computational overhead during hand-eye alignment, thereby significantly improving the first-time success rate of automated unlocking operations.

[0051] By implementing this method, this embodiment breaks down the data barriers between AGV parking control and robotic arm operation execution, realizing a closed loop of "perception-control-execution", which greatly improves the success rate and efficiency of automated operations. Example 2:

[0052] Based on the above embodiments, in this embodiment, both the first sensor and the second sensor are lidar, the first identification position is the front face of the container, the second identification position is the rear face of the container, the first sensor and the second sensor acquire three-dimensional point clouds of the front face and the rear face of the container, and calibrate the position features of the first identification position and the second identification position.

[0053] Specifically, lidar features high ranging accuracy and strong resistance to ambient light interference, making it particularly suitable for open-air port operations. The first lidar is deployed at the entrance of the intelligent unlocking station, its scanning field covering the front face of the container when the AGV enters the work area, and the rear face of the container when the AGV is within the work area; the second lidar is deployed deep within the intelligent unlocking station, its scanning field covering the front face of the container when the AGV is within the work area.

[0054] When an AGV carrying a container enters, the lidar emits a laser beam and receives the reflected signal, generating dense 3D point cloud data. Using a random sampling consensus algorithm or a region growing algorithm, a set of 3D point clouds conforming to planar features is extracted from the effective point cloud clusters; this set represents the 3D point cloud of the container's front or rear face. The centroid of the extracted planar point cloud is calculated to obtain the geometric center point of the plane, which is then used as the position calibration point for that end face. This process transforms the complex spatial point cloud into precise coordinate parameters, providing a reliable data foundation for subsequent distance calculations.

[0055] Furthermore, considering that during the process of the automated guided vehicle (AGV) entering the intelligent unlocking station, its direction of travel may deviate from the centerline of the unlocking station due to uneven road surfaces, differences in tire wear, or initial alignment deviations. If not corrected in time, this attitude deviation will directly affect the subsequent accurate positioning and parking, and may even cause the container to exceed the working range of the robotic arm. Therefore, this embodiment introduces real-time deviation correction logic for the travel path after controlling the AGV to decelerate from the first speed to the second speed, and before detecting the positional features of the second identification position.

[0056] Specifically, during this phase, the automated guided vehicle (AGV) maintains a stable second speed, providing favorable dynamic conditions for path correction. Meanwhile, a first lidar unit deployed at the entrance continuously scans the front face of the container at high frequency, acquiring multiple frames of 3D point cloud data to calculate the normal vector of that plane. Since the front face of the container should ideally be perpendicular to the direction of travel, the projection direction of this normal vector onto the horizontal plane represents the container's current orientation. The system tracks the movement trajectory of the center point of the front face in multiple consecutive frames of point cloud data to fit the actual travel path of the AGV.

[0057] The system then compares the actual travel path with the preset path. The preset path is typically defined as a straight line at a fixed distance from the smart unlocking station, or the central axis of two smart unlocking stations arranged opposite each other. The system calculates the angle between the actual travel path and the preset path, defining it as the deviation angle; it also calculates the vertical distance between the actual travel path and the plane (or side wall reference plane) where the smart unlocking station is located, calculating the distance deviation value. For example, if a 2-degree deviation angle is detected between the normal direction of the container's front face and the central axis of the unlocking station, it means the vehicle is traveling at an angle. If the distance deviation value shows the vehicle deviating 5 centimeters from the central axis, it means the vehicle is laterally offset.

[0058] Based on the calculated deviation angle and distance deviation, the central controller generates corresponding steering control commands and sends them to the automated guided vehicle (AGV) control system. These steering control commands can be generated using a proportional-integral-derivative (PID) control algorithm, aiming to gradually converge the deviation angle to zero and the distance deviation to a preset distance tolerance range by adjusting the AGV's steering angle. Through this closed-loop real-time correction control, the AGV can move along a preset path, ensuring that its attitude is adjusted to an ideal positive alignment state during the fine alignment stage at the third speed stage, thus greatly improving the success rate and accuracy of final parking. It should be understood that the above correction process is dynamically and continuously performed while the AGV is traveling at the second speed, until the first sensor detects the three-dimensional point cloud position features of the container's rear end face. At this point, the correction stage ends, and the system enters the fine deceleration stage.

[0059] After the AGV slows down to the third speed, the central controller acquires the first distance between the rear end of the container and the first target position, which is collected by the first lidar, and the second distance between the front end of the container and the second target position, which is collected by the second lidar.

[0060] Specifically, the first target position and the second target position are the calibration positions of the unlocking robotic arm station in the point cloud space; the first distance is the distance from the center point of the rear face of the container to the first target position, and the second distance is the distance from the center point of the front face of the container to the second target position.

[0061] It should be understood that the first and second target positions are the workstation positions of the unlocking robot arm, or the corresponding optimal grasping points of the unlocking robot arm, pre-determined in the point cloud spatial coordinate system of the LiDAR using high-precision calibration tools during the initial construction of the intelligent unlocking station. By accurately locating the optimal workstation point of the robot arm in the point cloud space constructed by the LiDAR, the deviation measured in real time by the sensors can be accurately mapped to the translation correction amount of the robot arm coordinate system under a unified coordinate system reference. The accuracy of this correction amount depends on the precise determination of the target position as the "reference zero point." Furthermore, by pre-calibrating and fixing the target position in the point cloud spatial coordinate system, a standardized workstation reference data file can be formed. When the unlocking robot arm undergoes minor position adjustments due to maintenance or replacement, only the calibration data needs to be updated, without modifying the core control algorithm, making the deployment and migration of the system between different intelligent unlocking stations more convenient and efficient.

[0062] Following the above embodiment, during the parking control process, the system calculates the first distance D1 and the second distance D2 in real time, and the preset condition for controlling the automated guided vehicle to brake and stop is D1=D2. The calculation formula is as follows: D1 = |P back- Target1| Where P back Target1 is the real-time coordinate of the center point of the back end face, and Target2 is the calibration coordinate of the first target position. D2 = |P front - Target2| Where P front Target1 represents the real-time coordinates of the center point of the front face, and Target2 represents the calibrated coordinates of the second target position.

[0063] After the AGV comes to a complete stop, the measured values ​​D1 and D2 are fed back to the unlocking robotic arm control system to revise the gripping coordinates of the unlocking robotic arm. Example 3:

[0064] Building upon the above embodiments, and further considering the diverse specifications and models of containers in automated terminal operations, commonly including 20-foot (double 20-foot), 40-foot, and 45-foot lengths, the spacing between the corner fittings (i.e., twist lock installation positions) varies depending on the container length. If the system were designed only for a single specification, it would severely limit the equipment's versatility. Therefore, this embodiment provides a control logic for container length recognition and workstation adaptation.

[0065] Please see Figure 2 Specifically, there are at least three unlocking robotic arm stations (single-sided), arranged at intervals along the direction of the intelligent unlocking station 1. For example, the intelligent unlocking station can have three stations: station A, station B, and station C, with station A at the front, station C at the rear, and station B in the middle. Before the automated guided vehicle (AGV) reaches a braking stop at a third speed (i.e., crawling speed), the system uses the three-dimensional point cloud data of the front and rear faces of the container 13 collected by the first sensor 11 and the second sensor 12. Using the principle of triangulation or directly calculating through the difference in point cloud coordinates, the system can accurately determine the distance between the front and rear faces of the container and execute the container length identification procedure. It should be understood that due to possible deformation of the container end face or sensor measurement errors, the system usually sets a certain tolerance range. For example, a calculated length between 11.5 meters and 12.5 meters is identified as a 40-foot container.

[0066] After identifying the length and specifications of the container, the system selects the corresponding unlocking robotic arm station based on these specifications and defines the calibrated position of the selected unlocking robotic arm station in the point cloud space as the first target position and the second target position. This process achieves the dynamic definition of target positions, meaning that the first and second target positions are not fixed physical coordinates, but rather logical coordinates dynamically mapped according to the container specifications.

[0067] For example, when a 20-foot container is detected, its corner fittings have a shorter distance between them, so the system selects the foremost workstations A and B as the working stations. At this time, the system maps the robotic arm calibration point corresponding to workstation A to the second target position and the robotic arm calibration point corresponding to workstation B to the first target position. When a 40-foot container is detected, its corner fittings have a longer distance between them, so the system selects the front workstation A and the rear workstation C as the working stations, or selects workstations B and C, depending on the physical layout spacing of the workstations and the container model, while adjusting the target position calibration accordingly.

[0068] Through the above logic, the system has already completed the adaptive matching of the workstation before the automated guided vehicle (AGV) comes to a complete stop. Subsequently, in the subsequent parking control steps, the system uses these dynamically defined first and second target positions as a reference to calculate the first and second distances, and controls the AGV to brake and stop. This design allows the same intelligent unlocking station system to seamlessly accommodate the handling of containers of various sizes without the need for manual intervention to switch modes, significantly improving the system's automation level and operational efficiency. Furthermore, since the target position is dynamically generated based on the actual detected container length, this also avoids the risk of robotic arm grasping failure or collision due to incorrect workstation selection. Example 4:

[0069] Please see Figure 3 This embodiment provides a docking control system for an intelligent unlocking station of an automated guided vehicle (AGV) at a dock. This system is used to implement the docking control method described in any of the above embodiments. The system includes an intelligent unlocking station 1, a sensor module 2, an unlocking robotic arm module 3, a spatial perception module 4, and a central controller 5.

[0070] Specifically, the intelligent unlocking station 1 is the physical carrier of the entire system. It is usually built in a specific operating area of ​​the automated terminal. Its structural design needs to meet the passage requirements of the automated guided vehicles and the working space requirements of the unlocking robotic arm.

[0071] Sensor module 2 includes a first sensor and a second sensor, which are respectively positioned at the rear and front of the automated guided vehicle (AGV) along the direction of travel on the intelligent unlocking station 1. These sensors are used to acquire the first and second identification positions on the container carried by the AGV. In this embodiment, the first and second sensors are preferably lidar. Sensor module 2 is communicatively connected to spatial perception module 4 via Ethernet or fiber optic cable, transmitting the collected raw point cloud data to spatial perception module 4 in real time.

[0072] The unlocking robotic arm module 3 includes several unlocking robotic arms, and the intelligent unlocking station 1 has unlocking robotic arm workstations spaced at intervals corresponding to the number of unlocking robotic arms. The unlocking robotic arms are the actuators that perform the torsion lock installation and removal actions, typically using industrial robots with high-precision motion control capabilities. Each workstation corresponds to the working range of one robotic arm, and the layout of the workstations needs to be optimized according to the specifications and model of the container. The robotic arm module is connected to the central controller 5 via a fieldbus (such as EtherCAT or Profinet), receiving control commands and providing feedback on the execution status.

[0073] The spatial perception module 4 is used to construct and calculate the three-dimensional spatial position information of the container and its positional relationship with the smart unlocking station 1 based on the data acquired by the first and second sensors. This module is typically implemented by a high-performance industrial control computer or embedded computing unit, and internally runs point cloud processing algorithms (such as the RANSAC algorithm and plane fitting algorithm in the aforementioned embodiments). The spatial perception module 4 receives the raw point cloud data transmitted by the sensor module 2, performs filtering, segmentation, feature extraction and other processing, calculates the position coordinates of the front and rear ends of the container, the trajectory deviation, the container length and other key parameters, and packages these parameters and sends them to the central controller 5.

[0074] The central controller 5 is connected to the space perception module 4 and the unlocking robotic arm module 3, and also communicates with the external automated guided vehicle (AGV) control system 6. The central controller is the logical core of the entire system, typically employing a programmable logic controller (PLC) or a high-performance industrial computer. It has a pre-set parking control logic program responsible for determining the current state of the AGV based on parameters provided by the space perception module 4 and generating corresponding control commands.

[0075] The system's signal flow and control logic are as follows: When the first sensor detects the positional characteristics of the first identification position on the container being transported by the AGV, the central controller sends a second speed deceleration signal to the external AGV control system. The AGV's onboard controller receives the signal and executes the deceleration action. When the first sensor detects the positional characteristics of the second identification position on the container, the central controller sends a third speed deceleration signal to the external AGV control system and enables the second sensor. At this time, the second sensor powers on and begins collecting data. The central controller acquires the first distance collected by the first sensor and the second distance collected by the second sensor. When the first and second distances meet preset conditions, a braking stop signal is sent to the external AGV control system. After the AGV stops, the central controller receives the stop signal from the external AGV control system and then sends detection commands to the first and second sensors to acquire the first and second distances in the stopped state of the AGV. This determines the actual positional offset relative to the first and second target positions, and the actual positional offset is sent to the unlocking robotic arm module to revise the grasping coordinates.

[0076] Furthermore, the intelligent unlocking stations are arranged in two symmetrical rows, forming a passageway for the automated guided vehicle (AGV) to simultaneously install and remove the locking pins at the four corners of the container. The first and second sensors are located on one of the intelligent unlocking stations. This symmetrical layout allows the robotic arms on both sides to simultaneously operate on the four corner pieces on both sides of the container, eliminating the need for AGV rotation or lateral movement and significantly shortening the operation cycle. Positioning the sensors on one side avoids cross-interference between the scanning fields of the sensors on both sides and simplifies wiring and synchronization logic. It should be understood that in other embodiments, sensors can also be deployed on both sides separately, further improving perception accuracy through data fusion.

[0077] In addition, the system includes a visual monitoring module and a manual control module. The visual monitoring module is connected to the spatial perception module and is used to display the point cloud spatial images of the AGV, containers, and intelligent unlocking station in real time. Monitoring personnel can intuitively see the real-time position of the AGV, the posture outline of the container, and the working status of the robotic arm through the display screen. When the system detects abnormalities (such as abnormal point cloud data or excessive deviation of the AGV from the path), the visual interface will issue an alarm through highlighting, flashing, or pop-up windows. The manual control module is connected to the external AGV control system through the central controller and is used to receive input from the monitoring personnel to manually control the movement of the AGV. When the automatic control mode fails or special working conditions are encountered, the monitoring personnel can manually control the start and stop, speed adjustment, and micro-positioning of the AGV through the operating handle or touch screen to ensure operational safety. This redundant design of "automatic as the main mode and manual as the auxiliary mode" greatly improves the robustness and emergency handling capability of the system. Example 5:

[0078] This embodiment provides a computer program product, including a computer program or instructions, which, when executed by a processor, implements the steps of the dock automation vehicle intelligent unlocking station berthing control method described in any of the above embodiments.

[0079] Specifically, the computer program product can be carried by various forms of storage media. For example, the storage media can be read-only memory (ROM), random access memory (RAM), flash memory, hard disk, optical disk, or cloud server, etc., and this embodiment does not limit this. The storage medium stores a computer program, which includes at least a piece of executable code or instructions for being called and executed by a controller or processor.

[0080] In a specific hardware architecture, the processor can be a microprocessor (MCU), digital signal processor (DSP), or field-programmable gate array (FPGA) within the central controller of the intelligent unlocking station. When the processor reads and executes the computer program in the storage medium, it triggers a series of logical operations, which correspond one-to-one with the steps in the aforementioned method embodiments. For example, when the processor executes the program, it reads the detection data from the first and second sensors through the interface circuit and generates control commands based on preset logical judgment rules (such as whether the first distance and the second distance meet preset conditions). These control commands are sent to the control system of the automated guided vehicle or the control system of the unlocking robotic arm through a communication interface (such as an Ethernet interface, a CAN bus interface, or a wireless communication module), thereby realizing automated control of the automated guided vehicle's deceleration, stopping, and robotic arm coordinate correction.

[0081] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A method for controlling the berthing of automated guided vehicles (AGVs) at intelligent unlocking stations in docks, characterized in that, include: In response to the first sensor detecting the positional characteristics of a first identification location on the container being transported by the automated guided vehicle, the automated guided vehicle is controlled to decelerate from a first speed to a second speed; In response to the first sensor detecting the positional characteristics of the second identification position on the container, the automated guided vehicle is controlled to decelerate from the second speed to the third speed, and the second sensor is enabled; wherein, the second identification position is located behind the first identification position in the direction of container movement, and the second sensor is located in front of the first sensor in the direction of container movement; At the third speed, the first distance between the second identification position of the container and the first target position collected by the first sensor, and the second distance between the first identification position of the container and the second target position collected by the second sensor are obtained. When the first distance and the second distance meet the preset conditions, the automated guided vehicle is controlled to brake and stop. After receiving the signal that the automated guided vehicle has come to a complete stop, the first distance and the second distance in the stopped state of the automated guided vehicle are obtained to determine the actual position offset with respect to the first target position and the second target position. The actual position offset is then sent to the unlocking robotic arm control system to revise the grasping coordinates.

2. The method for controlling the berthing of an automated guided vehicle (AGV) at a dock according to claim 1, characterized in that, Both the first sensor and the second sensor are lidar. The first identification position is the front face of the container, and the second identification position is the rear face of the container. The first sensor and the second sensor acquire three-dimensional point clouds of the front face and the rear face of the container, and calibrate the positional features of the first identification position and the second identification position.

3. The method for controlling the berthing of an automated guided vehicle (AGV) at a dock according to claim 1, characterized in that, The location feature is a trigger signal when the first or second identification position of the container enters a preset warning position within the range of the smart unlocking station.

4. The docking control method for an intelligent unlocking station of an automated guided vehicle at a wharf according to claim 2, characterized in that, In S200, after controlling the automated guided vehicle to decelerate from the first speed to the second speed, and before detecting the position feature of the second identified position, the following is also included: The first sensor acquires the three-dimensional point cloud trajectory of the front face of the container in real time to form the actual travel path; The actual travel path is compared with the preset path to obtain the deviation angle between the two, and the distance deviation value between the travel path and the plane where the smart unlocking station is located is obtained. Based on the deviation angle and the distance deviation value, the automated guided vehicle is controlled to turn, so that the automated guided vehicle moves along a preset path at a preset distance.

5. The docking control method for an intelligent unlocking station of an automated guided vehicle (AGV) at a wharf according to claim 2, characterized in that, The first target position and the second target position are the calibration positions of the unlocking robot arm in the point cloud space where the three-dimensional point cloud is located; the first distance is the distance from the center point of the rear face of the container to the first target position, and the second distance is the distance from the center point of the front face of the container to the second target position; the preset condition is that the first distance and the second distance are equal.

6. The docking control method for an intelligent unlocking station of an automated guided vehicle at a wharf according to claim 5, characterized in that: There are at least three unlocking robotic arm stations, arranged at intervals along the direction of the intelligent unlocking station; Before the automated guided vehicle travels at the third speed to a braking stop, the following is also included: The length and model of the container are determined by the three-dimensional point cloud of the front and rear faces of the container. Based on the length and model of the container, select the unlocking robotic arm station with the corresponding spacing, and define the calibration point of the selected unlocking robotic arm station in the point cloud space as the first target position and the second target position.

7. A dockside automated guided vehicle (AGV) intelligent unlocking station berthing control system, characterized in that, It includes an intelligent unlocking station, a sensor module, an unlocking robotic arm module, a spatial perception module, and a central controller; The sensor module includes a first sensor and a second sensor, which are respectively installed on the intelligent unlocking station. The first sensor is located behind the container in the direction of travel, and the second sensor is located in front of the container in the direction of travel. It is used to acquire the first identification position and the second identification position data on the container carried by the automated guided vehicle. The unlocking robotic arm module includes a plurality of unlocking robotic arms, and the intelligent unlocking station is provided with unlocking robotic arm workstations at intervals, the number of which corresponds to the number of unlocking robotic arms. The spatial perception module is used to construct and calculate the three-dimensional spatial position information of the container and its positional relationship with the smart unlocking station based on the data acquired by the first sensor and the second sensor. The sensor module is connected to the spatial perception module. The central controller is connected to both the spatial perception module and the unlocking robotic arm module, and is also communicatively connected to the external automated guided vehicle control system. The central controller is configured as follows: When the first sensor detects the positional characteristics of the first identification position on the container being transported by the automated guided vehicle, a second speed deceleration signal is sent to the external automated guided vehicle control system. When the first sensor detects the positional characteristics of the second identification position on the container being transported by the automated guided vehicle, a third speed deceleration signal is sent to the external automated guided vehicle control system, and the second sensor is enabled. The system acquires a first distance collected by a first sensor and a second distance collected by a second sensor. When the first distance and the second distance meet a preset condition, it sends a braking stop signal to the external automated guided vehicle control system. The system receives a stop signal from the external automated guided vehicle (AGV) control system, then sends detection commands to the first and second sensors to obtain the first and second distances when the AGV is stopped, in order to determine the actual position offsets of the first and second target positions. The actual position offsets are then sent to the unlocking robotic arm module to revise the grasping coordinates.

8. The intelligent unlocking station berthing control system for automated guided vehicles at a dock according to claim 7, characterized in that, The intelligent unlocking stations are arranged in two symmetrical rows, forming a passage for the automated guided vehicle to simultaneously install and remove the locking pins at the four corners of the container. The first sensor and the second sensor are installed on one of the intelligent unlocking stations.

9. A dockside automated guided vehicle (AGV) intelligent unlocking station berthing control system according to claim 7, characterized in that, It also includes a visual monitoring module and a manual control module; The visualization monitoring module is connected to the spatial perception module and is used to display the point cloud spatial images of the automated guided vehicle, the container, and the smart unlocking station in real time. The manual control module is connected to the external automated guided vehicle (AGV) control system via the central controller. It is used to receive input from monitoring personnel and manually take over the movement control of the AGV.

10. A computer program product, comprising a computer program or instructions, characterized in that, When the computer program or instructions are executed by the processor, they implement the steps of the dock automation vehicle intelligent unlocking station berthing control method according to any one of claims 1 to 6.

Images