A remote full-automatic multi-machine collaborative intelligent control system based on ship unloaders

Abstract

The present invention discloses a remote full-automatic multi-machine collaborative intelligent control system based on a ship unloader, which relates to the technical field of port machinery automation and intelligent control. It includes an initial alignment module that matches the hatch position based on three-dimensional model data, analyzes the optimal operation area of each ship unloader through a path planning algorithm combined with the three-dimensional model data, and adjusts the ship unloader to move to a specified position; a grasping module that combines the three-dimensional model data and real-time scanning data, uses a multi-objective genetic algorithm to generate the next grasping point of the grab bucket, and sends the next optimal grasping point to the ship unloader through a communication protocol for grasping. The present invention uses the three-dimensional model data generated by 3D laser scanning technology, combines it with a path planning algorithm to determine the optimal operation area of each ship unloader, ensures that the equipment is accurately positioned to the working position, and improves the operation accuracy and safety.

Description

Technical Field

[0001] The present invention relates to the technical field of port machinery automation and intelligent control, and particularly to a remote full-automatic multi-machine collaborative intelligent control system based on a ship unloader. Background Art

[0002] In recent years, with the acceleration of the global economic integration process and the continuous growth of international trade volume, the port cargo handling efficiency has become one of the key factors restricting the efficiency of the logistics chain. As an indispensable core equipment in large bulk cargo terminals, ship unloaders play an irreplaceable role in improving port operation efficiency. However, the operation of traditional ship unloaders still relies too much on manual driving, resulting in problems such as low operation efficiency, poor safety, and susceptibility to environmental impacts. Therefore, automation and intelligent technologies have gradually been applied to the operation control of ship unloaders in the hope of solving the above problems. For example, the application of three-dimensional scanning technology enables ship unloaders to more accurately identify target positions; and task scheduling based on artificial intelligence algorithms has improved the working efficiency of ship unloaders to a certain extent. Nevertheless, the existing technologies still face many challenges, especially in the generation and optimization of multi-machine collaborative operation schemes, which are somewhat insufficient.

[0003] Currently, although the existing remote control systems for ship unloaders can achieve the automatic operation of single equipment to a certain extent, their performance is not satisfactory when facing multi-machine collaborative operations under complex working conditions. First, for the generation of multi-machine collaborative operation schemes, most use preset rules or simple mathematical models, which are difficult to dynamically adjust operation strategies according to actual working conditions, restricting flexibility and adaptability. Second, there is currently a lack of an effective real-time data interaction mechanism, and there is no way to timely update key information such as operation status and remaining material quantity, which will lead to unreasonable task allocation and affect the overall operation efficiency. In view of these deficiencies, the present invention proposes a remote full-automatic multi-machine collaborative intelligent control system based on a ship unloader, which effectively solves the problems of self-adaptability and high efficiency of multi-machine collaborative operation schemes by introducing the Q-learning algorithm to generate multi-machine collaborative operation schemes and using the genetic algorithm for optimization. Summary of the Invention

[0004] In view of the above existing problems, the present invention is proposed.

[0005] Therefore, the present invention provides a remote full-automatic multi-machine collaborative intelligent control system based on a ship unloader to solve the problems of imperfect optimization of multi-machine collaborative operation schemes and real-time data interaction mechanism.

[0006] To solve the above technical problems, the present invention provides the following technical solutions:

[0007] The present invention provides a remote full-automatic multi-machine collaborative intelligent control system based on a ship unloader, which includes,

[0008] A task generation module generates a multi-machine collaborative operation plan according to the ship berthing information through the Q-learning algorithm and optimizes it through the genetic algorithm to obtain the optimal multi-machine collaborative operation plan.

[0009] A 3D modeling module, based on the optimal multi-machine collaborative operation plan, uses 3D laser scanning to collect hull data and in-cabin material data, and splices and fuses them to generate 3D model data.

[0010] An initial alignment module matches the hatch position based on the 3D model data, analyzes the best operation area of each ship unloader through the path planning algorithm combined with the 3D model data, and sends a movement instruction through the communication protocol to adjust the ship unloader to move to the specified position.

[0011] A grasping module combines the 3D model data and the real-time scanning data, uses the multi-objective genetic algorithm to generate the next grasping point of the grab bucket, and sends the next best grasping point to the ship unloader through the communication protocol for grasping.

[0012] A safety protection module monitors the hull inclination and safety distance through a camera and dynamically adjusts the operation position of the ship unloader.

[0013] A dynamic resource scheduling module adjusts the task priority and resource allocation based on the dynamically adjusted operation position of the ship unloader.

[0014] As a preferred solution of the ship unloader remote full-automatic multi-machine collaborative intelligent control system described in the present invention, wherein: the multi-machine collaborative operation plan is generated through the Q-learning algorithm according to the ship berthing information, and the specific steps are as follows.

[0015] Obtain the ship berthing information, remove the noise data by the standard deviation method, and supplement the missing values by the median filling method.

[0016] Through the Q-learning algorithm, calculate the processed ship berthing information to obtain a multi-machine collaborative operation plan.

[0017] As a preferred solution of the ship unloader remote full-automatic multi-machine collaborative intelligent control system described in the present invention, wherein: the optimal multi-machine collaborative operation plan is obtained through optimization by the genetic algorithm, and the specific steps are as follows.

[0018] Use the genetic algorithm to optimize the generated multi-machine collaborative operation plan, and use the multi-objective weighted scoring method to define the fitness function to measure the optimization effect of the multi-machine collaborative operation.

[0019] Obtain the optimal multi-machine collaborative operation plan through the iterative selection, crossover and mutation processes.

[0020] As a preferred solution of the remote full-automatic multi-machine collaborative intelligent control system based on ship unloaders of the present invention, wherein: based on the optimal multi-machine collaborative operation plan, 3D laser scanning is used to collect hull data and in-cabin material data, and the data is stitched and fused to generate three-dimensional model data. The specific steps are as follows:

[0021] Through 3D laser scanning and using a grid scanning strategy, the hull and its interior are scanned multiple times from multiple angles for the same cabin area to obtain the original 3D laser scanning data;

[0022] The original 3D laser scanning data is input into Geomagic Wrap, and the ICP algorithm is used for point cloud registration. Matching is performed by minimizing the distance error between two groups of points to obtain the registered point cloud data, which is then input into an improved Poisson surface reconstruction algorithm to be merged into a complete three-dimensional model data

[0023] As a preferred solution of the remote full-automatic multi-machine collaborative intelligent control system based on ship unloaders of the present invention, wherein: the hatch position is matched based on the three-dimensional model data. By combining the path planning algorithm with the three-dimensional model data, the optimal working area of each ship unloader is analyzed. The specific steps are as follows:

[0024] Use Geomagic Wrap to analyze the complete three-dimensional model data, locate each hatch, and record the coordinate information;

[0025] Based on the hatch coordinate information, the moving range and boom length of the ship unloader, use the path planning algorithm to obtain the optimal path;

[0026] Analyze the optimal path through the area division method to obtain the optimal working area of each ship unloader.

[0027] As a preferred solution of the remote full-automatic multi-machine collaborative intelligent control system based on ship unloaders of the present invention, wherein: the movement instructions are sent using the communication protocol to adjust the ship unloader to move to the specified position. The specific steps are as follows:

[0028] Based on the optimal path and the optimal working area, send the instructions of the target position coordinates and operation parameters to each ship unloader. After receiving the instructions, each ship unloader performs a self-diagnosis status, accurately moves to the specified position and starts working.

[0029] As a preferred solution of the remote full-automatic multi-machine collaborative intelligent control system based on ship unloaders of the present invention, wherein: the three-dimensional model data and the real-time scanning data are combined, and the multi-objective genetic algorithm is used to generate the next grabbing point of the grab. The specific steps are as follows:

[0030] The static 3D model data is synchronized and integrated with the 3D laser scanning data obtained in real time, and the ICP algorithm is used for data alignment to obtain the integrated 3D model data;

[0031] Based on the integrated 3D model data, a target function for the next grasping point is defined through a multi-objective optimization method of material volume and obstacle distance;

[0032] The multi-objective genetic algorithm is used to optimize and solve the target function to obtain the next best grasping point.

[0033] As a preferred solution of the remote full-automatic multi-machine collaborative intelligent control system based on the ship unloader of the present invention, wherein: the next best grasping point is sent to the ship unloader through the communication protocol and grasping is performed. The specific steps are as follows.

[0034] Based on the calculated best grasping position, a grasping instruction including the three-dimensional coordinates of the target position and operation parameters is first generated, and the grasping instruction is sent to the ship unloader. The ship unloader moves to the specified position according to the grasping instruction and performs grasping.

[0035] As a preferred solution of the remote full-automatic multi-machine collaborative intelligent control system based on the ship unloader of the present invention, wherein: the ship unloader operation position is dynamically adjusted by monitoring the hull tilt and safety distance through a camera. The specific steps are as follows.

[0036] The hull tilt angle and the distance data between the hull and the ship unloader are collected in real time through a camera, and the Kalman filter is used for denoising and smoothing processing;

[0037] Based on the processed hull tilt angle and the distance data between the hull and the ship unloader, a safety evaluation index is calculated using a safety evaluation method, and the best operation area of each ship unloader is recalculated through a multi-objective genetic algorithm;

[0038] Instructions are sent to each ship unloader according to the new best operation area allocation scheme, adjusted to the specified position to continue operation, and the environmental and ship unloader state changes are continuously monitored.

[0039] As a preferred solution of the remote full-automatic multi-machine collaborative intelligent control system based on the ship unloader of the present invention, wherein: based on the dynamically adjusted ship unloader operation position, the task priority and resource allocation are adjusted. The specific steps are as follows.

[0040] Statistical analysis of historical operation data is performed to obtain a comprehensive performance threshold;

[0041] After the ship unloader adjusts its position according to the new operation area, the performance state is analyzed by combining the ship unloader state and the safety evaluation index. The priority of the ship unloader below the comprehensive performance threshold is reduced and re-sorted to generate a new task priority list.

[0042] Reallocate resources to each ship unloader according to the new task priority list and the total available resources;

[0043] Allocate tasks to the ship unloaders according to the resource allocation results and the task priority list, and send operation instructions to start the operation.

[0044] The beneficial effects of the present invention are as follows: The Q-learning algorithm is used to generate a multi-machine collaborative operation plan, which is optimized by the genetic algorithm, realizing a customized ship unloading operation plan. This not only improves the operation efficiency and resource utilization rate, but also reduces manual intervention and operating costs. At the same time, based on the three-dimensional model data generated by the 3D laser scanning technology, combined with the path planning algorithm, the best operation area of each ship unloader is determined to ensure that the equipment is accurately positioned to the working position, improving the operation accuracy and safety. This method effectively avoids the risk of equipment collision, reduces energy consumption and maintenance costs, bringing significant economic and environmental benefits. In summary, the intelligent scheduling and precise operation work together to greatly improve the overall performance, realizing efficient, safe and economic ship unloading operations, providing a solid foundation for remote full-automatic multi-machine collaborative control. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for the description of the embodiments will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0046] Figure 1 It is a schematic diagram of the remote full-automatic multi-machine collaborative intelligent control system based on the ship unloader in Embodiment 1.

[0047] Figure 2 It is a schematic diagram of the task generation and optimization process in Embodiment 1.

[0048] Figure 3 It is a schematic diagram of three-dimensional modeling and data processing in Embodiment 1.

[0049] Figure 4 It is a schematic diagram of the safety monitoring and dynamic adjustment process in Embodiment 1. DETAILED DESCRIPTION OF THE INVENTION

[0050] In order to make the above objects, features and advantages of the present invention more obvious and understandable, the specific embodiments of the present invention will be described in detail below with reference to the drawings in the specification.

[0051] In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present invention. However, the present invention may be practiced in other ways different from those described herein. Those skilled in the art can make similar generalizations without departing from the spirit of the present invention. Therefore, the present invention is not limited by the specific embodiments disclosed below.

[0052] Secondly, the so-called "one embodiment" or "embodiment" herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation manner of the present invention. The appearances of "in one embodiment" in different places in this specification do not all refer to the same embodiment, nor are they separate or alternative embodiments that are mutually exclusive of other embodiments.

[0053] Embodiment 1, referring to Figures 1 to 4 , this embodiment provides a remote full-automatic multi-machine collaborative intelligent control system based on a ship unloader, including the following steps:

[0054] A task generation module that generates a multi-machine collaborative operation plan through the Q-learning algorithm according to the ship berthing information.

[0055] Obtain the ship berthing information, remove the noise data by the standard deviation method, and supplement the missing values by the median filling method;

[0056] It should be noted that AIS provides real-time data such as the size, shape, and berthing position of the ship to ensure an understanding of the specific parameters and current position of each ship. At the same time, the port sends information about the types and quantities of goods to be loaded and unloaded to the ship unloader. The real-time data such as the size, shape, and berthing position of the ship have been entered before the ship arrives and are updated according to the scheduling plan. Integrate the information about the types and quantities of goods to be loaded and unloaded, obtain the ship berthing information. After obtaining the ship berthing information, first preprocess the information about the types and quantities of goods to be loaded and unloaded to ensure accuracy and integrity. Use the standard deviation method to remove the noise data. The specific steps are to calculate the standard deviation of each item of the information about the types and quantities of goods to be loaded and unloaded, and identify the outliers that exceed a certain multiple of the set standard deviation as noise and eliminate them. Then, for the missing values caused by removing the noise or the original data collection process, use the median filling method to supplement them. Use the median of the existing information about the types and quantities of goods to be loaded and unloaded to fill the missing values by the standard deviation method to ensure the continuity and consistency of the information set about the types and quantities of goods to be loaded and unloaded. For example, when processing key parameters such as the ship size and berthing position, this method can effectively avoid the overall deviation caused by individual incorrect ship berthing data, thereby improving the accuracy and reliability of subsequent path planning and operation allocation, and ultimately achieving the purpose of efficiently and accurately obtaining the ship berthing information.

[0057] Through the Q-learning algorithm, the processed ship berthing information is calculated to obtain a multi-machine collaborative operation plan.

[0058] It should be noted that the processed ship berthing information is calculated through the Q-learning algorithm to obtain a multi-machine collaborative operation plan. First, the state space is defined, including the port and the ship automatic identification method, providing the real-time position information of the ship. By combining the processing and analysis of the ship berthing information, the position coordinates of the ship unloader, the material quantity at the current grabbing point, and the distances to the surrounding obstacles are obtained to comprehensively describe the working environment. Then, the action space is defined. Each action includes the displacement change along each axis and the grab / release operation, which is used to control the operation of the ship unloader. Next, a reward function is designed to evaluate the effect of each operation, and the Q-value table is updated using the Q-learning algorithm to learn which action to take in a specific state to obtain the best effect. Finally, based on the trained Q-value table, an optimal strategy is generated to guide the efficient cooperation of multiple ship unloaders, ensuring that the overall operation efficiency and safety reach the optimum, and realizing an automated and intelligent multi-machine collaborative operation plan.

[0059] The task generation module is optimized through the genetic algorithm to obtain the optimal multi-machine collaborative operation plan.

[0060] The genetic algorithm is used to optimize the generated multi-machine collaborative operation plan. By using the method of multi-objective weighted scoring, a fitness function is defined to measure the optimization effect of multi-machine collaborative operation.

[0061] It should be noted that the genetic algorithm is used to optimize the generated multi-machine collaborative operation plan. First, a fitness function is defined to measure the effect of multi-machine collaborative operation. The fitness function comprehensively considers multiple key performance indicators such as grabbing efficiency, operation safety, and energy consumption, and balances the importance of each objective through the method of weighted scoring. When initializing the population, a group of potential multi-machine collaborative operation plans are randomly generated; then the fitness value of each plan is calculated, and based on these fitness values, the roulette wheel selection method is used to select excellent individuals to enter the next generation. Next, new offspring individuals are created through the crossover operation. Specifically, the method is to exchange some genes of two parent individuals to produce new combinations; and the mutation operation is applied to introduce random changes, such as randomly changing the action sequence or position coordinates of some individuals, to increase the population diversity and explore a better solution space.

[0062] The optimal multi-machine collaborative operation plan is obtained through the iterative process of selection, crossover, and mutation.

[0063] It should be noted that the whole process is continuously optimized through iterative selection, crossover, and mutation steps until a predetermined stopping condition is reached, such as the maximum number of iterations or the fitness no longer improving significantly. By this method, leveraging the powerful search ability of the genetic algorithm and combining specific evaluation criteria, it ensures the realization of efficient, safe, and flexible multi-machine collaborative operations in a complex environment, and finally obtains the optimal multi-machine collaborative operation plan;

[0064] It should also be noted that the fitness function has the following expression:

[0065] ;

[0066] where, is the fitness function, is the weight of the grasping efficiency target, is the weight of the operation safety target, is the weight of the energy consumption target, represents the reciprocal of the minimum safety distance, is a small positive value to prevent the denominator from being zero, represents the amount of materials grasped per unit time, represents the energy consumption.

[0067] Based on the optimal multi-machine collaborative operation plan, the 3D modeling module uses 3D laser scanning to collect hull data and in-cabin material data, and splices and fuses them to generate 3D model data.

[0068] Through 3D laser scanning, using a grid scanning strategy, the hull and its interior are scanned multiple times from multiple angles for the same cabin area to obtain the original 3D laser scanning data.

[0069] It should be noted that through 3D laser scanning technology and using a grid scanning strategy, detailed 3D model data of the hull and its internal structure are collected. First, the hull surface or internal space is divided into multiple small, regular grid areas to ensure that each grid can be covered. Then, the same grid area is scanned multiple times from different angles to obtain comprehensive and high-resolution data. This method of multi-angle repeated scanning can effectively reduce the shadow effect and the lack of 3D model data, and improve the accuracy of the final 3D model. The 3D laser scanning device emits laser beams and measures the time it takes for the beams to reflect back, thereby calculating the distance of each point on the object surface and generating point cloud data. During the scanning process, the results of each scan are recorded to ensure that all point cloud data are accurately captured. These original point cloud data, after preliminary processing, can provide detailed information about the hull and its internal structure, laying a solid foundation for subsequent analysis and operations. By this method, the original 3D laser scanning data is finally obtained.

[0070] Input the original 3D laser scanning data into Geomagic Wrap, perform point cloud registration using the ICP algorithm, match by minimizing the distance error between two sets of points, obtain the registered point cloud data, and input it into the improved Poisson surface reconstruction algorithm to merge into a complete three-dimensional model data.

[0071] It should be noted that by inputting the original point cloud data obtained by 3D laser scanning into Geomagic Wrap and using the ICP algorithm for point cloud registration, the matching is achieved by minimizing the distance error between two sets of points, generating the registered point cloud data. Then, the improved Poisson surface reconstruction algorithm is used to process the point cloud data. The improved Poisson surface reconstruction algorithm enhances noise filtering, adaptive resolution control, and optimized boundary processing to ensure high-quality data for surface reconstruction. During the reconstruction process, first, fine noise filtering is performed on the point cloud data, then the grid resolution is automatically adjusted according to the point cloud density, a preliminary three-dimensional surface model is constructed using the improved Poisson surface reconstruction algorithm, and special processing is carried out on the open boundary to ensure the smoothness and continuity of the model surface. Finally, all processed point cloud data is merged into a complete and high-quality three-dimensional model data, obtaining a complete three-dimensional model data.

[0072] The initial alignment module matches the hatch position based on the three-dimensional model data, and analyzes the best working area of each ship unloader through the path planning algorithm combined with the three-dimensional model data.

[0073] Use Geomagic Wrap to analyze the complete three-dimensional model data, locate each hatch, and record the coordinate information.

[0074] It should be noted that when using Geomagic Wrap to analyze the complete three-dimensional model data, first import the previously generated high-quality three-dimensional model data. In the software, feature extraction methods are used to identify and locate the position of each hatch. The specific operations include using the automatic surface detection function in the software to initially identify the possible hatch areas, and then using the boundary definition tool to carefully analyze these areas to ensure that the edges of each hatch are accurately captured. Next, use the coordinate measurement tool to record the three-dimensional coordinates of the center point of each hatch. In addition, these coordinate data can be exported as standard format files such as CSV or TXT for subsequent processing and reference. Through this method, not only can each hatch be located, but also the coordinate information can be detailedly recorded, finally generating a complete dataset containing all hatch coordinates, providing an important reference basis for the operation of the ship unloader, thus obtaining a complete data containing all hatch coordinate information.

[0075] Based on the hatch coordinate information, the moving range and boom length of the ship unloader, use the path planning algorithm to obtain the optimal path.

[0076] It should be noted that by importing the coordinate information of the hatches and combining with the moving range and boom length of the ship unloader, the optimal path is calculated using a path planning algorithm. First, the three-dimensional coordinate data of each hatch is input into the path planning. Based on these input data, the A algorithm evaluates the costs of different paths and selects an optimal path that not only meets the operation requirements but also minimizes energy consumption and time. During this process, the movement trajectory of the ship unloader is simulated in real time to ensure the safety and feasibility of the path. Finally, a detailed path plan is generated to guide the ship unloader to reach each hatch efficiently and safely to perform the operation tasks, thus obtaining an optimized optimal path.

[0077] The optimal path is analyzed by the area division method to obtain the optimal operation area of each ship unloader.

[0078] It should be noted that in the future, to analyze the optimal path by the area division method, first, the optimal path data of each ship unloader and the three-dimensional model data of the ship need to be imported. Next, area division algorithms such as Voronoi diagrams and K-means clustering are used to subdivide the operation area. The specific steps are as follows: First, a preliminary partition is generated using the Voronoi diagram based on the optimal path of the ship unloader to ensure that each ship unloader has an independent working area. The K-means clustering is applied to further optimize these areas, making the task volume in each area balanced and avoiding cross-interference. During this process, factors such as the material distribution density and obstacle positions extracted from the three-dimensional model data and the operation efficiency of the ship unloader are considered to ensure that the optimal operation area of each ship unloader is both efficient and safe. Finally, a specific operation area allocation plan for each ship unloader is generated, providing detailed guidance for multi-machine collaborative operations, thus obtaining the optimal operation area of each ship unloader.

[0079] The initial alignment module uses the communication protocol to send movement instructions to adjust the ship unloader to move to the specified position.

[0080] Based on the optimal path and the optimal operation area, instructions for the target position coordinates and operation parameters are sent to each ship unloader. After receiving the instructions, each ship unloader conducts a self-diagnosis of its status, precisely moves to the specified position, and starts operating.

[0081] It should be noted that in the process of realizing the automated ship unloading based on the optimal path and the optimal operation area, first, data including the optimal path coordinates, operation parameters, and self-diagnosis status of each ship unloader need to be collected. Next, instructions for the target position coordinates and necessary operation parameters are sent to each ship unloader. This process applies the communication protocol to ensure the accurate transmission of the instructions. After receiving the instructions, each ship unloader uses its own self-diagnosis algorithm to check the equipment status. After confirming that there is no error, it precisely moves to the specified position according to the received coordinate data and starts operating.

[0082] The grasping module combines the 3D model data and the real-time scanning data, and uses a multi-objective genetic algorithm to generate the next grasping point of the grab bucket.

[0083] The static 3D model data is synchronized and integrated with the 3D laser scanning data obtained in real time, and the ICP algorithm is used for data alignment to obtain the integrated 3D model data.

[0084] It should be noted that in order to synchronize and integrate the static 3D model data with the 3D laser scanning data obtained in real time, two sets of data need to be imported first: one is the high-quality static 3D model data generated by Geomagic Wrap, and the other is the fresh point cloud data obtained by the real-time 3D laser scanning device, which contains detailed geometric information and spatial coordinates. Next, the ICP algorithm is used for data alignment to achieve matching by minimizing the distance error between the two sets of point clouds. In the specific operation, the static model data is used as the reference benchmark, and the real-time scanning data is gradually adjusted to achieve the best alignment effect. Finally, the integrated 3D model data is obtained.

[0085] Based on the integrated 3D model data, the objective function of the next grasping point is defined by a multi-objective optimization method of material volume and obstacle distance.

[0086] It should be noted that based on the integrated 3D model data, key information is first extracted, including material volume distribution and obstacle distance, etc., which comes from the integration result of the static 3D model data and the real-time 3D laser scanning data. Next, a multi-objective optimization method is used to define an objective function for determining the next grasping point of the grab bucket.

[0087] The multi-objective genetic algorithm is used to optimize and solve the objective function to obtain the next best grasping point.

[0088] It should be noted that the multi-objective genetic algorithm (MOGA) is used to optimize and solve the objective function to determine the next best grasping point. It provides the basic input for multi-objective optimization. Then, the multi-objective genetic algorithm is used for processing. The multi-objective genetic algorithm can find a balance between multiple conflicting objectives, such as maximizing the grasping material volume and minimizing the distance from obstacles. During the optimization process, the multi-objective genetic algorithm iteratively improves the solutions in the population through selection, crossover, and mutation operations, gradually approaching the optimal solution. Each generation of the population contains a set of potential grasping point positions. After multiple iterations, one or more Pareto optimal solutions are finally obtained, representing the best compromise between different objectives. Based on these Pareto optimal solutions, the most suitable grasping point is selected as the next operating point. The whole process not only considers the maximum collection of materials but also ensures the safety of the operation and avoids collisions with obstacles, and obtains the next best grasping point.

[0089] The grasping module sends the next best grasping point to the ship unloader through a communication protocol and performs grasping.

[0090] Based on the calculated best grasping points, first generate a grasping instruction containing the three-dimensional coordinates of the target position and operation parameters, and send the grasping instruction to the ship unloader. The ship unloader moves to the specified position according to the grasping instruction and performs grasping.

[0091] It should be noted that based on the calculated best grasping points, first generate a grasping instruction containing the three-dimensional coordinates of the target position and operation parameters. The operation parameters include the coordinates of the best grasping point obtained from the multi-objective genetic algorithm, as well as the required grab attitude adjustment angle and grasping force setting value, etc. Use the communication protocol to send these grasping instructions to the ship unloader. After receiving the instruction, the ship unloader first performs self-diagnosis to ensure that all mechanical components are in normal working condition, which is achieved through a built-in self-diagnosis algorithm. Once confirmed, the ship unloader accurately moves to the specified position according to the three-dimensional coordinate data in the instruction, and adjusts the attitude and force of the grab according to the operation parameters to start the grasping operation. During this process, continuously monitor the state of the ship unloader to ensure the safety and accuracy of the operation.

[0092] The safety protection module monitors the hull inclination and safety distance through a camera, and dynamically adjusts the operating position of the ship unloader.

[0093] Real-time collect the hull inclination angle and the distance data between the hull and the ship unloader through a camera, and use a Kalman filter for denoising and smoothing processing.

[0094] It should be noted that real-time collect the hull inclination angle and the distance data between the hull and the ship unloader through a camera. Specifically, the camera is used to capture the visual information of the hull, and the inclination angle and the distance data between the hull inclination angle and the hull and the ship unloader are usually noisy and fluctuating, so denoising and smoothing processing are required. Here, the Kalman filter is used as the main processing method. The Kalman filter can estimate the current state based on the current measurement value and the predicted value of the previous state, effectively reducing noise and smoothing the hull inclination angle and the distance data between the hull and the ship unloader.

[0095] Based on the processed hull inclination angle and the distance data between the hull and the ship unloader, use a safety assessment method to calculate safety assessment indicators, and recalculate the best operating area of each ship unloader through a multi-objective genetic algorithm.

[0096] It should be noted that based on the processed hull tilt angle and the distance data between the hull and the ship unloader, the safety assessment index is first calculated using the safety assessment method, which usually includes the quantitative analysis of potential risk factors, such as whether the tilt angle is within the safe range and whether the distance is sufficient to avoid collisions. Next, the multi-objective genetic algorithm is used to recalculate the optimal working area for each ship unloader. The multi-objective genetic algorithm can find a balance between multiple conflicting objectives, such as maximizing the working efficiency and minimizing the safety risk. During the optimization process, the genetic algorithm iteratively improves the solutions in the population through selection, crossover, and mutation operations, gradually approaching the optimal solution. Each generation of the population contains a set of potential working area allocation plans. After multiple iterations, one or more Pareto optimal solutions are finally obtained, representing the best compromise among different objectives. Based on these Pareto optimal solutions, a most suitable working area is allocated to each ship unloader, realizing the recalculation of the optimal working area for each ship unloader and obtaining an optimized working area allocation plan.

[0097] Send instructions to each ship unloader according to the new optimal working area allocation plan, adjust to the specified position to continue the operation, and continuously monitor the changes in the environment and the state of the ship unloader.

[0098] It should be noted that according to the new optimal working area allocation plan, instructions are sent to each ship unloader to adjust to the specified position to continue the operation, and the changes in the environment and the state of the ship unloader are continuously monitored. First, a grasping instruction including the three-dimensional coordinates of the target position and operation parameters is generated, including the specific coordinates and operation requirements from the optimal working area allocation plan optimized by the multi-objective genetic algorithm. These instructions are sent to each ship unloader using the communication protocol. After receiving the instructions, the ship unloader performs a status check through the built-in self-diagnosis algorithm to ensure that all mechanical components are working properly. Once confirmed, the ship unloader moves to the new specified position according to the instructions and starts the operation according to the operation parameters. During this process, environmental data such as the hull tilt angle and the distance between the hull and the ship unloader are collected in real time using cameras and sensors, and denoising and smoothing processing are performed through the Kalman filter to ensure accuracy and reliability. The cameras and sensors continuously monitor these environmental and ship unloader state changes to dynamically adjust the operation plan when necessary to ensure the safety and efficiency of the operation.

[0099] Based on the dynamically adjusted working position of the ship unloader, adjust the task priority and resource allocation.

[0100] Statistically analyze the historical operation data to obtain the comprehensive performance threshold.

[0101] It should be noted that data including operation efficiency, equipment operating status and safety index data are collected from multiple dimensions of the ship unloader historical operation database. The representative characteristic efficiency, equipment operation and safety characteristics and comprehensive characteristics are extracted after cleaning and preprocessing by the mean filling method. Descriptive statistical analysis, distribution fitting test, machine learning algorithm-assisted analysis and other methods are used to deeply analyze the characteristics to determine the threshold. The threshold is preliminarily set based on the comprehensive results and experience. The reserved verification operation efficiency, equipment operating status and safety index data sets are used for verification and adjustment. After multiple rounds of iterative optimization, a comprehensive performance threshold that can reflect the actual performance and has stable generalization capabilities is determined, and a threshold update mechanism is established.

[0102] After the ship unloader is adjusted to the new operating area, the performance status is analyzed in combination with the ship unloader status and safety assessment indicators. The ship unloaders below the comprehensive performance threshold are prioritized and re-ordered to generate a new task priority list.

[0103] It should be noted that after the ship unloader adjusts its position according to the new operating area, the status data of the ship unloader at this time is collected, including various parameters of the equipment operation, such as motor temperature, vibration frequency, current value, etc., to reflect the operating health of the equipment. At the same time, based on the information such as the hull inclination angle monitored by the safety protection module, the safe distance between the ship unloader and the surrounding facilities, combined with the pre-set safety assessment index calculation method, a detailed safety assessment result is obtained. Then, the ship unloader status data and safety assessment index results are compared and analyzed with the previously determined comprehensive performance threshold. For those ship unloaders that are below the comprehensive performance threshold under the comprehensive evaluation of various indicators, the priority is reduced according to the established rules. After reducing the priority, all ship unloaders are re-sorted according to the new priority order, thereby generating a new task priority list, providing an accurate basis for subsequent task allocation.

[0104] Re-allocate resources to each ship unloader based on the new task priority list and total available resources;

[0105] It should be noted that first, a new task priority list of the order and importance of each ship unloader task is obtained, and then resources are allocated to the ship unloaders according to priority from high to low. The specific allocation amount for each ship unloader is determined based on task requirements, current operating conditions and the total available resources. When allocating, the reasonable matching and utilization efficiency of resources are comprehensively considered. The dynamic programming algorithm is used to decompose the resource allocation problem into sub-problems. The global optimal solution is gradually obtained by solving the sub-problems. After completing the resource allocation of one ship unloader, the remaining total available resources are continuously updated and allocated to subsequent ship unloaders in order of priority until all ship unloaders are allocated or the resources are allocated.

[0106] According to the resource allocation results and task priority list, the task is assigned to the ship unloader, and the operation instruction is sent to start the operation.

[0107] It should be noted that, based on the previously determined resource allocation results, the specific resources allocated to each ship unloader are understood, and with reference to the new task priority list, the importance and sequence of each task are clarified. Then, the tasks are matched with the ship unloaders. In the order from the highest to the lowest task priority, a suitable ship unloader is searched for each task in turn. For the task with the highest priority, a ship unloader that has the resources required to complete the task and is in a suitable state will be searched for in the resource allocation results. Here, "suitable state" means that the ship unloader is not currently executing other tasks with higher priorities and its own equipment state is good and it can work normally. After finding the matching ship unloader, operation instructions are generated according to the specific requirements of the task and the actual situation of the ship unloader. These operation instructions contain key information such as the task objectives, operation steps, parameter settings, etc., such as the position to grab materials, the quantity to grab, the time limit for the operation, etc. Finally, the generated operation instructions are sent to the corresponding ship unloader through the communication protocol. After receiving the operation instructions, the ship unloader starts the corresponding equipment and programs and begins to perform the operation according to the instruction requirements.

[0108] In summary, the present invention: uses the Q-learning algorithm to generate a multi-machine collaborative operation plan and optimizes it with the genetic algorithm, realizing a customized ship unloading operation plan. This not only improves the operation efficiency and resource utilization rate, but also reduces manual intervention and operating costs. At the same time, based on the three-dimensional model data generated by the 3D laser scanning technology, combined with the path planning algorithm, the best operation area of each ship unloader is determined, ensuring that the equipment is accurately positioned to the working position, improving the operation accuracy and safety. This method effectively avoids the risk of equipment collision, reduces energy consumption and maintenance costs, bringing significant economic and environmental benefits. In summary, the intelligent scheduling and precise operation work together, greatly improving the overall performance, realizing efficient, safe and economic ship unloading operations, and providing a solid foundation for remote fully automatic multi-machine collaborative control.

[0109] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical solutions of the present invention, and they should all be covered within the scope of the claims of the present invention.

Claims

1. A remote full-automatic multi-machine collaborative intelligent control system based on a ship unloader, characterized in that: Including, A task generation module that generates a multi-machine collaborative operation plan according to the ship berthing information through the Q-learning algorithm and optimizes it through the genetic algorithm to obtain the optimal multi-machine collaborative operation plan; The specific steps for generating a multi-machine collaborative intelligent control operation plan through the Q-learning algorithm according to the ship berthing information are as follows: Obtain the ship berthing information, remove noise data by the standard deviation method, and supplement missing values by the median filling method; Calculate the processed ship berthing information through the Q-learning algorithm to obtain a multi-machine collaborative operation plan; The specific steps for optimizing through the genetic algorithm to obtain the optimal multi-machine collaborative operation plan are as follows: Use the genetic algorithm to optimize the generated multi-machine collaborative operation plan, and use the multi-objective weighted scoring method to define the fitness function to measure the optimization effect of multi-machine collaborative operation; Obtain the optimal multi-machine collaborative operation plan through iterative selection, crossover, and mutation processes; A 3D modeling module that, based on the optimal multi-machine collaborative operation plan, uses 3D laser scanning to collect hull data and in-cabin material data, and splices and fuses them to generate 3D model data; An initial alignment module that matches the hatch position based on the 3D model data, analyzes the best operation area of each ship unloader through the path planning algorithm combined with the 3D model data, and adjusts the ship unloader to move to the specified position; A grasping module that combines the 3D model data and the real-time scanning data, uses the multi-objective genetic algorithm to generate the next grasping point of the grab bucket, and sends the next best grasping point to the ship unloader through the communication protocol for grasping; The specific steps for combining the 3D model data and the real-time scanning data and using the multi-objective genetic algorithm to generate the next grasping point of the grab bucket are as follows: Synchronize and integrate the static 3D model data with the real-time acquired 3D laser scanning data, and use the ICP algorithm for data alignment to obtain the integrated 3D model data; Based on the integrated 3D model data, define the objective function of the next grasping point through the multi-objective optimization method of material volume and obstacle distance; Use the multi-objective genetic algorithm to optimize and solve the objective function to obtain the next best grasping point; A safety protection module that monitors the hull tilt and safety distance through a camera and dynamically adjusts the operation position of the ship unloader; A dynamic resource scheduling module that adjusts the task priority and resource allocation based on the dynamically adjusted operation position of the ship unloader.

2. The remotely automatic multi-machine collaborative intelligent control system based on a ship unloader according to claim 1, wherein: The specific steps for generating 3D model data by using 3D laser scanning to collect hull data and in-cabin material data and splicing and fusing them based on the optimal multi-machine collaborative operation plan are as follows: Through 3D laser scanning, adopt a grid scanning strategy to scan the hull and its interior from multiple angles for the same cabin area multiple times to obtain the original 3D laser scanning data; Input the original 3D laser scanning data into Geomagic Wrap, use the ICP algorithm for point cloud registration, match by minimizing the distance error between two groups of points to obtain the registered point cloud data, and input it into the improved Poisson surface reconstruction algorithm to merge into the complete 3D model data.

3. The remote full-automatic multi-machine collaborative intelligent control system based on a ship unloader according to claim 2, characterized in that: The hatch position is matched based on the 3D model data. By combining the path planning algorithm with the 3D model data, the optimal working area of each ship unloader is analyzed. The specific steps are as follows: Use Geomagic Wrap to analyze the complete 3D model data, locate each hatch, and record the coordinate information. Based on the hatch coordinate information, the moving range and boom length of the ship unloader, use the path planning algorithm to obtain the optimal path. Analyze the optimal path through the area division method to obtain the optimal working area of each ship unloader.

4. The remote fully automatic multi-machine collaborative intelligent control system for ship unloaders according to claim 3, characterized in that: The adjustment of the ship unloader to move to the specified position is as follows: Based on the optimal path and the optimal working area, send the instructions of the target position coordinates and operation parameters to each ship unloader. After each ship unloader receives the instructions, it conducts a self-diagnosis status, moves to the specified position, and starts working.

5. The remote full-automatic multi-machine collaborative intelligent control system based on a ship unloader according to claim 1, characterized in that: The next optimal grasping point is sent to the ship unloader through the communication protocol and grasping is carried out. The specific steps are as follows: Based on the calculated next optimal grasping point, generate a grasping instruction containing the 3D coordinates of the target position and operation parameters, send the grasping instruction to the ship unloader, and the ship unloader moves to the specified position according to the grasping instruction and conducts grasping.

6. The remote full-automatic multi-machine collaborative intelligent control system based on a ship unloader according to claim 1, wherein: The hull tilt and safety distance are monitored through the camera, and the working position of the ship unloader is dynamically adjusted. The specific steps are as follows: Real-time collect the hull tilt angle and the distance data between the hull and the ship unloader through the camera, and use the Kalman filter for denoising and smoothing processing. Based on the processed hull tilt angle and the distance data between the hull and the ship unloader, use the safety assessment method to calculate the safety assessment index, and recalculate the optimal working area of each ship unloader through the multi-objective genetic algorithm. Send instructions to each ship unloader according to the new optimal working area allocation scheme, adjust to the specified position to continue working, and continuously monitor the environmental and ship unloader status changes.

7. The remote full-automatic multi-machine collaborative intelligent control system based on a ship unloader according to claim 6, characterized in that: Based on the dynamically adjusted working position of the ship unloader, adjust the task priority and resource allocation. The specific steps are as follows: Conduct statistical analysis on the historical operation data to obtain the comprehensive performance threshold. After the ship unloader adjusts its position according to the new working area, analyze the performance status by combining the ship unloader status and the safety assessment index, lower the priority of the ship unloader below the comprehensive performance threshold, and reorder to generate a new task priority list. According to the new task priority list and the total available resources, reallocate resources to each ship unloader. Allocate tasks to the ship unloader according to the resource allocation result and the task priority list, and send the operation instruction to start working.

Images