Ship intelligent workshop crane hoisting path planning method

Abstract

The invention relates to the technical field of control system data processing, in particular to a ship intelligent workshop crane hoisting path planning method, which comprises the following steps: collecting crane coordinates, obstacle distribution, load swing angle and material weight, and inputting a physical constraint model to generate a structured environment state data set; initializing wolf pack algorithm parameters based on the structured environment state data set, calculating a compressed detection wolf walking step length according to an obstacle distance, generating a global task queue by an industrial cloud platform according to a dynamic parameter set, and outputting an execution state data set by a crane end; counting emergency braking times from an execution state data set to optimize a walk step length function, comparing task time to optimize a wolf investigation proportion, analyzing historical swing data to establish a step length and swing mapping library, and optimizing a parameter set to feed back and update a physical constraint model. According to the invention, cooperation of environment perception, dynamic path planning and closed-loop verification is realized, and accident risks and production delay in a ship workshop narrow space dynamic obstacle scene are reduced.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of control system data processing, and in particular to a ship intelligent workshop ship-hangar hoisting path planning method. BACKGROUND

[0002] The ship intelligent workshop is a modern shipbuilding core unit that integrates Internet of Things, artificial intelligence and automation technology. Through the integration of manufacturing execution systems, industrial networks and intelligent equipment, it realizes the digital management and collaborative optimization of production processes. Various types of equipment in the workshop collect and interact data in real time, rely on digital twin models for dynamic simulation and decision reasoning, thereby autonomously scheduling resources and adjusting process parameters to improve manufacturing precision and production efficiency. The highly integrated nature of the ship intelligent workshop enables the production process to have self-adaptive and self-optimizing capabilities, significantly reducing the dependence on manpower and operating costs, and supporting the evolution of shipbuilding towards flexibility and intelligence.

[0003] In the ship intelligent workshop, the ship-hangar hoisting operation relies on an intelligent system that integrates multiple types of sensors and advanced control units to achieve high-precision cargo handling and positioning. Through real-time data collection and analysis, the system automatically optimizes the hoisting trajectory and load distribution, thereby improving operational efficiency and enhancing safety levels. Reducing manual intervention makes the hoisting process adaptive to complex working conditions, reduces the probability of errors, and supports the continuity and coordination of the production process.

[0004] The existing ship intelligent workshop ship-hangar hoisting has the following technical pain points in actual application, specifically: manual direct participation in ship-hangar hoisting operations in complex workshop environments, such as narrow working spaces or dynamic obstacle interference scenarios, is prone to collision or falling accidents due to operator visual blind spots or judgment delays. For example, when handling heavy ship structure parts, manual operation cannot respond in real time to changes in the position of mobile equipment, resulting in a significant increase in the risk of part and equipment collisions. At the same time, path planning relying on experience rules lacks a dynamic optimization mechanism based on intelligent data processing, causing path redundancy or task conflicts in batch material handling tasks. Specifically, when multiple ships are working together, the scheduling waiting time accumulates, significantly prolonging the hoisting cycle and delaying subsequent production links. Under the synergistic effect, the high accident risk and production delay are mutually reinforced, threatening operational safety and reducing overall efficiency, which does not meet the requirements of intelligent manufacturing for control precision and data processing capabilities. SUMMARY

[0005] To address the deficiencies in the prior art, the present application provides a ship intelligent workshop ship-hangar hoisting path planning method, which solves the technical problem of high accident risk and production delay in the ship workshop dynamic obstacle and narrow working scenario caused by manual direct participation in ship-hangar hoisting operations and path planning relying on experience rules.

[0006] To solve the above technical problems, the specific content of the present application is as follows: The application provides a ship intelligent workshop navigation vehicle hoisting path planning method, comprising: Step 1: Obtain the three-dimensional coordinates of the navigation vehicle through the position sensor, scan the dynamic obstacle space distribution through the laser sensor, monitor the load swing angle and the change of the hook height through the tilt sensor, and record the material weight through the load sensor. The three-dimensional coordinates of the navigation vehicle, the dynamic obstacle space distribution, the load swing angle, the change of the hook height and the material weight are input into the physical constraint model to output a structured environment state data set. The structured environment state data set includes a coordinate set, a swing angle, a hook height change value, a material weight value, a step upper limit and a safety distance matrix; Step 2: Receive the structured environment state data set output in step 1, initialize the wolf swarm algorithm parameters based on the structured environment state data set, the wolf swarm algorithm parameters include the total number of wolf swarms, the proportion of scout wolves and the step factor, calculate the obstacle distance based on the safety distance matrix in the structured environment state data set, calculate the scout wolf walking step according to the obstacle distance, calculate the wolf rush step based on the swing angle and the navigation vehicle deceleration model in the structured environment state data set, map the odor concentration gradient based on the hook height change value in the structured environment state data set, correct the position of the alpha wolf by using the odor concentration gradient, synchronize the algorithm iteration period with the sensor sampling clock, integrate the walking step, the rush step and the odor concentration gradient, and output a dynamic parameter set. The dynamic parameter set includes the walking step, the rush step and the odor concentration gradient; Step 3: Receive the dynamic parameter set output in step 2, pre-generate a global task queue by an industrial cloud platform, and download the initial parameters to the navigation vehicle end through a 5G network. Based on the dynamic parameter set and the laser sensor real-time scanning, the dynamic obstacle coordinates are integrated and obtained. Based on the dynamic parameter set, the dynamic obstacle coordinates are encoded as the position offset of the alpha wolf. According to the position offset of the alpha wolf, a local optimization trajectory is generated. The position offset of the alpha wolf is returned to the cloud end to reconstruct the task queue. The walking mechanism is controlled to perform horizontal or vertical movement, the hoisting mechanism is controlled to dynamically adjust the lifting speed according to the swing angle, the trajectory execution state and the actual swing angle are recorded, and an execution state data set is output. The execution state data set includes the trajectory completion degree and the actual swing angle; Step 4: Receive the execution state data set output in step 3, count the number of navigation vehicle emergency braking from the execution state data set, update the walking step calculation function when the response delay exceeds the threshold value, compare the actual task completion time with the preset threshold value from the execution state data set, optimize the proportion of scout wolves when the difference exceeds the threshold value, analyze the historical swing angle data from the execution state data set to establish a wolf rush step and swing mapping library, generate an optimization parameter set, and feed back the optimization parameter set to step 1 to update the physical constraint model. The optimization parameter set includes the updated walking step calculation function, the optimized proportion of scout wolves and the wolf rush step and swing mapping library.

[0007] Further, the ship intelligent workshop ship crane hoisting path planning method disclosed by the application, step 1 comprises: continuously acquiring the ship crane three-dimensional coordinates through the position sensor; generating a dynamic obstacle point cloud coordinate set through the laser sensor; acquiring the load swing angle and the hook height change value through the tilt sensor; acquiring the material weight value through the load sensor; inputting the ship crane three-dimensional coordinates, the point cloud coordinate set, the load swing angle, the hook height change value and the material weight value into the physical constraint model; in the physical constraint model: the ship crane control system maps the maximum deceleration to the upper limit of the algorithm step length, the swing angle threshold triggers the generation of the path curvature constraint function, the obstacle coordinate set is converted into the safety distance matrix, and the ship crane three-dimensional coordinates, the point cloud coordinate set, the load swing angle, the hook height change value, the material weight value, the step length upper limit and the safety distance matrix are fused to generate the structured environment state data set.

[0008] Further, the ship intelligent workshop ship crane hoisting path planning method disclosed by the application, step 2 comprises: receiving the structured environment state data set output in step 1; initializing the total number of wolves, the proportion of scout wolves and the step length factor based on the structured environment state data set; performing real-time parameter adjustment operations: scout wolf wandering behavior: based on the safety distance matrix in the structured environment state data set, the obstacle distance is acquired, when the obstacle distance is less than the safety distance, the wandering direction is reset and the compressed step length is calculated; wolf attack behavior: based on the swing angle in the structured environment state data set, combined with the ship crane deceleration model, when the swing angle exceeds the threshold value in the physical constraint model, the reduced attack step length is calculated; alpha wolf decision mechanism: based on the hook height change value in the structured environment state data set, the smell concentration gradient is mapped, and the alpha wolf position is corrected based on the smell concentration gradient; the compressed step length, the reduced step length and the corrected alpha wolf position data are integrated under the condition that the algorithm iteration period is synchronized with the sensor sampling clock, and a dynamic parameter set including the wandering step length, the attack step length and the smell concentration gradient is output.

[0009] Further, the ship intelligent workshop ship crane hoisting path planning method disclosed by the application, step 3 comprises: the industrial cloud platform receives the dynamic parameter set output in step 2, and pre-generates a global task queue based on the dynamic parameter set; downloading initial parameters to the ship crane end through the 5G network; on the ship crane end: Input dynamic parameter set, obtain dynamic obstacle coordinates based on real-time scanning of laser sensor and integration of dynamic parameter set, encode dynamic obstacle coordinates as head wolf position offset, generate local optimization trajectory according to head wolf position offset, and return position offset to cloud; The cloud reconstructs the task queue; The walking mechanism performs lateral or longitudinal movement, the lifting mechanism dynamically adjusts the lifting speed according to the swing angle, records the trajectory execution state and the actual swing angle, and outputs the execution state data set including the trajectory completion degree and the actual swing angle.

[0010] Further, the ship intelligent workshop ship hoisting path planning method comprises the following steps: Receiving the execution state data set output in step 3; Extracting the ship emergency braking times from the execution state data set, updating the wandering step calculation function when the response delay exceeds the preset threshold; Extracting the actual task completion time from the execution state data set, comparing with the preset threshold, and optimizing the investigation wolf proportion when the difference exceeds the allowed range; Extracting historical swing angle data from the execution state data set, and establishing a fierce wolf raid step and swing mapping library; Integrating the updated wandering step calculation function, the optimized investigation wolf proportion and the fierce wolf raid step and swing mapping library into an optimization parameter set; Feeding back the optimization parameter set to step 1, and updating the path curvature constraint function and the step upper limit in the physical constraint model.

[0011] Further, the ship intelligent workshop ship hoisting path planning method comprises the following steps: The wandering step compression calculation rule of the investigation wolf in step 2 is: Step = dynamic coefficient obtained from the physical constraint model × × obstacle distance, k is a workshop environment factor obtained from the structured environment state data set, and the compressed step is output to the dynamic parameter set in step 3; In step 3, the cloud receives the compressed step data, and when the compressed step data is less than the set threshold in the safety distance matrix, the cloud triggers the task queue priority redistribution operation, and the cloud reconstructs the task queue according to the position offset.

[0012] Further, the ship intelligent workshop ship hoisting path planning method comprises the following steps: The fierce wolf raid step reduction proportion calculation rule in step 2 is: Reduction proportion = α × sin (swing angle) + β × deceleration coefficient in ship deceleration model, α, β are weight parameters; output the reduced step size to the dynamic parameter set of step 3; In step 4, the reduced step size data is received, the reduced step size and the historical swing angle data are input into the wolf rush step size and swing mapping library to generate the mapping relationship between the wolf rush step size change and the swing angle, and the mapping relationship is sent to the hoisting mechanism to control the hoisting mechanism to adjust the lifting speed according to the mapping relationship.

[0013] Further, the ship intelligent workshop ship hoisting path planning method provided by the application comprises the following steps: receive the position offset, receive the compressed step size data output in step 2, and receive the cooperative optimization factor generated in step 4; calculate the dynamic priority weight based on the compressed step size data and the cooperative optimization factor; perform a position correction operation: encode the position offset into a head wolf position correction amount, apply a safety distance matrix to constrain the head wolf position correction amount calculation, the constraint strength is positively correlated with the dynamic priority weight, and output the corrected position correction amount; send the corrected position correction amount to the wolf rush step size calculation step; in the wolf rush step size calculation process: receive the corrected position correction amount, dynamically adjust the alpha and beta weight parameters according to the position correction amount, calculate the reduced step size using the adjusted parameters, and output the reduced step size to the stability verification step; Wherein, the calculation of the dynamic priority weight W satisfies: W = μ × compressed step size + ν × cooperative optimization factor, μ, ν are balance coefficients obtained from the physical constraint model.

[0014] Further, the ship intelligent workshop ship hoisting path planning method provided by the application comprises the following steps: the proportion increment = γ × (the actual task completion time-the preset threshold), wherein γ is a scaling coefficient obtained from the physical constraint model; When the difference exceeds the allowed range, the proportion of the investigation wolf is updated to: the original proportion + the proportion increment; couple the updated proportion of the investigation wolf with the corrected position amount to generate a cooperative optimization factor; correct the dynamic coefficient in the compressed step size calculation formula based on the cooperative optimization factor; integrate the corrected dynamic coefficient and the updated proportion of the investigation wolf into the optimization parameter set.

[0015] Further, the ship intelligent workshop ship hoisting path planning method provided by the application comprises the following steps: receive the optimization parameter set fed back from step 4; extract the updated scout wolf proportion, the updated walking step calculation function and the hunting wolf rush step and swing mapping relationship from the optimization parameter set; The updated scout wolf proportion is adopted when initializing the wolf swarm algorithm parameters. The updated walking step calculation function is adopted when calculating the walking step of the scout wolf. The hunting wolf rush step and swing mapping relationship is adopted when calculating the hunting wolf rush step. The walking step, the rush step and the smell concentration gradient calculated by using the updated parameters are integrated into the dynamic parameter set and output. The optimization parameter set is generated by counting the emergency braking times of the vehicle, comparing the actual task completion time and analyzing the historical swing angle data in step 4, forming a closed-loop feedback mechanism from path execution to algorithm optimization, so that the wolf swarm algorithm parameters can be adaptively adjusted according to the actual hoisting operation effect.

[0016] The present application has the following advantages: The present application replaces manual monitoring with multi-sensor fusion environment perception, laser sensor scans dynamic obstacle spatial distribution combined with position sensor vehicle three-dimensional coordinates to construct real-time topology of the workshop, tilt sensor monitors load swing angle and load sensor material weight data input physical constraint model to generate structured environment state data set, eliminating collision risks caused by artificial visual blind area; wolf swarm algorithm parameters are dynamically initialized based on environment data set, scout wolf walking step is compressed according to obstacle distance index, hunting wolf rush step is reduced according to swing angle proportion, smell concentration gradient corrects the position of the leader wolf, breaking through the global search efficiency limit of experience rule path planning; the industrial cloud platform issues a task queue to the vehicle end through the 5G network, dynamic obstacle coordinates are encoded as the leader wolf position offset to trigger cloud-side rescheduling, the walking mechanism and the hoisting mechanism execute collaboratively according to the dynamic parameter set, solving production delays caused by multi-vehicle task conflicts; the execution state data set drives closed-loop verification, the emergency braking times are counted to update the walking step function, the task completion time is compared to optimize the scout wolf proportion, the historical swing angle analysis establishes the step and swing mapping library, the optimization parameter set feedback updates the physical constraint model, forming a self-enhancing closed loop of perception, decision-making, execution and verification, which collaboratively reduces the accident rate and production delay rate in the narrow space and dynamic obstacle scene of the shipyard. BRIEF DESCRIPTION OF DRAWINGS

[0017] In order to more clearly illustrate the technical solutions of the present application, the following will briefly introduce the drawings needed in the embodiments. Obviously, for those skilled in the art, other drawings can also be obtained without creative labor based on the drawings.

[0018] Fig. 1 A flowchart of a ship intelligent workshop vehicle hoisting path planning method provided by the present application.

[0019] Fig. 2 The wolf pack algorithm hunting model block diagram of a ship intelligent workshop ship hoisting path planning method is provided.

[0020] Fig. 3 The improved wolf pack optimization algorithm experiment diagram of a ship intelligent workshop ship hoisting path planning method is provided. DETAILED DESCRIPTION

[0021] In order to make the technical solutions of the present application clearer, specific embodiments of the present application and corresponding drawings will be used below to clearly and completely describe the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor belong to the scope of protection of the present application. The present application provided by the embodiments of the present application will be described in detail below with reference to the drawings. In order to better understand the purpose of the present application, the present application will be further described in detail.

[0022] Please refer to Figs. 1 to 3 The ship intelligent workshop ship hoisting path planning method provided by the present application comprises: Step 1: Obtain the ship three-dimensional coordinates through the position sensor, scan the dynamic obstacle space distribution through the laser sensor, monitor the load swing angle and the change of hook height through the tilt sensor, record the material weight through the load sensor, input the ship three-dimensional coordinates, dynamic obstacle space distribution, load swing angle, change of hook height and material weight into the physical constraint model, and output the structured environment state data set. The structured environment state data set includes the coordinate set, swing angle, hook height change value, material weight value, step upper limit and safety distance matrix; Step 2: Receive the structured environment state data set output by step 1, initialize the wolf pack algorithm parameters based on the structured environment state data set, the wolf pack algorithm parameters include the total number of wolf packs, the proportion of scout wolves and the step factor, calculate the obstacle distance based on the safety distance matrix in the structured environment state data set, calculate the scout wolf walking step according to the obstacle distance, calculate the wolf pack attack step based on the swing angle and ship deceleration model in the structured environment state data set, map the smell concentration gradient based on the hook height change value in the structured environment state data set, correct the position of the head wolf using the smell concentration gradient, synchronize the algorithm iteration period with the sensor sampling clock, integrate the walking step, the attack step and the smell concentration gradient, and output the dynamic parameter set. The dynamic parameter set includes the walking step, the attack step and the smell concentration gradient; Step 3: Receive the dynamic parameter set output in step 2, pre-generate a global task queue by the industrial cloud platform, and distribute the initial parameters to the vehicle end through the 5G network. Based on the dynamic parameter set and the real-time scanning integrated laser sensor, the dynamic obstacle coordinates are obtained. Based on the dynamic parameter set, the dynamic obstacle coordinates are encoded into the head wolf position offset, the local optimization trajectory is generated according to the head wolf position offset, the head wolf position offset is returned to the cloud end to reconstruct the task queue, the walking mechanism is controlled to execute horizontal or vertical movement, the lifting mechanism is controlled to dynamically adjust the lifting speed according to the swing angle, the trajectory execution state and the actual swing angle are recorded, and the execution state data set is output. The execution state data set includes the trajectory completion degree and the actual swing angle. Step 4: Receive the execution state data set output in step 3, and count the number of emergency braking of the vehicle from the execution state data set. When the response delay exceeds the threshold, update the walking step calculation function. Compare the actual completion time of the task with the preset threshold from the execution state data set. When the difference is too large, optimize the detection wolf ratio. Analyze the historical swing angle data from the execution state data set to establish a mapping library of the fierce wolf rush step and the swing, generate an optimization parameter set, and feed back the optimization parameter set to step 1 to update the physical constraint model.

[0023] The ship intelligent workshop vehicle hoisting path planning method provided by the application realizes closed-loop optimization from environment perception to path execution through multi-sensor fusion and intelligent algorithm cooperation. The core of the method is to dynamically combine physical constraints and wolf algorithm to cope with the path planning challenges in the narrow space of the ship workshop and the dynamic obstacle scene. The overall process includes four key steps, which are closely connected through data flow, forming a self-enhancing mechanism of perception, decision-making, execution and verification.

[0024] In step 1, the three-dimensional coordinates of the vehicle in the workshop coordinate system are continuously obtained by the position sensor, and the laser sensor scans the dynamic obstacle profile using the time-of-flight ranging principle to generate millimeter-level point cloud spatial distribution data. The tilt sensor monitors the pitch and roll angle changes of the hook load, and records the hook height change value, and the load sensor collects the material weight value. After these raw data are input into the physical constraint model, the vehicle control system maps the maximum deceleration parameter to the upper limit of the algorithm step, and correlates the load swing angle safety threshold with the path curvature constraint function to suppress resonance, and the dynamic obstacle coordinate set is converted into a safety distance matrix through Euclidean distance calculation. Finally, a structured environment state data set is output, which integrates the coordinate set, the swing angle, the hook height change value, the material weight value, the step upper limit and the safety distance matrix, providing the environment topology and dynamics boundary reference for path planning.

[0025] Step 2 receives the structured environment state data set output by step 1, initializes the wolf swarm algorithm parameters based on the data set, including the total number of wolves, the proportion of scout wolves, and the step factor. In the wandering behavior of the scout wolf, the obstacle distance is calculated in real time based on the safety distance matrix, and when the distance is less than the safety threshold, the wandering direction is reset, and the step is compressed using an exponential decay model to enhance the obstacle avoidance response. The hunting behavior of the wolf is based on the swing angle data combined with the vehicle deceleration model, and when the angle exceeds the limit, the hunting step is reduced in proportion to control the swing amplitude of the load. The head wolf decision mechanism maps the hook height change value to a nonlinear odor concentration gradient field, and corrects the head wolf position coordinates to optimize the global path direction. All parameters are integrated synchronously within the sensor sampling clock cycle, and the dynamic parameter set including the wandering step, the hunting step and the odor concentration gradient is output to realize the iteration of the algorithm and the matching of the real-time environment change.

[0026] Step 3 receives the dynamic parameter set output by step 2, and the industrial cloud platform pre-generates a global task queue based on the dynamic parameter set and issues initial parameters to the vehicle edge through the 5G network. The vehicle end encodes the dynamic obstacle coordinates obtained by real-time scanning of the laser sensor into the head wolf position offset, generates a Bezier curve optimized trajectory combined with local environment data. The position offset returns to the cloud to trigger task queue reconstruction, for example, when multiple vehicle path conflicts occur, the Hungarian algorithm is used to dynamically reallocate priorities. The walking mechanism executes lateral and longitudinal movement instructions, and the hoisting mechanism dynamically adjusts the lifting speed using PID control based on the swing angle data. The trajectory tracking error and the actual swing angle are recorded throughout the process, and the execution state data set including the trajectory completion degree and the actual swing angle is output for subsequent verification.

[0027] Step 4 receives the execution state data set output by step 3, and counts the number of vehicle emergency braking from the data set to evaluate the response efficiency of obstacle avoidance, and updates the wandering step calculation function when the number exceeds the threshold to enhance the dynamic obstacle avoidance capability. Compare the actual task completion time with the preset threshold of the production rhythm, and optimize the proportion of scout wolves when the difference is out of tolerance to improve the global search efficiency. Analyze the historical swing angle spectrum characteristics, establish a mapping relationship library between the hunting step of the wolf and the swing angle, and generate an optimized parameter set. The optimized parameter set is fed back to the physical constraint model of step 1 to update the path curvature constraint function coefficient and the step upper threshold, forming a closed loop of adaptive adjustment from path execution to algorithm parameters.

[0028] The present application drives the wolf swarm algorithm parameter iteration by environment perception data, and reversely optimizes the physical constraints based on the execution verification results, thereby cooperatively reducing the accident risk and production delay in the dynamic obstacle and narrow operation scene of the shipyard.

[0029] Specifically, the shipyard intelligent vehicle hoisting path planning method of the present application comprises the following steps: The three-dimensional coordinates of the vehicle are continuously obtained by the position sensor; A dynamic obstacle point cloud coordinate set is generated by a laser sensor; A load swing angle and a hook height change value are obtained by a tilt sensor; A material weight value is obtained by a load sensor; The trolley three-dimensional coordinates, the point cloud coordinate set, the load swing angle, the hook height change value and the material weight value are input into a physical constraint model; In the physical constraint model: the trolley control system maps the maximum deceleration to the upper limit of the algorithm step length, the swing angle threshold triggers the generation of the path curvature constraint function, the obstacle coordinate set is converted into a safety distance matrix, and the trolley three-dimensional coordinates, the point cloud coordinate set, the load swing angle, the hook height change value, the material weight value, the step length upper limit and the safety distance matrix are fused to generate a structured environment state data set.

[0030] Step 1 involves collecting environmental data through multi-sensor fusion and generating a structured environment state data set, providing a basis for path planning. The position sensor uses GNSS and UWB fusion positioning technology to continuously obtain the trolley's three-dimensional coordinates in the workshop coordinate system at the millimeter level, and updates the trolley's position information in real time to support dynamic path adjustment. The laser sensor scans the dynamic obstacle profile of mobile devices and stacking areas based on the time-of-flight ranging principle, generating high-precision point cloud coordinate sets, constructing the real-time topological distribution of the workshop space, and identifying obstacle boundaries and safety areas. The tilt sensor monitors the changes in the load's pitch and roll angles through the MEMS inertial unit, records the hook height change value, and captures the load's kinetic state to suppress swing risks. The load sensor measures the material weight value using a strain gauge bridge, providing load quality data for calculating mechanical motion boundaries.

[0031] The collected trolley three-dimensional coordinates, dynamic obstacle point cloud coordinate sets, load swing angles, hook height change values and material weight values are input into a physical constraint model for integrated processing. In the physical constraint model, the trolley control system maps the maximum deceleration parameter to the upper limit of the algorithm step length according to the motor torque characteristics, defining the mechanical constraints of the wolf pack algorithm search range. The swing angle safety threshold triggers the generation of the path curvature constraint function, which optimizes the turning trajectory through cubic spline interpolation to avoid resonance phenomena. The dynamic obstacle coordinate set is converted into a safety distance matrix through Euclidean distance calculation, quantifying the minimum safety interval between the obstacle and the trolley. Finally, all parameters are fused to generate a structured environment state data set, including the coordinate set, the swing angle, the hook height change value, the material weight value, the step length upper limit and the safety distance matrix, forming a unified data interface for environmental perception and physical constraints.

[0032] The whole process updates the physical constraint model parameters in real time through sensor data, realizes that the structured environment state data set dynamically reflects the workshop environment changes, and realizes that each sensor data stream is collected in parallel and input into the model synchronously, eliminates the blind area of artificial monitoring, and provides an accurate environment state benchmark for subsequent wolf pack algorithm initialization.

[0033] Specifically, the ship intelligent workshop ship-hoist lifting path planning method comprises the following steps: Receiving the structured environment state data set output by step 1; Initializing the total number of wolf packs, the proportion of scout wolves and the step factor based on the structured environment state data set; Performing real-time parameter adjustment operations: Scout wolf wandering behavior: based on the safety distance matrix in the structured environment state data set, the obstacle distance is obtained, when the obstacle distance is less than the safety distance, the wandering direction is reset and the compression step is calculated; Wolf attack behavior: based on the swing angle in the structured environment state data set, combined with the ship deceleration model, when the swing angle exceeds the threshold value in the physical constraint model, the reduced attack step is calculated; Alpha wolf decision mechanism: based on the hook height change value in the structured environment state data set, the smell concentration gradient is mapped, and the alpha wolf position is corrected based on the smell concentration gradient; The compression step, the reduced step and the corrected alpha wolf position data are integrated under the condition that the algorithm iteration period is synchronized with the sensor sampling clock, and a dynamic parameter set including the wandering step, the attack step and the smell concentration gradient is output.

[0034] Step 2 receives the structured environment state data set output by step 1, which includes coordinate set, swing angle, hook height change value, material weight value, step upper limit and safety distance matrix and other environment state parameters. Based on the structured environment state data set, the wolf pack algorithm parameters are initialized, the total number of wolf packs is dynamically set according to the proportion of workshop space area, the proportion of scout wolves is allocated according to the task complexity, and the step factor is related to the maximum acceleration characteristics of the ship, so that the algorithm parameters match the physical environment constraints.

[0035] When performing the real-time parameter adjustment operation, the wolf exploration behavior acquires the obstacle distance in real time based on the safety distance matrix in the structured environment state data set, triggers the exploration direction reset when the obstacle distance is less than the safety distance threshold, and calculates the compression step length using a decay model to strengthen the obstacle avoidance response. The wolf sprint behavior is based on the swing angle data in the structured environment state data set, combined with the vehicle deceleration model, and when the swing angle exceeds the threshold in the physical constraint model, the sprint step length is proportionally reduced to suppress the load swing amplitude. The head wolf decision mechanism is based on the hook height change value in the structured environment state data set, generates an odor concentration gradient field through nonlinear mapping, and corrects the head wolf position coordinates to optimize the global path direction.

[0036] The compression step length, the reduced step length and the corrected head wolf position data are integrated and processed in the algorithm iteration period, which is synchronized with the sensor sampling clock, and the timing interrupt mechanism is used to realize the timing alignment of parameter update and sensor data acquisition. The output dynamic parameter set includes the exploration step length, the sprint step length and the odor concentration gradient, and the dynamic parameter set is used as the input of step 3 to support the generation of the industrial cloud platform task queue and the execution of the vehicle trajectory. Each adjustment behavior is based on real-time feedback of the environment state data, forming a coherent logic chain from parameter initialization to dynamic optimization, enhancing the adaptability of the algorithm to narrow space dynamic obstacle scenes.

[0037] Specifically, the ship intelligent workshop vehicle hoisting path planning method provided by the application comprises the following steps: The industrial cloud platform receives the dynamic parameter set output by step 2, and pre-generates a global task queue based on the dynamic parameter set; The initial parameters are downloaded to the vehicle end through the 5G network; At the vehicle end: The dynamic parameter set is input, the dynamic obstacle coordinates are obtained by integrating the real-time scanning of the laser sensor and the dynamic parameter set, the dynamic obstacle coordinates are encoded as the head wolf position offset, the local optimization trajectory is generated according to the head wolf position offset, and the position offset is returned to the cloud end; The cloud end reconstructs the task queue; The walking mechanism performs horizontal or vertical movement, the hoisting mechanism dynamically adjusts the lifting speed according to the swing angle, records the trajectory execution state and the actual swing angle, and outputs the execution state data set including the trajectory completion degree and the actual swing angle.

[0038] Step 3 involves the industrial cloud platform receiving the dynamic parameter set output in step 2, which includes parameters such as wandering step length, raid step length, and odor concentration gradient. The industrial cloud platform pre-generates a global task queue based on the dynamic parameter set, optimizes the task allocation sequence using a graph search algorithm, considers the path topology between the current position of the vehicle and the target point, and generates a preliminary scheduling scheme. The initial parameters are sent to the vehicle end through the low-latency and high-reliability communication characteristics of the 5G network. The network transmission process uses an encryption protocol to ensure data security, and the parameters are accurately delivered.

[0039] At the vehicle end, the dynamic parameter set is input and integrated with real-time scanning data from the laser sensor to dynamically obtain the coordinates of dynamic obstacles in the workshop environment. The laser sensor updates the obstacle position information through continuous ranging, and after fusion with the dynamic parameter set, encodes the dynamic obstacle coordinates as the position offset of the head wolf. The position offset represents the direction adjustment amount in path planning, and a curve fitting algorithm is used based on the offset to generate a locally optimized trajectory. The trajectory generation process takes into account the kinematic constraints of the vehicle and the dynamic characteristics of the load to avoid sharp turns or sudden speed changes. The position offset is returned to the cloud through the communication module, triggering the cloud task queue reconstruction mechanism.

[0040] After the cloud receives the position offset, it reconstructs the task queue in combination with the real-time workshop state. The reconstruction process uses a dynamic priority scheduling algorithm to reassign the task sequence of multiple vehicles to avoid conflicts. After the task queue is updated, it is sent to the vehicle end to drive the walking mechanism to perform horizontal or vertical movement. The walking mechanism controls the vehicle to move along the optimized trajectory through a servo motor, while the hoisting mechanism dynamically adjusts the lifting speed according to the swing angle data, using a closed-loop control strategy to suppress load swing. The trajectory execution state, including position deviation and speed curve, and actual swing angle data, is recorded throughout the process, and an execution state data set is output, including trajectory completion index and actual swing angle value, providing input for step 4 verification and optimization.

[0041] Each sub-step is connected through data flow, forming a closed loop from global planning in the cloud platform to local execution in the vehicle end. The logic chain is represented by dynamic parameter-driven trajectory generation, real-time feedback-triggered queue reconstruction, and mechanism control to ensure motion stability.

[0042] Specifically, the ship intelligent workshop vehicle hoisting path planning method described in the present application comprises the following steps: Receiving the execution state data set output in step 3; Extracting the number of emergency braking times of the vehicle from the execution state data set, and updating the wandering step length calculation function when the response delay exceeds the preset threshold; Extracting the actual task completion time from the execution state data set and comparing it with the preset threshold. When the difference exceeds the allowed range, optimize the scout wolf proportion. Extract historical swing angle data from the execution state data set, and establish the mapping library of wolf rush step and swing; Integrate the updated walk step calculation function, the optimized wolf ratio and the mapping library of wolf rush step and swing into the optimization parameter set;

[0043] Step 4 receives the execution state data set output by step 3, and the execution state data set includes parameters such as trajectory completion degree and actual swing angle, which are used for closed-loop verification and parameter optimization. The number of emergency braking of the vehicle is extracted from the execution state data set, and the number of emergency braking reflects the response timeliness of obstacle avoidance in path planning. When the response delay exceeds the preset threshold, the walk step calculation function update mechanism is triggered, and the dynamic obstacle avoidance ability of the wolf is strengthened by adjusting the function parameters.

[0044] The actual task completion time is extracted from the execution state data set, and compared with the preset threshold of the production rhythm. When the difference exceeds the allowed range, the wolf ratio is optimized, the global search efficiency is improved by increasing the number of wolves, and the task delay is reduced. The historical swing angle data is extracted from the execution state data set, the swing amplitude and frequency characteristics are analyzed, and the mapping relationship library of wolf rush step and swing angle is established. The mapping relationship library records the corresponding relationship between the step adjustment amount and the swing amplitude, and is used to guide the crane mechanism control.

[0045] The updated walk step calculation function, the optimized wolf ratio and the mapping library of wolf rush step and swing are integrated into the optimization parameter set, and the optimization parameter set is fed back to the physical constraint model of step 1 as feedback data package. The feedback process updates the path curvature constraint function coefficient and the step upper limit threshold in the physical constraint model, so that the next path planning inherits the optimization effect of the previous period, forming an adaptive closed loop from execution verification to environment perception.

[0046] Each sub-step is driven by parameter adjustment based on the execution state data set, and the logic chain is embodied as data extraction, threshold judgment, function update, ratio optimization, mapping establishment and integrated feedback. The technical features such as execution state data set, walk step calculation function and wolf ratio are used as the original text terms, so as to avoid introducing undefined concepts and enhance the coherence and implementability of the scheme. The closed-loop verification mechanism iteratively optimizes the algorithm parameters through real-time data, improves the path planning accuracy and safety in the dynamic obstacle scene of the shipyard.

[0047] Specifically, the ship intelligent workshop vehicle hoisting path planning method provided by the application further comprises: The walk step calculation rule of the wolf in step 2 is compressed as: Step = dynamic coefficient obtained from the physical constraint model x ​X obstacle distance, k is the inter-vehicle environment factor obtained from the structured environment state dataset, and the compression step is output to the dynamic parameter set of step 3; In step 3, the cloud receives the compression step data, and when the compression step data is less than a set threshold in the safety distance matrix, the cloud triggers the task queue priority redistribution operation, and the cloud reconstructs the task queue according to the position offset.

[0048] The compression step calculation rule for investigating the wolf wandering behavior in step 2 involves dynamically adjusting the step based on environmental data to enhance obstacle avoidance response. Specifically, a dynamic coefficient is obtained from a physical constraint model, which is related to the vehicle deceleration characteristics, and is used to adjust the step change rate. The obstacle distance is extracted from the safety distance matrix in the structured environment state dataset, reflecting the real-time obstacle proximity. The inter-vehicle environment factor k is obtained from the structured environment state dataset, representing the complexity of the vehicle layout, such as lane width or obstacle density, affecting the degree of step attenuation. The compression step calculation uses an exponential decay model, and the dynamic coefficient is multiplied by the negative k power of e and then multiplied by the obstacle distance to generate a step value that adapts to environmental risks. The calculated compression step is integrated into the dynamic parameter set and output to step 3 to support subsequent path planning decisions.

[0049] In step 3, the cloud receives the compression step data from step 2 as part of the dynamic parameter set. The cloud compares the compression step data with a set threshold in the safety distance matrix, which is preset based on historical collision data or safety standards. When the compression step data is less than the set threshold, it indicates that the obstacle risk is high, and the cloud triggers the task queue priority redistribution operation. Priority redistribution uses a dynamic scheduling algorithm, such as weighted allocation based on urgency, to reorder the multi-vehicle task sequence. The cloud reconstructs the task queue according to the position offset, which is generated by the dynamic obstacle coordinate encoding, reflecting real-time environmental changes. The reconstructed task queue is issued to the vehicle end through the 5G network, realizing timely adaptation of the path planning to the dynamic obstacle scene, reducing the risk of collision and production delay.

[0050] Throughout the process, the compression step calculation and cloud response form a closed loop, the step adjustment is based on physical constraints and environmental factors, the cloud decision is based on real-time data threshold comparison, and the logic chain reflects the coordination from local step optimization to global task scheduling, enhancing the path safety and efficiency in narrow space of shipyard.

[0051] Specifically, the ship intelligent yard vehicle hoisting path planning method described in the present application further comprises: The wolf attack step reduction ratio calculation rule in step 2 is: Reduction ratio = alpha * sin (swing angle) + beta * deceleration coefficient in vehicle deceleration model, alpha and beta are weight parameters; The reduced step size is output to the dynamic parameter set of step 3; In step 4, the reduced step size data is received, and the reduced step size and historical swing angle data are input into the wolf sprint step size and swing mapping library to generate a mapping relationship between the wolf sprint step size change and the swing angle. The mapping relationship is sent to the hoisting mechanism to control the hoisting mechanism to adjust the lifting speed according to the mapping relationship.

[0052] The wolf sprint step size reduction ratio calculation rule in step 2 involves dynamically adjusting the step size based on real-time sensor data to optimize path execution stability. The weight parameters α and β are obtained from the physical constraint model, the α parameter is associated with the swing angle influence weight, and the β parameter is associated with the deceleration coefficient influence weight, and the two parameters are used to adjust the reduction ratio. The swing angle data is monitored by the tilt sensor, reflecting the load swing amplitude, and the deceleration coefficient is derived from the trolley deceleration model, representing the braking characteristics. The reduction ratio calculation integrates the sine value of the swing angle and the deceleration coefficient to generate a step size adjustment amount that adapts to the current working condition, aiming to suppress load swing and improve motion smoothness.

[0053] The calculated reduced step size is output to the dynamic parameter set of step 3, and the dynamic parameter set integrates parameters such as walking step size, sprint step size, and odor concentration gradient as input for the industrial cloud platform to generate a global task queue. The reduced step size data is sent to the trolley end through the 5G network, supporting local trajectory optimization and real-time control decision-making.

[0054] In step 4, the reduced step size data is received, and the reduced step size and historical swing angle data are input into the wolf sprint step size and swing mapping library. The historical swing angle data is extracted from the execution state data set, recording past swing patterns. The mapping library uses regression analysis or machine learning algorithms to generate a mapping relationship between the wolf sprint step size change and the swing angle, which captures the quantitative impact of step size adjustment on swing suppression. The mapping relationship is sent to the hoisting mechanism control unit, and the hoisting mechanism dynamically adjusts the lifting speed according to the mapping relationship, achieving continuous optimization of load stability through closed-loop control strategies, reducing swing risk during hoisting. The entire process forms a data flow from step size calculation to execution feedback, enhancing the safety and efficiency of shipyard operation in dynamic obstacle scenarios.

[0055] Specifically, the shipyard intelligent vehicle hoisting path planning method described in the present application, the cloud reconstructs the task queue in step 3, which includes: Receiving position offset, receiving compressed step size data output in step 2, and receiving collaborative optimization factors generated in step 4; Based on the compressed step size data and the collaborative optimization factor, the dynamic priority weight is calculated; Performing a position correction operation: Encode the position offset as the head wolf position correction, apply the safety distance matrix to constrain the head wolf position correction calculation, the constraint strength is positively correlated with the dynamic priority weight, and the output is the corrected position correction; Send the corrected position correction to the wolf raid step length calculation step; In the wolf raid step length calculation process: Receive the corrected position correction, dynamically adjust the alpha and beta weight parameters according to the position correction, and calculate the reduced step length using the adjusted parameters, and output the reduced step length to the stability verification step; Wherein, the calculation of dynamic priority weight W satisfies: W=μ×compression step length+ν×cooperative optimization factor, μ,ν are balance coefficients obtained from the physical constraint model.

[0056] The cloud reconstruction task queue in step 3 involves multi-data source integration and dynamic weight calculation to optimize task scheduling response. The cloud receives the position offset, the compression step length data output by step 2, and the cooperative optimization factor generated by step 4. The position offset is generated by encoding the dynamic obstacle coordinates, the compression step length data is output from the scout wolf wandering behavior, and the cooperative optimization factor is generated based on the execution state data set in step 4. Based on the compression step length data and the cooperative optimization factor, the dynamic priority weight is calculated. The compression step length data reflects the emergency degree of obstacle avoidance, and the cooperative optimization factor represents the historical performance optimization effect. The dynamic priority weight W is calculated using a linear weighting model, which satisfies W=μ×compression step length+ν×cooperative optimization factor. μ and ν are balance coefficients obtained from the physical constraint model, used to adjust the contribution proportion of the compression step length and the cooperative optimization factor.

[0057] When performing the position correction operation, encode the position offset as the head wolf position correction, and the encoding process uses a coordinate transformation algorithm to map two-dimensional or three-dimensional offset to a position correction that the algorithm can handle. Apply the safety distance matrix to constrain the head wolf position correction calculation, the safety distance matrix is obtained from the structured environment state data set, defines the minimum safety interval, the constraint strength is positively correlated with the dynamic priority weight, the higher the dynamic priority weight, the greater the constraint strength, to realize path safety in high-risk scenarios, and output the corrected position correction.

[0058] Send the corrected position correction to the wolf raid step length calculation step, and transmit data through the internal communication interface. In the wolf raid step length calculation process, receive the corrected position correction, dynamically adjust the alpha and beta weight parameters according to the position correction, the position correction indicates the degree of environmental change, the alpha and beta parameters respectively control the weight of the swing angle and the deceleration coefficient in the reduction ratio calculation, calculate the reduced step length using the adjusted parameters, the reduced step length calculation integrates real-time sensor data and adjustment parameters, and output to the stability verification step for subsequent swing suppression control.

[0059] The whole process realizes the cooperation from task reassignment to path execution through a data flow closed loop, adjusts the constraint strength through dynamic priority weight, adjusts the algorithm parameters through position correction amount feedback, enhances the adaptive ability of the system to dynamic obstacles, and reduces the collision risk and production delay in the hoisting operation between ships and workshops.

[0060] Specifically, the ship intelligent workshop ship-hoist hoisting path planning method provided by the present application has the following calculation rule for optimizing the wolf ratio in step 4: The proportion increment is γ times the difference between the actual task completion time and the preset threshold, where γ is a scaling coefficient obtained from the physical constraint model. When the difference exceeds the allowed range, the wolf ratio is updated to the original ratio plus the proportion increment. The updated wolf ratio is coupled with the corrected position amount to generate a cooperative optimization factor. The dynamic coefficient in the compressed step length calculation formula is corrected based on the cooperative optimization factor. The corrected dynamic coefficient and the updated wolf ratio are integrated into the optimization parameter set.

[0061] The calculation rule for optimizing the wolf ratio in step 4 involves dynamically adjusting the algorithm parameters based on the deviation between the actual task completion time in the execution state data set and the preset threshold, in order to improve the efficiency of path planning. The actual task completion time is extracted from the execution state data set, which records the actual time taken by the ship-hoist to complete the hoisting task. The preset threshold obtained from the physical constraint model represents the ideal task completion time benchmark. When the difference between the actual task completion time and the preset threshold exceeds the allowed range, the wolf ratio optimization mechanism is triggered.

[0062] The proportion increment calculation uses the scaling coefficient γ obtained from the physical constraint model to multiply the difference between the actual task completion time and the preset threshold. The scaling coefficient γ is used to adjust the sensitivity of the proportion increment, adapting to different workshop operation rhythms. When the difference exceeds the allowed range, the wolf ratio is updated to the original ratio plus the proportion increment, increasing the number of wolves to enhance the global search ability and reduce task delays.

[0063] The updated wolf ratio is coupled with the corrected position amount, which is derived from the head wolf position offset in step 3 or the position correction operation in step 4, through linear weighting or nonlinear mapping methods to generate a cooperative optimization factor. The cooperative optimization factor represents the synergistic effect of the wolf ratio and the position adjustment, which is used for subsequent parameter correction.

[0064] The dynamic coefficient in the compression step calculation formula is corrected based on the cooperative optimization factor, the dynamic coefficient is obtained from the physical constraint model, the calculation of the scout wolf's walking step is affected, and the size of the dynamic coefficient is adjusted in the correction process to match the current environmental requirements. The corrected dynamic coefficient and the updated scout wolf ratio are integrated into the optimization parameter set, the optimization parameter set is used as the feedback data packet, sent to the physical constraint model in step 1, the model parameters are updated, and a closed-loop optimization cycle from execution verification to perception input is formed. The whole process is driven by real-time data to adjust the parameters, and the adaptive ability of the algorithm to the dynamic obstacle scene of the ship workshop is enhanced.

[0065] Specifically, the ship intelligent workshop crane hoisting path planning method provided by the application further comprises the following steps of: receiving the optimization parameter set fed back from step 4; extracting the updated scout wolf ratio, the updated walking step calculation function and the attack step and swing mapping relationship from the optimization parameter set; using the updated scout wolf ratio when initializing the wolf swarm algorithm parameters; using the updated walking step calculation function when calculating the scout wolf walking step; using the attack step and swing mapping relationship when calculating the attack step of the wolf; integrating the walking step, the attack step and the smell concentration gradient calculated by using the updated parameters into the dynamic parameter set and outputting; The optimization parameter set is generated by step 4 by counting the number of emergency braking of the crane, comparing the actual completion time of the task and analyzing the historical swing angle data, forming a closed-loop feedback mechanism from path execution to algorithm optimization, so that the wolf swarm algorithm parameters can be adaptively adjusted according to the actual hoisting operation effect.

[0066] Step 2 receives the optimization parameter set fed back from step 4, the optimization parameter set is generated by step 4 by counting the number of emergency braking of the crane, comparing the actual completion time of the task and analyzing the historical swing angle data, including the updated scout wolf ratio, the updated walking step calculation function and the attack step and swing mapping relationship. The optimization parameter set is used as the output of the closed-loop feedback, so that the wolf swarm algorithm parameters can be adaptively adjusted according to the actual hoisting operation effect.

[0067] The updated scout wolf ratio, the updated walking step calculation function and the attack step and swing mapping relationship are extracted from the optimization parameter set. The updated scout wolf ratio reflects the optimization demand of the task completion efficiency, the updated walking step calculation function integrates the obstacle avoidance response improvement, and the attack step and swing mapping relationship includes the step adjustment rule driven by the historical swing data. The extraction process is based on parameter identifier matching to realize data accuracy and consistency.

[0068] The updated scout wolf proportion is adopted when initializing the wolf swarm algorithm parameters, and the scout wolf proportion is used to set the proportion of the number of scout wolves in the wolf swarm, which affects the global search ability. After adopting the updated proportion, the initial state of the algorithm matches the latest working environment, and the adaptability of the path planning starting point is improved.

[0069] The updated walking step calculation function is adopted when calculating the walking step of the scout wolf, and the walking step calculation function is dynamically generated based on the physical constraint model and the safety distance matrix. The updated function integrates the emergency braking statistical results, enhances the step compression logic to deal with dynamic obstacles. The calculation process calls the function output in real time, and realizes the optimization of the walking behavior.

[0070] The mapping relationship between the wolf rush step and the swing is adopted when calculating the wolf rush step, the mapping relationship establishes the correlation rules between the rush step and the swing angle, and is generated based on the analysis of the historical swing angle. After adopting the mapping relationship, the rush step is dynamically adjusted according to the real-time swing data, the swing amplitude of the load is controlled, and the motion stability is improved.

[0071] The walking step, the rush step and the smell concentration gradient calculated by adopting the updated parameters are integrated into the dynamic parameter set and output. The walking step represents the moving distance of the scout wolf, the rush step represents the moving distance of the wolf, and the smell concentration gradient represents the position correction amount of the head wolf. The integration process is completed in the algorithm iteration period, and the dynamic parameter set is used as the input of step 3 to support the task queue generation and trajectory execution.

[0072] The whole process forms a closed-loop feedback mechanism from path execution to algorithm optimization, and the optimized parameter set is fed back to update the algorithm parameters of step 2, so that the wolf swarm algorithm can continuously adapt to the actual lifting operation effect, and reduce the accident risk and production delay in the narrow space of the ship workshop under the dynamic obstacle scene.

[0073] The present application replaces artificial direct monitoring through the multi-sensor fusion environment perception system, solves the collision risk caused by the visual blind area and judgment delay of the operator. The position sensor continuously acquires millimeter-level three-dimensional coordinates of the vehicle, the laser sensor scans the dynamic obstacle to generate point cloud spatial distribution, the tilt sensor monitors the load swing angle and hook height change, and the load sensor records the material weight data. These sensor data are input into the physical constraint model to generate a structured environment state data set, which integrates the coordinate set, the swing angle, the hook height change value, the material weight value, the step upper limit and the safety distance matrix, constructs the real-time topology and dynamics boundary of the workshop, and eliminates the blind area of artificial monitoring.

[0074] The wolf swarm algorithm parameters are dynamically initialized based on a structured environment state data set, breaking through the path planning limitations relying on empirical rules. The wolf patrol step length adopts an exponential compression strategy according to the obstacle distance, and when the obstacle distance is less than a safety threshold, the direction is reset and the compressed step length is calculated, thereby strengthening the dynamic obstacle avoidance response. The wolf attack step length is proportionally reduced according to the swing angle, and the attack step length is adjusted when the angle exceeds the limit by combining the vehicle deceleration model, thereby suppressing the load swing amplitude. The head wolf decision mechanism maps the smell concentration gradient through the hook height change value, optimizes the global path direction by correcting the head wolf position, and improves the search efficiency in narrow spaces.

[0075] The industrial cloud platform pre-generates a global task queue based on a dynamic parameter set, and distributes parameters to the vehicle end through a 5G network. The dynamic obstacle coordinates are real-time encoded as the head wolf position offset to trigger cloud-side rescheduling, and a Bezier curve is used to generate a locally optimized trajectory. The walking mechanism performs lateral and longitudinal movement, and the lifting mechanism dynamically adjusts the lifting speed according to the swing angle, solving the production delay caused by multi-vehicle task conflicts. The execution state data set drives closed-loop verification, updates the patrol step length function by counting the number of emergency brakes, optimizes the proportion of the wolf by comparing the actual completion time of the task, and analyzes historical swing data to establish a step length and swing mapping library.

[0076] The optimized parameter set is fed back to the physical constraint model to update the path curvature constraint function and the step length upper limit, forming a self-enhancing closed loop of environment perception, dynamic planning, and execution verification. Through sensor data-driven algorithm parameter iteration, the execution result reversely optimizes the physical constraints, cooperatively reduces the accident rate and production delay rate in narrow spaces and dynamic obstacle scenarios in the shipyard, and realizes the technical transformation from manual intervention to intelligent self-adaptation.

[0077] The specific embodiment of the present application relates to the actual operation process of the shipyard intelligent vehicle hoisting path planning method, and aims at the problem that manual direct participation in operation in narrow spaces and dynamic obstacle scenarios easily causes collisions and production delays in the background art. Through the cooperation of multi-sensor fusion and intelligent algorithms, automatic path planning is realized.

[0078] In the shipyard environment, the three-dimensional coordinates of the vehicle are continuously obtained by the position sensor, the point cloud coordinate set of the dynamic obstacle is generated by the laser sensor, the load swing angle and the hook height change value are monitored by the tilt sensor, and the material weight value is recorded by the load sensor. After these data are input into the physical constraint model, the vehicle control system maps the maximum deceleration to the upper limit of the algorithm step length, the swing angle threshold triggers the generation of the path curvature constraint function, the obstacle coordinate set is converted into a safety distance matrix, and finally a structured environment state data set including the coordinate set, the swing angle, the hook height change value, the material weight value, the step length upper limit, and the safety distance matrix is generated.

[0079] The wolf swarm algorithm parameters are initialized based on the structured environment state data set, including the total number of wolves, the proportion of scout wolves and the step factor. The scout wolf wandering behavior calculates the obstacle distance according to the safety distance matrix, and when the distance is less than the safety threshold, the direction is reset and the compression step is calculated; the wolf rush behavior combines the swing angle and the model of the swing car deceleration, and when the angle is over the limit, the step is reduced; the head wolf decision mechanism corrects the head wolf position through the hook height change value mapping the odor concentration gradient. All parameters are integrated synchronously within the sensor sampling period, and the dynamic parameter set including the wandering step, the rush step and the odor concentration gradient is output.

[0080] The industrial cloud platform receives the dynamic parameter set to pre-generate a global task queue, and sends the parameters to the swing car end through the 5G network. The swing car end obtains dynamic obstacle coordinates based on real-time scanning of the laser sensor and the dynamic parameter set, encodes the head wolf position offset to generate a local optimization trajectory, and returns the position offset to the cloud end to trigger task queue reconstruction. The walking mechanism performs horizontal and vertical movement, the hoisting mechanism dynamically adjusts the lifting speed according to the swing angle, records the trajectory execution state and the actual swing angle to output the execution state data set.

[0081] The number of emergency braking is counted from the execution state data set, and the wandering step calculation function is updated when the response delay is over the threshold; the actual completion time is compared with the preset threshold, and the proportion of scout wolves is optimized when the difference is over the threshold; the wolf rush step and swing mapping library are analyzed based on historical swing angle data. The optimization parameter set including the updated wandering step calculation function, the optimized proportion of scout wolves and the wolf rush step and swing mapping library is fed back to the physical constraint model to update the path curvature constraint function and the step upper limit, forming a closed-loop adaptive mechanism from environment perception to algorithm optimization.

[0082] The present application iterates the wolf swarm algorithm parameters driven by sensor data, optimizes the physical constraints based on the execution results, and realizes the autonomous planning and continuous optimization of the swing car hoisting path in the dynamic obstacle and narrow space scene, effectively reducing the risk of shipyard accidents and production delay.

[0083] The technical features of the present application are explained as follows: The structured environment state data set is an integrated data set generated by processing the environment data collected by multi-sensor fusion through the physical constraint model, including parameters such as coordinate set, swing angle, hook height change value, material weight value, step upper limit and safety distance matrix, providing environment topology and dynamic boundary reference for path planning.

[0084] The hook height change value monitors the height displacement during the lifting of the load through the inclination sensor, reflects the position change in the vertical direction of the hook, and is used to evaluate the load space state and participate in the odor concentration gradient mapping calculation.

[0085] The upper limit of the step size is generated by a vehicle control system according to a maximum deceleration parameter mapping, defines a mechanical constraint boundary of a search range of the wolf swarm algorithm, and limits a maximum value of an algorithm step size to realize motion safety.

[0086] The safety distance matrix is generated by converting a dynamic obstacle coordinate set through Euclidean distance calculation, quantifies a minimum safety interval between the obstacle and the vehicle, and provides spatial relationship data support for obstacle avoidance decision-making.

[0087] The wolf swarm algorithm is a kind of swarm intelligence optimization algorithm simulating the hunting behavior of a wolf swarm, and is used to solve a path planning problem in the application, realizes global optimization by simulating a social hierarchy system and a cooperation mechanism of the wolf swarm.

[0088] The wolf swarm algorithm parameters include adjustable variables such as a total number of wolf swarms, a scout wolf proportion and a step size factor, and are dynamically initialized based on a structured environment state data set, so that the algorithm parameters are matched with physical environment constraints.

[0089] The total number of wolf swarms is dynamically set according to a space area proportion between vehicles, determines a total number of artificial wolves in the algorithm, and affects a coverage range and a calculation complexity of global search.

[0090] The scout wolf proportion is allocated according to a task complexity, sets a proportion of a number of scout wolves in the wolf swarm, and affects a balance between global exploration ability and local development ability.

[0091] The step size factor is related to a maximum acceleration characteristic of the vehicle, adjusts a size of a search step size of the algorithm, and affects a convergence speed and accuracy.

[0092] The scout wolf wandering step size is dynamically generated based on an obstacle distance calculated by the safety distance matrix, and is compressed by using an exponential decay model when the distance is less than a safety threshold, so as to strengthen an obstacle avoidance response ability.

[0093] The wolf rush step size is calculated according to the swing angle data in combination with a vehicle deceleration model, is reduced by a certain proportion when the angle exceeds a limit, and suppresses a swing amplitude of the load.

[0094] The smell concentration gradient is generated by a nonlinear mapping of a hook height change value, forms a gradient field guiding a position update of a lead wolf, and optimizes a global path direction.

[0095] The position of the lead wolf represents position information of an optimal solution in the wolf swarm, is constantly updated by the smell concentration gradient correction, and guides the entire wolf swarm to move to an optimal area.

[0096] The dynamic parameter set integrates algorithm output parameters such as the wandering step size, the wolf rush step size and the smell concentration gradient, and is used as an input data interface for generating a task queue of an industrial cloud platform.

[0097] The wandering step size represents a moving distance of the scout wolf in a search process, is dynamically adjusted according to the obstacle distance, and realizes adaptive step size control.

[0098] The sprint step represents the size of the wolf's moving step towards the prey, which is dynamically adjusted according to the swing angle, balancing the moving efficiency and stability requirements.

[0099] The global task queue is pre-generated by the industrial cloud platform based on a dynamic parameter set, and the initial parameters are delivered to the vehicle end through the 5G network, which is used to schedule multiple vehicle task sequences, avoid path conflicts and production delays, and support dynamic priority redistribution.

[0100] The dynamic obstacle coordinates are obtained by real-time scanning and integration of laser sensors, representing the real-time position data of moving devices and stacking areas and other obstacles in the workshop environment, which are used to encode the head wolf position offset to trigger path adjustment.

[0101] The head wolf position offset is generated by encoding the dynamic obstacle coordinates, representing the direction correction amount in path planning, which is used to generate a locally optimized trajectory and return to the cloud to reconstruct the task queue, reflecting the impact of environmental changes on the path.

[0102] The locally optimized trajectory is generated by curve fitting algorithm according to the head wolf position offset, considering the kinematics constraints of the vehicle and the dynamics characteristics of the load, optimizing the turning path to avoid sharp turns or sudden speed changes, and improving the motion smoothness.

[0103] The walking mechanism executes the vehicle's horizontal or vertical movement instructions, and controls the vehicle to move along the locally optimized trajectory through servo motors, achieving precise position control and supporting dynamic obstacle avoidance and path tracking.

[0104] The lifting mechanism dynamically adjusts the lifting speed according to the swing angle data, uses a closed-loop control strategy to suppress load swing, achieves stability in the lifting process, and reduces the risk of swing.

[0105] The execution state data set records the trajectory execution state and actual swing angle, including parameters such as trajectory completion degree and actual swing angle, which is used for subsequent closed-loop verification and parameter optimization.

[0106] The trajectory completion degree index evaluates the path execution effect, including position deviation, speed curve and task completion progress, which is used to measure the path planning accuracy and efficiency.

[0107] The actual swing angle is monitored by the tilt sensor, recording the changes in load pitch angle and roll angle, reflecting the dynamics state of the load, and used to analyze the swing amplitude and frequency.

[0108] The number of emergency braking times is counted from the execution state data set, reflecting the response efficiency of the vehicle when avoiding obstacles, and triggering the wandering step calculation function update when the number exceeds the threshold, to enhance the dynamic obstacle avoidance capability.

[0109] The walking step calculation function is dynamically generated based on the obstacle distance and the safety distance matrix, and is used to calculate the walking step of the scout wolf. After being updated, the function is integrated with the emergency braking data and the step compression logic is strengthened.

[0110] The optimization of the scout wolf ratio is adjusted according to the comparison result of the actual completion time of the task and the preset threshold value. When the difference exceeds the allowed range, the number of scout wolves is increased to improve the global search efficiency and reduce the task delay.

[0111] The historical swing angle data is extracted from the execution state data set, records the past swing pattern, and is used to analyze the spectral characteristics and establish the mapping relationship between the step and the swing.

[0112] The wolf attack step and swing mapping library is generated using regression analysis or machine learning algorithm, establishing the mapping relationship between the attack step change and the swing angle, and is used to guide the lifting mechanism control lifting speed.

[0113] The optimization parameter set integrates the updated walking step calculation function, the optimized scout wolf ratio, and the wolf attack step and swing mapping library, and is sent to the physical constraint model as feedback data package to support parameter adaptive update.

[0114] The path curvature constraint function is generated in the physical constraint model by triggering the swing angle safety threshold, and the path curvature is optimized by spline interpolation to avoid resonance phenomenon and improve path safety.

[0115] The physical constraint model constructs an environmental constraint framework by integrating multi-sensor data, receives the vehicle three-dimensional coordinates provided by the position sensor, the point cloud coordinate set generated by the laser sensor, the load swing angle and hook height change value monitored by the tilt sensor, and the material weight value recorded by the load sensor. The model maps the maximum deceleration parameter of the vehicle control system to the upper limit of the algorithm step, converts the swing angle safety threshold to the path curvature constraint function, and converts the obstacle coordinate set to the safety distance matrix through Euclidean distance calculation. Finally, all input data are fused to generate a structured environmental state data set, and an integrated data package including coordinate set, swing angle, hook height change value, material weight value, step upper limit and safety distance matrix is output, providing environmental topology and dynamics boundary reference for path planning.

[0116] The vehicle deceleration model establishes the braking characteristic mapping relationship based on vehicle dynamics principles, receives the deceleration parameter from the physical constraint model and the swing angle data provided by the tilt sensor. The model calculates the deceleration coefficient suitable for different working conditions by analyzing the relationship between the vehicle motor torque characteristics and the load dynamics, which is used to adjust the reduction ratio of the wolf attack step. The model outputs the deceleration coefficient to the wolf attack behavior calculation module, dynamically adjusts the attack step combined with real-time swing angle data, suppresses the load swing amplitude, and realizes the motion stability and safety.

Claims

1. A ship intelligent workshop navigation crane hoisting path planning method, characterized in that, Comprise: Step 1: obtain the three-dimensional coordinates of the trolley through the position sensor, scan the spatial distribution of dynamic obstacles through the laser sensor, monitor the load swing angle and hook height change through the tilt sensor, record the material weight through the load sensor, input the three-dimensional coordinates of the trolley, the spatial distribution of dynamic obstacles, the load swing angle, the hook height change and the material weight into the physical constraint model, output the structured environment state data set, and the structured environment state data set includes the coordinate set, the swing angle, the hook height change value, the material weight value, the step upper limit and the safety distance matrix; Step 2: receive the structured environment state data set output in step 1, initialize the wolf swarm algorithm parameters based on the structured environment state data set, the wolf swarm algorithm parameters include the total number of wolf swarm, the proportion of scout wolves and the step factor, calculate the obstacle distance based on the safety distance matrix in the structured environment state data set, calculate the scout wolf walking step according to the obstacle distance, calculate the wolf rush step based on the swing angle and the trolley deceleration model in the structured environment state data set, map the smell concentration gradient based on the hook height change value in the structured environment state data set, correct the alpha position using the smell concentration gradient, synchronize the algorithm iteration period with the sensor sampling clock, integrate the walking step, the rush step and the smell concentration gradient, output the dynamic parameter set, and the dynamic parameter set includes the walking step, the rush step and the smell concentration gradient; Step 3: receive the dynamic parameter set output in step 2, pre-generate the global task queue by the industrial cloud platform, download the initial parameters to the trolley end through the 5G network, obtain the dynamic obstacle coordinates based on the dynamic parameter set and the laser sensor real-time scanning integration, encode the dynamic obstacle coordinates into the alpha position offset based on the dynamic parameter set, generate the local optimization trajectory according to the alpha position offset, return the alpha position offset to the cloud end to reconstruct the task queue, control the walking mechanism to execute horizontal or vertical movement, control the hoisting mechanism to dynamically adjust the lifting speed according to the swing angle, record the trajectory execution state and the actual swing angle, output the execution state data set, and the execution state data set includes the trajectory completion degree and the actual swing angle; Step 4: receive the execution state data set output in step 3, count the trolley emergency braking times from the execution state data set, update the walking step calculation function when the response delay exceeds the threshold, compare the actual task completion time with the preset threshold from the execution state data set, optimize the proportion of scout wolves when the difference is too large, analyze the historical swing angle data from the execution state data set to establish the wolf rush step and swing mapping library, generate the optimization parameter set, and the optimization parameter set includes the updated walking step calculation function, the optimized proportion of scout wolves and the wolf rush step and swing mapping library, feed back the optimization parameter set to step 1, and update the physical constraint model.

2. The shipyard intelligent navigation path planning method according to claim 1, characterized in that, Step 1 includes: continuously obtain the three-dimensional coordinates of the trolley through the position sensor; generate the dynamic obstacle point cloud coordinate set through the laser sensor; obtain the load swing angle and hook height change value through the tilt sensor; obtain the material weight value through the load sensor; Input the trolley three-dimensional coordinates, point cloud coordinate set, load swing angle, hook height change value and material weight value into the physical constraint model; In the physical constraint model: the trolley control system maps the maximum deceleration to the upper limit of the algorithm step, the swing angle threshold triggers the generation of the path curvature constraint function, the obstacle coordinate set is converted into a safety distance matrix, and the trolley three-dimensional coordinates, point cloud coordinate set, load swing angle, hook height change value, material weight value, step upper limit and safety distance matrix are fused to generate a structured environment state data set.

3. The shipyard intelligent navigation path planning method of claim 2, wherein, Step 2 includes: Receive the structured environment state data set output by step 1; Initialize the total number of wolves, the proportion of scout wolves and the step factor based on the structured environment state data set; Perform real-time parameter adjustment operations: Scout wolf wandering behavior: based on the safety distance matrix in the structured environment state data set, the obstacle distance is obtained, and when the obstacle distance is less than the safety distance, the wandering direction is reset and the compressed step is calculated; Wolf attack behavior: based on the swing angle in the structured environment state data set, combined with the trolley deceleration model, when the swing angle exceeds the threshold in the physical constraint model, the reduced attack step is calculated; Alpha wolf decision mechanism: map the smell concentration gradient based on the hook height change value in the structured environment state data set, and correct the alpha wolf position based on the smell concentration gradient; Integrate the compressed step, reduced step and corrected alpha wolf position data under the condition that the algorithm iteration period is synchronized with the sensor sampling clock, and output the dynamic parameter set including the wandering step, attack step and smell concentration gradient.

4. The shipyard intelligent navigation path planning method of claim 3, wherein, Step 3 includes: The industrial cloud platform receives the dynamic parameter set output by step 2, and pre-generates a global task queue based on the dynamic parameter set; Distribute initial parameters to the trolley end through the 5G network; At the trolley end: Input the dynamic parameter set, obtain the dynamic obstacle coordinates based on real-time scanning of the laser sensor and integration of the dynamic parameter set, encode the dynamic obstacle coordinates into alpha wolf position offset, generate a local optimization trajectory according to the alpha wolf position offset, and return the position offset to the cloud end; The cloud end reconstructs the task queue; The walking mechanism performs horizontal or vertical movement, the hoisting mechanism dynamically adjusts the lifting speed according to the swing angle, records the trajectory execution state and actual swing angle, and outputs the execution state data set including the trajectory completion degree and actual swing angle.

5. The shipyard intelligent navigation path planning method of claim 4, wherein, Step 4 includes: Receive the execution state data set output by step 3; Extract the trolley emergency braking times from the execution state data set, and update the wandering step calculation function when the response delay exceeds the preset threshold; Extract the actual task completion time from the execution state data set, and compare it with the preset threshold, and optimize the proportion of scout wolves when the difference exceeds the allowed range; Extract historical swing angle data from the execution state data set, and establish a wolf attack step and swing mapping library; Integrate the updated wandering step calculation function, optimized proportion of scout wolves and wolf attack step and swing mapping library into an optimization parameter set; Feed back the optimization parameter set to step 1, and update the path curvature constraint function and step upper limit in the physical constraint model.

6. The shipyard intelligent navigation path planning method of claim 5, wherein, Also includes: The scout wolf wandering behavior compression step calculation rule in step 2 is: Step size = dynamic coefficient obtained from physics constraint model x x obstacle distance, k is the inter-vehicle environment factor obtained from the structured environment state dataset, and the compressed step size is output to the dynamic parameter set of step 3; In step 3, the cloud receives the compression step data, and when the compression step data is less than a set threshold in the safety distance matrix, the cloud triggers the task queue priority redistribution operation, and the cloud reconstructs the task queue according to the position offset.

7. The shipyard intelligent navigation path planning method of claim 6, wherein, Also includes: The reduction ratio calculation rule of the wolf attack step in step 2 is: The reduction ratio = α × sin (swing angle) + β × deceleration coefficient in the vehicle deceleration model, α and β are weight parameters; Output the reduced step to the dynamic parameter set in step 3; In step 4, receive the reduced step data, input the reduced step and the historical swing angle data into the wolf attack step and swing mapping library to generate the mapping relationship between the wolf attack step change and the swing angle, and send the mapping relationship to the hoisting mechanism to control the hoisting mechanism to adjust the lifting speed according to the mapping relationship.

8. The shipyard intelligent navigation path planning method of claim 7, wherein, The cloud reconstructs the task queue in step 3, including: Receive the position offset, receive the compression step data output in step 2, and receive the cooperative optimization factor generated in step 4; Calculate the dynamic priority weight based on the compression step data and the cooperative optimization factor; Perform position correction operation: Encode the position offset as the head wolf position correction amount, apply the safety distance matrix to constrain the head wolf position correction amount calculation, the constraint strength is positively correlated with the dynamic priority weight, and output the corrected position correction amount; Send the corrected position correction amount to the wolf attack step calculation step; In the wolf attack step calculation process: Receive the corrected position correction amount, dynamically adjust the α and β weight parameters according to the position correction amount, and calculate the reduced step using the adjusted parameters, and output the reduced step to the stability verification step; Wherein, the calculation of the dynamic priority weight W satisfies: W = μ × compression step + ν × cooperative optimization factor, μ and ν are balance coefficients obtained from the physical constraint model.

9. The shipyard intelligent navigation path planning method of claim 8, wherein, The calculation rule of the optimization scout wolf ratio in step 4 is: The ratio increment = γ × (actual task completion time - preset threshold), where γ is the scaling coefficient obtained from the physical constraint model; When the difference exceeds the allowed range, update the scout wolf ratio to: original ratio + ratio increment; Couple the updated scout wolf ratio with the corrected position amount to generate a cooperative optimization factor; Based on the cooperative optimization factor, correct the dynamic coefficient in the compression step calculation formula; Integrate the corrected dynamic coefficient and the updated scout wolf ratio into the optimization parameter set.

10. The shipyard intelligent navigation path planning method of claim 9, wherein, Step 2 also includes: Receive the optimization parameter set fed back from step 4; Extract the updated scout wolf ratio, the updated wandering step calculation function, and the wolf attack step and swing mapping relationship from the optimization parameter set; Use the updated scout wolf ratio when initializing the wolf swarm algorithm parameters; Use the updated wandering step calculation function when calculating the scout wolf wandering step; Use the wolf attack step and swing mapping relationship when calculating the wolf attack step; Integrate the calculated wandering step, attack step and odor concentration gradient using the updated parameters into the dynamic parameter set and output; Among them, the optimization parameter set is generated by step 4 by counting the number of emergency braking of the navigation vehicle, comparing the actual completion time of the task and analyzing the historical swing angle data, forming a closed-loop feedback mechanism from path execution to algorithm optimization, so that the wolf swarm algorithm parameters can be adaptively adjusted according to the actual lifting operation effect.

Images