A ship loader anti-collision method and system

Abstract

The application is a kind of ship loader anti-collision method and system, the method comprises the following steps: S1: multi-source perception and data fusion; S2: digital twin synchronization and trajectory prediction; S3: dynamic risk field calculation and conflict research; S4: intelligent collaborative decision based on priority; S5: hierarchical flexible control execution; S6: effect evaluation and closed-loop optimization. The system adopts a three-layer intelligent architecture of perception layer, cognition layer and execution layer, and realizes predictive collaborative protection based on digital twin. The application realizes the fundamental change from passive alarm to active prediction and collaborative anti-collision, while ensuring safety, significantly improving operation efficiency and system intelligence level.

Description

Technical Field

[0001] This invention relates to the field of port machinery automation and safety control technology, and in particular to a collision avoidance method and system for ship loaders. Background Technology

[0002] Ship loaders are core equipment for efficient loading operations in large bulk cargo ports. Their operating environment is complex, requiring them to work collaboratively with adjacent equipment, ship superstructures, and dock facilities within a confined space. Currently, collision avoidance for ship loaders mainly relies on operator visual observation, fixed limit switches, and monitoring systems based on single sensors (such as laser scanners or ultrasonic sensors). Existing technologies have the following main shortcomings:

[0003] Limitations in perception capabilities: Reliance on manual observation has blind spots, especially at night or in inclement weather. Single sensors are prone to failure or decreased accuracy in the typical dust and vibration interference environment of ports, and cannot provide complete three-dimensional information about the surrounding environment.

[0004] Lack of predictive capability: Existing systems are mostly passive responses based on the current instantaneous distance, unable to predict the future movement trajectory of the ship loader and adjacent equipment. When a risk is detected, a collision is often imminent, and only emergency braking can be implemented, causing mechanical impact on the equipment and production interruption.

[0005] The decision-making strategy is too simplistic: collision avoidance strategies are typically limited to "alarm-deceleration-stop," lacking flexibility. In multi-machine collaborative operation scenarios, the inability to intelligently schedule tasks and coordinate collision avoidance based on work priorities leads to low overall operational efficiency.

[0006] The system is isolated and lacks coordination: each device's collision avoidance system works independently, and information is not shared, making it impossible to achieve coordinated actions and overall efficiency optimization among multiple devices.

[0007] Therefore, there is an urgent need for a collision avoidance solution for ship loaders that can achieve forward-looking perception, predictive early warning, intelligent collaborative decision-making, and flexible control, so as to fundamentally improve the safety, continuity, and intelligence of port operations. Summary of the Invention

[0008] The present invention aims to address the shortcomings of the prior art by providing a method and system for preventing collisions with ship loaders.

[0009] To achieve the above objectives, the present invention adopts the following technical solution: a collision prevention method for ship loaders, comprising the following steps:

[0010] S1: Multi-source perception and data fusion: Simultaneously collect the body posture data of the ship loader, the relative pose data between it and adjacent targets, and the three-dimensional contour data of the environment, and perform fusion processing to obtain accurate anti-interference environmental status information.

[0011] S2: Digital Twin Synchronization and Trajectory Prediction: Based on the fused data, the digital twin model of the ship loader is driven to achieve state synchronization; at the same time, based on the acquired real-time control commands and the kinematic model in the digital twin model, the motion trajectory of the machine and adjacent equipment in the future short time range is predicted in parallel, and a cluster of predicted trajectories is generated.

[0012] S3: Dynamic Risk Field Calculation and Conflict Assessment: Based on the predicted motion trajectory, combined with the physical dimensions of the equipment and the dynamic safety margin, calculate the spatiotemporal risk field that evolves over time, and assess whether there are potential conflicts between the machine and adjacent equipment.

[0013] S4: Priority-based intelligent collaborative decision-making: When a potential conflict is determined, based on the task priority information input by the terminal production management system, a comprehensive evaluation of multiple preset avoidance strategies is conducted to generate a collaborative avoidance scheme that enables multiple devices to work together to minimize global operation interruption.

[0014] S5: Hierarchical Flexible Control Execution: Convert the cooperative avoidance scheme into hierarchical flexible control commands and execute them to achieve a balance between safety and efficiency;

[0015] S6: Effect Evaluation and Closed-Loop Optimization: Collect actual operating data after the implementation of the plan, calculate the prediction deviation and avoidance effect, and use the data to adaptively optimize the digital twin model and decision logic.

[0016] Specifically, in step S1, the relative pose data is obtained by fusing the displacement data of the absolute encoder of the ship loader's trolley with the ranging data of the millimeter-wave radar using Kalman filtering, and is used to output the real-time distance and velocity vectors between adjacent devices.

[0017] Specifically, in step S2, the future short time horizon is 3 to 5 seconds; the kinematic model is integrated into the digital twin model to predict the future state based on the current state and control commands.

[0018] Specifically, in step S3, the calculation of the spatiotemporal risk field is as follows: taking the predicted trajectory as the central axis, and based on the equipment size, safety margin, and current environmental parameters, a three-dimensional dynamic envelope that expands outward with the predicted time point is constructed.

[0019] Specifically, in step S4, multiple preset avoidance strategies include at least two of the following: complete stop, trajectory fine-tuning, and speed negotiation; the comprehensive evaluation includes a quantitative analysis of safety costs, efficiency losses, and task priorities.

[0020] Specifically, in step S5, the hierarchical flexible control instructions include:

[0021] Suggestion-level instructions are used to provide operators with early warnings and visual guidance;

[0022] Cooperative-level commands are used to inject fine-tuning quantities into the equipment control system through control signal couplers to achieve cooperative obstacle avoidance.

[0023] Protection-level commands are used to take over control and perform controlled deceleration or stopping under the highest risk conditions.

[0024] A collision avoidance system for ship loaders is provided to implement a collision avoidance method. It employs a three-layer intelligent architecture consisting of a perception layer, a cognition layer, and an execution layer, and achieves predictive collaborative protection based on digital twins.

[0025] The perception layer is used to collect environmental and device status data from all dimensions, including:

[0026] The body attitude sensing unit includes a high-precision tilt sensor installed on the boom of the ship loader, a length encoder for measuring the extension length of the boom, and a rotary encoder for measuring the slewing angle. It is used to collect raw data of boom movement and calculate the real-time accurate three-dimensional coordinates of the boom end.

[0027] The relative pose sensing unit includes a travel absolute encoder installed on the traveling mechanism of the ship loader trolley and a millimeter-wave radar installed on the traveling beam or gantry structure, which are used to collect the traveling displacement of the machine and the absolute distance with adjacent equipment, and obtain the real-time distance and speed information between the two through subsequent fusion processing.

[0028] The environmental contour perception unit includes at least two lidars deployed at the head of the ship loader boom and key parts of the main structure, used to perform three-dimensional point cloud scanning of adjacent equipment and complex obstacles to obtain their three-dimensional contour information.

[0029] The data fusion processing unit is communicatively connected to each of the above-mentioned sensing units. It is used to perform time synchronization, coordinate system unification and filtering fusion processing on the raw data collected by each unit, and output stable, consistent environmental and equipment status information with a unified spatiotemporal reference for direct access by the cognitive layer.

[0030] The cognitive layer, which connects to the perception layer via communication, serves as the system's decision-making center and includes:

[0031] The digital twin module is used to build and drive a high-fidelity virtual model that is synchronized with physical devices based on physical modeling and sensing data, and integrates the kinematic model of the ship loader.

[0032] The trajectory prediction module, integrated into or calling the digital twin module, is used to predict the future motion trajectory of the local device and adjacent devices based on real-time control commands and kinematic models.

[0033] The collaborative decision-making module is used to dynamically calculate the spatiotemporal risk field based on the predicted trajectory, and when a spatiotemporal conflict is detected, it dynamically generates a collaborative avoidance strategy based on the task priority input by the terminal production management system.

[0034] The execution layer, which communicates with the cognitive layer, is used to translate cooperative avoidance strategies into hierarchical control signals, including:

[0035] Multi-level human-computer interaction interface, used to provide gradient warnings and operation guidance;

[0036] An adaptive control coupler is used to flexibly inject hierarchical control signals into the ship loader control system according to a cooperative avoidance strategy.

[0037] The system also includes a self-learning optimization module, which is connected to the data of the perception layer, cognition layer and execution layer respectively, and is used to optimize the prediction accuracy of the digital twin module and the strategy library of the collaborative decision-making module based on actual operating data.

[0038] Specifically, the data fusion processing unit is configured to fuse encoder displacement data and millimeter-wave radar ranging data from the relative pose sensing unit using a Kalman filter algorithm to output interference-resistant and accurate real-time distance and velocity vectors between adjacent devices in vibration and dust environments.

[0039] Specifically, the trajectory prediction module is configured to use a kinematic model to calculate the predicted positions of the local device and adjacent devices at multiple consecutive moments within the next 3 to 5 seconds in parallel, and generate a cluster of predicted trajectories.

[0040] Specifically, the adaptive control coupler is configured to support three intervention modes: suggestion, cooperation, and protection; among them,

[0041] In recommended mode, the adaptive control coupler provides operators with audible, visual, or tactile warnings and visual risk alerts through a multi-level human-machine interface, without directly intervening in the control of the equipment;

[0042] In cooperative mode, the adaptive control coupler is configured to inject continuous fine-tuning commands for speed or trajectory into the ship loader control system to achieve smooth, non-stop cooperative avoidance.

[0043] In protection mode, the adaptive control coupler is configured to temporarily take over or override equipment control when the system determines that there is an emergency collision risk, and execute controlled smooth deceleration or emergency stop commands.

[0044] The beneficial effects of this invention are:

[0045] Fundamental Enhancement in Safety: Through multi-source fusion sensing and high-precision digital twin synchronization, 360-degree, interference-resistant monitoring of the operating environment and equipment status is achieved. Based on kinematic model-based short-term trajectory prediction and dynamic spatiotemporal risk field calculation, the timing of collision avoidance is advanced from "collision is about to occur" to "conflict may form," realizing a fundamental shift from passive reaction to proactive prevention and greatly reducing the probability of collision.

[0046] Significantly improved operational efficiency: The intelligent collaborative decision-making mechanism based on task priority has transformed the traditional, crude approach of emergency shutdown of single equipment. The system can evaluate multiple avoidance strategies within milliseconds based on scheduling instructions from the terminal production management system and generate the optimal solution that enables multiple pieces of equipment to coordinate their actions (such as trajectory fine-tuning and speed negotiation). Combined with hierarchical flexible control execution (especially the "collaborative mode"), it can achieve "non-stop" or "minimized downtime" avoidance while ensuring safety, minimizing production interruptions and improving overall terminal operational efficiency.

[0047] Extremely strong environmental adaptability and robustness: The perception layer adopts a Kalman filter fusion scheme of encoder and millimeter-wave radar, which makes full use of the advantages of encoder high update rate and radar absolute ranging, effectively overcoming the interference of common port dust, rain fog, vibration and other factors, and ensuring the continuity and reliability of key distance and speed information under harsh working conditions.

[0048] System intelligence and self-evolution capabilities: The built-in self-learning optimization module constructs a complete closed loop of "perception-decision-execution-evaluation". The system can continuously utilize actual operating data to automatically correct the prediction bias of the digital twin model and optimize the decision strategy library, making the system more accurate and intelligent with use, achieving continuous self-improvement of performance and reducing long-term maintenance and optimization costs.

[0049] User-friendly human-machine interaction with clear responsibilities: A clear human-machine co-driving interface is constructed through multi-level human-machine interaction interfaces and a three-level (suggestion, collaboration, protection) control mode. In most cases, the system acts as an intelligent assistant, providing operators with sufficient risk warnings and decision-making suggestions; it provides flexible collaborative intervention when necessary; and it only takes over control in extreme emergency situations. This leverages the continuous monitoring and rapid decision-making advantages of artificial intelligence while preserving the final judgment and situational handling capabilities of the human operator. Attached Figure Description

[0050] Figure 1 This is a flowchart of the method of the present invention;

[0051] Figure 2 This is a system structure block diagram of the present invention;

[0052] The following will describe in detail, with reference to the accompanying drawings, embodiments of the present invention. Detailed Implementation

[0053] The present invention will be further described below with reference to embodiments:

[0054] like Figure 1 As shown, a collision avoidance method for ship loaders includes the following steps:

[0055] S1: Multi-source sensing and data fusion: Simultaneously collect the body attitude data of the ship loader, the relative pose data between it and adjacent targets, and the three-dimensional contour data of the environment, and perform fusion processing to obtain accurate anti-interference environmental state information; Among them, the relative pose data is obtained by fusing the displacement data of the ship loader's trolley travel absolute encoder and the ranging data of the millimeter-wave radar through Kalman filtering, and is used to output the real-time distance and velocity vector between adjacent devices.

[0056] In practice:

[0057] Body posture data acquisition: Raw signals are acquired in real time through high-precision tilt sensors installed at the joints of the ship loader boom, length encoders that measure the extension and retraction of the chute, and rotary encoders of the slewing mechanism.

[0058] Relative pose data acquisition: The absolute encoder installed on the side of the trolley's traveling motor continuously measures the displacement of the ship loader on the track; at the same time, the millimeter-wave radar installed on the side of the traveling beam or gantry with the beam aligned with the track direction directly measures the absolute distance to adjacent ship loaders or fixed obstacles.

[0059] Environmental contour data acquisition: By using two or more multi-line lidars installed at the head and middle of the boom to perform 360-degree or fan-shaped scans at a frequency of 10-20Hz, high-density three-dimensional point clouds of surrounding equipment, ship masts, dock facilities, etc. are obtained.

[0060] Fusion Processing: Data from all the aforementioned sensors are timestamped with high precision and then sent to the data fusion processing unit. For relative pose data, a Kalman filter algorithm is used as the core fusion engine. This algorithm uses the displacement increment calculated by the encoder as the prediction step and the absolute ranging value from the millimeter-wave radar as the observation update step. Through iterative calculation, it optimally estimates the true, smooth, and interference-resistant real-time distance and velocity vectors. Simultaneously, all sensing data is uniformly transformed into a global coordinate system with the ship loader's rotation center as the origin.

[0061] This step, through the complementarity of heterogeneous sensors and the fusion of Kalman filtering, fundamentally solves the problem of unreliability of single sensors under complex operating conditions, providing an accurate, consistent, and interference-resistant environmental "snapshot" for subsequent decision-making, and is the first guarantee of high system reliability.

[0062] S2: Digital Twin Synchronization and Trajectory Prediction: Based on the fused data, the digital twin model of the ship loader is driven to achieve state synchronization; at the same time, based on the acquired real-time control commands and the kinematic model in the digital twin model, the motion trajectory of the machine and adjacent equipment in the future short time range is predicted in parallel, generating a cluster of predicted trajectories; where the future short time range is 3 to 5 seconds; the kinematic model is integrated into the digital twin model and is used to infer the future state based on the current state and control commands.

[0063] In practice:

[0064] Digital Twin Construction and Synchronization: A high-fidelity 3D model corresponding to the physical ship loader is created on an industrial computer or edge server using physical modeling tools, at a 1:1 scale. Its kinematic model (including motion equations for each degree of freedom of the loader's movement, rotation, pitch, and extension) is then integrated. The fused data from S1 drives this virtual model through real-time communication, achieving millisecond-level state synchronization.

[0065] Trajectory prediction: The system monitors the speed, angle, and other control commands issued by the ship loader control system in real time. The trajectory prediction module calls the kinematic model in the digital twin, takes the current synchronization state as the initial value, substitutes the future continuous control commands (assuming the current command is maintained), calculates the state of the local machine and adjacent devices obtained through the wireless network in parallel, and deduces the predicted position for each moment (e.g., every 0.1 seconds) within the next 3 to 5 seconds, forming a series of "predicted trajectory clusters".

[0066] This step upgrades the traditional "status monitoring" to "status + future trajectory prediction". By using a digital twin that integrates a precise kinematic model for simulation and prediction, the system can "predict" the device's position within seconds, providing a valuable time window for early intervention. This is the key to achieving "proactive" collision avoidance.

[0067] S3: Dynamic Risk Field Calculation and Conflict Assessment: Based on the predicted motion trajectory, combined with the equipment's physical dimensions and dynamic safety margin, the spatiotemporal risk field that evolves over time is calculated, and the potential conflict between the machine and adjacent equipment is assessed accordingly. Specifically, the calculation of the spatiotemporal risk field involves constructing a three-dimensional dynamic envelope that expands outward with the predicted time point, using the predicted trajectory as the central axis and based on the equipment dimensions, safety margin, and current environmental parameters.

[0068] In practice:

[0069] Risk field modeling: Instead of using simple fixed safety distances, the collaborative decision-making module creates a dynamic three-dimensional risk envelope for each predicted trajectory during risk calculation. This envelope, centered on the trajectory, includes not only the physical contours of the equipment itself (such as the envelope of a boom or chute) but also a dynamic safety margin. This margin is dynamically adjusted based on factors such as current wind speed (obtained through meteorological sensors), equipment load, and historical motion stability; the higher the wind speed, the larger the margin.

[0070] Conflict assessment: The system calculates the risk field of all active predicted trajectories frame by frame (e.g., 10 times per second). When the risk fields of two different devices (or a device and an obstacle) intersect at the same time and spatial location in the future (i.e., envelope overlap), it is determined to be a "potential spatiotemporal conflict". This method of determination based on spatiotemporal overlap is more scientific and accurate than simply judging whether the current distance is less than a threshold.

[0071] This step introduces the concept of a "spatiotemporal risk field," upgrading safety assessment from a static, distance-based "circle" to a dynamic, time-evolving "pipeline." It comprehensively considers equipment movement trends, physical dimensions, and environmental disturbances, enabling earlier, more accurate, and more realistic collision detection.

[0072] S4: Priority-based intelligent collaborative decision-making: When a potential conflict is determined, based on the task priority information input from the terminal production management system, a comprehensive evaluation of multiple preset avoidance strategies is conducted to generate a collaborative avoidance scheme that enables multiple devices to work together to minimize global operation interruption; among them, multiple preset avoidance strategies include at least two of the following: complete stop, trajectory fine-tuning, and speed negotiation; the comprehensive evaluation includes a quantitative analysis of safety costs, efficiency losses, and task priorities.

[0073] In practice:

[0074] Strategy library: The system has a variety of pre-set avoidance strategies, such as: complete stop (highest safety, lowest efficiency); fine-tuning of the local trajectory (such as raising or deflecting the boom); adjustment of the local speed (decelerating or accelerating through) or sending negotiation requests to adjacent equipment to request them to fine-tune their trajectory or speed, etc.

[0075] Decision Engine: Once a potential conflict is identified in step S3, the decision engine is immediately activated. First, it queries the terminal production management system in real time via a standard interface to obtain the operational task priorities of the equipment involved in the conflict. Then, it performs a millisecond-level comprehensive evaluation of all feasible strategies in the strategy library. The evaluation factors include: safety cost (residual risk after strategy execution); efficiency loss (expected operational delays or energy consumption); and task priority (prioritizing ensuring that high-priority tasks are not affected).

[0076] Solution Generation: After evaluation, the system selects the strategy with the lowest overall cost. For example, if the local task has a low priority, a "local trajectory fine-tuning" solution may be generated; if both tasks are critical, a "coordinated speed adjustment" solution may be generated. The solution specifies the avoidance action, magnitude, executing equipment, and time sequence.

[0077] This step represents a leap from "single-machine instinctive reaction" to "multi-machine global intelligent scheduling." The decision-making process integrates production scheduling information to ensure that collision avoidance actions serve optimal overall production efficiency, rather than local safety, truly achieving a synergy between safety and efficiency.

[0078] S5: Hierarchical Flexible Control Execution: This involves converting the cooperative avoidance scheme into hierarchical flexible control commands and executing them to achieve a balance between safety and efficiency. The hierarchical flexible control commands include:

[0079] Suggestion-level instructions are used to provide operators with early warnings and visual guidance;

[0080] Cooperative-level commands are used to inject fine-tuning quantities into the equipment control system through control signal couplers to achieve cooperative obstacle avoidance.

[0081] Protection-level commands are used to take over control and perform controlled deceleration or stopping under the highest risk conditions.

[0082] In practice:

[0083] Command conversion and issuance: The cooperative avoidance scheme generated in step S4 is converted into specific hierarchical flexible control commands.

[0084] Level 3 execution mode:

[0085] Recommended Level: The system only highlights the risk area and predicted trajectory on the HMI interface and issues a voice warning through the speaker, allowing the operator to handle the situation independently. This mode respects operator autonomy and is suitable for low-risk or complex situations where the operator is more proficient.

[0086] Collaborative Level: The system injects rigorously safety-verified fine-tuning commands into the ship loader's existing control system via an adaptive control coupler. For example, a small speed correction value is superimposed on the original handle control signal, or a smooth angular offset path is inserted. The entire process resembles "lane-keeping assist," where the equipment remains under operator control, but the system assists in achieving precise obstacle avoidance. This mode achieves smooth obstacle avoidance without downtime, significantly reducing production interruptions and mechanical shocks, directly demonstrating improved efficiency.

[0087] Protection Level: When the system predicts an impending collision and there are no other possible avoidance maneuvers, the adaptive control coupler will issue a highest priority signal, temporarily overriding the original control commands, triggering the equipment's safety loop, and executing controlled, smooth deceleration to a stop (superior to emergency stop). This mode serves as the ultimate safety baseline, mitigating emergency hazards in the smoothest way possible and maximizing the protection of equipment and personnel.

[0088] This step, through three levels of flexible control, achieves a smooth transition and precise implementation of safety intervention, ensuring absolute safety while minimizing interference with normal operations.

[0089] S6: Effect Evaluation and Closed-Loop Optimization: Collect actual operating data after the implementation of the plan, calculate the prediction deviation and avoidance effect, and use the data to adaptively optimize the digital twin model and decision logic.

[0090] In practice:

[0091] Data collection: The self-learning optimization module continuously collects "actual trajectory data" from the perception layer, "predicted trajectory data" from the cognition layer, and "control commands and execution results" from the execution layer.

[0092] Effectiveness evaluation: Calculate key indicators such as "root mean square error of trajectory prediction deviation", "actual safety margin after the avoidance strategy is implemented", and "operation time loss caused by avoidance".

[0093] Model optimization: The above evaluation data is used as training feedback to: correct the parameters of the kinematic model in the digital twin and reduce prediction bias; optimize the weight coefficients of the strategy evaluation in the collaborative decision-making module so that the strategy selection is more in line with the balance between safety and efficiency under actual working conditions.

[0094] This step transforms the system from a static, "factory-set" product into an intelligent agent capable of adapting to specific dock environments and equipment characteristics, and continuously improving itself over time, thereby reducing long-term maintenance costs and enhancing performance throughout its entire lifecycle.

[0095] like Figure 2 As shown, a ship loader collision avoidance system is used to implement a ship loader collision avoidance method. It adopts a three-layer intelligent architecture consisting of a perception layer, a cognition layer, and an execution layer, and achieves predictive collaborative protection based on digital twins. Specifically:

[0096] The perception layer is used to collect environmental and device status data from all dimensions, including:

[0097] The body attitude sensing unit includes a high-precision tilt sensor installed on the boom of the ship loader, a length encoder for measuring the extension length of the boom, and a rotary encoder for measuring the slewing angle. It is used to collect raw data of boom movement and calculate the real-time accurate three-dimensional coordinates of the boom end.

[0098] The relative pose sensing unit includes a travel absolute encoder installed on the traveling mechanism of the ship loader trolley and a millimeter-wave radar installed on the traveling beam or gantry structure, which are used to collect the traveling displacement of the machine and the absolute distance with adjacent equipment, and obtain the real-time distance and speed information between the two through subsequent fusion processing.

[0099] The environmental contour perception unit includes at least two lidars deployed at the head of the ship loader boom and key parts of the main structure, used to perform three-dimensional point cloud scanning of adjacent equipment and complex obstacles to obtain their three-dimensional contour information.

[0100] The data fusion processing unit is communicatively connected to each of the above-mentioned sensing units. It is used to perform time synchronization, coordinate system unification and filtering fusion processing on the raw data collected by each unit, and output stable, consistent environmental and equipment status information with a unified spatiotemporal reference for direct access by the cognitive layer.

[0101] The data fusion processing unit is configured to fuse encoder displacement data and millimeter-wave radar ranging data from the relative pose sensing unit using a Kalman filter algorithm to output interference-resistant and accurate real-time distance and velocity vectors between adjacent devices in vibration and dust environments.

[0102] The cognitive layer, which connects to the perception layer via communication, serves as the system's decision-making center and includes:

[0103] The digital twin module is used to build and drive a high-fidelity virtual model that is synchronized with physical devices based on physical modeling and sensing data, and integrates the kinematic model of the ship loader.

[0104] The trajectory prediction module, integrated into or calling the digital twin module, is used to predict the future motion trajectory of the local device and adjacent devices based on real-time control commands and kinematic models.

[0105] The collaborative decision-making module is used to dynamically calculate the spatiotemporal risk field based on the predicted trajectory, and when a spatiotemporal conflict is detected, it dynamically generates a collaborative avoidance strategy based on the task priority input by the terminal production management system.

[0106] The trajectory prediction module is configured to use a kinematic model to calculate the predicted positions of the local device and adjacent devices at multiple consecutive moments within the next 3 to 5 seconds in parallel, and generate a cluster of predicted trajectories.

[0107] The execution layer, which communicates with the cognitive layer, is used to translate cooperative avoidance strategies into hierarchical control signals, including:

[0108] Multi-level human-computer interaction interface, used to provide gradient warnings and operation guidance;

[0109] An adaptive control coupler is used to flexibly inject hierarchical control signals into the ship loader control system according to a cooperative avoidance strategy. The adaptive control coupler is configured to support three intervention modes: suggestion, cooperation, and protection.

[0110] In recommended mode, the adaptive control coupler provides operators with audible, visual, or tactile warnings and visual risk alerts through a multi-level human-machine interface, without directly intervening in the control of the equipment;

[0111] In cooperative mode, the adaptive control coupler is configured to inject continuous fine-tuning commands for speed or trajectory into the ship loader control system to achieve smooth, non-stop cooperative avoidance.

[0112] In protection mode, the adaptive control coupler is configured to temporarily take over or override equipment control when the system determines that there is an emergency collision risk, and execute controlled smooth deceleration or emergency stop commands.

[0113] The system also includes a self-learning optimization module, which is connected to the data of the perception layer, cognition layer and execution layer respectively, and is used to optimize the prediction accuracy of the digital twin module and the strategy library of the collaborative decision-making module based on actual operating data.

[0114] This invention achieves a fundamental shift from passive alarm to proactive predictive collaborative collision avoidance, significantly improving operational efficiency and system intelligence while ensuring safety.

[0115] In the description of this invention, it should be understood that the terms "center," "longitudinal," "lateral," "length," "width," "thickness," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," "outer," "clockwise," "counterclockwise," "axial," "radial," and "circumferential" indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are used only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this invention.

[0116] In this invention, unless otherwise explicitly specified and limited, the terms "installation," "connection," "linking," and "fixing," etc., should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral part; they can refer to a mechanical connection, an electrical connection, or a connection that allows communication between them; they can refer to a direct connection or an indirect connection through an intermediate medium; they can refer to the internal communication of two components or the interaction between two components, unless otherwise explicitly limited. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0117] The present invention has been described above by way of example. Obviously, the specific implementation of the present invention is not limited to the above-described manner. Any improvements made by adopting the inventive concept and technical solution of the present invention, or direct application to other occasions without modification, are all within the protection scope of the present invention.

Claims

1. A method for preventing collisions with a ship loader, characterized in that, Includes the following steps: S1: Multi-source perception and data fusion: Simultaneously collect the body posture data of the ship loader, the relative pose data between it and adjacent targets, and the three-dimensional contour data of the environment, and perform fusion processing to obtain accurate anti-interference environmental status information. S2: Digital Twin Synchronization and Trajectory Prediction: Based on the fused data, the digital twin model of the ship loader is driven to achieve state synchronization; at the same time, based on the acquired real-time control commands and the kinematic model in the digital twin model, the motion trajectory of the machine and adjacent equipment in the future short time range is predicted in parallel, and a cluster of predicted trajectories is generated. S3: Dynamic Risk Field Calculation and Conflict Assessment: Based on the predicted motion trajectory, combined with the physical dimensions of the equipment and the dynamic safety margin, calculate the spatiotemporal risk field that evolves over time, and assess whether there are potential conflicts between the machine and adjacent equipment. S4: Priority-based intelligent collaborative decision-making: When a potential conflict is determined, based on the task priority information input by the terminal production management system, a comprehensive evaluation of multiple preset avoidance strategies is conducted to generate a collaborative avoidance scheme that enables multiple devices to work together to minimize global operation interruption. S5: Hierarchical Flexible Control Execution: Convert the cooperative avoidance scheme into hierarchical flexible control commands and execute them to achieve a balance between safety and efficiency; S6: Effect Evaluation and Closed-Loop Optimization: Collect actual operating data after the implementation of the plan, calculate the prediction deviation and avoidance effect, and use the data to adaptively optimize the digital twin model and decision logic.

2. The collision prevention method for a ship loader according to claim 1, characterized in that, In step S1, the relative pose data is obtained by fusing the displacement data of the absolute encoder of the ship loader's trolley with the ranging data of the millimeter-wave radar using Kalman filtering, and is used to output the real-time distance and velocity vectors between adjacent devices.

3. The collision prevention method for a ship loader according to claim 1, characterized in that, In step S2, the future short time horizon is 3 to 5 seconds; the kinematic model is integrated into the digital twin model to predict the future state based on the current state and control commands.

4. The collision prevention method for a ship loader according to claim 1, characterized in that, In step S3, the calculation of the spatiotemporal risk field is specifically as follows: taking the predicted trajectory as the central axis, and based on the equipment size, safety margin, and current environmental parameters, a three-dimensional dynamic envelope that expands outward with the predicted time point is constructed.

5. A collision prevention method for a ship loader according to claim 1, characterized in that, In step S4, multiple preset avoidance strategies include at least two of the following: complete stop, trajectory fine-tuning, and speed negotiation; the comprehensive evaluation includes a quantitative analysis of safety costs, efficiency losses, and task priorities.

6. A collision avoidance method for a ship loader according to claim 1, characterized in that, In step S5, the hierarchical flexible control instructions include: Suggestion-level instructions are used to provide operators with early warnings and visual guidance; Cooperative-level commands are used to inject fine-tuning quantities into the equipment control system through control signal couplers to achieve cooperative obstacle avoidance. Protection-level commands are used to take over control and perform controlled deceleration or stopping under the highest risk conditions.

7. A ship loader collision avoidance system, used to implement the ship loader collision avoidance method according to any one of claims 1-6, characterized in that, It adopts a three-layer intelligent architecture consisting of a perception layer, a cognition layer, and an execution layer, and achieves predictive collaborative protection based on digital twins, wherein: The perception layer is used to collect environmental and device status data from all dimensions, including: The body attitude sensing unit includes a high-precision tilt sensor installed on the boom of the ship loader, a length encoder for measuring the extension length of the boom, and a rotary encoder for measuring the slewing angle. It is used to collect raw data of boom movement and calculate the real-time accurate three-dimensional coordinates of the boom end. The relative pose sensing unit includes a travel absolute encoder installed on the traveling mechanism of the ship loader trolley and a millimeter-wave radar installed on the traveling beam or gantry structure, which are used to collect the traveling displacement of the machine and the absolute distance with adjacent equipment, and obtain the real-time distance and speed information between the two through subsequent fusion processing. The environmental contour perception unit includes at least two lidars deployed at the head of the ship loader boom and key parts of the main structure, used to perform three-dimensional point cloud scanning of adjacent equipment and complex obstacles to obtain their three-dimensional contour information. The data fusion processing unit is communicatively connected to each of the above-mentioned sensing units. It is used to perform time synchronization, coordinate system unification and filtering fusion processing on the raw data collected by each unit, and output stable, consistent environmental and equipment status information with a unified spatiotemporal reference for direct access by the cognitive layer. The cognitive layer, which connects to the perception layer via communication, serves as the system's decision-making center and includes: The digital twin module is used to build and drive a high-fidelity virtual model that is synchronized with physical devices based on physical modeling and sensing data, and integrates the kinematic model of the ship loader. The trajectory prediction module, integrated into or calling the digital twin module, is used to predict the future motion trajectory of the local device and adjacent devices based on real-time control commands and kinematic models. The collaborative decision-making module is used to dynamically calculate the spatiotemporal risk field based on the predicted trajectory, and when a spatiotemporal conflict is detected, it dynamically generates a collaborative avoidance strategy based on the task priority input by the terminal production management system. The execution layer, which communicates with the cognitive layer, is used to translate cooperative avoidance strategies into hierarchical control signals, including: Multi-level human-computer interaction interface, used to provide gradient warnings and operation guidance; An adaptive control coupler is used to flexibly inject hierarchical control signals into the ship loader control system according to a cooperative avoidance strategy. The system also includes a self-learning optimization module, which is connected to the data of the perception layer, cognition layer and execution layer respectively, and is used to optimize the prediction accuracy of the digital twin module and the strategy library of the collaborative decision-making module based on actual operating data.

8. A ship loader anti-collision system according to claim 7, characterized in that, The data fusion processing unit is configured to fuse encoder displacement data and millimeter-wave radar ranging data from the relative pose sensing unit using a Kalman filter algorithm to output interference-resistant and accurate real-time distance and velocity vectors between adjacent devices in vibration and dust environments.

9. A ship loader anti-collision system according to claim 7, characterized in that, The trajectory prediction module is configured to use a kinematic model to calculate the predicted positions of the local device and adjacent devices at multiple consecutive moments within the next 3 to 5 seconds in parallel, and generate a cluster of predicted trajectories.

10. A ship loader anti-collision system according to claim 7, characterized in that, The adaptive control coupler is configured to support three intervention modes: suggestion, cooperation, and protection; among them, In recommended mode, the adaptive control coupler provides operators with audible, visual, or tactile warnings and visual risk alerts through a multi-level human-machine interface, without directly intervening in the control of the equipment; In cooperative mode, the adaptive control coupler is configured to inject continuous fine-tuning commands for speed or trajectory into the ship loader control system to achieve smooth, non-stop cooperative avoidance. In protection mode, the adaptive control coupler is configured to temporarily take over or override equipment control when the system determines that there is an emergency collision risk, and execute controlled smooth deceleration or emergency stop commands.

Images