General-purpose unmanned trolley multi-target path planning and multi-machine cooperative scheduling method and system

Abstract

This invention discloses a general-purpose multi-objective path planning and multi-machine cooperative scheduling method and system for unmanned overhead cranes, comprising: acquiring task and crane status; constructing a multi-objective normalized comprehensive cost model including completion time, energy consumption, and waiting penalties, wherein energy consumption adopts a differential model coupled with load and acceleration; the upper layer uses dynamic tokens to improve the contract network protocol to achieve anti-starvation task allocation, and the lower layer uses spatiotemporal safe corridor constraints to improve the hybrid A* generation spatiotemporal path to achieve conflict pre-resolution; calculating an integral-type conflict risk function, and combining tokens and remaining costs for dynamic priority resolution and local replanning when the time limit is exceeded, forming a closed loop. The system includes perception, multi-objective configuration, cooperative scheduler, path planner, conflict resolution, and execution monitoring modules. This invention improves the versatility, energy efficiency, and dynamic robustness of multi-crane systems through deep collaboration of load adaptive energy consumption coupling, dynamic token anti-starvation, spatiotemporal corridor pre-resolution, and risk prediction and resolution.

Description

Technical Field

[0001] This invention relates to the field of automated scheduling technology, and in particular to a general-purpose method and system for multi-objective path planning and multi-machine collaborative scheduling of unmanned overhead cranes. Background Technology

[0002] As a key material handling equipment in industrial logistics systems, the operational efficiency and multi-machine collaboration level of unmanned overhead cranes directly affect the overall production rhythm of the workshop. With the in-depth advancement of intelligent manufacturing, scenarios in which multiple overhead cranes operate in parallel within the same or adjacent cross-area areas are becoming increasingly common, placing higher demands on the path planning and collaborative scheduling of multiple overhead cranes.

[0003] Currently, various methods exist for path planning and scheduling of unmanned overhead cranes. Path planning for a single crane often employs the A* algorithm based on grid maps, the artificial potential field method, or the fast exploration random tree algorithm, which can generate collision-free paths in static obstacle environments. In collaborative scenarios involving multiple cranes, common approaches include: centralized mathematical programming methods, which abstract multi-crane scheduling into a mixed-integer programming model and obtain the global optimum through a solver; time-window recursive methods, which allocate preset time windows to each crane to avoid interference; these methods are simple to implement but lack robustness against task insertion or disturbances; and agent-based distributed negotiation methods, where cranes adjust local paths through information exchange, but their optimization capability for global multi-objectives is limited.

[0004] However, the existing solutions mentioned above still have significant shortcomings in practical applications:

[0005] The optimization objectives are relatively singular, mostly aiming to minimize the maximum completion time or total travel distance, making it difficult to take into account multi-dimensional production requirements such as energy consumption and task waiting time. Even when some studies introduce multiple objectives, they only use linear combinations with fixed weights, lacking a refined energy consumption model coupled with the dynamic characteristics of the crane load. This makes it impossible to flexibly adapt to the differentiated needs for efficiency and energy saving at different process stages, and also fails to achieve truly load-adaptive energy-saving planning.

[0006] Multi-day traffic conflict resolution mechanisms are relatively rigid, typically relying on fixed priority rules or offline calculated time windows, lacking the ability to predict future conflict risks online and dynamically adjust priorities. When sudden disturbances or temporary order insertions occur, these fixed rules can easily lead to unnecessary waiting or deadlocks, significantly reducing the effectiveness of coordination.

[0007] The path planning layer and the scheduling layer are either too tightly coupled or too loosely coupled. Excessive coupling leads to model dimensionality explosion, making real-time solutions difficult; excessive looseness ignores kinematic constraints and actual energy costs, resulting in significant deviations from theoretical scheduling plans in actual execution, and even triggering residual collision risks. While existing layered architectures have made some attempts, they lack effective spatiotemporal constraint transmission and conflict feedback loops between layers, failing to achieve true collaboration.

[0008] Existing methods do not fully consider the impact of crane load changes on dynamic response and energy consumption. In path planning, energy consumption is mostly a linear function of distance, ignoring the energy consumption differences of the trolley acceleration and deceleration and the hoisting mechanism under different loads. This makes it difficult to achieve refined energy-saving operation while ensuring safety.

[0009] Therefore, there is an urgent need for a multi-crane path planning and collaborative scheduling scheme that can integrate multiple optimization objectives online, has the ability to predict and adaptively resolve conflict risks, and achieves layered and deep collaboration in its architecture, so as to significantly improve versatility and operational efficiency. Summary of the Invention

[0010] The purpose of this invention is to provide a general method and system for multi-objective path planning and multi-machine collaborative scheduling of unmanned overhead cranes, so as to solve the technical problems existing in the prior art.

[0011] The present invention aims to provide a general method and system for multi-objective path planning and multi-machine collaborative scheduling of unmanned overhead cranes, in order to solve the technical problems of existing solutions, such as single optimization objectives that are disconnected from the load, rigid conflict resolution, insufficient coordination of hierarchical architecture, and weak adaptability to dynamic disturbances.

[0012] The core improvement of this invention lies in proposing a deeply collaborative technical system: embedding a configurable multi-objective optimization model and a load-adaptive energy consumption differential model into a layered architecture. At the upper layer, an improved contract network protocol with dynamic priority tokens is used for starvation-resistant task allocation. At the lower layer, an improved hybrid A* algorithm with spatiotemporal safe corridor constraints is used to generate kinematically feasible energy-saving paths. An integral conflict risk function and a token-based dynamic resolution mechanism are designed to form a closed loop of prediction, resolution, and local replanning. This deep integration across layers brings synergistic benefits: multi-objective configuration guides task allocation and path preference; dynamic tokens enhance scheduling fairness and flexibility; spatiotemporal corridors ensure physical feasibility of paths and reduce subsequent conflict adjustments; and the risk function enables early smooth prediction and minimizes the replanning range. Overall, it achieves comprehensive performance that is unpredictable by existing technologies.

[0013] The present invention adopts the following technical solution:

[0014] A general method for multi-objective path planning and multi-machine cooperative scheduling of unmanned overhead cranes includes the following steps:

[0015] Obtain information on all pending tasks, real-time status information of each unmanned crane, and environmental model of the work area;

[0016] Based on the preset multi-objective weight configuration, a multi-objective optimization model is constructed, which includes the maximum completion time, total operating energy consumption, and cumulative task waiting penalty. The multi-objective evaluation is then normalized into a comprehensive cost index using the multi-objective optimization model. The total operating energy consumption is calculated based on a differential model coupled with load, acceleration, and velocity.

[0017] At the upper-level collaborative scheduling layer, based on the improved contract network protocol and the dynamic priority token mechanism, tasks are dynamically allocated to each crane and an initial collaborative task sequence is generated. The dynamic priority token is adjusted online according to the crane's cumulative load, waiting time, and task completion progress to avoid starvation in task allocation.

[0018] In the lower path planning layer, for overhead cranes assigned tasks, an improved hybrid A* algorithm constrained by a spatiotemporal safety corridor is used to generate conflict-free spatiotemporal paths. The spatiotemporal safety corridor is dynamically constructed based on the kinematic model of each overhead crane and the projection of future time windows, and serves as an intrusive state constraint in the path search process, thereby moving conflict resolution forward to the path generation stage.

[0019] Conflict risk prediction and dynamic priority resolution include: calculating the conflict risk value between crane pairs based on an integral conflict risk function; when the conflict risk value exceeds a preset threshold, dynamically adjusting the crane passage order according to real-time task priority, dynamic token level and remaining path cost; and triggering replanning only in local spatiotemporal segments for disturbed cranes, forming a closed loop of prediction-resolution-replanning until a globally feasible conflict-free collaborative solution is generated.

[0020] The final generated path instructions and control instructions are then sent to each crane for execution.

[0021] The objective function of the multi-objective optimization model is expressed in a weighted sum form based on reference point normalization as follows:

[0022]

[0023] In the formula, C max For the maximum completion time, E total For total energy consumption, W total To accumulate task waiting penalty values, C norm E norm W norm For the normalized reference value, α, β, and γ are configurable non-negative weight coefficients that satisfy α+β+γ=1.

[0024] The total energy consumption E totalAccording to the load adaptive energy consumption model, its differential expression is:

[0025]

[0026] Among them, dE i Let a be the energy consumption increment of crane i within dt. i For the combined acceleration of the large and small vehicles, v i v represents the combined speed of the large and small vehicles. z，i For the speed of the lifting mechanism, m 1oad，i Let g be the actual load mass of crane i, g be the gravitational acceleration, and c1, c2, and c3 be constant coefficients.

[0027] The conflict risk function is defined as follows: for cranes i and j within the future prediction time window t+ΔT:

[0028]

[0029] In the formula, d ij (τ) represents the minimum predicted distance between the outer contours of the vehicle over two days at time T, and d safe λ is the preset safe distance threshold, K is the risk gain coefficient, λ is the distance-sensitive attenuation coefficient, and I(·) is the indicator function.

[0030] This invention also provides a general-purpose unmanned overhead crane multi-objective path planning and multi-machine cooperative scheduling system, comprising:

[0031] The sensing and status acquisition module is used to acquire task information, real-time position, speed, load status, and environmental model of the crane;

[0032] A multi-objective optimization configuration module is used to store and adjust the weight coefficients α, β, γ, and calculate the comprehensive cost index, while providing load adaptive energy consumption model parameters.

[0033] The upper-level collaborative scheduler uses an improved contract network protocol and a dynamic priority token mechanism to allocate tasks to each crane and generate a preliminary scheduling sequence.

[0034] The lower-level path planner, with its built-in spatiotemporal safety corridor construction unit and path search unit based on improved hybrid A*, is used to generate conflict-free spatiotemporal paths for a single overhead crane; the conflict prediction and resolution module is used to calculate the conflict risk function value between overhead crane pairs in real time, execute dynamic priority resolution strategies, and drive local replanning.

[0035] It also includes an execution monitoring module, which is used to convert the final path into control commands for each crane drive and monitor the execution status.

[0036] Compared with the prior art, the beneficial effects of the present invention are:

[0037] Deep coupling of multi-objective optimization and load adaptive energy consumption model: The differential model of acceleration, speed and load related energy consumption shown in Equation (2) is embedded into the cost evaluation of path planning, so that the crane can automatically tend to reduce lifting and acceleration and deceleration actions when under heavy load, realizing true energy-saving optimization that changes with load, which is far superior to the existing energy consumption calculation based only on travel distance.

[0038] Dynamic priority tokens and improved contract network anti-starvation collaboration: The token value is dynamically adjusted based on load accumulation, waiting time, etc., and combined with contract network bidding to prevent high-load cranes from being continuously assigned tasks while low-load cranes are idle, or urgent tasks from waiting for a long time, thereby improving system fairness and dynamic adaptability.

[0039] The pre-resolution fusion of spatiotemporal safety corridors and hybrid A*: Spatiotemporal no-go zones are constructed and enforced during the path search phase, making the output path itself highly spatiotemporally conflict-free, greatly reducing the computational burden of subsequent global conflict resolution, and realizing the forward coordination of planning and resolution.

[0040] Robust closed loop of integral-type conflict risk function and local replanning: The smooth distance integral risk index can detect conflict trends in advance. Combined with dynamic priority and local replanning, only the disturbed segment needs to be fine-tuned for disturbances, maintaining global stability and overcoming the drawbacks of traditional hard collision detection and full-sequence rescheduling.

[0041] The overall solution achieves cross-level synergistic efficiency through the organic integration of the above features, and has significant versatility, energy saving and dynamic robustness. It is clearly different from existing single improvement or simple superposition solutions, and has outstanding substantive features and significant progress. Attached Figure Description

[0042] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0043] Figure 1 The flowchart is a general-purpose multi-objective path planning and multi-machine collaborative scheduling method and system for unmanned overhead cranes.

[0044] Figure 2 This is the first sub-flowchart of a general-purpose unmanned overhead crane multi-objective path planning and multi-machine collaborative scheduling method and system.

[0045] Figure 3 This is the second sub-flowchart of a general-purpose unmanned overhead crane multi-objective path planning and multi-machine collaborative scheduling method and system.

[0046] Figure 4 This is the third sub-flow diagram of a general-purpose unmanned overhead crane multi-objective path planning and multi-machine collaborative scheduling method and system.

[0047] Figure 5 This is a block diagram showing the components and structure of a general-purpose unmanned overhead crane multi-objective path planning and multi-machine cooperative scheduling method and system.

[0048] Figure 6 This is a schematic diagram illustrating a general-purpose unmanned overhead crane multi-objective path planning and multi-machine collaborative scheduling method and system. Detailed Implementation

[0049] To make the technical problems to be solved, the technical solutions, and the beneficial effects of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the present invention and are not intended to limit the present invention.

[0050] Example 1: Multi-crane collaborative scheduling under static task sets

[0051] This embodiment addresses a set of fixed handling tasks assigned by the workshop within a production cycle, and utilizes this method to achieve conflict-free scheduling and path planning for multiple overhead cranes.

[0052] First, the perception and status acquisition module obtains all task information to be executed from the manufacturing execution system or scheduling terminal, including picking point, delivery point, material type, target completion time limit, etc., and reads the real-time position, current speed, load status of each unmanned crane, as well as the environmental occupancy grid map constructed by LiDAR or multi-line scanning equipment.

[0053] The multi-objective optimization configuration module has preset weight combinations for high efficiency, such as α taking a relatively large value and β and γ taking smaller values, to highlight the importance of completion time. The comprehensive cost index is calculated according to equation (1), and the reference value is set according to the statistical characteristics of the workshop's historical operations for dimensionless normalization. At the same time, according to the load adaptive energy consumption model of equation (2), the total energy consumption E of each potential path scheme is estimated in real time. total This energy consumption model is introduced as a cost term during the node expansion stage of path planning, which naturally makes the search algorithm tend to choose paths with gentle acceleration and deceleration and low take-off energy consumption, thus forming energy-saving optimization at the planning level.

[0054] The upper-layer collaborative scheduler employs an improved contract network protocol for task allocation. Traditional contract networks operate with a task posting-bidding-awarding process. This embodiment improves upon this by introducing a dynamic priority token mechanism. Initially, each crane is assigned a basic priority token based on its current idle time, cumulative load, and expected remaining work time. The token's value indicates its priority in competing for task resources. The token is not fixed but dynamically updated during operation: for cranes continuously handling heavy loads, the token value gradually decreases to avoid excessive concentration of allocation leading to uneven energy consumption or equipment wear; for cranes waiting for tasks for extended periods, the token value gradually increases to ensure their chance of obtaining tasks, thus preventing "starvation." The scheduler treats each task in the task pool as a bidding target. Each crane submits a bid based on its location, existing task sequences, and the increase in the overall cost F resulting from matching the task. The scheduler evaluates bids by combining the weighted result of token priority and bidding cost, prioritizing allocation to cranes with smaller overall cost increments and moderate token priorities. After each round of allocation is completed, the token manager and task sequence generator update their status in tandem, thereby generating the initial collaborative task sequence for each crane.

[0055] The lower-level path planner generates specific operating trajectories for overhead cranes assigned tasks using spatiotemporal safety corridors and an improved hybrid A* algorithm. The spatiotemporal safety corridor is constructed by recursively predicting the occupied intervals for several future time steps along the assigned task sequence based on the kinematic models of the crane's main trolley, auxiliary trolley, and hoisting mechanism. This interval is then expanded according to a safety distance threshold, delineating spatiotemporal no-go zones on the cross-segment track projection that other cranes cannot intrude into. The improved hybrid A* algorithm, building upon the standard hybrid A* algorithm, expands the state space from a purely spatial dimension to include a temporal dimension and introduces a constraint cost term for the future spatiotemporal corridor into the heuristic function. During path search, node expansion always occurs within the kinematically feasible region and directly avoids all spatiotemporal corridor no-go zones, thus the generated spatiotemporal path itself already possesses a high degree of conflict-free guarantee. This pre-resolution design significantly reduces the computational load and adjustment range of subsequent global conflict resolution.

[0056] After the spatiotemporal paths of all overhead cranes are generated, the conflict prediction and resolution module performs conflict risk function calculations on each pair of overhead cranes. For cranes i and j, R is calculated according to equation (3). i，j (t). The integral form of this function smoothly reflects the duration and proximity of the collision, rather than a simple instantaneous collision determination. When there exists any R i，j (t) Exceeds the set risk acceptance threshold R thWhen an unacceptable risk of conflict is detected, the conflict predictor reports the involved crane pair and its spatiotemporal overlap section to the coordinating scheduler. The scheduler adjusts the passage order based on the dynamic priority tokens currently held by the cranes and the additional path cost required to bypass the conflict zone: the crane with the higher token priority or the greater detour cost is granted temporary priority passage, while the other crane slows down and waits at a designated waiting point before the conflict zone. This process is as follows: Figure 4 As shown in the diagram. Subsequently, the scheduler re-triggers lower-level path planning for the waiting overhead cranes, but the planning scope is limited to the conflict period and its adjacent local spatiotemporal segments, rather than a full sequence reconstruction. This local replanning strategy greatly reduces computational overhead while maintaining the locked state of the undisturbed parts, ensuring the stability and rapid response capability of the overall scheme. Finally, the final path command and control command are converted into speed and acceleration curves of the overhead crane PLC or motion controller through the execution monitoring module and then sent out to complete the collaborative operation of multiple overhead cranes.

[0057] This embodiment fully demonstrates how, in a static task environment, a globally conflict-free spatiotemporal path can be generated based on a configurable multi-objective model and a hierarchical deep collaborative architecture. Through the synergistic effects of token anti-starvation, spatiotemporal corridor pre-dissolution, integral risk prediction, and local replanning, a balance of high versatility, high efficiency, and energy saving is achieved.

[0058] Example 2: Online Adaptive Scheduling under Dynamic Task Insertion

[0059] In production, there are often situations where orders are inserted temporarily or tasks are changed. This embodiment focuses on explaining the response method of the present invention under dynamic events.

[0060] The perception module detects the sudden addition of a new task and immediately updates the task pool. At this time, some cranes may be executing existing paths and cannot be interrupted. The multi-objective optimization configuration module maintains the current weights. The upper-layer collaborative scheduler does not completely reschedule all tasks, but instead adopts an online incremental allocation strategy: only new tasks are published as bidding targets, without changing crane task segments already in an execution-locked state; for cranes waiting for tokens or whose remaining paths can still be adjusted, they are allowed to participate in the bidding for the new task. The improved contract network protocol and dynamic token mechanism are still used to generate supplementary allocation results. This dynamic token mechanism demonstrates its advantage in resisting starvation: if a new task is urgent and a particular crane in the bidding documents is the most suitable, but its token priority is low, the system compares the token with the task urgency weight, temporarily increasing the token value or granting exceptional passage rights when necessary, ensuring that urgent tasks are responded to promptly.

[0061] The addition of a new task may disrupt the existing, verified spatiotemporal coordination between overhead cranes. Therefore, the conflict prediction and resolution module quickly calculates the conflict risk function for the crane pairs affected by the newly assigned task. Since the function uses an integral form, it is predictive and can anticipate interference trends within the next few minutes. Once the risk exceeds a threshold, the resolution process is initiated. Unlike Example 1, the dynamic priority adjustment here considers not only tokens and detour costs but also the urgency weight of the new task itself, which can be marked in the task attributes. Urgent tasks can temporarily obtain higher passage priority and are thus preferentially inserted into the original scheduling plan.

[0062] If intervention is required on the cranes currently executing tasks during the resolution process, waiting or speed adjustment is only applied at the next safe decision point after the cranes have completed their current uninterruptible micro-actions, avoiding the risk of load swaying caused by sudden stops. A time relaxation factor is introduced into the local replanning process, appropriately relaxing the time constraints of the original path when planning a new path to absorb disturbances. After replanning, all cranes are still guaranteed to have conflict risk function values ​​below the threshold. Through this collaborative closed loop of "prediction-dynamic prioritization-local replanning," the system gracefully absorbs random tasks without global reconstruction, demonstrating good dynamic robustness.

[0063] Example 3: Differentiated Multi-Objective Scheduling Configurable for Energy Saving and Efficiency This example demonstrates another dimension of versatility—by adjusting the weighting coefficients to switch scheduling strategies based on the difference between daytime and nighttime peak and valley electricity prices or equipment maintenance status, an efficiency-first or energy-saving-first operation mode can be achieved.

[0064] Operator through Figure 6 The user scheduling and monitoring interface shown allows direct adjustment of the multi-objective weight slider. For example, during peak energy periods, the β weight can be increased, and α and γ can be decreased accordingly. The multi-objective optimization configuration module reconstructs the comprehensive cost function, and the upper and lower layer algorithms then re-evaluate the scheduling and path scheme according to the new criteria. The load adaptive energy consumption model in Equation (2) plays a key role in this process: when the crane lifts a heavy load, the model will significantly increase the energy consumption cost of the lifting speed in the total energy consumption calculation. Therefore, the lower layer path planning will naturally tend to choose the path that reduces the number of heavy load lifts or the lifting height, and even arrange the crane with a lighter load to undertake more tasks without affecting the overall completion time. This path replanning driven by the load-energy consumption coupling model, combined with the configurable weights of the upper layer, enables the same system to quickly switch between "efficiency mode" and "energy-saving mode" without redeploying programs or hard rules.

[0065] Furthermore, parameters such as safe distance thresholds and conflict risk acceptance thresholds can be configured through the interface. After a configuration change, the system can immediately adjust tasks in currently unlocked segments online, generating collaborative solutions that better suit the new preferences. This multi-objective switching achieved through weighting and parameterization gives the entire system strong versatility, enabling it to adapt to various production requirements, from rapid delivery to low-consumption and stable operation.

[0066] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0067] The above are merely preferred embodiments of the present invention and do not limit the scope of the patent. Any equivalent structural or procedural transformations made based on the description and drawings of the present invention, or direct or indirect applications in other related technical fields, are similarly included within the scope of patent protection of the present invention.

Claims

1. A general method for multi-objective path planning and multi-machine cooperative scheduling of unmanned overhead cranes, characterized in that, Includes the following steps: Acquire information on all pending tasks, the real-time status of each unmanned crane, and the environmental model of the work area; Based on the preset multi-objective weight configuration, a multi-objective optimization model is constructed, which includes the maximum completion time, total operating energy consumption, and cumulative task waiting penalty. The multi-objective evaluation is normalized into a comprehensive cost index. The total operating energy consumption is estimated in real time based on a differential model of load, acceleration, and velocity coupling. At the upper-level collaborative scheduling layer, based on the improved contract network protocol and the dynamic priority token mechanism, tasks are dynamically allocated to each crane to generate an initial collaborative task sequence; among them, the dynamic priority token is adjusted online according to the cumulative load and waiting time of the crane to avoid starvation; In the lower path planning layer, for overhead cranes assigned tasks, an improved hybrid A* algorithm constrained by a spatiotemporal safety corridor is used to generate conflict-free spatiotemporal paths. The spatiotemporal safety corridor is dynamically constructed based on the crane kinematic model and future time window projection, and serves as an intrusive constraint for path search, achieving conflict pre-resolution. Conflict risk prediction and dynamic priority resolution are performed, including calculating the risk value between crane pairs based on an integral conflict risk function. When the risk value exceeds a preset threshold, the passage order is dynamically adjusted according to real-time task priority, token level, and remaining path cost, and local spatiotemporal segment replanning is triggered for disturbed cranes, forming a prediction-resolution one-level planning closed loop until a globally feasible conflict-free collaborative solution is generated. The final path command and control command are issued to each crane for execution.

2. The method according to claim 1, characterized in that, The objective function of the multi-objective optimization model is expressed in a weighted sum form based on reference point normalization as follows: In the formula, C max For the maximum completion time, E total For total energy consumption, W total To accumulate task waiting penalty values, C norm E norm W norm For the normalized reference value, α, β, and γ are configurable non-negative weight coefficients that satisfy α+β+γ=1.

3. The method according to claim 2, characterized in that, The total energy consumption E total According to the load adaptive energy consumption model, its differential expression is: Among them, dE i Let a be the energy consumption increment of crane i within dt. i For the combined acceleration of the large and small vehicles, v i v represents the combined speed of the large and small vehicles. z，i For the speed of the lifting mechanism, m load，i Let g be the actual load mass of crane i, g be the gravitational acceleration, and c1, c2, and c3 be constant coefficients. The energy consumption model is introduced as a cost term during the path search node expansion stage to guide the generation of energy-saving paths.

4. The method according to claim 1, characterized in that, The improved contract network protocol introduces a dynamic priority token mechanism. The token value decreases as the cumulative load of the overhead crane increases and increases as the waiting time increases. During task evaluation, the bidding cost and token priority are combined to prevent overuse of high-load overhead cranes and long-term idleness of low-load overhead cranes.

5. The method according to claim 1, characterized in that, The spatiotemporal safety corridor is a series of space-time occupied intervals obtained recursively from the kinematic equations of the overhead crane at discrete time steps, and is expanded to embed a safety margin. The improved hybrid A* algorithm treats these corridors as unintrusive states during node expansion and incorporates the time dimension into the state space to ensure the spatiotemporal non-conflictability of the output path.

6. The method according to claim 1, characterized in that, The conflict risk function is defined as follows: for cranes i and j within the future prediction time window t+ΔT: Where, d ij (τ) represents the predicted minimum distance, d safe K is the safety distance threshold, λ is the risk gain coefficient, I(·) is the attenuation coefficient, and I(·) is the indicator function; this integral form smoothly reflects the duration and intensity of the conflict.

7. The method according to claim 6, characterized in that, The dynamic priority resolution includes: comparing the current token priority of the conflicting overhead cranes with the remaining path cost they need to pay to avoid the conflict zone, granting the right of passage in the conflict zone to the cost-sensitive and / or high-priority overhead cranes, while the other overhead crane waits or slows down in front of the conflict zone; and this resolution can be linked with local replanning, adjusting only the path during the affected time period.

8. The method according to claim 7, characterized in that, The local replanning only corrects the path of the affected cranes during the conflict period and its adjacent periods, while the unaffected task segments remain locked, and a time relaxation factor is introduced to absorb disturbances.

9. The method according to claim 1, characterized in that, The multi-objective weight configuration can be adjusted online through the human-machine interface. After adjustment, incremental rescheduling is initiated for tasks in the unlocked segment. Combined with the load adaptive energy consumption model, dynamic switching between efficiency-first and energy-saving-first modes is achieved.

10. A general-purpose unmanned overhead crane multi-objective path planning and multi-machine cooperative scheduling system, characterized in that, include: The sensing and status acquisition module is used to acquire task information, real-time position, speed, load status, and environmental model of the crane; A multi-objective optimization configuration module is used to store and adjust the weight coefficients α, β, γ, and calculate the comprehensive cost index. It also has a built-in load adaptive energy consumption model. The upper-level collaborative scheduler uses an improved contract network protocol and a dynamic priority token mechanism to allocate tasks to each crane and generate a preliminary scheduling sequence. The lower-level path planner, with a built-in spatiotemporal safety corridor construction unit and a path search unit based on improved hybrid A*, is used to generate conflict-free spatiotemporal paths for a single crane. The conflict prediction and resolution module is used to calculate the integral conflict risk function value between crane pairs in real time, execute dynamic priority resolution strategies, and drive local replanning. The execution monitoring module is used to convert the final path into control commands for each crane drive and monitor the execution status.

Images