An aerial rail type charging robot management method and system

Abstract

This invention provides a management method and system for aerial track-mounted charging robots, belonging to the field of automatic scheduling technology. The method includes: acquiring the status data of the charging robots in real time; calculating the dynamic priority evaluation value of each robot based on the status data, wherein the evaluation value is a weighted sum of energy urgency, task criticality, and space cost components, and the weighting coefficients can be dynamically and adaptively adjusted according to the real-time operating status of the system; allocating charging resources and planning paths according to the charging requests based on the dynamic priority evaluation value. This invention solves the problem that fixed weight strategies cannot achieve optimal scheduling in dynamic and changing scenarios by introducing a state-aware adaptive weight adjustment mechanism, realizing the coordinated optimization of charging resources, track resources, and time resources, and improving the overall efficiency and robustness of the system.

Description

Technical Field

[0001] This invention relates to the field of automatic scheduling technology, and in particular to a management method and system for an aerial track-mounted charging robot. Background Technology

[0002] With the increasing prevalence of automated warehousing and smart factories, mobile robots (hereinafter referred to as "rail robots") mounted on aerial rail networks for tasks such as material handling and inspection are becoming increasingly widely used. To ensure the continuous operation of these robots, a common solution is to deploy charging robots that move along the rails, or to set up charging stations at fixed locations to provide charging services for rail robots with low battery levels. Therefore, how to efficiently manage these charging robots and charging resources has become a key technical issue affecting the overall system's operating efficiency, reliability, and cost.

[0003] Existing methods for managing charging robots primarily focus on static or simple rule-based scheduling strategies. The most common strategies are "First-Come, First-Served" (FCFS) or the "nearest charging station" principle, whereby when a robot's battery level falls below a threshold, it moves to the nearest available charging station and waits in line. Another improved strategy is to introduce fixed priorities, determining the charging order based on the preset priority of the tasks performed by the robots, allowing robots with higher-priority tasks to jump the queue. Furthermore, some research attempts to combine path planning algorithms with charging scheduling to plan a shorter, congestion-avoiding path for requesting robots to reach the charging station.

[0004] However, in complex real-world operating environments, the aforementioned existing technical solutions have gradually revealed many limitations, specifically: Inflexible scheduling strategies cannot handle dynamic loads: FCFS or fixed-priority strategies do not adequately consider the real-time state of the system. For example, when multiple high-priority robots need charging simultaneously, congestion may still occur at their target charging stations; while a robot performing a low-priority task but about to run out of power may stop due to a long queue, causing job interruptions or even safety accidents. The spatiotemporal imbalance of system load makes it difficult for static rules to achieve global optimization.

[0005] The decision-making process suffers from a lack of holistic optimization: Existing methods often handle charging scheduling, path planning, and conflict resolution in isolation. The "nearest charging station" strategy only optimizes the travel distance but may turn that charging station into a hotspot, causing new congestion. Simple path planning algorithms may not consider the charging plans of other robots, leading to conflicts and waiting at track intersections. The three key dimensions of energy state (battery level), task attributes (urgency), and spatial resources (tracks, charging stations) are not placed within a unified framework for collaborative decision-making.

[0006] Fixed parameters and lack of adaptability: While some more advanced scheduling algorithms introduce weighted calculations to integrate multiple factors (such as battery level, task priority, and distance), the weight coefficients of each factor are usually pre-set and fixed. This makes the system unable to adapt to different operating phases and conditions. For example, during off-peak periods, energy economy should be prioritized (e.g., prioritizing charging robots with lower battery levels); while during peak periods or when a large number of urgent tasks suddenly appear, task continuity should be prioritized. Fixed weight coefficients cannot achieve this flexible switching of strategies, limiting the overall robustness and efficiency ceiling of the system.

[0007] Passive conflict resolution impacts system smoothness: Existing systems often employ passive, reactive methods to resolve path conflicts between robots on tracks, such as stopping one robot when an impending collision is detected. This approach increases robot travel time, reduces the efficiency of the track network, and can trigger cascading waits during periods of high charging demand, exacerbating system congestion.

[0008] Therefore, there is an urgent need for a management method and system for aerial track-mounted charging robots to solve the technical problems of rigid scheduling, multi-dimensional fragmentation, poor adaptability, and passive conflict resolution in existing aerial track-mounted charging robot management technologies. Summary of the Invention

[0009] The purpose of this invention is to overcome the shortcomings of the prior art and provide a management method and system for an aerial orbital charging robot. This new management method and system aims to perceive the system status in real time, dynamically adjust the decision focus, and proactively coordinate and optimize energy, tasks, and space resources to improve overall operational efficiency, ensure the execution of high-priority tasks, and ensure the stable operation of the system in changing environments.

[0010] To achieve the above objectives, this application proposes a management method for an aerial orbital charging robot, comprising: Step S1: Acquire the status data of each charging robot in real time. The status data includes location coordinates, remaining battery percentage, task priority, and estimated energy consumption. Step S2: Based on the state data, calculate the dynamic priority evaluation value for each charging robot. The dynamic priority evaluation value is obtained by weighted summation, and its calculation formula is: P_i = α·E_i + β·T_i + γ·S_i. Where P_i is the dynamic priority evaluation value of charging robot i, E_i is the energy urgency component, T_i is the task criticality component, S_i is the space cost component, and α, β, γ are weighting coefficients and satisfy α+β+γ=1. Step S3: Based on the dynamic priority evaluation value, allocate charging resources to each charging robot that makes a charging request and plan an aerial track path to the charging station. The charging resources include charging piles and charging time periods.

[0011] As a further solution, the energy urgency component E_i is calculated as follows: E_i = exp[-(B_i - B_min) / (τ·(1 - D_i / R_i))]; Where B_i is the current remaining power percentage of charging robot i, B_min is the preset safe power threshold, τ is the energy decay time constant related to the charging robot model, D_i is the aerial track distance from the current position of charging robot i to the nearest charging station, and R_i is the range of charging robot i when fully charged.

[0012] As a further solution, the formula for calculating the task criticality component T_i is: T_i = (TP_i / TP_max) · [1 + (T_remain_i / T_total_i)]; Where TP_i is the current task priority value of charging robot i, TP_max is the maximum value of the task priority value, T_remain_i is the remaining execution time of the current task of charging robot i, and T_total_i is the total budget time of the current task of charging robot i.

[0013] As a further solution, the spatial cost component S_i is calculated as: S_i = σ(W_i) ·(C_i / C_max); Where σ(·) is an S-shaped function of waiting time, W_i is the expected waiting time of charging robot i in the current charging queue, C_i is the real-time congestion degree on the planned path of charging robot i to the target charging station, and C_max is the maximum congestion degree.

[0014] As a further solution, the method for allocating a charging time period for the charging request in step S3 is as follows: based on the dynamic priority evaluation value P_i, allocate an elastic time window [Tre, Tre+ΔT_i], where Tre is the time when the charging request is issued, ΔT_i is the elastic window duration and ΔT_i = ΔT_max · (1 - P_i), and ΔT_max is the preset maximum elastic window duration.

[0015] As a further solution, the weighting coefficients α, β, and γ are non-fixed values ​​and are adaptive variables that are dynamically adjusted according to the real-time operating status of the system. The adjustment is based on the current system state vector X=(x1, x2, x3, x4), where x1 is the proportion of emergency charging requests, x2 is the proportion of high-priority tasks, x3 is the system track congestion index, and x4 is the variance of charging pile utilization.

[0016] As a further solution, the steps for dynamically adjusting the weighting coefficients α(t), β(t), and γ(t) include: Step S61: Monitor and calculate the system state vector X(t) at the current time t; Step S62: Calculate the change in system state from the previous adjustment time to the current time, dX = X(t) - X(t-Δt); Step S63: Based on the preset state-weight sensitivity matrix A and the state change amount dX, predict the weight adjustment direction ΔW_pred = η·(A·dX), where η is the learning rate; Step S64: Project the adjusted weight vector W_temp = [α(t-Δt), β(t-Δt), γ(t-Δt)] + ΔW_pred onto a constraint space that satisfies α+β+γ=1 and all coefficients are non-negative, to obtain the updated weighting coefficients.

[0017] As a further solution, the state-weight sensitivity matrix A is obtained by training with historical operating data, and its element a_ij represents the partial influence strength of the j-th system state variable on the i-th weighting coefficient.

[0018] On the other hand, the present invention also provides an aerial orbital charging robot management system for implementing an aerial orbital charging robot management method as described in any of the preceding claims, comprising: The status monitoring module is used to acquire and update the status data of each charging robot in real time; The priority calculation module is configured to calculate the dynamic priority evaluation value of each charging robot based on the status data using a dynamic priority evaluation function. The scheduling decision module is configured to allocate charging piles to charging requests and plan aerial track paths based on the dynamic priority evaluation value. The track control module is used to control the charging robot's movement on the aerial track according to the planned path; The charging station management module is used to manage the status and charging sequence of each charging pile.

[0019] As a further solution, a weight adaptation module is also included, which is connected to the priority calculation module and is configured as follows: Real-time monitoring of system operating status indicators; Based on the system operation status indicators and the pre-stored state-weight mapping relationship, the weighting coefficients in the dynamic priority evaluation function are dynamically adjusted. The adjusted weighting coefficients are sent to the priority calculation.

[0020] Compared with related technologies, the aerial track-type charging robot management method and system provided by this invention have the following advantages: 1. This invention introduces a dynamic priority evaluation value as the core scheduling basis and innovatively enables the weighting coefficients for calculating this evaluation value to be adaptively adjusted according to the real-time operating status of the system, thereby constructing a two-layer dynamic optimization framework. This system no longer relies on preset fixed rules but becomes a feedback control system capable of sensing the environment and making autonomous decisions. This allows the scheduling strategy to flexibly adapt to various operating conditions such as peak and off-peak periods and sudden emergencies, dynamically seeking optimization from a global perspective, and solving the fundamental problem of performance degradation of static strategies under dynamic loads. 2. This invention encodes energy urgency, task criticality, and space cost into a unified mathematical model and coordinates the relationship between the three through dynamic weights. When the system is congested, the algorithm can automatically increase the weight of space cost to prioritize traffic flow. When high-priority tasks are concentrated, the weight of task criticality is increased to ensure operational continuity. When the number of low-battery robots increases, the weight of energy urgency is increased to prevent downtime accidents. This collaborative mechanism based on a mathematical model enables the allocation of limited charging piles, aerial rail channels, and time windows to achieve overall optimization, which can improve the overall throughput of the system and significantly reduce the completion delay of high-priority tasks.

[0021] 3. The adaptive weight adjustment algorithm of this invention continuously monitors the system state vector and uses the state-weight sensitivity matrix to predict the optimal adjustment direction, enabling the system to actively avoid risks. When the proportion of robots with critical battery power increases, the algorithm automatically and quickly increases the weight coefficient of energy urgency to ensure that scheduling resources are tilted towards the most dangerous units, thereby reducing the incidence of unplanned shutdowns or emergency rescue events caused by power depletion, and providing a solid guarantee for continuous industrial operations.

[0022] 4. This invention allocates a flexible time window to each request based on a dynamic priority evaluation value; high-priority requests receive a narrow window to ensure timely response, while low-priority requests receive a wide window, allowing them to be appropriately delayed when resources are scarce. This design gives the scheduler more flexible scheduling space during peak hours, allowing it to arrange charging tasks and routes more compactly, like "puzzle pieces," thereby reducing the idle waiting time of charging piles and improving the overall traffic efficiency of the rail network.

[0023] 5. The core algorithm of this invention does not rely on complex "black box" artificial intelligence models, but is built on clear mathematical formulas and constrained optimization theory. Furthermore, the computational complexity of this mathematical framework is controllable, meeting the needs of real-time scheduling for large-scale systems. In addition, the design of the system state vector and sensitivity matrix has good scalability, facilitating the integration of new optimization dimensions (such as electricity prices and equipment health status) or adaptation to more complex track topologies in the future. Attached Figure Description

[0024] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0025] To more clearly illustrate the technical solutions in the embodiments of this application or related technologies, the accompanying drawings used in the description of the embodiments or related technologies will be briefly introduced below. Obviously, those skilled in the art can obtain other drawings based on these drawings without creative effort.

[0026] Figure 1 A schematic diagram illustrating the steps of a management method for an aerial track-mounted charging robot provided by the present invention; Figure 2 This invention provides a schematic diagram of the structure of an aerial track-type charging robot management system. The purpose, features, and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0027] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations.

[0028] Example 1 Please see Figure 1 This embodiment provides a management method for an aerial orbital charging robot, including: Step S1: Acquire the status data of each charging robot in real time. The status data includes location coordinates, remaining battery percentage, task priority, and estimated energy consumption. Step S2: Based on the state data, calculate the dynamic priority evaluation value for each charging robot. The dynamic priority evaluation value is obtained by weighted summation, and its calculation formula is: P_i = α·E_i + β·T_i + γ·S_i. Where P_i is the dynamic priority evaluation value of charging robot i, E_i is the energy urgency component, T_i is the task criticality component, S_i is the space cost component, and α, β, γ are weighting coefficients and satisfy α+β+γ=1. Step S3: Based on the dynamic priority evaluation value, allocate charging resources to each charging robot that makes a charging request and plan an aerial track path to the charging station. The charging resources include charging piles and charging time periods.

[0029] It should be noted that the key difference between this embodiment and existing static or simple rule-based scheduling lies in the introduction of a dynamic priority evaluation mechanism based on multi-dimensional state perception as the unified basis for all scheduling decisions. Its technical path is clearly divided into three logically rigorous steps: First, the system collects multi-dimensional state data for each charging robot in real time, including its location, battery level, task priority, and energy consumption prediction, providing accurate input for dynamic decision-making. Second, it innovatively summarizes the key factors affecting scheduling into three quantitative components: "energy urgency," "task criticality," and "space cost," and integrates them into a single, comparable dynamic priority evaluation value through a weighted summation model, thereby transforming the complex multi-objective optimization problem into a computable form. Finally, the system uses this dynamic evaluation value to perform integrated allocation and planning of charging resources (charging piles and time periods) and spatial resources (aerial tracks).

[0030] Furthermore, the formula for calculating the energy urgency component E_i is: E_i = exp[-(B_i - B_min) / (τ·(1 - D_i / R_i))]; Where B_i is the current remaining power percentage of charging robot i, B_min is the preset safe power threshold, τ is the energy decay time constant related to the charging robot model, D_i is the aerial track distance from the current position of charging robot i to the nearest charging station, and R_i is the range of charging robot i when fully charged.

[0031] Specifically, this implementation introduces the ratio of "the aerial track distance from the current location to the nearest charging station (D_i)" to "the range on a full charge (R_i)". The formula dynamically reflects the relative energy consumption pressure of the journey to charging. Simultaneously, the "energy decay time constant (τ)" accommodates the energy consumption characteristics of different robot models. Crucially, the mathematical form of the exponential function exp[-(B_i - B_min) / (τ·(1 - D_i / R_i))] causes the energy urgency value to rise sharply and non-linearly when the battery level approaches the safety threshold or the energy consumption ratio of the journey is high. This design ensures that the system can sensitively perceive and prioritize robots that are "not absolutely depleted of power, but pose a high actual risk due to their remote location," thereby proactively preventing downtime accidents caused by power depletion at the algorithm level and realizing a safe scheduling shift from "post-event emergency response" to "pre-event early warning."

[0032] Furthermore, the calculation formula for the task criticality component T_i is: T_i = (TP_i / TP_max) · [1 + (T_remain_i / T_total_i)]; Where TP_i is the current task priority value of charging robot i, TP_max is the maximum value of the task priority value, T_remain_i is the remaining execution time of the current task of charging robot i, and T_total_i is the total budget time of the current task of charging robot i.

[0033] Specifically, the innovation of the calculation formula in this embodiment lies in the introduction of a time dimension to dynamically correct the static task priority. It first normalizes the preset priority of the task by (TP_i / TP_max) as the basic weight; then, crucially, by multiplying by a factor [1 + (T_remain_i / T_total_i)], the criticality of the task increases linearly as the remaining execution time (T_remain_i) decreases.

[0034] This means that a task, even if initially low in priority but nearing its deadline, will have its criticality significantly increased, thus gaining higher charging priority in dynamic scheduling. This design effectively prevents task timeouts or interruptions due to charging waits, transforming the originally static and rigid task priority management into an adaptive mechanism that can sense task execution progress and dynamically ensure task continuity, thereby optimizing the system's task completion rate and reliability from a time perspective.

[0035] Furthermore, the spatial cost component S_i is calculated as follows: S_i = σ(W_i) · (C_i / C_max); Where σ(·) is an S-shaped function of waiting time, W_i is the expected waiting time of charging robot i in the current charging queue, C_i is the real-time congestion degree on the planned path of charging robot i to the target charging station, and C_max is the maximum congestion degree.

[0036] Specifically, the computational model in this embodiment creatively couples time waiting costs with spatial path costs. Its core consists of two parts: First, it quantifies the nonlinear impact of charging queue delays through an S-shaped function σ(W_i) about waiting time. The characteristics of this function make the cost of short-term waiting grow slowly, while long-term waiting will lead to a sharp increase in cost, thereby effectively avoiding the robot from getting stuck in an indefinite queue. Second, it introduces the ratio of real-time path congestion C_i to the system's maximum congestion C_max, which dynamically reflects the difficulty of track passage on the way to the target charging station.

[0037] The key innovation of this formula lies in its integration of "queue waiting time" and "congestion during travel" into a unified, comprehensive spatial cost metric, rather than assessing distance or congestion in isolation. This allows the scheduling system to intelligently identify charging station options that, while physically close, have high actual arrival costs due to long queues or route congestion. Consequently, it guides robots to choose the globally optimal charging solution with the lowest total time cost, significantly improving the overall traffic efficiency and resource utilization of the aerial rail network.

[0038] Furthermore, in step S3, the method for allocating a charging time period for the charging request is as follows: based on the dynamic priority evaluation value P_i, an elastic time window [Tre, Tre+ΔT_i] is allocated, where Tre is the time when the charging request is issued, ΔT_i is the duration of the elastic window and ΔT_i = ΔT_max · (1 - P_i), and ΔT_max is the preset maximum elastic window duration.

[0039] Specifically, in this embodiment, the dynamic priority evaluation value (P_i) is creatively transformed into an executable "elastic time window" scheduling mechanism. The core of this mechanism lies in establishing an inverse functional relationship between priority and time tolerance: ΔT_i = ΔT_max · (1 - P_i). That is, the higher the priority, the narrower the allocated elastic time window, and the more urgent the required response; conversely, the lower the priority, the larger the acceptable delay range. This design, for the first time, realizes "differentiated service" and "time resource pooling" in charging scheduling: high-priority tasks receive deterministic guarantees, while the elastic time of low-priority tasks becomes a buffer resource for the system to cope with sudden congestion and perform global optimization adjustments. By mapping abstract priority values ​​to specific and flexibly plannable time intervals, this scheme enables the system to significantly improve the time reuse rate of charging piles and tracks while ensuring the timeliness of critical tasks, fundamentally optimizing system throughput.

[0040] Furthermore, the weighting coefficients α, β, and γ are non-fixed values ​​and are adaptive variables that are dynamically adjusted according to the real-time operating status of the system. The adjustment is based on the current system state vector X=(x1, x2, x3, x4), where x1 is the proportion of emergency charging requests, x2 is the proportion of high-priority tasks, x3 is the system track congestion index, and x4 is the variance of charging pile utilization.

[0041] Specifically, this embodiment abandons the limitation of fixed weighting coefficients in existing technologies and innovatively proposes that α, β, and γ are adaptive variables that are dynamically adjusted based on the real-time operating state vector X=(x1, x2, x3, x4) of the system. Among them, the state vector accurately describes the macroscopic situation of the system, such as the proportion of emergency charging requests (x1) reflecting energy security pressure, and the track congestion index (x3) reflecting the tension of space resources.

[0042] This design means that the priority calculation model itself possesses "learning" and "adaptation" capabilities: when the system detects that a large number of robots are running low on power, it can automatically increase α to prioritize energy security; when it detects a high density of high-priority tasks, it dynamically increases β to ensure operational continuity. Thus, dynamic priority calculation evolves from a "good formula" with fixed parameters into an "intelligent kernel" capable of self-optimization and self-evolution based on system operating conditions. This represents a leap from single-point optimization to system-level closed-loop adaptation in scheduling strategies, fundamentally improving decision-making quality and system robustness in complex dynamic environments.

[0043] Further steps for dynamically adjusting the weighting coefficients α(t), β(t), and γ(t) include: Step S61: Monitor and calculate the system state vector X(t) at the current time t; Step S62: Calculate the change in system state from the previous adjustment time to the current time, dX = X(t) - X(t-Δt); Step S63: Based on the preset state-weight sensitivity matrix A and the state change amount dX, predict the weight adjustment direction ΔW_pred = η·(A·dX), where η is the learning rate; Step S64: Project the adjusted weight vector W_temp = [α(t-Δt), β(t-Δt), γ(t-Δt)] + ΔW_pred onto a constraint space that satisfies α+β+γ=1 and all coefficients are non-negative, to obtain the updated weighting coefficients.

[0044] Specifically, this embodiment constructs a complete and mathematically verifiable closed-loop control process. The core of this process is: by calculating the change (dX) of the system state vector between adjacent time steps, and using a pre-trained state-weight sensitivity matrix (A) to map it to the weight adjustment direction (ΔW_pred), and then using a projection operator to ensure that the adjusted weight coefficients always meet the normalization and non-negativity constraints.

[0045] The series of steps in this embodiment transforms the high-level intent of "adjusting weights according to the state" into a deterministic mathematical process based on gradient feedback and constraint optimization. This allows the system to move away from reliance on experience or blind trial and error, and instead self-adjust in a predictable and interpretable manner based on the real-time "trend" of system state changes. This ensures the stability and convergence of the adaptive process, which is the key algorithmic guarantee for the reliable operation of the adaptive mechanism of this invention in real-time industrial systems.

[0046] Furthermore, the state-weight sensitivity matrix A is obtained by training with historical operating data, and its element a_ij represents the partial influence strength of the j-th system state variable on the i-th weighting coefficient.

[0047] Specifically, this embodiment distinguishes itself from all schemes that use fixed matrices that are preset, guessed, or theoretically derived by limiting it to "historical running data training". It emphasizes its ability to continuously learn and iteratively optimize from actual system operation, thereby ensuring that the adaptive mechanism can continuously approach and adapt to the true optimality of specific application scenarios, and improving the exclusivity, effectiveness and convergence accuracy of the method.

[0048] Example 2 Please see Figure 2 The present invention also provides an aerial orbital charging robot management system for implementing the aerial orbital charging robot management method as described in any of the preceding claims, comprising: The status monitoring module is used to acquire and update the status data of each charging robot in real time; The priority calculation module is configured to calculate the dynamic priority evaluation value of each charging robot based on the status data using a dynamic priority evaluation function. The scheduling decision module is configured to allocate charging piles to charging requests and plan aerial track paths based on the dynamic priority evaluation value. The track control module is used to control the charging robot's movement on the aerial track according to the planned path; The charging station management module is used to manage the status and charging sequence of each charging pile.

[0049] Furthermore, it also includes a weight adaptation module, which is connected to the priority calculation module and is configured as follows: Real-time monitoring of system operating status indicators; Based on the system operation status indicators and the pre-stored state-weight mapping relationship, the weighting coefficients in the dynamic priority evaluation function are dynamically adjusted. The adjusted weighting coefficients are sent to the priority calculation.

[0050] It should be noted that the system in this embodiment mainly includes five core modules: a status monitoring module, a priority calculation module, a weight adaptive module, a scheduling decision module, and a charging station and track control interface module. Each module exchanges data via a system bus and is coordinated by a central processing unit.

[0051] Status monitoring module: Collects status data of all charging robots (hereinafter referred to as "robots") and the system environment in real time, and outputs robot status set {R_i}, track network topology and occupancy status diagram, and charging pile status set {P_m}.

[0052] Each robot periodically (e.g., every second) reports its unique ID, 3D coordinates (provided by the track positioning system), remaining battery percentage (B_i), current task priority (TP_i, levels 1-5), and task execution time via wireless communication. This module maintains a dynamic database storing the latest status of all robots. Simultaneously, this module obtains the real-time occupancy status of each track segment from the track control system and the busy / idle status and estimated idle time of each charging station (P_1, P_2, ..., P_M) from the charging station controller.

[0053] Priority Calculation Module: Based on the data from the Status Monitoring Module and the coefficients provided by the Weight Adaptation Module, calculates the dynamic priority evaluation value for each robot that submits a charging request.

[0054] Specifically, priority calculation is performed through the following steps: a. Receive a charging request, which includes the robot ID and the request time T_req.

[0055] b. Obtain the current state of robot i from the status monitoring module: pos_i, B_i, TP_i, total task time T_total_i, and executed time T_spent_i (then the remaining time T_remain_i = T_total_i - T_spent_i).

[0056] c. Obtain the weighting coefficients [α, β, γ] at the current time step from the weight adaptive module.

[0057] d. Calculate the priority evaluation value P_i according to the following formula: Calculate the energy urgency E_i: Query the map database to calculate the shortest track distance from robot i to each charging station, and take the minimum value as D_i.

[0058] Read the robot's full-charge range R_i and decay constant τ (e.g., τ=0.25 for the standard model) from the robot's performance table.

[0059] Set the safe power threshold B_min = 0.15 (15%).

[0060] Calculation: E_i = exp[-(B_i - B_min) / (τ (1 - D_i / R_i))].

[0061] Computation task criticality T_i: T_i = (TP_i / 5) [1 + (T_remain_i / T_total_i)].

[0062] Computational space cost S_i: The scheduling decision module (see below) needs to simulate a path to candidate charging station m for robot i in advance and evaluate the path congestion C_i (which is the weighted average of the current occupancy rate to capacity ratio of each segment on the path, and C_max is taken as an empirical value of 1.5).

[0063] Based on the current charging queue, estimate the expected waiting time W_i for robot i to go to charging station m.

[0064] Define the sigmoid function σ(x) = 1 / (1 + exp(-(x-μ) / σ_s)), where in this embodiment we take μ = 300 seconds and σ_s = 100 seconds.

[0065] Calculate: S_i = σ(W_i) (C_i / C_max).

[0066] The final dynamic priority evaluation value is calculated as: P_i = α E_i + β T_i + γ S_i.

[0067] Weight Adaptive Module: Based on the system's macroscopic state, the weighting coefficients α, β, and γ in the priority calculation module are dynamically adjusted. This is achieved through the following steps: a. Status monitoring: every Calculate the system state vector X(t) = [x1, x2, x3, x4] once every t time (e.g., 5 minutes).

[0068] x1: Number of robots with remaining battery power B_i < 0.2 / Total number of robots.

[0069] x2: Number of robots with task priority TP_i <= 2 / Total number of robots currently executing tasks.

[0070] x3: Number of track segments with a current occupancy rate exceeding 80% / Total number of track segments.

[0071] x4: Variance of utilization rate of all charging stations (utilization rate = peak time / total time).

[0072] b. Coefficient adjustment algorithm (every...) (t executes once): Read the weight W(t-) from the previous period t) = [α_old, β_old, γ_old] and state X(t- t).

[0073] Calculate the change in state: dX = X(t) - X(t-) t).

[0074] Read the preset 3x4 state-weight sensitivity matrix A from memory. This matrix can be obtained through offline regression analysis of historical system data. Initialization example: A = [[ 0.6, 0.0, -0.1, 0.0], # α row: highly positively sensitive to x1 (urgent request) [ 0.0,0.7, 0.0, 0.0], # β row: highly positively sensitive to x2 (high priority task) [-0.2, 0.0, 0.8, -0.1]] # γ row: highly positively sensitive to x3 (congestion), negatively sensitive to x1 Calculate the prediction adjustment: ΔW_pred = η (A · dX), where · represents matrix multiplication and η is the learning rate (initial value 0.3).

[0075] Calculate the temporary weights: W_temp = W(t- t) + ΔW_pred.

[0076] Projection into the constrained space: Since the weights must satisfy α+β+γ=1 and all are non-negative, the following projection algorithm is used: 1. Set all negative values ​​in W_temp to 0.

[0077] 2. Calculate the sum sum = α_temp + β_temp + γ_temp.

[0078] 3. If sum != 1, then let [α_new, β_new, γ_new] = [α_temp, β_temp, γ_temp] / sum.

[0079] The new weights W(t) = [α_new, β_new, γ_new] obtained after projection are sent to the priority calculation module.

[0080] Scheduling decision module: Based on dynamic priority evaluation values, it makes final decisions on charging pile allocation, route planning, and time window scheduling. Specifically, the scheduling decision is implemented through the following steps: a. Generation of candidate charging stations: For robot i that makes a request, select charging stations with a status of "idle" or "about to be idle".

[0081] b. Evaluation and Decision-Making: For each candidate charging station m: Call the path planning submodule to plan a conflict-free spatiotemporal path from pos_i to pos_m, and estimate the arrival time T_arrive and path congestion C_i.

[0082] Based on the job queue of charging pile m, calculate the estimated start charging time T_start and waiting time W_i = T_start - T_arrive after inserting the request.

[0083] W_i and C_i are fed back to the priority calculation module to calculate S_i and the final P_i,m for the charging pile.

[0084] Calculate the overall cost of selecting this charging station: Cost_{i,m} = (1 - P_i,m) + ω PathLength (ω is the path weight coefficient).

[0085] c. Decision and allocation: Select the charging pile m* that minimizes Cost_{i,m} as the target.

[0086] d. Allocate a flexible time window: Set ΔT_max = 1800 seconds. The charging time window allocated for this request is: [T_req, T_req + ΔT_max (1 - P_i,m*)]. The actual start time of charging required in the scheduling instruction must not be later than the end time of this window.

[0087] e. Send the final decision (robot ID, target charging station ID, planned path, and reservation time window) to the track control and charging station control interface module.

[0088] The charging station and track control interface module converts scheduling decisions into executable instructions for the underlying control system and monitors the execution status. Specifically, it decomposes the path into a series of movement instructions sent to the robot drive controller; sends charging reservations to the charging station controller, locking in the time slot for the corresponding charging pile. Simultaneously, it monitors the instruction execution status in real time, and immediately alerts the scheduling decision module to trigger rescheduling in case of any anomalies (such as robot malfunction or track blockage).

[0089] In a specific test implementation, a typical running scenario is shown below: Assume that at time t0, there are 20 robots running in the system, among which: Robot A: battery B_A=18%, executing a task with priority 2, which is nearing completion (T_remain_A / T_total_A = 0.1).

[0090] The system state X(t0) is calculated as [0.1, 0.3, 0.4, 0.05], indicating that the system is generally stable with slight congestion.

[0091] The current weighting coefficients are [α=0.5, β=0.3, γ=0.2].

[0092] At this time, robot A sends a charging request: the priority calculation module obtains its status and calculates E_A (because its power is low, the E_A value is high), T_A (because the task is about to end, the T_A value is high), and S_A.

[0093] Using the current weights, calculate P_A = 0.5*E_A + 0.3*T_A + 0.2*S_A to obtain a higher evaluation value, such as 0.85.

[0094] The scheduling decision module allocates a narrow elastic time window [t0, t0+270 seconds] (ΔT_A = 1800*(1-0.85)=270) and plans a fast path to the nearest charging station that will soon be available.

[0095] Suppose that a few minutes later, the system suddenly issues multiple low battery alarms (x1 rises sharply), and the weighted adaptive module detects that dx1 in the state vector change dX is significantly positive.

[0096] Based on matrix A (the sensitivity of the first row to x1 is +0.6), the module calculates that Δα is positive, and after projection, new weights are obtained, such as [α=0.7, β=0.2, γ=0.1].

[0097] Subsequently, for new charging requests, the weight of the energy urgency E_i is automatically increased, and the overall system strategy shifts towards "ensuring energy security" until the emergency is alleviated.

[0098] It should be noted that those skilled in the art can adjust specific parameters (such as the safety threshold B_min, time constant τ, initial value of matrix A, and adjustment period) according to the actual scenario. (e.g., t); Through the above implementation methods, the present invention successfully combines dynamic priority calculation with adaptive weight adjustment, realizing efficient, reliable, and adaptive scheduling management of the aerial track-type charging robot system.

[0099] The above are only some embodiments of this application and do not limit the patent scope of this application. All equivalent structural transformations made under the technical concept of this application and using the contents of the specification and drawings of this application, or direct / indirect applications in other related technical fields, are included in the patent protection scope of this application.

Claims

1. A management method for an aerial track-mounted charging robot, characterized in that, include: Step S1: Acquire the status data of each charging robot in real time. The status data includes location coordinates, remaining battery percentage, task priority, and estimated energy consumption. Step S2: Based on the state data, calculate the dynamic priority evaluation value for each charging robot. The dynamic priority evaluation value is obtained by weighted summation, and its calculation formula is: P_i = α·E_i + β·T_i + γ·S_i. Where P_i is the dynamic priority evaluation value of charging robot i, E_i is the energy urgency component, T_i is the task criticality component, S_i is the space cost component, and α, β, γ are weighting coefficients and satisfy α+β+γ=1. Step S3: Based on the dynamic priority evaluation value, allocate charging resources to each charging robot that makes a charging request and plan an aerial track path to the charging station. The charging resources include charging piles and charging time periods.

2. The management method for an aerial track-mounted charging robot according to claim 1, characterized in that, The formula for calculating the energy urgency component E_i is: E_i = exp[-(B_i - B_min) / (τ·(1 - D_i / R_i))]; Where B_i is the current remaining power percentage of charging robot i, B_min is the preset safe power threshold, τ is the energy decay time constant related to the charging robot model, D_i is the aerial track distance from the current position of charging robot i to the nearest charging station, and R_i is the range of charging robot i when fully charged.

3. The management method for an aerial track-mounted charging robot according to claim 1, characterized in that, The formula for calculating the critical component T_i of the task is: T_i = (TP_i / TP_max) · [1 + (T_remain_i / T_total_i)]; Where TP_i is the current task priority value of charging robot i, TP_max is the maximum value of the task priority value, T_remain_i is the remaining execution time of the current task of charging robot i, and T_total_i is the total budget time of the current task of charging robot i.

4. The management method for an aerial track-mounted charging robot according to claim 1, characterized in that, The spatial cost component S_i is calculated as follows: S_i = σ(W_i) · (C_i / C_max); Where σ(·) is an S-shaped function of waiting time, W_i is the expected waiting time of charging robot i in the current charging queue, C_i is the real-time congestion degree on the planned path of charging robot i to the target charging station, and C_max is the maximum congestion degree.

5. The management method for an aerial track-mounted charging robot according to claim 1, characterized in that, In step S3, the method for allocating a charging time period for the charging request is as follows: Based on the dynamic priority evaluation value P_i, an elastic time window [Tre, Tre+ΔT_i] is allocated, where Tre is the time when the charging request is issued, ΔT_i is the duration of the elastic window and ΔT_i = ΔT_max · (1 - P_i), and ΔT_max is the preset maximum elastic window duration.

6. The management method for an aerial track-mounted charging robot according to claim 1, characterized in that, The weighting coefficients α, β, and γ are non-fixed values ​​and are adaptive variables that are dynamically adjusted according to the real-time operating status of the system. The adjustment is based on the current system state vector X=(x1, x2, x3, x4), where x1 is the proportion of emergency charging requests, x2 is the proportion of high-priority tasks, x3 is the system track congestion index, and x4 is the variance of charging pile utilization.

7. The management method for an aerial track-mounted charging robot according to claim 6, characterized in that, The steps for dynamically adjusting the weighting coefficients α(t), β(t), and γ(t) include: Step S61: Monitor and calculate the system state vector X(t) at the current time t; Step S62: Calculate the change in system state from the previous adjustment time to the current time, dX = X(t) - X(t-Δt); Step S63: Based on the preset state-weight sensitivity matrix A and the state change amount dX, predict the weight adjustment direction ΔW_pred = η·(A·dX), where η is the learning rate; Step S64: Project the adjusted weight vector W_temp = [α(t-Δt), β(t-Δt), γ(t-Δt)] + ΔW_pred onto a constraint space that satisfies α+β+γ=1 and all coefficients are non-negative, to obtain the updated weighting coefficients.

8. A management method for an aerial track-mounted charging robot according to claim 7, characterized in that, The state-weight sensitivity matrix A is obtained by training with historical operating data, and its element a_ij represents the partial influence strength of the j-th system state variable on the i-th weighting coefficient.

9. A management system for an aerial track-mounted charging robot, used to implement the management method for an aerial track-mounted charging robot as described in any one of claims 1-8, characterized in that, include: The status monitoring module is used to acquire and update the status data of each charging robot in real time; The priority calculation module is configured to calculate the dynamic priority evaluation value of each charging robot based on the status data using a dynamic priority evaluation function. The scheduling decision module is configured to allocate charging piles to charging requests and plan aerial track paths based on the dynamic priority evaluation value. The track control module is used to control the charging robot's movement on the aerial track according to the planned path; The charging station management module is used to manage the status and charging sequence of each charging pile.

10. The system according to claim 9, characterized in that, It also includes a weight adaptation module, which is connected to the priority calculation module and is configured as follows: Real-time monitoring of system operating status indicators; Based on the system operation status indicators and the pre-stored state-weight mapping relationship, the weighting coefficients in the dynamic priority evaluation function are dynamically adjusted. The adjusted weighting coefficients are sent to the priority calculation.

Images