A resource scheduling method of a special energy supplement mode communication network

Abstract

The application discloses a resource scheduling method of a special energy supplement mode communication network, and comprises the following steps: step 1, constructing an environment and energy consumption perception model, quantifying the illumination and node power consumption in an extreme environment, and providing bottom layer perception data input for the whole network; step 2, based on the perception data, independently calculating the energy safety budget that will never touch the survival bottom line by each node in the communication network, decoupling the long-term survival hard constraint into a single cycle energy boundary; step 3, taking the energy boundary of each node as a hard constraint, and completing the global task and energy collaborative planning and allocation by the center node in a long cycle scale; and step 4, based on the planning instruction of the collaborative planning and allocation in step 3, guiding each communication node to perform adaptive resource allocation and action execution according to real-time channel and energy fluctuation in a micro time slot. The application improves the network communication utility and reduces the node failure risk under the condition of a long cycle without energy supplement window.

Description

Technical Field

[0001] This invention belongs to the field of wireless communication technology, specifically relating to a resource scheduling method for a special power replenishment mode communication network. Background Technology

[0002] In extreme application scenarios such as lunar exploration, polar scientific expeditions, deep space exploration, and high-latitude regional monitoring, communication networks generally exhibit intermittent, random, and fragmented characteristics in their energy replenishment. Such networks, which rely on renewable energy sources such as solar power, can be categorized as "special energy replenishment mode communication networks."

[0003] Due to extreme environments such as extremely low temperatures and long periods without recharging windows, network nodes need to continuously consume rigid survival power (e.g., thermal control systems, minimum operating power of core circuits) to maintain basic operation, and this power consumption must be prioritized. Once the energy is exhausted and the node can no longer sustain its survival power consumption, it will permanently fail, leaving network nodes in a tight balance of "intermittent recharge and continuous consumption" for a long time. Traditional resource scheduling methods that focus solely on throughput or energy efficiency are prone to overspending energy in pursuit of short-term communication performance, making it difficult to balance node survival and communication efficiency. Therefore, how to achieve efficient communication resource scheduling while ensuring the continuous survival of nodes has become a core technical challenge in this field.

[0004] To address the aforementioned issues, a representative category of existing research has laid the foundation for modeling optimal transmission strategies in energy harvesting communication systems. For example, Ozel et al. studied the optimal power control problem for energy harvesting nodes in fading channels, providing an offline optimal strategy and discussing online strategy modeling paths under energy causality and battery capacity constraints. However, this type of research mainly focuses on physical layer power sequence optimization for single-link / point-to-point networks, with objectives primarily aimed at maximizing throughput or minimizing completion time within the deadline. It typically does not involve the coordinated allocation of network layer and MAC layer resources such as service arrival, routing / forwarding, and bandwidth access in multi-node networks, and it fails to treat the rigid survival constraints of "rigid power consumption guarantee" and "permanent failure when energy falls below a threshold" as core hard constraints for cross-layer scheduling. Therefore, it is difficult to directly apply to extreme scenarios where long-term power replenishment interruptions occur and nodes must prioritize ensuring survival power consumption.

[0005] Another representative type of work focuses on state updates in energy harvesting sensors / communications, studying online update strategies under energy arrival fluctuations. For example, Arafa et al. provided a framework for online strategy design and performance analysis for the timely state update problem of energy harvesting sensors. However, the core of this type of method lies in the optimization of the application layer / information freshness index, usually limiting the system control variables to decisions such as "whether to send an update packet / when to send it," and paying insufficient attention to the joint optimization of multi-dimensional resources (power, bandwidth, access time slots, forwarding relays, etc.) across time scales in the communication network; moreover, its constraints are mostly energy causal or average performance constraints, which differ from the hard constraints of "must survive for a long time and rigid survival power consumption must be prioritized" in extreme environments. Therefore, it is also difficult to solve the core contradiction of "coupling long-term survival guarantee and short-term service scheduling" in communication networks with special power replenishment modes.

[0006] In summary, for energy-constrained communication networks with long-cycle, discontinuous power replenishment characteristics, existing technologies generally suffer from the following shortcomings in engineering implementation: First, existing resource scheduling and energy-saving optimization methods often focus on indicators such as throughput, energy efficiency, or latency, and often fail to use the requirement that node power levels must not fall below the survival threshold as a hard constraint throughout the planning and execution process. This leads to the risk of permanent failure of nodes due to insufficient power consumption to guarantee survival in scenarios such as power replenishment deviations, sudden service disruptions, or long periods without power replenishment windows. Second, existing solutions often adopt a relatively fragmented approach to "long-term planning" and "short-term scheduling" in terms of time scale. Long-term planning focuses on energy balancing or periodic strategies, while short-term scheduling focuses on the real-time channel and service queue status. There is a lack of an executable connection mechanism to stably transfer long-term energy safety boundaries to short-term resource allocation actions, making it difficult to simultaneously consider survival reliability and communication effectiveness in dynamic environments. Third, centralized real-time control usually requires high network-wide status interaction and computational costs. Purely distributed strategies, lacking global power replenishment prediction information, cannot coordinate the survival safety boundaries of each node, making it difficult for nodes to reach a unified survival guarantee standard and resource usage rules.

[0007] Therefore, there is an urgent need for a collaborative resource scheduling method that clarifies the overall network survival and security boundaries through centralized planning on a large time scale, allows nodes to flexibly adapt and execute within the boundaries on a small time scale, and dynamically adjusts the planning through deviation feedback, in order to improve the long-term operational reliability and communication efficiency per unit energy of such networks. Summary of the Invention

[0008] To overcome the shortcomings of the existing technology, the present invention aims to provide a resource scheduling method for a special power replenishment mode communication network. This method takes node survival guarantee as a hard constraint, generates an activity energy consumption budget based on power replenishment prediction and energy evolution model, completes the collaborative planning of tasks and link resources under budget constraints, and achieves rolling closed-loop optimization through distributed execution and event-triggered feedback, thereby improving network communication efficiency and reducing node failure risk under long-term power replenishment window conditions.

[0009] To achieve the above objectives, the technical solution adopted by the present invention is as follows: A resource scheduling method for a special power replenishment mode communication network includes the following steps; Step 1: Construct an environment and energy consumption perception model, and provide underlying perception data input for the entire network by quantifying the illumination and node power consumption in extreme environments; Step 2: Distributed energy budget calculation and reporting. Based on the sensed data, each node in the communication network independently calculates the energy security budget that will never trigger the survival bottom line, decoupling the long-term hard constraint of survival into a single-cycle energy boundary. Step 3: Large-scale task and energy coordination planning and scheduling, using the energy boundaries of each node as hard constraints, and the central node to complete the overall coordination and allocation of global tasks and energy on a long-term scale; Step 4: Small Time Scale Resource Scheduling and Execution Phase. Based on the planning instructions in Step 3, each communication node is guided to adaptively allocate resources and execute actions within a micro time slot according to real-time channel and energy fluctuations.

[0010] The four steps described above, tailored to the characteristics of special power replenishment modes, form a closed-loop resource scheduling framework in the communication network that is progressive and mutually supportive of perception, budgeting, planning, and execution. These four steps are interconnected, jointly ensuring the long-term absolute survival and efficient operation of the communication network under special power replenishment conditions.

[0011] Step 1 specifically involves: Step 1.1. Construct a mixed lighting environment model: To address the coexistence of long-period light-dark cycles and random terrain occlusion at the lunar south pole, a hybrid illumination intensity model combining deterministic trends and random perturbations is established. : Among them, the deterministic component Simulate the trend of solar altitude angle variation around the lunar south pole in approximately 28 days. Angular frequency; random occlusion component This is an indicator function; it takes the value of 1 when physical devices in the communication network, such as probes, relay base stations, etc. (hereinafter referred to as nodes), are obscured by local craters or terrain shadows, and otherwise takes the value of 0; this variable follows a Bernoulli distribution based on terrain complexity; Based on this, the output power of the energy harvesting component, namely the solar panel... for: in, For photoelectric conversion efficiency, This refers to the effective area of ​​the windsurfing board; Step 1.2: Construct a power consumption model; Suppose the network contains a set of nodes. This set represents the collection of all devices in the communication network, among which base station-type devices with global information collection and scheduling computing power are denoted as central nodes. The remaining terminal devices responsible for data collection and relay forwarding are referred to as ordinary nodes. The plan adopts a two-level time scale: the larger time scale planning period is [length missing]. Further divided into each planning cycle Hourly timescale time slots; small timescale time slot length is ; node The energy at the beginning of the planning period is The upper limit of energy is Survival threshold is In the time slot Within, the predicted energy replenishment for each node is: It can be obtained from illumination prediction or energy harvesting prediction; the survival power consumption is For example, thermal control, minimum operating temperature of core circuits, etc.; then the survival energy consumption is: Nodes in time slots Energy consumption related to communication and processing, i.e., active energy consumption. The timeframe is determined by the actual timing of data generation, reception, and local processing. Let's define a node. In the time slot Within that timeframe, the cumulative time allocated for sending, receiving, and processing actions are respectively , and The corresponding hardware operating power is respectively , and Ignoring the minute dormant leakage current, the time slot... The total active energy consumption is expressed as: The sum of the cumulative times of each action must satisfy the time slot length constraint: The evolution of battery energy is as follows: ; And it satisfies: ; Step 1.3: Channel fading model; Considering path loss and shadow fading caused by irregular lunar terrain, nodes To the node Channel gain; in, For communication distance, This is the path loss index. The shadowing fading factor follows a log-normal distribution. It is the transmit / receive antenna gain. For carrier frequency; Step 1.4: Calculation of survival safety factor and status determination; Step (1): Real-time acquisition of the battery's current state of charge. And obtain the estimated duration of the next effective energy replenishment in cyclotron mechanics from the energy replenishment prediction sequence. ; Step (2): Calculate the survival safety factor This coefficient represents the energy surplus of a node after deducting its basic survival needs. Step (3): According to The value categorizes nodes into three survival states to guide access control on a small timescale; if The energy is sufficient, indicating a safe state, and full service access is permitted; if Energy is low, indicating a state of alert; only high-priority services are allowed access. Survival is threatened, a critical state is declared, and the communication radio frequency module is forcibly shut down, maintaining only [operational function]. .

[0012] Step 2 specifically involves: Step (1): Local information acquisition and initialization; at the beginning of each large-scale planning cycle, the node Obtain and update its local energy state and prediction parameters: initial energy The upper limit of energy is Survival threshold is Survival power consumption is ,future Energy replenishment sequence for each time slot node and operating mode power consumption parameters, transmission power Received power Processing power Sleep power ; Step (2): Generate candidate active energy consumption budget sequence; the node first generates an initial candidate active energy consumption budget upper limit sequence. This sequence represents the energy estimate of each time slot available for non-survivability tasks before the node's survivability is verified; initialization can be achieved using a simple heuristic rule that makes it proportional to the predicted supplemental energy: in, This is a preset conservative coefficient used to reserve a safety margin for energy replenishment prediction errors; For time slots The predicted energy replenishment; this step provides an iterative starting point for subsequent feasibility verification and repair; Step (3): Forward extrapolation and default verification based on the energy evolution model; the node will generate candidate active energy consumption budget sequences. As the active energy consumption in the model defined in step 1.2 Input, combined with survival energy consumption Perform full-cycle forward simulation; Specifically, substituting the above parameters into the energy evolution equation: The predicted energy trajectory is obtained through recursive calculation. ; Subsequently, it is verified whether the survival hard constraints are satisfied throughout the entire lifecycle: If the constraint holds for all time slots, the current candidate budget sequence is feasible, and the process jumps to step (5) for output; if there is a first time slot that causes the constraint to be violated... If the default occurs, the adaptive repair process in step (4) is initiated. Step (4): A spatiotemporally aware adaptive default repair mechanism, when in a time slot Upon detection of a breach, a three-stage progressive strategy will be employed to rectify the situation. Step (5): Feasible budget output and reporting; Finally, the node obtains a rigorously validated and feasible active energy consumption budget sequence and calculates its sum; the node then outputs this sequence and sum. In addition, power consumption parameters are reported to the central node to complete the distributed budget generation task.

[0013] The three stages in step (4) are as follows: (1) First stage: Local backtracking and fine-tuning: Checking the breach of contract time slot The previous time slot Budget value If it is significantly higher than the average of neighboring time slots, such as being greater than the historical average within a sliding window; If the peak value is multiples of the previous value, it is determined to be a peak value that can be reduced. A portion of the energy is drawn from this peak. Transfer to the default time slot The aim is to resolve early defaults through local rebalancing; after this adjustment is completed, immediately return to step (3) to re-perform the forward deduction verification; (2) Second stage: Elastic weight contraction: If the first phase is not applicable or fails to eliminate the default, the following will be triggered: For each time slot within the default segment Assign a shrinkage weight The weighting design principle is that the less energy is added, the more contraction it bears, for example: Calculate a baseline contraction factor based on the total energy deficit. Then, differentiated adjustments are made to the budget for the defaulted segments: After this operation, return to step (3) to re-verify.

[0014] (3) Third stage: Proactive deep repair: 1) Set a higher recovery target; set The goal of time slot repair is ,in For forward-looking safety margins; 2) Reverse budgeting backtracking and enhanced compression; to achieve this goal, from The time slots will begin to be traced back, identifying and prioritizing the reduction of time slot budgets that have relatively abundant budgets and high flexibility for adjustment. The reduction amount must be sufficient to fill the current gap and achieve a safety margin. ; After completing this deep repair, return to step (3) to perform a full-cycle simulation; because The energy state of the time slot is reinforced, significantly reducing the risk of subsequent defaults; if a new default occurs in a later time slot, that time slot will be considered the new default. This step involves iteratively executing the three-stage repair process. Step 3 specifically involves: Step 3.1: Planning Inputs and Symbol Explanation: With a planning period length of Discretize it within the window. Each planning period, with the period index as follows: ; Step 3.2: Task Decomposition and Energy Consumption Calculation: To map tasks into executable instructions, this invention decomposes tasks on candidate routing paths into atomic actions such as sending, receiving, and processing at nodes along the route, and calculates energy consumption based on the power consumption parameters in step 2. Step 3.3: Task and Energy Joint Planning Model under Hard Constraints of Energy Budget The central node solves within the planning window, selects a path for each task, and determines the amount to be completed in different time periods, so as to maximize the benefits of task completion without exceeding the node's activity energy consumption budget. Step 3.4: Solve for fluency and output instruction format: This invention uses a priority-driven iterative solution strategy of feasible allocation and budget deduction update to generate feasible solutions. Its output naturally satisfies the above hard constraints and is easy to be issued as a node-parseable instruction packet. This instruction package contains the following three core components: (1) Energy budget constraint: The sequence of safe activity energy consumption budgets available for each time slot within the period confirmed by the center at this node. and total periodic budget This section represents the rigid performance limit for all communication activities of the node. (2) Atomic action instruction set: The instruction list of this node. Each instruction in the list; The specific time slots of the nodes are clearly defined. The atomic actions to be performed, the communication peer, the amount of data, and the reference rate; this is the direct executable result after task decomposition.

[0015] In step 4, each node executes the process autonomously and in a distributed manner within each small timescale scheduling slot; its core objective is to strictly follow the scheduling and energy management instruction packages issued by the central node. Under real-time channel changes and energy fluctuations, atomic instructions are reliably and efficiently transformed into communication actions through local fine-tuning and neighbor collaboration, while ensuring strict adherence to the hard energy budget constraint throughout the entire process.

[0016] Step 4.1: Instruction packet parsing and local state initialization: node First, analyze its instruction package. and initialize the local scheduling state; Energy Budget Loading: Reads the communication energy budget for this time slot. and total periodic budget Initializing this time slot has already consumed energy. ; Instruction queue construction: Extracting time slot labels from instruction packets All atomic action instructions constitute the local time slot instruction queue. Each instruction in the queue ; Survival status determination: Calculate the local survival safety factor in real time based on the model in step 1.4. And mapped to survival level ; Step 4.2: Real-time power budget constraint calculation: Resource allocation primarily targets transmit power and wireless spectrum; each node operates at each scheduling moment within each time slot. The remaining energy budget needs to be converted into an immediately available upper limit for transmit power; Calculate the upper limit of instantaneous transmit power; in, This is the maximum transmit power supported by the hardware. For time slots The remaining time; this constraint ensures that the expected power consumption at any time does not exceed the budget before the end of the time slot; Step 4.3: Survival Awareness Instruction Queue Management: To cope with real-time energy fluctuations, nodes preprocess the instruction queue at the beginning of each small scheduling cycle based on their survival level. like Clear All communication commands that are not essential for survival are retained, with only the minimum state maintenance commands remaining. like : Enable priority-aware random discarding; traverse Low-priority instructions are discarded or deferred with a probability inversely proportional to their priority. like The entire instruction queue participates in this round of scheduling; Step 4.4: Channel-aware joint resource allocation: Node in power budget Under constraints, the transmission resources are allocated in the most energy-efficient way to execute commands that need to be transmitted wirelessly; the resources for receiving and processing actions are represented by the allocated execution time slices. Step (1): Water injection power allocation. The system bandwidth is divided into... There are several sub-channels; for the command to be sent, its target node is... In the sub-channel The real-time channel gain on is The noise power is ; Node-based solution for optimal transmit power allocation : Its classic interpretation is ,in Water injection line to meet total power constraints; Step (2): Instruction-driven time-frequency resource mapping; Based on the power allocation results and instruction attributes, sub-channels and time resources are allocated to each instruction to be executed; Step 4.5: Task Execution and Node Collaboration: Step (1): Execution and Energy Consumption Recording. Nodes execute instructions according to the resource allocation results, and then accurately measure the actual energy consumed in this scheduling. And update the used energy: At the same time, update the remaining data amount of the corresponding instruction in the instruction queue. ; Step (2): Event-triggered lightweight coordination beacon; to achieve distributed synchronization and anomaly avoidance across nodes, the node immediately broadcasts this beacon when any of the following events occur: Step 4.6: Complete the judgment and trigger the exception feedback: Step (1): Intra-slot scheduling loop; repeat steps 4.2 to 4.5 until any of the following conditions are met, marking the end of the current small-scale scheduling cycle: (1) Time slot The clock time has ended; (2) Energy budget for this time slot Exhausted; (3) Instruction queue All instructions have been completed or processed according to the rules; Step (2): Anomaly monitoring and reporting; at the end of the time slot, the node performs final calculation; (1) Calculate energy consumption deviation: ; (2) Verify instruction completion rate: Statistically analyze instructions that were not completed or downgraded, and their priorities; if Exceeding the threshold If a high-priority instruction is not completed, or if the survival level changes unexpectedly, the node will package the above abnormal information and report it to the central node, thereby triggering the global replanning process.

[0017] The specific calculation process for step (2) of step 4.4 is as follows: (1) Calculate instruction performance weights; for instructions Priority If it is a sending command, its weight If it is a receiving or processing instruction, its weight is priority. ; (2) Allocate sub-channels; Each sub-channel is weighted according to the transmission command. The data is divided proportionally for parallel transmission; (3) Mapping between time slices and actual energy consumption. Within the current scheduling time slot, nodes process instructions in the queue sequentially. For a single instruction... Based on its actual allocated bandwidth and power, the actual achievable rate of the physical layer is calculated. The exact execution time slice required for this instruction is... .like If the remaining time in the current time slot is greater than the current time slot's remaining time, execution will be truncated, and the untransmitted data will be retained in the queue; if execution is allowed within the remaining time, the actual energy consumption increment will be... The specific calculation process for step (2) of step 4.5 is as follows: (1) To begin executing or complete an instruction involving neighbor cooperation; (2) Survival level Changes have occurred; (3) A high-priority instruction is continuously blocked due to insufficient local resources or extremely poor channel; Therefore, the beacon format can be described as a quadruple: in, As a unique identifier for the instruction, ; (4) Neighboring nodes Coordination strategy upon receiving a beacon: If the beacon indicates a neighbor If it enters a critical state, then Cancel or suspend all planned trips immediately Send data commands to avoid invalid transmissions; if the beacon indicates a neighbor... If a send command has already been executed and the sender is the receiver, then Prioritize resource allocation for receiving; if a beacon indicates that the predecessor instruction of a multi-hop task has been completed, the neighboring node, as the successor node, can start resource preparation in advance to achieve pipeline-style advancement.

[0018] The beneficial effects of this invention are: (1) Reliable task scheduling under hard constraints of node survival energy was achieved: By combining "distributed energy budget generation" with "centralized task energy collaborative planning", the rigid constraint that "node power must not be lower than the survival threshold" is given the highest priority and is implemented throughout the entire scheduling process for the first time. This fundamentally eliminates the risk of permanent node failure due to the pursuit of communication performance, and provides a fundamental guarantee for the long-term survival of the intermittent power replenishment network.

[0019] (2) Achieving synergistic optimization of global energy efficiency and task utility: Under the condition that the energy safety boundaries of each node are known, the central node selects paths and allocates resources across nodes with the task completion as the guide, so that the communication actions actively adapt to the energy situation of the entire network. Thus, while ensuring survival, it significantly improves the effective data delivery per unit of energy, especially ensuring the completion rate of high-priority tasks.

[0020] (3) A robust closed-loop mechanism of "centralized planning and distributed collaborative execution" has been constructed: The architecture employs a centralized planning approach on a large timescale to generate deterministic instructions, a distributed and autonomous scheduling approach on a small timescale, and lightweight collaboration. This design not only solves the problem of global resource coupling optimization through centralized planning, but also absorbs uncertainties such as real-time channel fluctuations through local execution and collaboration mechanisms. This reduces the system's dependence on centralized control and communication overhead, and enhances engineering feasibility and environmental adaptability. Attached Figure Description

[0021] Figure 1 This is a diagram showing the overall system framework and core stage interaction of the present invention.

[0022] Figure 2 This is a flowchart of the node energy sensing and distributed budget generation process of the present invention.

[0023] Figure 3 This is a flowchart of the large-timescale survival planning layer of the present invention.

[0024] Figure 4 This is a flowchart of the online decision-making process for the small-timescale resource scheduling layer of this invention. Detailed Implementation

[0025] The present invention will now be described in further detail with reference to the accompanying drawings.

[0026] To facilitate understanding and quantitative verification, this embodiment uses the most representative "lunar south polar region communication network" as a specific example. In this scenario, communication nodes face a polar night lasting approximately 14 days and complex terrain obstruction, exhibiting typical long-period discontinuous power replenishment characteristics, which fully demonstrates the core innovation and technical advantages of this invention.

[0027] This invention discloses a resource scheduling method for a communication network with a special power replenishment mode, the implementation process of which is as follows: Figure 1 As shown, the specific steps include: Step 1: Environmental and Energy Consumption Sensing Model: Combination Figure 2 The flowchart shown illustrates the node energy sensing and distributed budget generation process. The core of this stage and the subsequent second stage lies in the node completing a closed-loop assessment of the energy security budget through local sensing. The specific process is as follows: First, the node obtains the local initial energy and predicted energy replenishment sequence, then calculates the survival safety factor and defines the survival state; subsequently, it initializes the candidate active energy consumption budget sequence and performs a full-cycle forward extrapolation based on the energy evolution model; during the extrapolation process, if a survival threshold default is triggered, a three-stage adaptive default repair mechanism (including three sub-processes: local backtracking fine-tuning, elastic weight shrinkage, and forward-looking deep repair) is triggered for iteration until no default occurs, and finally, a safe and feasible activity energy consumption budget is output and reported.

[0028] 1.1: Mixed lighting environment model: To address the coexistence of long-period light-dark cycles and random terrain occlusion at the lunar south pole, a hybrid illumination intensity model combining deterministic trends and random perturbations is established. : Among them, the deterministic component Simulate the trend of solar altitude angle variation around the lunar south pole in approximately 28 days. Angular frequency. Random occlusion component. This is an indicator function. It takes a value of 1 when a node is obscured by a local crater or terrain shadow, and 0 otherwise. This variable follows a Bernoulli distribution based on terrain complexity.

[0029] Based on this, the output power of the energy harvesting component, namely the solar panel... for: in, For photoelectric conversion efficiency, This represents the effective area of ​​the windsurfing board.

[0030] 1.2: Power Consumption Model: Suppose the network contains a set of nodes. The central node is denoted as Ordinary nodes are denoted as The plan adopts a two-level time scale: the larger time scale has a planning period of [length missing]. Further divided into each planning cycle Hourly timescale time slots; small timescale time slot length is .

[0031] node The energy at the beginning of the planning period is The upper limit of energy is Survival threshold is In the time slot Within, the predicted energy replenishment for each node is: This can be predicted by illumination or energy harvesting. The survival power consumption is... Such as thermal control, minimum operating temperature of core circuits, etc. Therefore, the survival energy consumption is: Nodes in time slots Energy consumption related to communication and processing, i.e., active energy consumption. The timeframe is determined by the actual timing of data generation, reception, and local processing. Let's define a node. In the time slot Within that timeframe, the cumulative time allocated for sending, receiving, and processing actions are respectively , and The corresponding hardware operating power is respectively , and Ignoring the minute dormant leakage current, the time slot... The total active energy consumption is expressed as: The evolution of battery energy is as follows: And it satisfies: .

[0032] 1.3: Channel Fading Model: Considering path loss and shadow fading caused by irregular lunar terrain, nodes To the node Channel gain; in, For communication distance, This is the path loss index. The shadowing fading factor follows a log-normal distribution. It is the transmit / receive antenna gain. For carrier frequency.

[0033] 1.4: Calculation of Survival Safety Factor and Status Determination: Step (1): Real-time acquisition of the battery's current state of charge. And obtain the estimated duration of the next effective energy replenishment in cyclotron mechanics from the energy replenishment prediction sequence. .

[0034] Step (2): Calculate the survival safety factor This coefficient characterizes the energy surplus of a node after deducting its survival rigidity requirements. Step (3): According to The value categorizes nodes into three survival states to guide access control on small timescales. If The energy is sufficient, indicating a safe state, and full service access is permitted; if Energy is low, indicating a state of alert; only high-priority services are allowed access. Survival is threatened, a critical state is declared, and the communication radio frequency module is forcibly shut down, maintaining only [operational function]. .

[0035] Steps 1.1 to 1.4 above constitute a complete node sensing system. Specifically, the illumination model in step 1.1 determines the upper limit of the node's energy input, the power consumption model in step 1.2 limits the lower limit of the node's energy consumption, and the channel model in step 1.3 determines the expected energy consumption required for communication. Step 1.4 integrates these three parameters to calculate the survival safety factor at the current moment in real time. This provides a basis for state determination and data for generating the distributed energy budget in step 2. Step 2: Distributed Energy Budget Calculation and Reporting; This phase aims to address how each node can independently calculate the energy budget available for active tasks such as communication and processing within the current large-scale planning period without triggering a survival threshold, and report this budget sequence to the central node as a hard constraint input for subsequent global task and role collaborative planning.

[0036] Step (1): Local information acquisition and initialization. At the beginning of each large-scale planning cycle, the node... Obtain and update its local energy state and prediction parameters: initial energy The upper limit of energy is Survival threshold is Survival power consumption is ,future Energy replenishment sequence for each time slot node and operating mode power consumption parameters, transmission power Received power Processing power Sleep power .

[0037] Step (2): Generate a candidate active energy consumption budget sequence. The node first generates an initial candidate active energy consumption budget upper limit sequence. This sequence characterizes the energy estimates of nodes for non-survivability tasks in each time slot before survivability verification. Initialization can employ a simple heuristic, such as proportionalizing it to the predicted supplemental energy: in, This is a preset conservative coefficient used to reserve a safety margin for energy replenishment prediction errors; For time slots The predicted energy replenishment. This step provides an iterative starting point for subsequent feasibility verification and repair.

[0038] Step (3): Forward extrapolation and default verification based on the energy evolution model. Nodes will generate candidate active energy consumption budget sequences. As the active energy consumption in the model defined in step 1.2 Input, combined with survival energy consumption A full-cycle forward simulation was performed. Specifically, the above parameters were substituted into the energy evolution equation: The predicted energy trajectory is obtained through recursive calculation. .

[0039] Subsequently, it is verified whether the survival hard constraints are satisfied throughout the entire lifecycle: If the constraint holds for all time slots, the current candidate budget sequence is feasible, and the process proceeds to step 5 for output. If there exists a first time slot where the constraint is violated... If a breach occurs, the process proceeds to step 4, the adaptive repair process.

[0040] Step (4): Spatiotemporally Aware Adaptive Default Repair Mechanism. This step is one of the core innovations of this invention, aiming to overcome the "top-heavy" or overly conservative budget sequence problem caused by simple iterative shrinkage. When in a time slot... Upon detection of a default, a three-stage progressive strategy will be employed for remediation.

[0041] (1) First stage: Local backtracking and fine-tuning Objective: To attempt to eliminate defaults with minimal changes and maintain the smoothness of the budget curve.

[0042] Operation: Check for breach of contract slots The previous time slot Budget value If it is significantly higher than the average of neighboring time slots, such as being greater than the historical average within a sliding window. If the peak value is multiples of the previous value, it is determined to be a peak value that can be reduced.

[0043] Repair: Allocate a portion of the energy from this peak. Transfer to the default time slot The aim is to address early defaults through local rebalancing. After this adjustment is completed, immediately return to step 3 to re-perform the forward derivation verification.

[0044] (2) Second stage: Elastic weight contraction Triggered if Phase 1 is not applicable or if the default is not eliminated.

[0045] Objective: To address the initial stage of a breach of contract Intelligent contraction is implemented to match the degree of contraction with the energy supply resilience of each time slot.

[0046] Operation: For each time slot within the default segment Assign a shrinkage weight The weighting design principle is that the less energy is added, the more contraction it bears. For example: Calculate a baseline contraction factor based on the total energy deficit. Then, differentiated adjustments are made to the budget for the defaulted segments: After this operation, return to step 3 to re-verify.

[0047] (3) Third stage: Proactive deep repair Objective: Not only to resolve current defaults, but also to inject a safety margin into subsequent time slots, prevent cascading defaults, and improve sequence robustness.

[0048] operate: 1) Set a higher recovery target. The goal of time slot repair is ,in This provides a forward-looking safety margin.

[0049] 2) Reverse budgeting backtracking and enhanced compression. To achieve this goal, from... The time slots are then traced back to identify and prioritize budget cuts for those with relatively ample budgets and greater flexibility. The cuts must be sufficient to fill the current budget shortfall and achieve a safety margin. .

[0050] Next step: After completing this deep repair, return to step 3 to perform a full-cycle simulation. Because... The energy state of the time slot is reinforced, significantly reducing the risk of subsequent defaults. If a new default occurs in a later time slot, that time slot will be considered the new default. This step involves iteratively executing the three-stage repair process.

[0051] The entire repair process iterates in a closed loop of deduction, verification, and iteration until the survival constraint is satisfied in all time slots.

[0052] Step (5): Feasible Budget Output and Reporting. Finally, the node obtains a rigorously validated and feasible active energy consumption budget sequence and calculates its sum. The node then outputs this sequence and sum. In addition, power consumption parameters are reported to the central node to complete the distributed budget generation task.

[0053] Step 3: Large-scale task and energy coordination planning and scheduling: Combination Figure 3 The flowchart shown below illustrates the large-scale survival planning layer workflow. This stage details how the central node utilizes the resource ledger for collision-free task allocation. At the beginning of each planning cycle, the central node first initializes the global link capacity ledger and the energy budget ledgers of each node. Then, it sorts the entire network task set in descending order according to the priority evaluation function and extracts the highest priority tasks from the queues one by one. For each task, it calculates the energy pressure index of candidate paths to select the optimal feasible path, allocates the amount of data to be transmitted according to time slots, and simultaneously and strictly deducts the corresponding resource ledger. The allocation results are encapsulated into atomic operation instructions in a unified format. After the entire network task set has been traversed and allocated, the instruction lists of each node are aggregated, a structured instruction package is generated, and sent to the corresponding nodes.

[0054] 3.1: Planning Inputs and Symbol Explanation: With a planning period length of Discretize it within the window. Each planning period, with the period index as follows: .

[0055] The central node gathers the following inputs at this stage: (1) Node status information Time-slot-level security activity energy consumption budget sequence: This is used to constrain the node in the time slot. The maximum energy consumption limit for activities within the area.

[0056] Total energy budget for safety activities: It is used to constrain the cumulative energy consumption of activities within a cycle.

[0057] Device power consumption parameters: .

[0058] (2) Network topology and link capabilities The global link information required for planning comes entirely from data preset before network deployment.

[0059] Link set Determined based on a predefined static network topology, representing the time slot. The set of available link pairs within the network.

[0060] Equivalent transmission rate This is directly taken from the pre-configured link budget table. This table is pre-calculated based on node hardware performance, communication distance, and an environmental model, representing the time slot... The average available transmission rate between internal nodes.

[0061] (3) Collection of task requirements across the entire network The central node builds a task set at the beginning of each planning cycle. Each task The unified representation is: in, This represents the source node, the node where data is generated or collected; Indicates the destination node, center, or specified receiving node; This indicates the amount of data expected to be transmitted and processed this week; Indicates task weight or priority; Indicates the deadline slot, requiring it to be no later than Time slot completed; Indicates the task type, such as data collection and reporting, forwarding and transmission, collaborative relay, and on-network processing.

[0062] Task sources may include periodic tasks generated by pre-set plan instructions, routine detection tasks generated by fixed network rules, and bursty backhaul tasks formed by buffer status requests reported by nodes through the control channel.

[0063] 3.2: Task Decomposition and Energy Consumption Calculation: To map tasks into executable instructions, this invention decomposes tasks on candidate routing paths into atomic actions such as sending, receiving, and processing at nodes along the route, and calculates energy consumption based on the power consumption parameter in step 2.

[0064] Step 1: Candidate Path Generation and Action Deployment. For each task... The central node is based on Generate several candidate path sets Any candidate path is denoted as: in, This represents the number of hops. For the path... any adjacent link In the time slot The estimated transmission time is: in, Indicates in time slot For the task The amount of data to be transmitted is a variable for subsequent decision-making (it can be completed in time slots and can be partially completed).

[0065] Step 2: Node energy consumption calculation.

[0066] In a given task ,path With time slot Arrangement volume The new energy consumption of nodes along the path is estimated separately for sending, receiving, and processing.

[0067] For the sending node : For the receiving node : For nodes that require local processing, caching, encoding, etc. : in, For the task at the node The processing time requirement can be given by the unit data processing latency coefficient or task rules.

[0068] Therefore, the task In the time slot For any node The total increase in energy consumption is: 3.3: Task and Energy Joint Planning Model under Hard Constraints of Energy Budget The central node solves the problem within the planning window, selects a path for each task, and determines the amount to be completed in different time periods, so as to maximize the benefits of task completion without exceeding the node's activity energy consumption budget.

[0069] Decision variables: ,Task In the time slot The amount of data to be completed as arranged.

[0070] Path selection variable: for task Select a path .

[0071] Objective function: That is, to maximize the weighted completion amount within the deadline slot.

[0072] Constraints: For any node Arbitrary time slot Any task satisfy For the task In the path relay node In any time slot The cumulative number of forwards does not exceed the cumulative number of receptions: in Indicates task In the link In the time slot The allocation amount.

[0073] 3.4: Solving for fluency and output instruction format: This invention uses a priority-driven iterative solution strategy of feasible allocation and budget deduction update to generate feasible solutions. Its output naturally satisfies the above hard constraints and is easy to be issued as a node-parseable instruction packet.

[0074] Step (1): Initialization. The central node establishes a budget ledger for each node.

[0075] Remaining budget for time slots: .

[0076] Remaining budget for the period: .

[0077] And initialize the instruction list of each node to be empty: .

[0078] Step (2): Iteratively assign tasks according to priority. For the task set... according to Sort and process each task in turn. .

[0079] For each candidate path Based on the current Calculate the upper limit of feasible completion for the task in each time slot. That is, the upper limit of feasible completion for the task in each time slot. Upper bound of allocatable quantity: in, For unit data volume in the path With time slot Next pair of nodes The energy consumption coefficient is directly obtained from the formula in section 3.2. Based on this, the achievable amount of the path before the deadline is obtained: and from satisfying Choose the path with the least energy pressure from the available paths. The energy pressure index is acceptable: choose The shortest path allows tasks to avoid budget-constrained nodes as much as possible, improving overall sustainability.

[0080] If a feasible path exists Then, the slots will be allocated sequentially according to the time slot before the deadline: like If the task is only partially completed within this cycle, the uncompleted portion can be postponed to the next planning cycle or downgraded according to the task rules.

[0081] For each allocated time slot With path Calculate the corresponding energy consumption and update the budget ledger: Simultaneously send instructions to the node list Add structured instruction elements. Instruction elements use standardized field descriptions for easy parsing and execution by nodes, such as: in, ; For the communication peer node; This refers to the amount of data that will be processed or transmitted in this time slot. For reference rate; Inherited task priority .

[0082] Step (3): Summarize the duty cycle parameters of each node task. After completing all task iterations, the central node summarizes the duty cycle parameters of each node. instruction list By combining and statistically analyzing the data, the activity time requirements for each time slot are obtained as follows: Sending time requirements: Reception time requirements: Processing time requirements: Based on this, node slot-level duty cycle parameters are given, which are used as task intensity boundaries when executing at lower-level small scales: Step (4): Encapsulation and Distribution of Planning Results. After the central node completes the iterative allocation of all tasks and the generation of instructions, it needs to encapsulate the planning results and distribute them to each node as the basis for their execution within the planning cycle. For each node... Generate a structured scheduling and energy management instruction package This instruction package contains the following three core components: (1) Energy budget constraint: The sequence of safe activity energy consumption budgets available for each time slot within the period confirmed by the center at this node. and total periodic budget This section represents the rigid performance limit for all communication activities of the node.

[0083] (2) Atomic action instruction set: The instruction list of this node. Each instruction in the list; The specific time slots of the nodes are clearly defined. The atomic actions to be performed, the communication peer, the amount of data, and the reference rate. This is the directly executable result after task decomposition.

[0084] Step (5): Command packet distribution. Before the start of the planning period, the central node distributes command packets from each node via the control channel. The command packet is reliably sent to the corresponding node. After receiving the complete command packet, the node must return an acknowledgment message to ensure successful synchronization of the planning information, and then enter the command-based execution phase.

[0085] Step 4: Small-scale resource scheduling and execution phase: Combination Figure 4 The online decision-making flowchart of the small-scale resource scheduling layer shown in this section illustrates the distributed autonomous closed-loop execution process of nodes within a micro-time slot. At the start of a scheduling time slot, each node first parses the instruction packet issued by the center and initializes the energy usage ledger for that time slot; then, it performs real-time filtering of the instruction queue based on survivability levels; based on the remaining energy budget of the time slot, it calculates the physical upper limit of the instantaneous maximum transmit power; subsequently, combined with the real-time channel state, it performs joint resource allocation, power truncation, and rate degradation, executes communication actions based on the allocation results, and accurately measures and backfills the actual energy consumption; during the execution loop, if an abnormal event is triggered (such as state transition or instruction blocking), a lightweight cooperative beacon is immediately broadcast to notify neighboring nodes; when the time slot ends or the budget for that time slot is exhausted, a final calculation is performed, and the abnormal information is packaged and reported to the central node. 4.1: Instruction Packet Parsing and Local State Initialization: node First, analyze its instruction package. And initialize the local scheduling state.

[0086] Energy Budget Loading: Reads the communication energy budget for this time slot. and total periodic budget Initializing this time slot has already consumed energy. .

[0087] Instruction queue construction: Extracting time slot labels from instruction packets All atomic action instructions constitute the local time slot instruction queue. Each instruction in the queue .

[0088] Survival status determination: The local survival safety factor is calculated in real time based on model 1.4. And mapped to survival level .

[0089] 4.2: Real-time power budget constraint calculation: Resource allocation primarily targets transmit power and wireless spectrum. Nodes allocate resources at each scheduling moment within each time slot. The remaining energy budget needs to be converted into an immediately available upper limit for transmit power.

[0090] Step 1: Calculate the instantaneous transmit power limit.

[0091] in, This is the maximum transmit power supported by the hardware. For time slots The remaining time. This constraint ensures that the expected power consumption at any time does not exceed the budget before the end of the time slot.

[0092] 4.3: Survival Awareness Instruction Queue Management: To cope with real-time energy fluctuations, nodes preprocess the instruction queue at the beginning of each small scheduling cycle based on their survival level.

[0093] like Clear All communication commands that are not essential for survival are retained, with only the minimum required state maintenance commands.

[0094] like : Enable priority-aware random discarding. Iterate through the process. It discards or postpones low-priority instructions with a probability inversely proportional to the instruction priority.

[0095] like The entire instruction queue participates in this round of scheduling.

[0096] 4.4: Channel-Aware Joint Resource Allocation: Node in power budget Under constraints, the transmission resources are allocated in the most energy-efficient way to execute commands that need to be transmitted wirelessly. Resources for actions such as receiving and processing are represented by allocated execution time slices.

[0097] Step 1: Water Injection Power Allocation. The system bandwidth is divided into... There are several sub-channels. For the command to be sent, its target node is... In the sub-channel The real-time channel gain on is The noise power is Node-based solution for optimal transmit power allocation : Its classic interpretation is ,in Water injection line designed to meet total power constraints.

[0098] Step 2: Instruction-driven time-frequency resource mapping. Based on the power allocation results and instruction attributes, sub-channels and time resources are allocated to each instruction to be executed.

[0099] (1) Calculate instruction performance weights. For instructions... Priority If it is a sending command, its weight If it is a receiving or processing instruction, its weight is priority. .

[0100] (2) Allocate sub-channels. Each sub-channel is weighted according to the transmission command. The data is divided proportionally for parallel transmission.

[0101] (3) Allocate time slices. Within the current small scheduling cycle, allocate time according to the weights of all instructions. The instructions are divided proportionally and executed sequentially. Higher-weighted instructions receive better channel access or longer execution time.

[0102] 4.5: Task Execution and Node Collaboration: Step 1: Execution and Energy Consumption Recording. Nodes execute instructions according to resource allocation results, and then accurately measure the actual energy consumed in this scheduling. And update the used energy: At the same time, update the remaining data amount of the corresponding instruction in the instruction queue. .

[0103] Step 2: Event-triggered lightweight cooperative beacon. To achieve distributed synchronization and anomaly avoidance across node tasks, this invention defines a structured lightweight cooperative beacon. A node immediately broadcasts this beacon when any of the following events occur: (1) Start executing or complete an instruction involving neighbor cooperation.

[0104] (2) Survival level Things have changed.

[0105] (3) A high-priority instruction is continuously blocked due to insufficient local resources or extremely poor channel.

[0106] Therefore, the beacon format can be described as a quadruple: in, As a unique identifier for the instruction, .

[0107] (4) Neighboring nodes Coordination strategy upon receiving a beacon: If the beacon indicates a neighbor If it enters a critical state, then Cancel or suspend all planned trips immediately Send data commands to avoid invalid transmissions; if the beacon indicates a neighbor... If a send command has already been executed and the sender is the receiver, then Prioritize resource allocation for receiving; if a beacon indicates that the predecessor instruction of a multi-hop task has been completed, the neighboring node, as the successor node, can start resource preparation in advance to achieve pipeline-style advancement.

[0108] 4.6: Completion of judgment and exception feedback triggering: Step 1: Intra-slot scheduling loop. Repeat steps 4.2 to 4.5 until any of the following conditions are met, marking the end of the current small-scale scheduling cycle: (1) Time slot The clock time has ended.

[0109] (2) Energy budget for this time slot Exhausted.

[0110] (3) Instruction queue All instructions have been completed or processed according to the rules.

[0111] Step 2: Anomaly Monitoring and Reporting. At the end of the time slot, the node performs final calculations.

[0112] (1) Calculate energy consumption deviation: .

[0113] (2) Verify instruction completion rate: Statistically analyze instructions that were not completed or downgraded, and their priorities. If Exceeding the threshold If a high-priority instruction is not completed, or if the survival level changes unexpectedly, the node will package the above abnormal information and report it to the central node, thereby triggering the global replanning process.

Claims

1. A resource scheduling method for a special power replenishment mode communication network, characterized in that, Includes the following steps; Step 1: Construct an environment and energy consumption perception model, and provide underlying perception data input for the entire network by quantifying the illumination and node power consumption in extreme environments; Step 2: Based on the sensed data, each node in the communication network independently calculates an energy security budget that will never trigger the survival baseline, thus decoupling the long-term hard constraint of survival and transforming it into a single-cycle energy boundary. Step 3: Use the energy boundaries of each node as hard constraints, and have the central node coordinate and allocate global tasks and energy on a long-term scale; Step 4: Based on the planning instructions for coordination and allocation in Step 3, guide each communication node to adaptively allocate resources and execute actions according to real-time channel and energy fluctuations within the micro-time slot.

2. The resource scheduling method for a special power replenishment mode communication network according to claim 1, characterized in that, Step 1 specifically involves: Step 1.

1. Considering the coexistence of long-period light and dark alternation and random terrain occlusion at the lunar south pole, a hybrid illumination intensity model combining deterministic trends and random perturbations is established. : Among them, the deterministic component Simulate the trend of solar altitude angle variation around the lunar south pole in approximately 28 days. Angular frequency; random occlusion component This is an indicator function; it takes the value of 1 when physical devices in the communication network, such as probes or relay base stations (hereinafter referred to as nodes), are obscured by local craters or terrain shadows, and otherwise takes the value of 0; this variable follows a Bernoulli distribution based on terrain complexity; Energy harvesting components, i.e., the output power of solar panels for: in, For photoelectric conversion efficiency, This refers to the effective area of ​​the windsurfing board; Step 1.2: Assume the network contains a set of nodes. This set represents the collection of all devices in the communication network, among which base station-type devices with global information collection and scheduling computing power are denoted as central nodes. The remaining terminal devices responsible for data collection and relay forwarding are referred to as ordinary nodes. The plan adopts a two-level time scale: the larger time scale planning period is [length missing]. Further divided into each planning cycle Hourly timescale time slots; small timescale time slot length is ; node The energy at the beginning of the planning period is The upper limit of energy is Survival threshold is In the time slot Within, the predicted energy replenishment for each node is: It can be obtained from illumination prediction or energy harvesting prediction; the survival power consumption is Then the energy consumption for survival is: Nodes in time slots Energy consumption related to communication and processing, i.e., active energy consumption. The timeframe for the actual execution of data generation, reception, and local processing is determined by the timing of the nodes. In the time slot Within that timeframe, the cumulative time allocated for sending, receiving, and processing actions are respectively , and The corresponding hardware operating power is respectively , and Ignoring the tiny dormant leakage current, the time slot The total active energy consumption is expressed as: The sum of the cumulative times of each action must satisfy the time slot length constraint: The evolution of battery energy is as follows: ； And it satisfies: ; Step 1.3: Considering path loss and shadow fading caused by irregular lunar terrain, nodes To the node Channel gain: in, For communication distance, This is the path loss index. The shadowing fading factor follows a log-normal distribution. It is the transmit / receive antenna gain. For carrier frequency; Step 1.4: Step (1): Real-time acquisition of the battery's current state of charge. And obtain the estimated duration of the next effective energy replenishment in cyclotron mechanics from the energy replenishment prediction sequence. ; Step (2): Calculate the survival safety factor This coefficient represents the energy surplus of a node after deducting its basic survival needs. Step (3): According to The value categorizes nodes into three survival states to guide access control on a small timescale; if The energy is sufficient, indicating a safe state, and full service access is permitted; if Energy is low, indicating a state of alert; only high-priority services are allowed access. Survival is threatened, a critical state is declared, and the communication radio frequency module is forcibly shut down, maintaining only [operational function]. .

3. The resource scheduling method for a special power replenishment mode communication network according to claim 2, characterized in that, Step 2 specifically involves: Step (1): At the beginning of each large-scale planning cycle, the nodes Obtain and update its local energy state and prediction parameters: initial energy The upper limit of energy is Survival threshold is Survival power consumption is ,future Energy replenishment sequence for each time slot node and operating mode power consumption parameters, transmission power Received power Processing power Sleep power ; Step (2): The node first generates an initial sequence of candidate active energy consumption budget upper limits. ; This sequence characterizes the energy estimates of nodes for non-survivability tasks in each time slot before survival feasibility verification; initialization can be achieved using a simple heuristic rule that makes it proportional to the predicted supplemental energy: in, This is a preset conservative coefficient used to reserve a safety margin for energy replenishment prediction errors; For time slots The predicted energy replenishment; this step provides an iterative starting point for subsequent feasibility verification and repair; Step (3): The node will select the candidate active energy consumption budget sequence. As the active energy consumption in the model defined in step 1.2 Input, combined with survival energy consumption Perform full-cycle forward simulation; Specifically, substituting the above parameters into the energy evolution equation: The predicted energy trajectory is obtained through recursive calculation. ; Subsequently, it is verified whether the survival hard constraints are satisfied throughout the entire lifecycle: If the constraint holds for all time slots, the current candidate budget sequence is feasible, and the process jumps to step (5) for output; if there is a first time slot that causes the constraint to be violated... If the default occurs, the adaptive repair process in step (4) is initiated. Step (4): A spatiotemporally aware adaptive default repair mechanism, when in a time slot Upon detection of a breach, a three-stage progressive strategy will be employed to rectify the situation. Step (5): Feasible budget output and reporting; Finally, the node obtains a rigorously validated and feasible active energy consumption budget sequence and calculates its sum; the node then outputs this sequence and sum. In addition, power consumption parameters are reported to the central node to complete the distributed budget generation task.

4. The resource scheduling method for a special power replenishment mode communication network according to claim 3, characterized in that, The three stages in step (4) are as follows: (1) First stage: Checking the time slots for breach of contract The previous time slot Budget value If it is significantly higher than the average of neighboring time slots, such as being greater than the historical average within a sliding window; If the peak value is multiples of the previous value, it is determined to be a peak value that can be reduced. A portion of the energy is drawn from this peak. Transfer to the default time slot The aim is to resolve early defaults through local rebalancing; after this adjustment is completed, immediately return to step (3) to re-perform the forward deduction verification; (2) Second phase: If the first phase is not applicable or the default is not eliminated, the trigger will be implemented. For each time slot within the default segment Assign a shrinkage weight The weighting design principle is that the less energy is added, the more contraction it bears, for example: Calculate a baseline contraction factor based on the total energy deficit. Then, differentiated adjustments are made to the budget for the defaulted segments: After this operation, return to step (3) to re-verify; (3) Third stage: 1) Settings The goal of time slot repair is ,in For forward-looking safety margins; 2) To achieve this goal, from The time slots will begin to be traced back, identifying and prioritizing the reduction of time slot budgets that have relatively abundant budgets and high flexibility for adjustment. The reduction amount must be sufficient to fill the current gap and achieve a safety margin. ; After completing this deep repair, return to step (3) to perform a full-cycle simulation; because The time slot energy state is reinforced, significantly reducing the risk of subsequent default. If a new default occurs in a later time slot, that time slot will be considered the new default. This step involves iteratively executing the three-stage repair process.

5. The resource scheduling method for a special power replenishment mode communication network according to claim 4, characterized in that, Step 3 specifically involves: Step 3.1: When the planning period is... Discretize it within the window. Each planning period, with the period index as follows: ; Step 3.2: Decompose the task on the candidate routing path into atomic actions of sending, receiving and processing at nodes along the way, map the task into executable instructions, and calculate the energy consumption based on the power consumption parameters in Step 2. Step 3.3: The central node solves within the planning window, selects a path for each task, and determines the amount to be completed in different time periods, so as to maximize the benefits of task completion without exceeding the node's activity energy consumption budget; Step 3.4: A feasible solution is generated by using a priority-driven iterative solution strategy of feasible allocation and budget deduction update. Its output naturally satisfies the above hard constraints and is easy to be issued as a node-parseable instruction packet.

6. The resource scheduling method for a special power replenishment mode communication network according to claim 4, characterized in that, The instruction package contains the following three core components: (1) Energy budget constraint: The sequence of safe activity energy consumption budgets available for each time slot within the period confirmed by the center at this node. and total periodic budget This section represents the rigid performance limit for all communication activities of the node. (2) Atomic action instruction set: The instruction list of this node. Each instruction in the list; The specific time slots of the nodes are clearly defined. The atomic actions to be performed, the communication peer, the amount of data, and the reference rate; this is the direct executable result after task decomposition.

7. The resource scheduling method for a special power replenishment mode communication network according to claim 5, characterized in that, In step 4, each node autonomously and in a distributed manner executes the process within each small timescale scheduling slot, strictly following the scheduling and energy management instruction packages issued by the central node. Under real-time channel changes and energy fluctuations, atomic instructions are reliably and efficiently transformed into communication actions through local fine-tuning and neighbor collaboration, while ensuring strict adherence to the hard energy budget constraint throughout the entire process.

8. The resource scheduling method for a special power replenishment mode communication network according to claim 7, characterized in that, Step 4.1: Node First, analyze its instruction package. and initialize the local scheduling state; Read the communication energy budget for this time slot and total periodic budget Initializing this time slot has already consumed energy. ; Extract the time slot label from the instruction packet. All atomic action instructions constitute the local time slot instruction queue. Each instruction in the queue ; The local survival safety factor is calculated in real time based on the model in step 1.

4. And mapped to survival level ; Step 4.2: Nodes at each scheduling moment in each time slot The remaining energy budget needs to be converted into an immediately available upper limit for transmit power; Calculate the upper limit of instantaneous transmit power; in, This is the maximum transmit power supported by the hardware. For time slots The remaining time; this constraint ensures that the expected power consumption at any time does not exceed the budget before the end of the time slot; Step 4.3: At the beginning of each small scheduling cycle, the node preprocesses the instruction queue according to the survival level; like Clear All communication commands that are not essential for survival are retained, with only the minimum state maintenance commands remaining. like : Enable priority-aware random discarding; traverse Low-priority instructions are discarded or deferred with a probability inversely proportional to their priority. like The entire instruction queue participates in this round of scheduling; Step 4.4: Node in power budget Under constraints, the transmission resources are allocated in the most energy-efficient way to execute commands that need to be transmitted wirelessly; the resources for receiving and processing actions are represented by the allocated execution time slices. Step 4.5: Task Execution and Node Collaboration: Step (1): Execution and energy consumption recording; the node executes instructions according to the resource allocation results, and then accurately measures the actual energy consumed in this scheduling. And update the used energy: At the same time, update the remaining data amount of the corresponding instruction in the instruction queue. ; Step (2): Event-triggered lightweight coordination beacon; to achieve distributed synchronization and anomaly avoidance across nodes, the node immediately broadcasts this beacon when any of the following events occur: (1) To begin executing or complete an instruction involving neighbor cooperation; (2) Survival level Changes have occurred; (3) A high-priority instruction is continuously blocked due to insufficient local resources or extremely poor channel; Therefore, the beacon format can be described as a quadruple: in, As a unique identifier for the instruction, ; (4) Neighboring nodes Coordination strategy upon receiving a beacon: If the beacon indicates a neighbor If it enters a critical state, then Cancel or suspend all planned trips immediately Send data commands to avoid invalid transmissions; if the beacon indicates a neighbor... If a send command has already been executed and the sender is the receiver, then Prioritize resource allocation for receiving; if the beacon indicates that the predecessor instruction of a multi-hop task has been completed, the neighboring node, as the successor node, can start resource preparation in advance to achieve pipeline-style progress. Step 4.6: Complete the judgment and trigger the abnormal feedback.

9. The resource scheduling method for a special power replenishment mode communication network according to claim 8, characterized in that, Step 4.4 specifically involves: Step (1): Water injection power allocation. The system bandwidth is divided into... There are several sub-channels; for the command to be sent, its target node is... In the sub-channel The real-time channel gain on is The noise power is ; Node-based solution for optimal transmit power allocation : Its classic interpretation is ,in Water injection line to meet total power constraints; Step (2): Instruction-driven time-frequency resource mapping; Based on the power allocation results and instruction attributes, sub-channels and time resources are allocated to each instruction to be executed; (1) Calculate instruction performance weights; for instructions Priority If it is a sending command, its weight If it is a receiving or processing instruction, its weight is priority. ; (2) Allocate sub-channels; Each sub-channel is weighted according to the transmission command. The data is divided proportionally for parallel transmission; (3) Mapping between time slices and actual energy consumption. Within the current scheduling time slot, nodes process instructions in the queue sequentially. For a single instruction... Based on its actual allocated bandwidth and power, the actual achievable rate of the physical layer is calculated. The exact execution time slice required for this instruction is... .like If the remaining time in the current time slot is greater than the current time slot's remaining time, execution will be truncated, and the untransmitted data will be retained in the queue; if execution is allowed within the remaining time, the actual energy consumption increment will be... .

10. The resource scheduling method for a special power replenishment mode communication network according to claim 8, characterized in that, Step 4.6 specifically involves: Step (1): Intra-slot scheduling loop; repeat steps 4.2 to 4.5 until any of the following conditions are met, marking the end of the current small-scale scheduling cycle: (1) Time slot The clock time has ended; (2) Energy budget for this time slot Exhausted; (3) Instruction queue All instructions have been completed or processed according to the rules; Step (2): Anomaly monitoring and reporting; at the end of the time slot, the node performs final calculation; (1) Calculate energy consumption deviation: ; (2) Verify instruction completion rate: Statistically analyze instructions that were not completed or downgraded, and their priorities; if Exceeding the threshold If a high-priority instruction is not completed, or if the survival level changes unexpectedly, the node will package the above abnormal information and report it to the central node, thereby triggering the global replanning process.

Images