Task scheduling method and device, electronic equipment and storage medium

Abstract

One or more embodiments of the present specification provide a task scheduling method and device, electronic equipment and storage medium. The task scheduling method provided by the present specification is applied to a management node of a node network, and the method comprises: acquiring a plurality of communication tasks to be scheduled in a scheduling period of the node network; for each communication task, generating a TA envelope according to a propagation delay drift interval between a sending end and a receiving end, wherein the TA envelope is a union of all sending time slots required by the sending end to occupy under all possible time delay values in the propagation delay drift interval when the target of the communication task is to complete reception in a specified receiving time slot; determining whether there is a conflict between any two communication tasks based on the TA envelope; and determining a node scheduling scheme that meets all conflict constraints according to the result of the conflict determination.

Description

Technical Field

[0001] This application relates to the field of wireless communication network scheduling technology, and in particular to a task scheduling method, apparatus, electronic device and storage medium. Background Technology

[0002] Wireless communication networks typically consist of multiple nodes that exchange data via radio waves. In a network of multiple mobile wireless nodes, these nodes may share a portion of the wireless channel. To ensure orderly communication without collisions among the numerous nodes on this shared channel, at least one management node is usually present. The management node is responsible for overall planning, allocating specific data transmission and reception time slices to each node, thus ensuring that each node executes its communication tasks according to the assigned schedule.

[0003] In related technologies, when nodes in a network are in a state of high-speed relative movement, the communication quality between nodes often becomes poor, and even packet loss may occur. Summary of the Invention

[0004] In view of the above, one or more embodiments of this specification provide the following technical solutions: This specification provides a task scheduling method in its first aspect, applied to a management node in a node network, the method comprising: Obtain multiple communication tasks to be scheduled within the scheduling period of the node network; For each communication task, a TA envelope is generated based on the propagation delay drift interval between the sender and receiver, wherein the TA envelope is: When the goal of the communication task is to complete reception in a specified reception time slot, the union of all transmission time slots required by the transmitting end under all possible delay values ​​within the propagation delay drift interval; Based on the TA envelope, determine whether there is a conflict between any two of the communication tasks; Based on the results of the conflict determination, a node scheduling scheme that satisfies all conflict constraints is determined.

[0005] Optionally, determining whether a conflict exists between any two communication tasks based on the TA envelope includes: Based on the TA envelope, determine whether there is a half-duplex conflict and / or signal interference conflict between any two communication tasks.

[0006] Optionally, the half-duplex conflict includes at least one of the following: The two communication tasks are sent from the same node, and their TA envelopes overlap on the time axis. The receiving end of the two communication tasks is the same node, and their designated receiving time slots are the same; The sender of one communication task is the receiver of another communication task, and the TA envelope of the sender covers the designated receive time slot of the other communication task.

[0007] Optionally, the method further includes: Based on the propagation delay drift interval of the first communication task and the propagation delay drift interval of the second communication task, the time overlap ratio on the time axis between the transmission signal of the first communication task arriving at the receiving end of the second communication task and the designated receiving time slot of the second communication task is determined, wherein the first communication task and the second communication task are any two communication tasks among the plurality of communication tasks. Based on the time overlap ratio, the signal-to-interference-plus-noise ratio (SINR) of the receiving end of the second communication task when it is interfered with by the first communication task is determined. If the SINR is lower than a preset demodulation threshold, then it is determined that there is a signal interference conflict between the first communication task and the second communication task.

[0008] Optionally, determining the time overlap ratio on the time axis between the arrival of the transmission signal of the first communication task at the receiving end of the second communication task and the specified receiving time slot of the second communication task includes: Based on the current position and reference speed of the sending end of the first communication task, and the current position and reference speed of the receiving end of the second communication task, the propagation delay drift interval between the sending end of the first communication task and the receiving end of the second communication task is determined. Multiple delay combinations are obtained, the delay combination including a first delay and a second delay, the first delay being located within the propagation delay drift interval of the first communication task, and the second delay being located within the propagation delay drift interval between the sending end of the first communication task and the receiving end of the second communication task; For each delay combination, based on the specified receive time slot of the first communication task and the first delay in the delay combination, the transmit time window of the first communication task under the delay combination is determined; And, based on the transmission time window and the second delay, determine the arrival time window of the transmission signal of the first communication task at the receiving end of the second communication task; Furthermore, based on the overlap duration between the arrival time window and the specified reception time slot of the second communication task on the time axis, the candidate overlap ratio under the current delay combination is determined; The maximum value among the candidate overlap ratios corresponding to each delay combination is determined as the time overlap ratio on the time axis between the sending signal of the first communication task and the designated receiving time slot of the second communication task when the sending signal of the first communication task arrives at the receiving end of the second communication task.

[0009] Optionally, determining the node scheduling scheme that satisfies all conflict constraints based on the conflict determination result includes: A set of mutually exclusive tasks is determined among the plurality of communication tasks, wherein any two communication tasks included in the set of mutually exclusive tasks conflict with each other. Based on a preset solution model, a preset optimization objective is solved under constraints to obtain the node scheduling scheme of the node network within the scheduling period. The constraints include: In each set of mutually exclusive tasks, at most one communication task is active.

[0010] Optionally, determining the set of mutually exclusive tasks among the plurality of communication tasks includes: Using each of the multiple communication tasks as a vertex, conflict edges are established between the vertices corresponding to two conflicting communication tasks to obtain a conflict graph. The set of vertices contained in each maximal clique in the conflict graph is determined as a set of mutually exclusive tasks.

[0011] Optionally, the step of solving a preset optimization objective based on a preset solution model under constraints to obtain a node scheduling scheme for the node network within the scheduling period includes: Based on the preset solution model, the preset optimization objective is solved under the constraints of the current iteration round to obtain the temporary scheduling scheme for the current iteration round. For each task pair, determine whether there is signal interference conflict between the two communication tasks in the task pair, wherein the task pair is two communication tasks activated in the temporary scheduling scheme of the current iteration round, and the arrival time window of the transmission signal of one communication task may overlap with the designated reception time slot of the other communication task. In the event of signal interference conflict between two communication tasks within at least one task pair, a new set of mutually exclusive tasks is determined among the plurality of communication tasks, and the next iteration round is initiated. If there is no signal interference conflict between the two communication tasks within each task pair, the temporary scheduling scheme of the current iteration round is determined as the node scheduling scheme of the node network within the scheduling period.

[0012] Optionally, the propagation delay drift interval is derived based on the current positions of the transmitting end and the receiving end and a reference motion speed, wherein the lower limit of the propagation delay drift interval corresponds to the transmitting end and the receiving end moving towards each other at the reference motion speed, and the upper limit corresponds to moving away from each other at the reference motion speed.

[0013] Optionally, before determining the node scheduling scheme of the node network within the scheduling period, the method further includes: For each of the plurality of communication tasks, the upper limit of the latency corresponding to the communication task is obtained, and the expected communication delay is determined according to the propagation delay drift interval corresponding to the communication task. If the sum of the queuing waiting time of the communication task at the sending end and the expected communication delay is greater than the upper limit of the latency, it is determined that the communication task will not enter the current scheduling cycle.

[0014] A second aspect of this specification provides a task scheduling device applied to a management node in a node network, the device comprising: The task acquisition module is used to acquire multiple communication tasks to be scheduled in the node network within the scheduling period; The envelope generation module is used to generate a TA envelope for each communication task based on the propagation delay drift interval between the sender and receiver, wherein the TA envelope is: When the goal of the communication task is to complete reception in a specified reception time slot, the union of all transmission time slots required by the transmitting end under all possible delay values ​​within the propagation delay drift interval; The conflict determination module is used to determine whether there is a conflict between any two communication tasks based on the TA envelope; The node scheduling module is used to determine a node scheduling scheme that satisfies all conflict constraints based on the result of the conflict determination.

[0015] A third aspect of this specification provides an electronic device, comprising: processor; Memory used to store processor-executable instructions; The processor implements the steps of the method as described in the first aspect by running the executable instructions.

[0016] A fourth aspect of this specification provides a computer-readable storage medium having computer instructions stored thereon that, when executed by a processor, implement the steps of the method as described in the first aspect.

[0017] This specification provides a computer program product, including a computer program / instructions, which, when executed by a processor, implement the steps of the method as described in the first aspect.

[0018] As can be seen from the above embodiments, this specification describes an embodiment that obtains multiple communication tasks to be scheduled within the node network during the scheduling period. For each communication task, based on the propagation delay drift interval between the sender and receiver, it generates the union of all possible transmission time slots required by the sender under all possible delay values ​​when the goal of the communication task is to complete reception in the specified reception time slot. The node scheduling scheme is then determined directly based on this union to determine conflicts between tasks. Based on this, the management node essentially plans and reserves a safe transmission time window covering all delay uncertainties for the sender in advance through this union, ensuring that the sender is not mistakenly assigned to other tasks and collides during the period when the communication task may need to be executed. Regardless of how the actual distance between nodes in physical space drifts over time, the sender can always find a suitable and precise transmission time within the legal time slots covered by the union, ensuring that the transmitted signal accurately reaches the receiver in the specified reception time slot. This scheme avoids bandwidth waste and congestion risks caused by frequent probe signaling interactions between nodes before each communication, while also preventing timing errors and task conflicts caused by high-speed relative movement of nodes in highly dynamic networks. Without significantly increasing the system idle protection interval, it greatly improves the reliability and robustness of node network task scheduling. Attached Figure Description

[0019] Figure 1 This is a flowchart of a task scheduling method provided in an exemplary embodiment.

[0020] Figure 2 This is an exemplary embodiment of an application scenario diagram of a task scheduling method.

[0021] Figure 3 This is a schematic diagram of a node network provided in an exemplary embodiment.

[0022] Figure 4 This is a block diagram of a task scheduling device provided in an exemplary embodiment.

[0023] Figure 5 This is a schematic diagram of the structure of an electronic device provided in an exemplary embodiment. Detailed Implementation

[0024] As described in the background section, in related technologies, when nodes in a network are in a state of high-speed relative movement, the communication quality between nodes often becomes poor, and even packet loss may occur.

[0025] In view of this, this specification provides a task scheduling method, apparatus, electronic device, and storage medium, which can be applied, for example, to communication task scheduling scenarios for dynamic multi-hop wireless networks.

[0026] The methods provided in this specification can be applied, for example, to management nodes such as gateways or spaceborne controllers (also known as network management nodes or central nodes). These network management nodes need to provide communication scheduling schemes for each node in the network (e.g., low-altitude drones, satellites, ground-based unmanned vehicles, etc.) within each scheduling cycle, enabling the nodes to communicate with each other without conflict. Generally, the network management node plans several time slots and notifies each node so that each node can operate according to a unified time slot schedule.

[0027] In this specification, a communication task refers to a communication request, i.e., a node needs to send a data packet to another node, and this request is considered a communication task. The hardware of each node may be half-duplex, meaning that the same node cannot be both sending and receiving in the same time slot, nor can it clearly receive signals from two different nodes at the same time (signals will interfere with each other). Therefore, the network management device needs to know the communication tasks of each node within a scheduling period (e.g., 0.01s) to plan a node scheduling scheme for that period and notify each node. This node scheduling scheme is used to instruct the actions of each node within the pre-divided time slots. For example, please refer to [link to relevant documentation]. Figure 2 If node A (satellite A) needs to send a data packet to node B (satellite B), the network management device needs to determine when node B should be in a receiving state to receive the data packet based on information such as the communication task involved in node B. This is the designated receiving time slot (also called the receiving time slot) for the communication task. By making both node A and node B aware of this node scheduling scheme, it can be ensured that within the designated receiving time slot allocated to the communication task, node B is only in a receiving state, and the signal sent by node A will fall within the time span of that time slot.

[0028] In large-scale networks such as inter-satellite or space-to-ground networks, the physical distance between nodes is vast, and the time required for signal propagation in the air is often non-negligible. To ensure that the signal arrives at node B precisely at the designated reception time slot specified by the network management equipment, node A cannot begin transmitting only at the designated reception time slot. Instead, it needs to calculate the corresponding air propagation delay based on the spatial distance between the two nodes just before reaching the designated reception time slot, and then advance the transmission action accordingly on the timeline.

[0029] In conventional static networks, the distance between nodes is fixed, and the delay is a constant value. The sender only needs to send the data at the corresponding time in advance. However, in highly dynamic networks such as inter-satellite or air-to-ground networks, the nodes are in a state of high-speed relative motion. Within a scheduling cycle, the distance between node A and node B may continuously change, causing the signal's time-of-flight to drift. If we still rely on real-time interaction and precise delay calculation only when approaching the designated reception time slot, there is a risk of timing errors due to queuing and computational time. Moreover, if the two nodes are moving in opposite directions at high speed (moving away from each other), and the current distance between the nodes is significantly greater than expected, then node A may know that the time slot it should have used to send the data packet has passed, and therefore the data packet may fail to be sent.

[0030] Therefore, related technologies often require significantly increasing the span (duration of coverage) of each time slot, or adding a large amount of idle waiting time between each time slot, in order to offset the aforementioned risks while wasting a lot of time and computing power. However, this technology comes at a high cost, especially when the node network is very large, the communication latency between nodes can become extremely high.

[0031] The first aspect of this specification provides a task scheduling method; please refer to [link / reference]. Figure 1 The method may include steps S101 to S104.

[0032] S101, Obtain multiple communication tasks to be scheduled in the node network within the scheduling period.

[0033] Communication tasks are typically reported by each node in the node network to the network management node, or uniformly issued by the network management node based on the service flow forwarding requirements in the network.

[0034] In some embodiments of this specification, before determining the node scheduling scheme of the above-mentioned node network within the scheduling period, the method may further include: for each of the aforementioned multiple communication tasks, obtaining the upper limit of the latency corresponding to the communication task, and determining the expected communication delay based on the propagation delay drift interval corresponding to the communication task; if the sum of the queuing waiting time of the communication task at the sending end and the above-mentioned expected communication delay is greater than the upper limit of latency, determining that the communication task will not enter the current scheduling period.

[0035] It is understandable that the above-mentioned latency limit refers to the maximum latency tolerance allowed for the business data carried by the communication task (for example, the latency limit of some real-time control signaling is relatively short, while the latency limit of ordinary background data is relatively long), while the queuing waiting time refers to the expected time that the task will stay in the buffer queue of the sending node, and the expected communication delay is the flight time of the signal in the air calculated based on the relative distance between the sending end and the receiving end.

[0036] In other words, if a data packet, even if scheduled immediately, has already exceeded the service's allowed latency limit due to the sum of its queuing time and flight time, then the task has effectively lost its scheduling value. Therefore, considering the limited communication resources and concurrent processing capabilities of nodes in the network, determining that these communication tasks will not enter the current scheduling cycle before entering the scheduling process can effectively prevent invalid tasks from occupying channel resources. At this point, the task can either be discarded (i.e., marked as invalid and notified to the corresponding node), or it can be postponed, for example, placed in the next scheduling cycle, or processed when the number of communication tasks in a certain scheduling cycle is less than the task number threshold (i.e., the communication bandwidth pressure is relatively low).

[0037] S102, for each communication task, generate a TA envelope based on the propagation delay drift interval between the sending end and the receiving end, wherein the TA envelope is: the union of all transmission time slots required by the sending end under all possible delay values ​​within the propagation delay drift interval when the goal of the communication task is to complete reception in a specified reception time slot.

[0038] In other words, the TA (Timing Advance) envelope represents the total transmission time window that the sender must reserve to ensure that the receiver receives the signal precisely within the specified reception time slot. Due to the highly dynamic nature of networks, the distance between nodes changes in real time, and the exact duration of a signal's flight through the air is often impossible to determine in advance, but its fluctuation range (i.e., the propagation delay drift range) is known. This means that there are many possible timings for the sender to accurately deliver the data packet to the receiver at the specified reception time slot. In S102, the network management node can extract all these time slots that may need to enter the transmission state and merge them (take the union) to form a time block sufficient to cover all possibilities; this time block is the TA envelope. During the period covered by the TA envelope, the sender is considered to be in an occupied state, thereby reserving sufficient safety advance for the sender to perform the packet transmission action locally.

[0039] To accurately define the aforementioned time delay uncertainty, in some embodiments of this specification, the propagation delay drift interval can be derived based on the current positions of the transmitting and receiving ends and a reference motion speed. Specifically, the lower limit of this propagation delay drift interval corresponds to the case where the transmitting and receiving ends move towards each other at the reference motion speed, and the upper limit corresponds to the case where they move away from each other at the reference motion speed.

[0040] It's understandable that the movement trajectory of nodes is often difficult to predict accurately within a scheduling cycle; therefore, extreme scenarios of relative node movement can be considered. The reference speed can refer to the pre-calibrated (or real-time acquired) maximum relative movement speed between nodes, or it can be an estimated speed amplified by a safety margin, or a reference value obtained through other means. In the most extreme close-up scenario, if the transmitter and receiver move directly towards each other at this reference speed, the distance between them will be shortened to its minimum possible value, resulting in the shortest signal propagation time—this is the lower limit of this interval. Conversely, in the most extreme far-away scenario, if the two move completely away from each other at this speed, the distance will be increased to its maximum possible value, resulting in the longest signal propagation time—this is the upper limit of this interval.

[0041] This time interval, inferred from extreme motion states, can cover any actual propagation delays that may occur in the future, thus ensuring high robustness of the system without consuming too many system resources.

[0042] S103, determine whether there is a conflict between any two communication tasks based on the TA envelope.

[0043] Once the network management node obtains the TA envelope of each communication task, it has the conditions to assess whether different tasks collide in the time domain.

[0044] In some embodiments of this specification, the above-mentioned determination of whether there is a conflict between any two communication tasks based on the TA envelope may include: determining whether there is a half-duplex conflict and / or signal interference conflict between any two communication tasks based on the TA envelope.

[0045] Half-duplex conflicts are primarily limited by the hardware transmit / receive capabilities of nodes. In some embodiments of this specification, the aforementioned half-duplex conflicts may include at least one of the following: First, the two communication tasks originate from the same node, and their TA (Transmission Time) envelopes overlap on the time axis. In other words, two different communication tasks need to invoke the same node (e.g., the same satellite) to transmit data. If the TA envelopes corresponding to these two tasks overlap in time, it means that during this period, the node may face the situation of simultaneously transmitting two independent signals. Since this often exceeds the single-channel processing capability of conventional half-duplex hardware, it can be considered as forming a half-duplex conflict at the transmitting end.

[0046] Second, the receiving end of both communication tasks is the same node, and their designated receiving time slots are the same. In other words, both tasks require the same node to receive data within the same time slot. Since it is generally difficult for the same node to receive and demodulate aliased signals from two transmitting ends without interference within the same time period, this leads to half-duplex collisions at the receiving end.

[0047] Third, the sender of one communication task is the receiver of another, and the sending TA envelope overlaps with the designated receive time slot of the other communication task. For example, suppose task A requests node A to send data, while task B requests node A to receive data. If the packet sending time window (TA envelope) planned by task A overlaps with the designated receive time slot for task B to receive packets, then node A must act as both sender and receiver during the overlapping period. Due to the limitation of half-duplex hardware, which can only perform one send / receive action at a time, node A cannot handle both sending and receiving simultaneously, thus potential conflicts may arise.

[0048] It should be added that, in addition to the half-duplex conflict caused by the hardware system, air interface communication may also have signal interference conflicts caused by the superposition of signals in space. That is, the signal beams of multiple transmitters happen to converge at a certain receiver and generate malicious interference. The specific calculation and judgment details of signal interference conflicts will be described in detail in subsequent embodiments.

[0049] S104. Based on the result of the conflict determination, determine a node scheduling scheme that satisfies all conflict constraints.

[0050] After clarifying the conflicts between various communication tasks, the network management node can schedule these tasks accordingly. Under the premise of satisfying constraints (e.g., no half-duplex conflict and / or signal interference conflict between any two allowed communication tasks), the network management node determines the final combination of communication tasks activated in the current scheduling cycle, as well as the specific execution sequence (time slot) of these tasks on each node, thus obtaining a node scheduling scheme that satisfies all conflict constraints. In the specific solution stage, the solution can be obtained using heuristic backtracking algorithms, greedy algorithms, or graph theory-based methods; this specification does not impose specific restrictions on this approach.

[0051] The above process can also be described as follows: For each of the at least one communication task, based on the current position and reference speed of the sending node of the communication task, and the current position and reference speed of the receiving node of the communication task, the communication delay interval (i.e., propagation delay drift interval) between the sending node and the receiving node is determined at the target completion time slot (i.e., the specified receiving time slot) of the communication task, and the time slots of the sending node that overlap with the communication delay interval in the scheduling period are marked as occupied; then, based on the occupancy status of each time slot of each node in the node network in the scheduling period, the task scheduling relationship (i.e., task scheduling scheme) of the node network in the scheduling period is determined.

[0052] Based on this, all time slots within the sending node that might be used to send the data packet have been marked as occupied (i.e., no new tasks can be allocated), so any method can be used to solve the problem subsequently.

[0053] Once the node scheduling scheme is determined, the network management device will distribute the time slot schedule in the scheme to the relevant nodes. Each node then only needs to perform local calculations and send / receive operations within the safe time slots specified in the scheme.

[0054] Through the above steps, at the beginning of each scheduling cycle, this scheme calculates the propagation delay drift interval based on the predicted speed and position of the relative movement of nodes, thereby obtaining a TA envelope covering all potential transmission times. After the network management node uses the TA envelope and the receiving time slot as constraints for conflict determination, it can prevent tasks that might cause transmission and reception errors from being activated on the same node in the same time slot, ensuring that each node has sufficient lead time to execute packet transmission. Compared to existing methods that involve frequent signaling interaction calibration and significantly increased idle waiting time, the scheme provided in this specification also significantly reduces time overhead in highly dynamic networks, effectively improving network throughput while greatly enhancing communication reliability.

[0055] In some embodiments of this specification, the above-mentioned determination of whether there is a conflict between any two communication tasks based on the TA envelope may further include: determining, based on the propagation delay drift interval of the first communication task and the propagation delay drift interval of the second communication task, the time overlap ratio on the time axis between the transmitted signal of the first communication task arriving at the receiving end of the second communication task and the designated receiving time slot of the second communication task, wherein the first communication task and the second communication task are any two of the plurality of communication tasks; determining, based on the time overlap ratio, the signal-to-interference plus noise ratio (SINR) when the receiving end of the second communication task is interfered with by the first communication task; if the SINR is lower than a preset demodulation threshold, then it is determined that there is a signal interference conflict between the first communication task and the second communication task.

[0056] It should be understood that in wireless communication, if the beam of a useful signal transmitted by the transmitter of the first communication task covers the receiver of the second communication task, then the signal of the first communication task can be considered a form of spatial interference wave for the second communication task. For example, please see... Figure 3 For example, when node A sends a signal to node B, the signal will also affect node D, interfering with node D's reception of the signal from node C. Since the flight times of both the interference wave and the useful signal are uncertain, the duration for which they meet and overlap at the receiver of the second communication task will also vary. This time overlap ratio is used to quantify the severity of the interference wave occupying the receiving time slot. Based on this overlap ratio, and parameters such as spatial attenuation between the transmitting and receiving ends, the network management node can calculate the SINR of the second communication task when facing this interference. SINR is a general indicator for measuring the quality of the received signal. If the calculated SINR is lower than a preset demodulation threshold (i.e., the minimum signal-to-noise ratio requirement for the receiver hardware to decode correctly, or other preset reference values), it means that the interference is too great, and the two tasks cannot be executed simultaneously, thus indicating a signal interference conflict between them.

[0057] Because the distance between nodes in a highly dynamic network changes constantly, the signal propagation delay may not be constant. Therefore, in some embodiments of this specification, determining the time overlap ratio between the arrival of the transmitted signal of the first communication task at the receiving end of the second communication task and the designated receiving time slot of the second communication task on the time axis may include: determining the propagation delay drift interval between the transmitting end of the first communication task and the receiving end of the second communication task based on the current position and reference motion speed of the transmitting end of the first communication task, and the current position and reference motion speed of the receiving end of the second communication task; obtaining multiple delay combinations, including a first delay and a second delay, wherein the first delay is located within the propagation delay drift interval of the first communication task, and the second delay is located at the transmitting end of the first communication task. Within the propagation delay drift interval between the first communication task and the receiving end of the second communication task; for each delay combination, based on the designated receiving time slot of the first communication task and the first delay in the delay combination, determine the transmission time window of the first communication task under the delay combination; and, based on the transmission time window and the second delay, determine the arrival time window of the transmitted signal of the first communication task at the receiving end of the second communication task; and, based on the overlap duration of the arrival time window and the designated receiving time slot of the second communication task on the time axis, determine the candidate overlap ratio under the current delay combination; and determine the maximum value among the candidate overlap ratios corresponding to each delay combination as the time overlap ratio of the transmitted signal of the first communication task with the designated receiving time slot of the second communication task on the time axis when the transmitted signal of the first communication task arrives at the receiving end of the second communication task.

[0058] To facilitate understanding, a concrete example is provided to illustrate the above situation. Assume that the designated receiving time slot for the first communication task (Task A) is... The sending end of Task A, in order to make the signal in To reach its corresponding receiver, the data needs to be sent in advance. However, the actual propagation delay of task A (i.e., the first delay) is a drifting value, denoted as . In a specific Under these circumstances, the launch time window (i.e., the start and end time of the launch action) for Task A can be approximately estimated as follows: Move forward .

[0059] Meanwhile, the signal emitted by Task A not only travels to its own receiver but also serves as an interference wave towards the receiver of the second communication task (Task B). The flight time of this interference wave (i.e., the second delay) is also a drift value, denoted as [missing value]. Therefore, the time window for the interference wave to reach the receiver of Task B is the aforementioned transmission time window plus... .

[0060] Subsequently, the network management node compared this arrival time window with the original designated receiving time slot for Task B. By comparing the two on a timeline, the overlap in duration was calculated. The percentage of total duration, which is the delay combination. The candidate overlap ratio. Considering and They all change irregularly within their respective propagation delay drift intervals. The network management node can traverse or optimize to take all possible delay combinations and find the extreme case with the largest overlap ratio, which will be used as the final judgment basis.

[0061] Based on the above implementation method, by performing an extreme traversal of all possible combinations of interference delay, the most severe time overlap situation in the high dynamic network can be accurately located, thereby ensuring that the system can still accurately determine the conflict and provide safety redundancy in the worst case, preventing sudden communication paralysis caused by high-speed movement of nodes.

[0062] After completing the conflict determination between all pairs of communication tasks, the network management node can carry out global scheduling. In some embodiments of this specification, the above-mentioned determination of a node scheduling scheme that satisfies all conflict constraints based on the conflict determination results may include: determining a set of mutually exclusive tasks among multiple communication tasks, wherein any two communication tasks included in the set of mutually exclusive tasks conflict with each other; and solving for a preset optimization objective under the constraints based on a preset solution model to obtain a node scheduling scheme for the node network within the scheduling period, wherein the constraints include: in each set of mutually exclusive tasks, at most one communication task is activated.

[0063] To more efficiently convert the above logical constraints into a mathematical model that the solver can recognize, in some embodiments of this specification, the above-mentioned determination of a set of mutually exclusive tasks among multiple communication tasks may include: taking each of the multiple communication tasks as a vertex, establishing conflict edges between the vertices corresponding to two conflicting communication tasks to obtain a conflict graph; and determining the set of vertices contained in each maximal clique in the conflict graph as a set of mutually exclusive tasks.

[0064] In this context, a clique is a subset of vertices in a graph where every two vertices are adjacent (i.e., connected by an edge), while a maximal clique is a clique that cannot be completely contained within any other clique. In a scheduling scenario, after constructing a conflict graph with tasks as vertices and conflict relationships as edges, each maximal clique represents a set of communication tasks that satisfy the condition that "all tasks within the same maximal clique are mutually exclusive." Within the same time period, no two or more tasks belonging to the same maximal clique can occur simultaneously; that is, the constraint can be that at most one communication task can be activated within each maximal clique.

[0065] For example, assuming tasks 1, 2, and 3 form a maximal clique, then the task scheduling constraint is: (in (This is a binary variable, where 1 represents activation and 0 represents deactivation). Based on this, the network management node can use a pre-defined solution model, such as Mixed Integer Linear Programming (MILP), to solve the problem with the optimization objective of "maximizing the total number of successfully scheduled tasks" or "maximizing the overall network throughput," thereby deriving the final scheduling scheme.

[0066] The above implementation method uses maximal cliques to extract constraint equations that are easier to compute, enabling large-scale node scheduling problems to be efficiently processed by mathematical solvers, thereby improving the speed and accuracy of task scheduling computation.

[0067] Considering that in some complex network topologies, interference is often the result of multiple tasks working together, calculating the interference between all task pairs in the initial stage would consume a lot of computing power. Therefore, a dynamic iterative solution strategy can be adopted. In some embodiments of this specification, the above-mentioned solution based on a preset solution model, under constraints, to solve for a preset optimization objective to obtain a node scheduling scheme for the node network within a scheduling period, may include: solving for a preset optimization objective under constraints of the current iteration round based on a preset solution model to obtain a temporary scheduling scheme for the current iteration round; for each task pair, determining whether there is signal interference conflict between the two communication tasks in the task pair, wherein the task pair consists of two communication tasks activated in the temporary scheduling scheme of the current iteration round, and the arrival time window of the transmission signal of one communication task may overlap with the specified reception time slot of the other communication task; if there is signal interference conflict between the two communication tasks in at least one task pair, re-determining the set of mutually exclusive tasks among multiple communication tasks and entering the next iteration round; if there is no signal interference conflict between the two communication tasks in each task pair, determining the temporary scheduling scheme of the current iteration round as the node scheduling scheme for the node network within a scheduling period.

[0068] In other words, after obtaining the temporary scheduling scheme for the current iteration, the network management node can extract the activated communication tasks in the temporary scheduling scheme as task pairs to determine whether there are signal interference conflicts between these communication tasks. In large-scale networks, the physical distance between nodes is large, and the propagation of signals in the air not only has a time delay, but this time delay also drifts with the relative movement of nodes. This time delay difference may cause the transmission signal of the first communication task in an earlier time slot to arrive in a delayed time window after a long flight, thus falling into the designated reception time slot of the second communication task in a later time slot. Therefore, as long as two communication tasks are activated in the temporary scheduling scheme, and the arrival time window of the transmission signal of one communication task (which can be estimated based on the current distance between the sender and receiver) may overlap with the designated reception time slot of the other communication task, the network management node can identify the two communication tasks as a task pair and perform the aforementioned signal interference conflict determination for the task pair. Of course, this specification does not exclude the possibility of setting the judgment condition to "being activated in the same time slot", that is, the task includes two communication tasks that are activated in the same time slot in the temporary scheduling scheme.

[0069] For example, in the first round, the network management node can generate a first version of the temporary scheduling scheme based solely on "half-duplex conflict" and a few signal interference constraints (at this point, the confirmed conflicts may not have been fully considered, so iteration can continue). Subsequently, the network management node performs the SINR calculation mentioned above only for task pairs scheduled to be sent in the same time slot in this temporary scheduling scheme. If it is found that although task A and task B do not have half-duplex conflict, their simultaneous activation would cause the SINR to be substandard (signal interference conflict exists), then the network management node can add a conflict edge between task A and task B, causing the set of mutually exclusive tasks (maximum clique) to change, that is, incorporating the condition "there is a conflict between task A and task B" into the constraints. Then, the updated constraints are carried into the second round of solution. This process is repeated until the interference conflicts between all tasks activated in the same time slot in a certain round of the temporary scheduling scheme are acceptable, which represents the convergence of the iteration, and this scheme is then issued as the final node scheduling scheme.

[0070] By implementing this iterative task scheduling method, calculations can be performed only on the actual possible interferences, instead of calculating all tasks. This significantly reduces the ineffective computing power overhead of network management nodes and ensures the execution efficiency and robustness of the scheduling scheme.

[0071] The following is a comprehensive discussion of the above implementation methods.

[0072] First, the node network can be considered as a millimeter-wave multi-hop backbone network composed of directional wireless links, which uses a superframe structure for periodic scheduling. Each superframe contains... There are several equal-length time slots, with a time slot length of [missing information]. All nodes in the network can maintain time slot alignment through a clock synchronization mechanism. For any directed link in the network... Its physical propagation delay Euclidean distance between nodes With signal propagation speed The ratio determines, that is Before the scheduling cycle begins, the control plane determines the location and maximum speed of each node. It is estimated that this link will be in the next scheduling cycle. The range of propagation delay values ​​within the range is shown in Equation 1.1.

[0073] (1.1) In Equation 1.1, This represents the link distance at the current moment. This range covers all possible delay values ​​when nodes move towards or away from each other at maximum speed in three dimensions.

[0074] In traditional static delay-aware scheduling, if the scheduling link In the time slot The set of transmission time slots required by the sending end to complete the reception. Based on the current propagation delay The only definitive answer is Equation 1.2.

[0075] (1.2) However, when When uncertainty exists, its spanning time slot length Integer multiple boundaries will lead to The value of changes abruptly. For example, when from jump to hour, When the hop count changes from 3 to 4, the corresponding transmission set will also change. If the scheduling table only locks two fixed transmission time slots based on the initial time delay, then half-duplex conflicts such as "send when it should receive" or "receive when it should send" will occur when subsequent time slots drift across each other.

[0076] To address this issue, the solutions provided in the embodiments of this specification introduce the concept of a TA envelope. For scheduling events... Sending end Not just locked Instead, it is locked to the union of the transmission time slots corresponding to all possible delay values ​​within the value range, i.e., Equation 1.3.

[0077] (1.3) This union is the TA envelope, and its physical meaning is: regardless of how the propagation delay changes within the period, as long as it does not exceed the predicted drift range, the actual required transmission time slots will necessarily be contained within the TA envelope. Therefore, as long as the transmitter conservatively locks all time slots in the TA envelope (i.e., does not schedule other transmission or reception tasks on these time slots), it can ensure that the receiver receives the required transmission time slots. (For example, it can be a specified receiving time slot, which can also be called the target receiving time slot) The receiving process is not affected by time delay drift.

[0078] In the static case (the value range degenerates into a single point). When a node moves, the propagation delay can traverse at most an integer multiple of its length within a given range. The number of boundaries is Therefore, the number of elements in the TA envelope does not exceed the sum of the static case and the number of cross-boundaries, as shown in Equation 1.4.

[0079] (1.4) For a typical backbone network system, there are usually: Obviously at this time This indicates that, under typical ground or low-altitude backbone network parameters, the TA envelope under this mechanism can cover all time delay drift uncertainties with a maximum of 3 time slots, and the amount of redundancy required is very small.

[0080] Optionally, a time-unfolded conflict diagram can be constructed. This integrates cross-timeslot scheduling constraints and delay uncertainty tolerance mechanisms into a graph theory framework. The vertex set is Each vertex Indicates activation of directed link And requires the data packet to be in the first... Each time slot is received. Therefore, the total size of the vertices is... .vertex The activation state corresponds to the Boolean scheduling variable. An edge is an undirected edge that connects two scheduling events that cannot occur simultaneously in physical terms. Depending on the source of the conflict, the edge set can be generated from two types of constraints (or more) namely half-duplex constraint edges and signal interference constraint edges.

[0081] Next, for any two vertices and If they involve a common physical node and their respective time slot sets overlap on the time axis, then a half-duplex conflict edge exists between them. The specific determination condition is: Let... The set of occupied time slots is , The set of occupied time slots is If any of the following conditions are true, then in and Add conflicting edges between them: (The same node acts as the sender of two different received events), and ; (The same node acts as the receiver for two different transmitted events), and ; (The sender of one event is the receiver of another event), and or .

[0082] Due to TA envelope This decision criterion, which already includes all possible occupied time slots within the delay drift interval, guarantees that as long as the time slot is in the initial scheduling table... and If there is no boundary between them (i.e. they are determined to be conflict-free), then no half-duplex conflict will occur between them during the entire scheduling cycle, regardless of how the actual propagation delay drifts.

[0083] In directional millimeter-wave transmission, although spatial beam isolation significantly reduces inter-link interference, concurrent transmission in beam intersection areas can still generate non-negligible fractional interference. (Fractional interference coefficient) Indicates link Transmitted signals and links The arrival time window at the receiving end The time overlap ratio at each point. Through derivation, when... and The fractional interference coefficient can be uniquely determined by the difference between the receiving time slots and the difference between the propagation delays of the two events, i.e., Equation 1.5.

[0084] (1.5) In dynamic scenarios, propagation delay and All of these vary within their respective drift intervals. To ensure the robustness of the scheduling table in dynamic scenarios, the scheme provided in the embodiments of this specification uses the fractional interference coefficient under the worst-case scenario. Equation 1.6 serves as the basis for generating interference edges.

[0085] (1.6) Based on this worst-case interference coefficient, the link In the time slot The lower bound of the signal-to-interference-plus-noise ratio (SIR) during reception is shown in Equation 1.7.

[0086] (1.7) like Below the preset demodulation threshold (i.e., the demodulation threshold mentioned earlier), then at the vertex With all related Strong interference conflict edges are added between them to indicate that the two cannot coexist in the same scheduling cycle. When multiple interference sources cause the SINR to fall below the threshold, the interfered vertex and the corresponding set of interference source vertices can be combined to form a mutual exclusion clique constraint.

[0087] In undirected conflict graphs In this context, a clique is a complete subgraph where every two vertices within the clique are connected by an edge. A maximal clique is a clique that is not contained within any larger clique. Since any two scheduling events within a clique are mutually exclusive, at most one vertex in each clique can be activated. Therefore, the necessary and sufficient condition for conflict-free scheduling across the entire network can be equivalently expressed as Equation 1.8.

[0088] (1.8) in for The set of all maximal cliques in the set. This set of inequalities constitutes the conflict-free constraint for maximal cliques, which strictly guarantees the feasibility of physical link scheduling.

[0089] Building upon the aforementioned physical layer constraints, the solutions provided in the embodiments of this specification further introduce model constraints for upper-layer deterministic services, constructing a global optimization problem of joint admission and scheduling. Assume the network carries a set of constant bit-rate service flows. Each stream Model as tuples These represent the single fixed routing path of the flow, end-to-end rate requirement, end-to-end worst-case latency requirement, priority weight, and access indication, respectively. Define continuous variables. Represents a stream In the link The The actual rate allocated to each receive time slot, and an auxiliary Boolean variable is introduced. Indicates whether the time slot has been spun. Occupancy. The optimization objective is to maximize the weighted utility sum of the access business flow, as shown in Equation 1.9.

[0090] (1.9) The constraints include: Service flow rate constraint: For admitted flows, the sum of the timeslot rates of each hop on the path must meet the end-to-end rate requirement, i.e. .

[0091] Single-link capacity constraint: The total rate allocated to each link in each time slot shall not exceed its physical rate limit, i.e. .

[0092] Transmission timing constraints: On a multi-hop path, data packets must be received at an intermediate node before being forwarded to the next hop. That is, for two consecutive hops... and If the data packet is in the time slot quilt Receive, in time slot Depend on To issue, it must meet the following requirements. .

[0093] Worst-case delay constraint: Considering propagation delay, queuing delay, and reception / processing delay, the stream... The total time slot span from the source node's transmission to the destination node's reception must not exceed its requested latency limit. This constraint can be uniformly expressed as: for a path The time slot index difference between the first hop and the last hop satisfies the upper bound limit.

[0094] The aforementioned optimization problem belongs to mixed-integer nonlinear programming, and the number of maximal cliques in the conflict graph increases exponentially with the number of vertices in the worst case, posing a serious computational bottleneck to direct solutions. The embodiments in this specification propose the following acceleration strategies.

[0095] Firstly, in the initial conflict graph construction phase, only conflict edges based on half-duplex constraints are generated, while all signal interference constraint edges are temporarily ignored. This results in a relaxed initial model. During the branch-and-bound search process of the solver (such as Gurobi), a post-hoc interference check is performed whenever a feasible solution is found. For all activated scheduling events in this solution, the actual signal-to-interference-plus-noise ratio (SNR) at the receiver is calculated one by one. If the SNR at all receivers is not lower than a threshold... If the feasible solution is retained as a candidate feasible solution, the algorithm terminates when the solver completes the global optimality proof and no further constraint violations are found, and outputs the final scheduling result. If the signal-to-interference-plus-noise ratio (SNR) at a receiver is below the threshold, the set of interfering vertices is identified. This set, together with the interfered vertices, constitutes a constraint-violation clique. This clique is dynamically added to the main model as a new maximal clique constraint, and the solver is triggered to resolve the problem.

[0096] This strategy leverages the high spatial isolation of directional antennas, meaning that most potential interference combinations will not be activated simultaneously in actual scheduling. By solving first, verifying later, and adding as needed, the constraint size and memory usage of the main model are significantly reduced.

[0097] Secondly, before constructing the optimization model, the state space of the variables is pre-pruned using deterministic time delay constraints. For business flows... If a certain intermediate node on the path receives a time slot With subsequent hops in the transmission time slot The queuing time between them, plus the accumulated propagation delay, has exceeded the flow. End-to-end latency limit Then the corresponding combination of variables In the preprocessing stage, they are marked as infeasible and removed from the variable domain, further shortening the solution time.

[0098] Next, the process of performing a complete superframe scheduling table calculation and distribution in a dynamic wireless backbone managed by a centralized controller (such as an SDN (Software-Defined Networking) controller or network management system) is described. This process can be executed once before the start of each scheduling period (e.g., 0.01 seconds) to generate the network-wide scheduling table for that period. The specific flow is as follows.

[0099] First, in the initial step, before scheduling begins, the central controller collects the following information via the control channel: the three-dimensional position coordinates of each node. (This can be obtained via GNSS (Global Navigation Satellite System) or network positioning); the maximum speed of each node. (This information can be obtained from node capability reports or historical trajectory statistics, or theoretical values ​​can be used); Channel status information for each link, including whether it is in the LoS (Line of Sight) state, path loss index, antenna beam pointing, etc.; Deterministic service flow information to be accessed, including source and destination node pairs, routing paths, rate requirements, latency requirements, and priority weights.

[0100] Next, in the second step, the controller can calculate all Loss of Position (LoS) links based on the current position and reference motion velocity information collected in the previous step. The propagation delay interval within the current scheduling cycle is given by Equation 1.10.

[0101] (1.10) Among them, if If , then it is truncated to 0.

[0102] Next, in the third step, we can traverse the entire network link set. and superframe time slot set Generate all Each vertex. Corresponding to a Boolean scheduling variable , indicating link Is it in the first Data reception is completed in one time slot.

[0103] Then, in the fourth step, we can do this for each vertex. Based on the value range obtained in the second step, calculate the TA envelope, i.e., Equation 1.11.

[0104] (1.11) In practice, calculations can be performed more efficiently using the two endpoints of the interval.

[0105] Next, in step five, half-duplex constraint edges can be generated. For example, by traversing all vertex pairs... ,in If one of the following conditions is met, add a half-duplex conflict edge between the vertex pairs: Condition 1 (Send conflict within the same node): and ; Condition 2 (Conflict in reception within the same node): and ; Condition 3 (Node send / receive conflict): ( or )and( or ).

[0106] Then, in step six, the initial relaxation model can be constructed and prepared for solution. In the initial stage, no signal interference constraint edges are added, and maximal clique constraints derived from the half-duplex conflict graph are constructed (all maximal cliques can be extracted using graph theory algorithms such as Bron-Kerbosch), and combined with traffic flow constraints to form an initial mixed-integer programming model. At the same time, state space pruning is performed: for each traffic flow path, all possible combinations of receiving time slots are traversed, and those time slot combinations that still exceed the end-to-end delay limit after considering the accumulated propagation delay, necessary receiving delay, and minimum queuing time between adjacent hops are eliminated.

[0107] In step seven, the main loop for solving the delayed constraints can be executed. At this point, the initial model can be fed into a mixed-integer programming solver (such as Gurobi). The solver searches for the optimal solution in the branch-and-bound tree. Whenever the solver finds a feasible solution (i.e., all...)... and When all integers are present, execute the following verification subroutine: for all feasible solutions... Calculate the vertices of the receiver one by one. In the time slot The signal-to-interference-plus-noise ratio (SIR) at the specified location. The worst-case fractional interference coefficient is used. (As mentioned above); if the signal-to-interference-plus-noise ratio (SIR) at all receivers is not lower than the threshold If the solution is accepted as the final scheduling result, proceed to step eight; if the signal-to-interference-plus-noise ratio (SNR) of a receiver is below a threshold, identify all interference source vertices that interfere with that receiver (i.e., the corresponding...). Vertices). These source vertices and the disturbed vertices together form a new clique constraint, which is added to the solver, and the solution continues.

[0108] Next, in step eight, the scheduling table can be generated and distributed. That is, based on the final obtained Boolean variable... The values ​​of are used to generate the superframe scheduling matrix, which is the node scheduling scheme mentioned earlier. For each node, the scheduling scheme indicates whether it should be in a transmitting, receiving, or idle state in each time slot, as well as the corresponding beam direction. The central controller distributes the scheduling table to each node through the control channel.

[0109] Finally, in the ninth step, during the next scheduling cycle, each node performs data transmission and reception operations according to the issued scheduling table. Since the scheduling table has already covered all possible delay drifts within the cycle through TA envelope locking and worst-case interference calculation during the generation phase, nodes do not need to reconstruct the MAC layer scheduling table for real-time changes in propagation delay during execution, thus ensuring conflict-free transmission and deterministic delay satisfaction. When approaching each TA envelope, the distance is calculated in real-time, and the packet transmission time is determined; at this point, the packet transmission time is naturally already included in the TA envelope.

[0110] In addition to the above-described implementation methods, for scenarios with a very large number of nodes or limited control channel bandwidth, the methodology of the solutions provided in the embodiments of this specification can also guide the design of distributed scheduling algorithms as an alternative implementation method.

[0111] In this variant, the network-wide time-unfolded conflict graph can be divided into spatial regions or node clusters. Each cluster head node is responsible for constructing and solving the local conflict graph within its cluster, and adjacent clusters coordinate through boundary constraints. The calculation of the TA envelope is decentralized to each transmitting node: after receiving the receive time slot allocated by the cluster head, each node calculates and locks its local transmit time slot based on the estimated distance variation range with the destination node. Specifically: First, the entire network is divided into multiple clusters, and each cluster elects a cluster head node. Next, the cluster head collects information about the nodes within the cluster and constructs a local time unfolding conflict graph within the cluster. Then, the clusters coordinate the time slot occupancy of the boundary links through message exchange. Then, each cluster head solves the local scheduling problem and sends the results to the members within the cluster. Finally, after receiving the allocated receive time slots, the member nodes locally calculate the TA envelope and configure the local transceiver state machine.

[0112] In summary, this specification, through dynamic latency drift tolerance and utilizing the TA envelope locking mechanism, covers continuous propagation latency drift within the scheduling cycle at the minimal cost of adding at most a few redundant time slots. Even under conditions of high-speed node movement and long-term non-reconstruction of the MAC layer, it maintains an extremely high on-time delivery rate for admission services, effectively solving the core problem of scheduling tables failing when drifting across time slot boundaries under static latency assumptions. Furthermore, the implementation method provided in this specification also possesses robust interference management capabilities. By deriving and employing the worst-case fractional interference coefficient to construct conflict edges, the scheduling table can still ensure that inter-link interference constraints are not breached in dynamic scenarios. This avoids transmission failures caused by inaccurate interference prediction due to delay drift. Furthermore, this specification implements deterministic end-to-end delay guarantees by unifying cross-slot scheduling and service flow end-to-end delay constraints through a time-expanded conflict graph model. It also introduces joint service admission optimization, enabling adaptive admission judgment in high-load dynamic networks and providing strict hard delay guarantees for admitted services. Finally, this specification also delivers higher resource efficiency and lower latency. Compared to traditional solutions that insert uniform protection intervals across the entire network, this specification significantly reduces average end-to-end latency and maintains deterministic on-time delivery at the cost of controllable redundant time slots.

[0113] Based on the same concept as the method described above, a second aspect of this specification provides a task scheduling device applied to a management node in a node network. Please refer to [link to relevant documentation]. Figure 4 The device may include: The task acquisition module 401 is used to acquire multiple communication tasks to be scheduled in the node network within the scheduling period; Envelope generation module 402 is used to generate TA envelope for each communication task based on the propagation delay drift interval between the sending end and the receiving end. The TA envelope is the union of all transmission time slots required by the sending end under all possible delay values ​​in the propagation delay drift interval when the goal of the communication task is to complete the reception in a specified reception time slot. The conflict determination module 403 is used to determine whether there is a conflict between any two communication tasks based on the TA envelope; The node scheduling module 404 is used to determine a node scheduling scheme that satisfies all conflict constraints based on the result of the conflict determination.

[0114] Optionally, when the conflict determination module determines whether a conflict exists between any two communication tasks based on the TA envelope, it is used to: Based on the TA envelope, determine whether there is a half-duplex conflict and / or signal interference conflict between any two communication tasks.

[0115] Optionally, the half-duplex conflict includes at least one of the following: The two communication tasks are sent from the same node, and their TA envelopes overlap on the time axis. The receiving end of the two communication tasks is the same node, and their designated receiving time slots are the same; The sender of one communication task is the receiver of another communication task, and the TA envelope of the sender covers the designated receive time slot of the other communication task.

[0116] Optionally, the device further includes a signal interference module for: Based on the propagation delay drift interval of the first communication task and the propagation delay drift interval of the second communication task, the time overlap ratio on the time axis between the transmission signal of the first communication task arriving at the receiving end of the second communication task and the designated receiving time slot of the second communication task is determined, wherein the first communication task and the second communication task are any two communication tasks among the plurality of communication tasks. Based on the time overlap ratio, the signal-to-interference-plus-noise ratio (SINR) of the receiving end of the second communication task when it is interfered with by the first communication task is determined. If the SINR is lower than a preset demodulation threshold, then it is determined that there is a signal interference conflict between the first communication task and the second communication task.

[0117] Optionally, the signal interference conflict module is used to determine the time overlap ratio between the transmitted signal of the first communication task arriving at the receiving end of the second communication task and the designated receiving time slot of the second communication task on the time axis, for the following purposes: Based on the current position and reference speed of the sending end of the first communication task, and the current position and reference speed of the receiving end of the second communication task, the propagation delay drift interval between the sending end of the first communication task and the receiving end of the second communication task is determined. Multiple delay combinations are obtained, the delay combination including a first delay and a second delay, wherein the first delay is located within the propagation delay drift interval of the first communication task, and the second delay is located within the propagation delay drift interval between the sending end of the first communication task and the receiving end of the second communication task; For each delay combination, based on the specified receive time slot of the first communication task and the first delay in the delay combination, the transmit time window of the first communication task under the delay combination is determined; And, based on the transmission time window and the second delay, determine the arrival time window of the transmission signal of the first communication task at the receiving end of the second communication task; Furthermore, based on the overlap duration between the arrival time window and the specified reception time slot of the second communication task on the time axis, the candidate overlap ratio under the current delay combination is determined; The maximum value among the candidate overlap ratios corresponding to each delay combination is determined as the time overlap ratio on the time axis between the sending signal of the first communication task and the designated receiving time slot of the second communication task when the sending signal of the first communication task arrives at the receiving end of the second communication task.

[0118] Optionally, when the node scheduling module determines a node scheduling scheme that satisfies all conflict constraints based on the conflict determination result, it is used to: A set of mutually exclusive tasks is determined among the plurality of communication tasks, wherein any two communication tasks included in the set of mutually exclusive tasks conflict with each other. Based on a preset solution model, a preset optimization objective is solved under constraints to obtain the node scheduling scheme of the node network within the scheduling period. The constraints include: In each set of mutually exclusive tasks, at most one communication task is active.

[0119] Optionally, when determining the set of mutually exclusive tasks among the plurality of communication tasks, the node scheduling module is used to: Using each of the multiple communication tasks as a vertex, conflict edges are established between the vertices corresponding to two conflicting communication tasks to obtain a conflict graph. The set of vertices contained in each maximal clique in the conflict graph is determined as a set of mutually exclusive tasks.

[0120] Optionally, the node scheduling module is used to solve a preset optimization objective based on a preset solution model under constraints to obtain a node scheduling scheme for the node network within the scheduling period, and is used to: Based on the preset solution model, the preset optimization objective is solved under the constraints of the current iteration round to obtain the temporary scheduling scheme for the current iteration round. For each task pair, determine whether there is signal interference conflict between the two communication tasks in the task pair, wherein the task pair is two communication tasks activated in the temporary scheduling scheme of the current iteration round, and the arrival time window of the transmission signal of one communication task may overlap with the designated reception time slot of the other communication task. In the event of signal interference conflict between two communication tasks within at least one task pair, a new set of mutually exclusive tasks is determined among the plurality of communication tasks, and the next iteration round is initiated. If there is no signal interference conflict between the two communication tasks within each task pair, the temporary scheduling scheme of the current iteration round is determined as the node scheduling scheme of the node network within the scheduling period.

[0121] Optionally, the propagation delay drift interval is derived based on the current positions of the transmitting end and the receiving end and a reference motion speed, wherein the lower limit of the propagation delay drift interval corresponds to the transmitting end and the receiving end moving towards each other at the reference motion speed, and the upper limit corresponds to moving away from each other at the reference motion speed.

[0122] Optionally, before determining the node scheduling scheme of the node network within the scheduling period, the device further includes a task rejection module, used for: For each of the plurality of communication tasks, the upper limit of the latency corresponding to the communication task is obtained, and the expected communication delay is determined according to the propagation delay drift interval corresponding to the communication task. If the sum of the queuing waiting time of the communication task at the sending end and the expected communication delay is greater than the upper limit of the latency, it is determined that the communication task will not enter the current scheduling cycle.

[0123] Based on the same concept as the above method, this specification also provides an electronic device, including: a processor; a memory for storing processor-executable instructions; wherein the processor performs the steps of the method as described in the first aspect by executing the executable instructions.

[0124] For example, Figure 5 This is a schematic structural diagram of a device provided in an exemplary embodiment. Please refer to... Figure 5 At the hardware level, the device includes a processor 502, an internal bus 504, a network interface 506, memory 508, and non-volatile memory 510, and may also include other hardware required for its functions. One or more embodiments of this specification can be implemented in software, for example, the processor 502 reads the corresponding computer program from the non-volatile memory 510 into memory 508 and then runs it. Of course, in addition to software implementation, one or more embodiments of this specification do not exclude other implementation methods, such as logic devices or a combination of hardware and software, etc. That is to say, the execution subject of the following processing flow is not limited to each logic unit, but can also be hardware or logic devices.

[0125] Based on the same concept as the methods described above, this specification also provides a computer-readable storage medium having computer instructions stored thereon that, when executed by a processor, implement the steps of the methods as described in any of the above embodiments.

[0126] Based on the same concept as the methods described above, this specification also provides a computer program product, including a computer program / instructions that, when executed by a processor, implement the steps of the methods as described in any of the above embodiments.

Claims

1. A task scheduling method, characterized in that, The method, applied to a management node in a node network, includes: Obtain multiple communication tasks to be scheduled within the scheduling period of the node network; For each communication task, a TA envelope is generated based on the propagation delay drift interval between the sender and receiver. The TA envelope is the union of all transmission time slots required by the sender under all possible delay values ​​within the propagation delay drift interval when the goal of the communication task is to complete reception in a specified reception time slot. Based on the TA envelope, determine whether there is a conflict between any two of the communication tasks; Based on the results of the conflict determination, a node scheduling scheme that satisfies all conflict constraints is determined.

2. The task scheduling method according to claim 1, characterized in that, The step of determining whether a conflict exists between any two communication tasks based on the TA envelope includes: Based on the TA envelope, determine whether there is a half-duplex conflict and / or signal interference conflict between any two communication tasks.

3. The task scheduling method according to claim 2, characterized in that, The half-duplex conflict includes at least one of the following: The two communication tasks are sent from the same node, and their TA envelopes overlap on the time axis. The receiving end of the two communication tasks is the same node, and their designated receiving time slots are the same; The sender of one communication task is the receiver of another communication task, and the TA envelope of the sender covers the designated receive time slot of the other communication task.

4. The task scheduling method according to claim 2, characterized in that, The method further includes: Based on the propagation delay drift interval of the first communication task and the propagation delay drift interval of the second communication task, the time overlap ratio on the time axis between the transmission signal of the first communication task arriving at the receiving end of the second communication task and the designated receiving time slot of the second communication task is determined, wherein the first communication task and the second communication task are any two communication tasks among the plurality of communication tasks. Based on the time overlap ratio, the signal-to-interference-plus-noise ratio (SINR) of the receiving end of the second communication task when it is interfered with by the first communication task is determined. If the SINR is lower than a preset demodulation threshold, then it is determined that there is a signal interference conflict between the first communication task and the second communication task.

5. The task scheduling method according to claim 4, characterized in that, The determination of the time overlap ratio on the time axis between the transmission signal of the first communication task arriving at the receiving end of the second communication task and the designated receiving time slot of the second communication task includes: Based on the current position and reference speed of the sending end of the first communication task, and the current position and reference speed of the receiving end of the second communication task, the propagation delay drift interval between the sending end of the first communication task and the receiving end of the second communication task is determined. Multiple delay combinations are obtained, the delay combination including a first delay and a second delay, wherein the first delay is located within the propagation delay drift interval of the first communication task, and the second delay is located within the propagation delay drift interval between the sending end of the first communication task and the receiving end of the second communication task; For each delay combination, based on the specified receive time slot of the first communication task and the first delay in the delay combination, a transmit time window for the first communication task under the delay combination is determined; and, Based on the transmission time window and the second delay, determine the arrival time window of the transmission signal of the first communication task at the receiving end of the second communication task; and, Based on the overlap duration between the arrival time window and the specified reception time slot of the second communication task on the time axis, the candidate overlap ratio under the current delay combination is determined. The maximum value among the candidate overlap ratios corresponding to each delay combination is determined as the time overlap ratio on the time axis between the sending signal of the first communication task and the designated receiving time slot of the second communication task when the sending signal of the first communication task arrives at the receiving end of the second communication task.

6. The task scheduling method according to claim 1, characterized in that, The step of determining a node scheduling scheme that satisfies all conflict constraints based on the conflict determination result includes: A set of mutually exclusive tasks is determined among the plurality of communication tasks, wherein any two communication tasks included in the set of mutually exclusive tasks conflict with each other. Based on a preset solution model, a preset optimization objective is solved under constraints to obtain a node scheduling scheme for the node network within the scheduling period. The constraints include: in each set of mutually exclusive tasks, at most one communication task is activated.

7. The task scheduling method according to claim 6, characterized in that, Determining the set of mutually exclusive tasks among the plurality of communication tasks includes: Using each of the multiple communication tasks as a vertex, conflict edges are established between the vertices corresponding to two conflicting communication tasks to obtain a conflict graph. The set of vertices contained in each maximal clique in the conflict graph is determined as a set of mutually exclusive tasks.

8. The task scheduling method according to claim 6, characterized in that, The method, based on a preset solution model, solves for a preset optimization objective under constraints to obtain a node scheduling scheme for the node network within the scheduling period, including: Based on the preset solution model, the preset optimization objective is solved under the constraints of the current iteration round to obtain the temporary scheduling scheme for the current iteration round. For each task pair, determine whether there is signal interference conflict between the two communication tasks in the task pair, wherein the task pair is two communication tasks activated in the temporary scheduling scheme of the current iteration round, and the arrival time window of the transmission signal of one communication task may overlap with the designated reception time slot of the other communication task. In the event of signal interference conflict between two communication tasks within at least one task pair, a new set of mutually exclusive tasks is determined among the plurality of communication tasks, and the next iteration round is initiated. If there is no signal interference conflict between the two communication tasks within each task pair, the temporary scheduling scheme of the current iteration round is determined as the node scheduling scheme of the node network within the scheduling period.

9. The task scheduling method according to claim 1, characterized in that, The propagation delay drift interval is derived based on the current positions of the transmitting end and the receiving end and the reference motion speed. The lower limit of the propagation delay drift interval corresponds to the transmitting end and the receiving end moving towards each other at the reference motion speed, and the upper limit corresponds to moving away from each other at the reference motion speed.

10. The task scheduling method according to claim 1, characterized in that, Before determining the node scheduling scheme of the node network within the scheduling period, the method further includes: For each of the plurality of communication tasks, the upper limit of the latency corresponding to the communication task is obtained, and the expected communication delay is determined according to the propagation delay drift interval corresponding to the communication task. If the sum of the queuing waiting time of the communication task at the sending end and the expected communication delay is greater than the upper limit of latency, it is determined that the communication task will not enter the current scheduling cycle. The not entering the current scheduling cycle includes discarding the communication task or postponing the scheduling of the communication task.

11. A task scheduling device, characterized in that, A management node applied to a node network, the device comprising: The task acquisition module is used to acquire multiple communication tasks to be scheduled in the node network within the scheduling period; An envelope generation module is used to generate a TA envelope for each communication task based on the propagation delay drift interval between the sender and receiver. The TA envelope is the union of all transmission time slots required by the sender under all possible delay values ​​within the propagation delay drift interval when the goal of the communication task is to complete reception in a specified reception time slot. The conflict determination module is used to determine whether there is a conflict between any two communication tasks based on the TA envelope; The node scheduling module is used to determine a node scheduling scheme that satisfies all conflict constraints based on the result of the conflict determination.

12. An electronic device, characterized in that, include: processor; A memory for storing processor-executable instructions; wherein the processor implements the steps of the method as described in any one of claims 1-10 by executing the executable instructions.

13. A computer-readable storage medium, characterized in that, It stores computer instructions that, when executed by a processor, implement the steps of the method as described in any one of claims 1-10.

14. A computer program product, characterized in that, Includes a computer program / instructions that, when executed by a processor, implement the steps of the method as described in any one of claims 1-10.

Images