A drone-assisted integrated resource allocation method for forest fire fighting.

Abstract

This invention provides a method and apparatus for drone-assisted integrated sensing, computing, and communication resource allocation in forest fire fighting, relating to the field of communication technology. The method includes: establishing a drone-assisted ISCC system scenario; constructing a drone-assisted ISCC system model based on the established system scenario; determining the total delay and energy consumption for the drone-assisted ISCC system model to complete its tasks based on the model; constructing an optimization objective for the drone-assisted ISCC system model based on the total delay and energy consumption; constructing an optimization problem for task scheduling and resource allocation within the drone-assisted ISCC system model based on the optimization objective; modeling the optimization problem as a Markov decision process, and using the TD3 algorithm to optimize the task scheduling and resource allocation strategies to obtain the optimal task scheduling and resource allocation scheme. Using this invention can enhance real-time monitoring capabilities of fire situations and improve fire fighting speed and success rate.

Description

Technical Field

[0001] This invention relates to the field of communication technology, and in particular to a method and apparatus for resource allocation using drones to assist in forest fire fighting, integrating communication, sensing, and computing. Background Technology

[0002] The rapid development of drones marks a new era for the transportation and communication industries. Drones have demonstrated unique advantages, particularly in forest firefighting. Drones can be flexibly equipped with various sensors, including communication modules, lidar, and optical cameras. These sensors can collect vast amounts of information during a fire. Drones can serve as aerial communication base stations, providing temporary communication networks in disaster areas to ensure uninterrupted communication between rescue personnel and command centers. Furthermore, modern drones possess powerful onboard computing capabilities, enabling them to share the computing tasks of other IoT devices. The antennas on drones can also transmit signals to fire targets via radar and receive echoes, achieving precise wireless sensing and positioning. Drone-assisted integrated sensing, computing, and communication (ISCC) systems are becoming a cutting-edge field in critical tasks such as forest firefighting. Drones assisting ISCC systems are equipped with high-performance onboard servers for data analysis, advanced communication antennas for real-time information transmission, and highly sensitive wireless sensors for environmental monitoring, collectively forming a wireless platform capable of performing multiple tasks over a wide area. Drone-assisted ISCC systems can quickly locate fire sources and assess the speed and direction of fire spread when a fire breaks out, providing crucial decision support for rescue teams. Furthermore, after a fire is brought under control, it can continuously monitor the fire site to ensure no fire spots are missed, preventing reignition and effectively protecting forest resources and reducing ecological damage. In addition, drone-assisted ISCC systems can also play a role in routine forest fire prevention management. By regularly patrolling forest areas, they can promptly identify and report potential fire risks, such as abnormal temperature increases or illegal activities, enabling preventative measures to be taken to avoid fires.

[0003] Currently, joint task scheduling and resource allocation in UAV-assisted ISCC systems for forest fire fighting scenarios is a crucial and challenging problem. Existing methods for joint task scheduling and resource allocation in UAV-assisted ISCC systems include: time division multiple access (TDMA), the first-sense-then-offload protocol, and space division multiple access (SDMA). These methods neglect the mutual interference between task computation and offloading and fire target perception. The simultaneous execution of task scheduling and perception functions may lead to resource contention and performance degradation, affecting the overall system functionality. Summary of the Invention

[0004] To address the technical challenges of joint task scheduling and resource allocation in existing UAV-assisted ISCC systems for forest fire fighting scenarios, this invention provides a UAV-assisted integrated sensor-computer resource allocation method and apparatus for forest fire fighting. The technical solution is as follows:

[0005] On the one hand, a method for drone-assisted integrated sensory-computing resource allocation for forest fire fighting is provided. This method is implemented by a drone-assisted integrated sensory-computing resource allocation device for forest fire fighting, and includes:

[0006] S1. Set up a scenario for a drone-assisted ISCC system;

[0007] S2. Construct a drone-assisted ISCC system model based on the aforementioned drone-assisted ISCC system scenario;

[0008] S3. Based on the UAV-assisted ISCC system model, determine the total delay and energy consumption for the UAV-assisted ISCC system model to complete the task;

[0009] S4. Based on the total delay and energy consumption, construct the optimization objective of the UAV-assisted ISCC system model;

[0010] S5. Based on the optimization objective, construct the optimization problem of task scheduling and resource allocation in the unmanned aerial vehicle-assisted ISCC system model;

[0011] S6. Model the optimization problem as a Markov decision process, and use the TD3 algorithm to optimize the task scheduling and resource allocation strategies in the UAV-assisted ISCC system model to obtain the optimal task scheduling and resource allocation scheme.

[0012] On the other hand, a drone-assisted integrated sensory-computing resource allocation device for forest fire fighting is provided. This device is applied to a drone-assisted integrated sensory-computing resource allocation method for forest fire fighting. The device includes:

[0013] The first building unit is used to build a drone-assisted ISCC system scenario;

[0014] The second construction unit is used to construct a drone-assisted ISCC system model based on the drone-assisted ISCC system scenario.

[0015] The determining unit is used to determine the total delay and energy consumption of the UAV-assisted ISCC system model in completing the task, based on the UAV-assisted ISCC system model.

[0016] The third construction unit is used to construct the optimization objective of the UAV-assisted ISCC system model based on the total delay and energy consumption;

[0017] The fourth construction unit is used to construct the optimization problem of task scheduling and resource allocation in the unmanned aerial vehicle-assisted ISCC system model according to the optimization objective.

[0018] The acquisition unit is used to model the optimization problem as a Markov decision process, and to optimize the task scheduling and resource allocation strategies in the UAV-assisted ISCC system model using the TD3 algorithm to obtain the optimal task scheduling and resource allocation scheme.

[0019] On the other hand, a drone-assisted integrated sensory computing resource allocation device for forest fire fighting is provided. The drone-assisted integrated sensory computing resource allocation device for forest fire fighting includes: a processor; a memory, wherein the memory stores computer-readable instructions, and when the computer-readable instructions are executed by the processor, any one of the methods described above for drone-assisted integrated sensory computing resource allocation for forest fire fighting is implemented.

[0020] On the other hand, a computer-readable storage medium is provided, wherein at least one instruction is stored in the storage medium, the at least one instruction being loaded and executed by a processor to implement any of the methods in the above-described drone-assisted integrated sensory computing resource allocation method for forest fire fighting.

[0021] The beneficial effects of the technical solutions provided by the embodiments of the present invention include at least the following:

[0022] This invention first establishes a UAV-assisted ISCC system scenario; based on the UAV-assisted ISCC system scenario, a UAV-assisted ISCC system model is constructed; based on the UAV-assisted ISCC system model, the total latency and energy consumption for the UAV-assisted ISCC system model to complete the task are determined; secondly, based on the total latency and energy consumption, an optimization objective for the UAV-assisted ISCC system model is constructed; based on the optimization objective, an optimization problem for task scheduling and resource allocation in the UAV-assisted ISCC system model is constructed; finally, the optimization problem is modeled as a Markov decision process, and the TD3 algorithm is used to optimize the task scheduling and resource allocation strategies in the UAV-assisted ISCC system model to obtain the optimal task scheduling and resource allocation scheme.

[0023] This invention addresses the applicability of uplink ISCC in handling computationally intensive, data-diverse, and latency-sensitive forest firefighting missions by constructing an optimization model. It focuses on minimizing the weighted sum of computational offloading and perception penalty indices, while also considering various practical operational constraints, including feasible flight zones and UAV energy constraints, thus optimizing the efficiency and effectiveness of uplink ISCC. Furthermore, this invention reconstructs the task scheduling and resource allocation problem as a Markov decision process, employing the TD3 algorithm to address the joint scheduling and resource allocation of tasks in UAV-assisted ISCC systems, effectively improving task processing speed and environmental perception quality. Finally, this invention provides robust aerial support for forest firefighting, enhancing real-time fire monitoring capabilities, data processing speed, and resource optimization, thereby improving the response speed and success rate of firefighting operations. Attached Figure Description

[0024] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0025] Figure 1 This is a flowchart of a drone-assisted integrated sensory computing resource allocation method for forest fire fighting provided by an embodiment of the present invention;

[0026] Figure 2 This is a structural schematic diagram of an example scenario of unmanned aerial vehicle-assisted ISCC provided by an embodiment of the present invention;

[0027] Figure 3 This is a flowchart of a forest fire fighting algorithm based on TD3 provided in an embodiment of the present invention;

[0028] Figure 4 This is a diagram illustrating the iteration of the objective function provided in an embodiment of the present invention;

[0029] Figure 5 This is an iterative diagram of the cumulative reward item provided in an embodiment of the present invention;

[0030] Figure 6 This is a 2D flight trajectory diagram of a drone in a forest fire fighting scenario provided by an embodiment of the present invention;

[0031] Figure 7 This is a 3D flight trajectory diagram of a drone in a forest fire fighting scenario provided by an embodiment of the present invention;

[0032] Figure 8This is a block diagram of a drone-assisted integrated sensory computing resource allocation device for forest fire fighting provided by an embodiment of the present invention;

[0033] Figure 9 This is a schematic diagram of the structure of a drone-assisted integrated sensory computing resource allocation device for forest fire fighting provided by an embodiment of the present invention. Detailed Implementation

[0034] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0035] In embodiments of the present invention, words such as "exemplarily," "for example," etc., are used to indicate that something is an example, illustration, or description. Any embodiment or design described as "exemplary" in the present invention should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the word "exemplary" is intended to present the concept in a concrete manner. Furthermore, in embodiments of the present invention, the meaning expressed by "and / or" can be both, or either one.

[0036] In the embodiments of this invention, the terms "image" and "picture" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning. Similarly, the terms "of," "corresponding (relevant)," and "corresponding" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning.

[0037] In this embodiment of the invention, sometimes a subscript such as W1 may be written in a non-subscript form such as W1. When the difference is not emphasized, the meaning they express is the same.

[0038] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0039] This invention provides a method for drone-assisted integrated sensory computing resource allocation in forest fire fighting. This method can be implemented using a drone-assisted integrated sensory computing resource allocation device for forest fire fighting, which can be a terminal or a server. Figure 1 The flowchart shown is for a drone-assisted integrated sensor-computer resource allocation method for forest fire fighting. The processing flow of this method may include the following steps:

[0040] 1. A method for resource allocation using drones assisted by integrated sensing and computing for forest fire fighting, characterized in that the method includes:

[0041] S1. Set up a scenario for an unmanned aerial vehicle-assisted ISCC system.

[0042] Optionally, the scenarios for drone-assisted ISCC systems include: base stations, IoT devices, sensing targets, and drones;

[0043] The base station is used to receive the results of the drone's computational tasks.

[0044] IoT devices are used to collect environmental information, generate tasks, and send tasks to drones; IoT devices are also used to communicate with drones and coordinate task unloading.

[0045] Within a time slot, IoT devices can generate tasks that need to be processed and offload these tasks to the drone. These IoT devices are responsible for collecting environmental data, generating tasks, and sending them to the drone. IoT devices can also participate in the drone's communication to coordinate the task offloading and result transmission process.

[0046] In one feasible implementation, IoT devices can collect environmental information, including temperature, wind speed, humidity, and wind direction. Due to the limited onboard computing and storage capacity of IoT devices, ground-deployed IoT devices need to offload some computationally intensive tasks to drones performing tasks within the area when processing large amounts of environmental monitoring data. Each IoT device is equipped with an omnidirectional antenna to ensure communication with the drone in all directions; while each drone is equipped with a directional antenna to more accurately establish communication links with specific IoT devices or ground control centers.

[0047] Target perception, used to provide perception tasks for drones;

[0048] Drones are used to collect and aggregate information from Internet of Things (IoT) devices.

[0049] Within each time slot, the drone can perform a series of coordinated tasks, including receiving computational tasks from IoT devices, moving to approach different IoT devices as needed, and handling unloading tasks. After processing, the drone transmits the computational results to the base station. The drone is also responsible for receiving echo signals from IoT devices, which may contain environmental monitoring data to assist in real-time decision-making during forest fire fighting.

[0050] Among them, the scenario for the drone-assisted ISCC system is forest fire fighting. In the forest fire fighting scenario, drones work in collaboration with ground-deployed IoT devices to collect environmental information quickly and cost-effectively. This collaborative operation is crucial for the early detection and real-time monitoring of forest fires.

[0051] In this embodiment of the invention, a fixed-size monitoring area is set up. Within the fixed monitoring area, drones and IoT devices jointly perform monitoring tasks. In order to control and manage more effectively, this embodiment of the invention adopts a time discretization strategy, which divides the entire monitoring time range into multiple equal time periods or time slots corresponding to the duration of a fire event, which helps to achieve continuous tracking and precise control of fire dynamics.

[0052] In this embodiment of the invention, 5G next-generation wireless communication technology and vehicle-to-everything (V2X) technology are adopted. 5G next-generation wireless communication technology has high-speed and low-latency communication capabilities. V2X technology allows vehicles to communicate directly with other vehicles, pedestrians and infrastructure, and can be used to command and coordinate rescue vehicles, personnel and equipment in forest fire fighting.

[0053] In this invention, the drone continuously emits directional radar pulse signals to detect fire targets in the forest, including trees, bushes, houses, and sheds, all of which can affect the fire's spread. This embodiment of the invention separates the transmitting and receiving antennas on the drone, ensuring that the receiving antenna can always be used to receive transmissions from sensing echoes and IoT devices. Regarding the sensing task, this embodiment of the invention aims to maximize the success rate of sensing and fairness among fire targets. The drone will tend to visit areas with a high density of fire targets to improve the success rate of sensing; to ensure fairness, the drone will also sense less frequently visited fire targets. Less frequently visited fire targets include: fire targets in remote areas: fires located deep in forests or mountainous areas may be less frequently visited and monitored due to their remote geographical location; and hard-to-reach areas: certain areas that are difficult to access or reach, such as swamps and cliff edges, may be less frequently visited in fire monitoring due to their danger or difficulty in access.

[0054] In this context, drones and IoT devices can work closely together within a time slot. Drones utilize computing resources and mobility to support the data processing needs of IoT devices, while IoT devices provide data input and communication interaction with drones, jointly supporting emergency response missions such as forest fire fighting.

[0055] In one feasible implementation, multiple drones can be pre-deployed to a specific monitoring area at the start of a forest fire fighting event. These drones are prepared to respond to any unforeseen circumstances. The drones are programmed to continuously perform tasks throughout the monitoring period, without leaving the designated monitoring area, and no new drones can join the monitoring area during their missions. By planning unified departure and return points for the drones, charging is convenient, allowing them to remain operational throughout the mission period. The docking points are selected at the edge of the monitoring area to maximize the drones' coverage and enable them to respond quickly to emergencies within the area.

[0056] In one feasible implementation, in a drone-assisted ISCC system, drones and IoT devices can interact and offload tasks using 5G next-generation wireless communication technology and vehicle-to-everything (V2X) technology. During task offloading, each task of an IoT device will be offloaded to a maximum of one drone for processing. Each IoT device can select the most suitable drone to offload tasks from the IoT device based on the drone's computing power, current task load, distance, and other factors. In this way, drones can act as mobile edge computing nodes, providing additional computing resources and helping IoT devices process data, thus providing more comprehensive data support and decision-making basis for emergency tasks such as forest fire fighting.

[0057] In one feasible implementation, it is assumed that each IoT device collects environmental data at the same sampling rate, but the tasks generated by each IoT device differ in data volume and required computing resources. Since the tasks generated by each IoT device are independent, they can be processed in parallel, improving processing speed. This embodiment considers the differences in CPU and battery capacity among IoT devices, while drones have the same CPU and battery capacity. Therefore, this embodiment reserves necessary resources for the local task execution of IoT devices and uses the remaining resources for computational offloading. Task scheduling and resource allocation decisions can be centrally made by the cellular base station connected to the drone. Information sharing between the drone and the cellular base station, including the status information of the drone and nearby IoT devices, is achieved through 5G next-generation wireless communication technology and vehicle-to-everything (V2X) technology. The goal of this embodiment is to maximize the success rate of computational offloading and fairness among IoT devices. To improve the success rate of computational offloading, drones will tend to choose IoT devices with more tasks to offload. Simultaneously, to ensure fair service for all IoT devices, this embodiment introduces a penalty index to incentivize drones to access less frequently accessed IoT devices.

[0058] S2. Based on the scenario of the UAV-assisted ISCC system, construct a UAV-assisted ISCC system model.

[0059] Based on the drone-assisted ISCC system model, the location information of the drone, the fire target, and the IoT device can be obtained within the time slot.

[0060] The received signals of the UAV include: all communication signals of the unloaded IoT devices and the sensing echo signals emitted by the UAV and reflected by the fire target in the time slot. The received signals of the UAV are a linear combination of the communication signals and the sensing echo signals, which can be expressed by the following formula (1):

[0061] (1)

[0062] in, Indicates time, This indicates whether an IoT device offloads a task to a drone in time slot tn; it is assumed that each IoT device can only offload a computing task to one drone within an acceptable communication distance in each time slot. This indicates the channel power gain between IoT devices and drones; This represents the transmitted signal of an IoT device in time slot tn; This indicates whether the drone sensed the fire target within time slot tn; This indicates the channel power gain between the IoT device and the fire target; This represents the radar pulse signal of the UAV in time slot tn; Indicates transmission time; Indicates the round-trip testing time; among which, and The time is very short and can usually be ignored; Tolerate the maximum unloading delay.

[0063] S3. Based on the UAV-assisted ISCC system model, determine the total delay and energy consumption for the UAV-assisted ISCC system model to complete the task.

[0064] Optionally, the total latency of the S3 drone-assisted ISCC system model to complete the task includes: local latency, transmission latency, and drone computing latency;

[0065] The total latency is the maximum value between the local computing latency, the transmission latency, and the drone computing latency.

[0066] The local latency is determined by the local computing power and the computing speed of the IoT devices; the local computing power is the data size of the computing tasks that are not offloaded to the drone.

[0067] The computing speed is related to the CPU frequency of the IoT device and the number of CPU revolutions required to compute 1 bit for the IoT device's computing task.

[0068] The transmission delay depends on the amount of data offloaded and the data rate transmitted; the data rate transmitted is related to the available bandwidth allocated to each IoT device, the transmission power of the IoT device, the channel power gain between the IoT device and the drone, interference terms, and the power of noise.

[0069] The interference items include inter-user interference and perceived interference of echo signals.

[0070] The computational latency of the drone is determined by the amount of computation unloaded and the computational rate of the drone.

[0071] In this system, the CPU frequency of the drone edge server is evenly distributed to each IoT device that is offloading its tasks. The computational load is the size of the computational task offloaded to the drone, and the drone's computational speed is related to the CPU frequency of the drone edge server allocated to the task and the number of CPU revolutions required to compute 1 bit for that task.

[0072] Optionally, the energy consumption of the S3 drone-assisted ISCC system model in completing the task includes: the energy consumption of the IoT device and the energy consumption of the drone;

[0073] The energy consumption of IoT devices includes: the transmission energy of IoT devices and the computing energy of IoT devices;

[0074] Among them, the transmission energy of IoT devices is the product of the transmission power of IoT devices and the transmission delay of IoT devices;

[0075] Among them, the computing power of IoT devices is the product of constants related to the chip structure of IoT devices, the square of the CPU frequency of IoT devices, and the local computing latency.

[0076] The energy consumption of drones includes: energy consumption for mission computing, energy consumption for sensing fire targets, and energy consumption for propulsion.

[0077] The energy consumption of the UAV can be expressed by the following formula (2):

[0078] (2)

[0079] in, Indicates the energy consumption of the drone; This indicates the energy consumption of the task calculation; Indicates the energy consumption of the perceived fire target; This refers to the propulsion energy consumption of a drone while it is flying or hovering in the air.

[0080] Within a single time slot, the total number of radar pulses is a fixed value. Regardless of the number of detected fire targets, the UAV sends only one radar pulse to each detected fire target at a time. The energy consumption for detecting fire targets is the product of the UAV's radar pulse power, radar pulse duration, and the total number of radar pulses.

[0081] Among them, propulsion energy consumption can be expressed as ;in, The incremental correction factor represents the induced power. Indicates the weight of the drone. Indicates air density, This indicates the area of ​​the drone's rotor disk.

[0082] Among them, the energy consumption of a single UAV's task computing is the product of a constant related to the UAV chip architecture, the square of the CPU frequency of the UAV edge server, and the computing latency. The task computing energy consumption is the sum of the computing energy consumption of all UAV tasks.

[0083] S4. Based on the total delay and energy consumption, construct the optimization objective of the UAV-assisted ISCC system model;

[0084] Optionally, the optimization objectives of the S4 drone-assisted ISCC system model include: calculating the unloading penalty index and the perception penalty index.

[0085] In this embodiment of the invention, the goal of maximizing the fair successful computational unloading and perception ratio under constraints is considered. By employing a penalty function method for constraint optimization, the objective of maximizing the fair successful computational unloading and perception ratio is transformed into minimizing the computational unloading and perception penalty. The constraints include: feasible area, UAV mobility, radar power, transmission power, computing power, unloading ratio, and energy. To ensure the performance of computational unloading, the computational unloading penalty index is used as one of the optimization objectives. The computational unloading penalty index can be expressed by the following formula (3):

[0086] (3)

[0087] in, Indicates the total delay. This represents the threshold corresponding to the maximum delay. This indicates the successful computation and unloading status of the computation task. for , This indicates taking the minimum value. Represented as from 1 to The sum of whether IoT devices unload tasks from drones during time slots. This represents the energy consumption of the computational task. This represents the threshold corresponding to the maximum energy consumption; A value of 1 indicates Successfully offloaded computational data to the drone; This represents the penalty exponent for calculating the task in time slot tn-1. This represents the maximum value of the penalty index.

[0088] The smaller the penalty index, the more timely the offloading of computational tasks. IoT devices offload some computational tasks to drones, while the remainder is computed locally. This embodiment of the invention sets up local computational tasks and offloading tasks to be generated and executed in parallel simultaneously.

[0089] The perceived penalty index can be expressed by the following formula (4):

[0090] (4)

[0091] in, This represents the perception penalty index of the fire target in time slot tn-1; Indicates the duration of the radar pulse; Indicates the radar pulse count; This represents the threshold corresponding to the maximum delay; This indicates the successful detection status of the fire target. for , This represents the radar estimated information rate of the fire target in time slot tn. express A specific threshold; when A fire target is considered successfully detected only if this threshold is exceeded; when hour, =0; when hour, for ,in A set of indices representing drones that detect fire targets. This index represents the drone with the best perception performance among all drones served. This represents the radar estimated information rate of the drone's detection of fire targets. Indicates the total bandwidth. This represents the maximum number of orthogonal channels for IoT devices. This indicates the available bandwidth allocated to each IoT device; This indicates the radar pulse power of the drone. This indicates the two-ray round-trip power gain between the drone and the fire target.

[0092] Since the radar pulse echoes arrive at the drones at different times, the total perceived interference for all users only includes communication interference from all IoT devices. The interference for a single user is the product of the transmission power of the IoT device and the channel power gain between the IoT device and the drone. The total perceived interference is the sum of the perceived interference for all users.

[0093] S5. Based on the optimization objective, construct the optimization problem of task scheduling and resource allocation in the unmanned aerial vehicle-assisted ISCC system model.

[0094] In the field of forest fire fighting, the design goal of the UAV-assisted ISCC system is to minimize the weighted sum of the computational offloading penalty index and the perception penalty index through precise task scheduling and resource allocation. Achieving this goal ensures that latency and energy consumption are controlled within preset maximum thresholds during computational offloading, while providing equal computational offloading opportunities for all IoT devices, guaranteeing system fairness and energy efficiency. Optimization of the perception penalty index ensures that the information rate of radar perception tasks exceeds a minimum threshold, providing a fair execution opportunity for each perception task. In this way, the UAV-assisted ISCC system can efficiently support real-time monitoring and rapid response to forest fires, improving the accuracy of firefighting decisions and the success rate of firefighting operations.

[0095] In this optimization problem, the objective is to minimize the weighted sum of the computational unloading penalty index and the perception penalty index. The optimization problem includes six decision variables: the IoT device transmission power set is represented as... UAV radar pulse power set The drone trajectory set is represented as The unload association indicator set is represented as The set of CPU frequencies for IoT devices is represented as follows: And the IoT offloading ratio set is represented as .

[0096] The optimization problem is subject to 11 constraints, including: transmission power constraints of IoT devices, radar pulse power constraints of drones, drone boundary constraints, drone movement distance constraints within a time slot, drone energy constraints, offload association indicator constraints, CPU frequency constraints of IoT devices, and offload ratio constraints of IoT devices. The drone movement distance constraints within a time slot include: movement distance of the drone in the X-axis direction, movement distance of the drone in the Y-axis direction, and movement distance of the drone in the Z-axis direction within a time slot. All drones, IoT devices, and time slots must satisfy these 11 constraints.

[0097] S6. Model the optimization problem as a Markov decision process, and use the TD3 algorithm to optimize the task scheduling and resource allocation strategies in the UAV-assisted ISCC system model to obtain the optimal task scheduling and resource allocation scheme.

[0098] Among them, the TD3 algorithm is an improvement based on the Deep Deterministic Policy Gradient (DDPG) algorithm. The improvements of the TD3 algorithm based on the DDPG algorithm include: (1) Dual Q-network; The TD3 algorithm uses two independent Q-networks to estimate the action value function to reduce overestimation bias. Specifically, the minimum value of the two Q-networks is selected as the target Q-value, so as to estimate the true Q-value more accurately; (2) Delayed policy update; The TD3 algorithm reduces the update frequency of the policy network. Usually, the policy network is updated only once after the Q-network has been updated several times, thereby reducing the policy network's dependence on unstable Q-value estimation and improving the stability of the algorithm; (3) Target policy smoothing; When calculating the target Q-value, noise is added to the target action to make the target value smoother and reduce the instability caused by sudden changes in the policy.

[0099] In order to efficiently implement deep reinforcement learning and solve the task scheduling and resource allocation problems of the UAV-assisted ISCC system in forest fire fighting, the optimization problem is formalized into a Markov decision process. The Markov decision process consists of four basic elements: agent, environment, state, action, policy, and reward.

[0100] In this invention, an intelligent agent is an entity responsible for making decisions and taking actions; in this embodiment of the invention, all cellular base stations within the reach of all drones are designated as intelligent agents to determine all variables.

[0101] The environment includes all external factors that the intelligent agent interacts with, constituting the scenario of the unmanned aerial vehicle-assisted integrated ISCC; the environment involves the dynamic interaction between the unmanned aerial vehicle, IoT devices and sensing targets, which work together to contribute to the forest fire fighting mission.

[0102] The state describes the current state of the environment or the environmental conditions in which the agent is located. Each state represents a specific configuration of the environment. In time slot tn, the state space set contains the locations of all UAVs, IoT devices and sensing targets, the computational offload penalty index and computational tasks of all IoT devices, the sensing penalty index of all sensing targets, and the energy consumption of all UAVs. The state space set can be represented by the following formula (5):

[0103] (5)

[0104] in, Represents the state space set; Represents a set of drone trajectories; Indicates the location of IoT devices; Indicates a computational task. The data size for the computation task is measured in bits. To calculate the number of CPU cycles required to complete 1 bit of the task, For the maximum tolerable delay, for For any value in , tn is Any value in; Indicates the energy consumption of the drone; This indicates the CPU frequency of the drone server.

[0105] Among them, an action is an operation or behavior that an agent can choose to perform in a certain state; the action determines the agent's intervention in the environment. In time slot tn, the action space set includes the flight distance and radar pulse power, unloading rate, CPU clock frequency, transmission power of IoT devices, and unloading association indicators of all UAVs. The action space set can be represented by the following formula (6):

[0106] (6)

[0107] in, Represents the action space set; This represents the distance the drone moves in the x-direction within time slot tn; This represents the distance the drone moves in the y-direction within time slot tn; This represents the distance the drone moves in the z-direction within time slot tn; Indicates the first The radar pulse power of a UAV in time slot tn; Indicates the first CPU frequency of an IoT device in time slot tn; Indicates the first The offloading ratio of an IoT device in time slot tn; Indicates the first The transmission power of an IoT device within time slot tn; This indicates the number of times within time slot tn. Does the first IoT device send to the first A drone unloading task.

[0108] In this context, a policy is a rule or method by which an agent selects actions in different states; the policy determines how the agent chooses the optimal action based on the current state. Under a deterministic policy, the agent always chooses the same specific action for each given state. Under a stochastic policy, the agent selects actions according to a probability distribution, increasing flexibility and potentially exploring more state-action pairs. In this embodiment of the invention, the policy is defined as the probability of selecting a specific action in a given state.

[0109] In this context, reward is the immediate feedback an agent receives from the environment after performing an action. Reward is a quantifiable metric used to evaluate the effectiveness of the agent's actions. A positive reward typically indicates that the action is beneficial to the goal, while a negative reward indicates that the action may be detrimental or should be avoided. The agent's goal is to maximize its cumulative reward by learning to select actions, which in a forest firefighting scenario might mean improving firefighting efficiency, reducing resource waste, or accelerating response time. In time slot tn, the reward is represented by the following formula (7):

[0110] (7)

[0111] in, It is a proportionality constant. As weight, This is the bias constant; This represents the perceived penalty index of the fire target in time slot tn; This represents the calculated unloading penalty index at time slot tn.

[0112] The TD3 algorithm is an online heterogeneous deep reinforcement learning algorithm for solving continuous control problems, improved upon the DDPG (Deep Deterministic Policy Gradient) algorithm. Essentially, the TD3 algorithm incorporates the ideas of the Double Q-Learning algorithm into the DDPG algorithm.

[0113] Optionally, S6 uses the TD3 algorithm to optimize the task scheduling and resource allocation strategies in the UAV-assisted ISCC system model, including:

[0114] S61. Initialize the policy network, the first Q-value network, the second Q-value network, the target network of the policy network, the target network of the first Q-value network, and the target network of the second Q-value network;

[0115] The target network parameters are set to be the same as those of its corresponding main network.

[0116] S62. Sample an action from the environment, execute the action, and obtain the next state and reward;

[0117] S63. Store the tuple of the next state into the experience replay pool.

[0118] S64. Randomly drawing a mini-batch of tuples from the experience replay pool can be represented as: ;

[0119] S65. Calculate the target value based on the obtained mini-batch of tuples;

[0120] The formula for calculating the target value can be expressed by the following formula (8):

[0121] (8)

[0122] in, Indicates the next state; The target network represents the policy network; Indicates a reward; Indicates the target value; This represents the target network of the first Q-value network; This represents the target network of the second Q-value network; This is a discount factor used to balance the relative importance of immediate and future rewards, and its value ranges between 0 and 1. It is the added noise used for smoothing the target policy.

[0123] S66. Based on the target value, update the parameters of the first Q-value network and the second Q-value network by minimizing the loss function;

[0124] The parameters of the first Q-value network and the second Q-value network are updated using the following formula (9):

[0125] (9)

[0126] in, represents the loss function of the first Q-value network and the second Q-value network; M represents the sample or batch size, i.e., the number of samples drawn from the empirical replay buffer in one parameter update. Indicates the current state; Indicates an action; This represents the Q-value network, where j=1 is the first Q-value network and j=2 is the second Q-value network.

[0127] S67. Based on the number of times the parameters of the first Q-value network and the second Q-value network are updated, update the parameters of the policy network every certain number of steps.

[0128] The process of updating the parameters of the policy network can be represented by the following formula (10):

[0129] (10)

[0130] in, Represents a given state The following action, The objective function of the policy network is represented by For parameters gradient, The policy network represents the parameters. The gradient; This represents the gradient of the Q-value network with respect to A.

[0131] S68. Update the parameters of the target network simultaneously according to the parameters of the update strategy network;

[0132] The target network includes: the target network of the policy network, the target network of the first Q-value network, and the target network of the second Q-value network;

[0133] The process of updating the parameters of the target network can be represented by the following formulas (11) and (12):

[0134] (11)

[0135] (12)

[0136] in, Indicates the learning rate. The parameters represent the policy network. The parameters represent the target policy network. This represents the parameters of the Q-network for the j-th target. This represents the parameters of the j-th Q-network.

[0137] S69. Based on the parameters of the target network, perform a soft update of the target network;

[0138] The process of softly updating the target network can be expressed by the following formulas (13) and (14):

[0139] (13)

[0140] (14)

[0141] in, Indicates the update rate; The target network represents the policy network; Represents a policy network; This represents the target network, which is the first Q-value network and the second Q-value network. This represents the first Q-value network and the second Q-value network.

[0142] S70. Repeat steps S62-S68 until the algorithm reaches the preset convergence value and then stops.

[0143] In one feasible implementation, Table 1 shows a joint task scheduling and resource allocation algorithm based on deep reinforcement learning; such as Figure 3 The flowchart shown is a forest fire fighting algorithm based on TD3.

[0144] Table 1

[0145]

[0146] In one feasible implementation, this embodiment of the invention utilizes synthetic data to construct a simulation model for a forest fire fighting scenario. Based on this, the algorithm of the proposed UAV-assisted ISCC system will be evaluated to determine whether it achieves the expected results and to ensure its efficient task scheduling and resource allocation capabilities in a simulated forest fire environment, thus verifying the algorithm's effectiveness in handling emergency fire fighting tasks. In the context of forest fire fighting, this embodiment of the invention constructs a small-scale simulation scenario set within a rectangular area, deploying three UAVs to perform assisted ISCC tasks. A periodic task is simulated, divided into TN=100 time slots to facilitate detailed analysis and evaluation of the UAVs' actions and decisions in each time slot. To ensure the safety of the UAVs during task execution and avoid collisions with trees or other obstacles on the ground, this embodiment of the invention sets a safe flight altitude range for all UAVs. Within the simulation area, 10 IoT devices were deployed to collect critical environmental data, such as temperature, humidity, and wind speed. This environmental data is crucial for predicting fire spread and developing firefighting strategies. This embodiment of the invention sets four sensing targets, which may include specific terrain features or potential fire sources that need to be monitored. The location and characteristics of these sensing targets will be considered in detail during the simulation to evaluate the drone's sensing capabilities. Since both the IoT devices and the sensing targets are located on the ground, this applies to all drones, IoT devices, and time slots. The computation task size is set to be randomly distributed between 20kb and 30kb. The CPU cycles required to process one bit of a computation task are randomly distributed between 500cyc / bit and 1000cyc / bit. The maximum latency threshold is randomly distributed between 40ms and 60ms; any computation task that fails will be allocated an 80ms latency. Other parameter settings are shown in Table 2.

[0147] Table 2

[0148]

[0149] In a simulated forest fire fighting scenario, the changes in the objective function were observed through an iterative process, such as... Figure 4 As shown in the figure, after 10,000 iterations, the objective function value curve generally exhibits a trend of first rapidly decreasing and then stabilizing, further confirming the effectiveness of the proposed algorithm. Specifically, in the first few thousand iterations, the objective function value decreases significantly, indicating that the algorithm can quickly adapt to the specific needs of forest fire fighting scenarios and effectively find optimal task scheduling and resource allocation strategies. The rapid decreasing trend shows that the algorithm can learn and understand the environment in a short time, thus providing an efficient solution for UAV-assisted ISCC systems in forest fire fighting. As the iterations continue, the objective function value curve gradually becomes stable, indicating that the algorithm has gradually approached the optimal solution and can maintain this performance level in subsequent iterations. The trend of first rapidly decreasing and then stabilizing verifies the effectiveness and stability of the TD3 algorithm in handling the task scheduling and resource allocation problems of UAV-assisted ISCC systems in forest fire fighting.

[0150] Among them, such as Figure 5 The diagram illustrates the increasing trend of cumulative reward items with the number of iterations. This phenomenon indicates that the algorithm, through continuous learning and self-improvement, is gradually becoming more efficient, thereby enhancing the performance of task scheduling and resource allocation. In the early stages of iteration, the algorithm may still be exploring effective strategies, but with accumulated experience, it begins to quickly adapt to the complex environment of forest fire fighting and finds more suitable ways to allocate tasks and utilize resources. As the number of iterations increases, the algorithm exhibits stronger adaptability and flexibility, enabling it to more accurately predict fire development and respond quickly. This invention can better handle the computation and perception tasks in UAV-assisted ISCC systems, especially in the face of variable and urgent forest fire situations.

[0151] The continuous growth of the accumulated reward verifies the algorithm's practicality and convergence in synthetic scenarios, providing confidence and a basis for its future application in real-world forest firefighting missions. Through iterative learning and optimization, the algorithm ensures that drones achieve maximum efficiency and effectiveness in both computational unloading and environmental perception during firefighting operations, thus providing strong support for the success of firefighting efforts. To more intuitively demonstrate the operation of drones in forest firefighting operations, such as... Figure 6 and Figure 7The simulation visualized the drone flight paths, including both 2D and 3D views. During the designated service time, three drones flew above IoT devices in the forest, playing a dual role: firstly, providing computational offloading services to assist in processing large amounts of data from the IoT devices; secondly, hovering over multiple key monitoring areas to perform environmental perception tasks, enhancing real-time monitoring of potential fire risks. The drones fully leveraged their maneuverability and flexibility, quickly and efficiently moving between different task points, ensuring that IoT devices could promptly transmit data to the drones for processing, while continuously monitoring key areas in the forest. Flexible task scheduling and resource allocation strategies enabled the drones to respond quickly to the complex and ever-changing situation of forest fires, adjusting their flight paths and monitoring focus in a timely manner to adapt to real-time changes in the fire environment.

[0152] This invention first establishes a UAV-assisted ISCC system scenario; based on the UAV-assisted ISCC system scenario, a UAV-assisted ISCC system model is constructed; based on the UAV-assisted ISCC system model, the total latency and energy consumption for the UAV-assisted ISCC system model to complete the task are determined; secondly, based on the total latency and energy consumption, an optimization objective for the UAV-assisted ISCC system model is constructed; based on the optimization objective, an optimization problem for task scheduling and resource allocation in the UAV-assisted ISCC system model is constructed; finally, the optimization problem is modeled as a Markov decision process, and the TD3 algorithm is used to optimize the task scheduling and resource allocation strategies in the UAV-assisted ISCC system model to obtain the optimal task scheduling and resource allocation scheme.

[0153] This invention addresses the applicability of uplink ISCC in handling computationally intensive, data-diverse, and latency-sensitive forest firefighting missions by constructing an optimization model. It focuses on minimizing the weighted sum of computational offloading and perception penalty indices, while also considering various practical operational constraints, including feasible flight zones and UAV energy constraints, thus optimizing the efficiency and effectiveness of uplink ISCC. Furthermore, this invention reconstructs the task scheduling and resource allocation problem as a Markov decision process, employing the TD3 algorithm to address the joint scheduling and resource allocation of tasks in UAV-assisted ISCC systems, effectively improving task processing speed and environmental perception quality. Finally, this invention provides robust aerial support for forest firefighting, enhancing real-time fire monitoring capabilities, data processing speed, and resource optimization, thereby improving the response speed and success rate of firefighting operations.

[0154] Figure 8This is a block diagram illustrating a drone-assisted integrated sensory-computing resource allocation device for forest fire fighting, according to an exemplary embodiment. This device is used in a drone-assisted integrated sensory-computing resource allocation method for forest fire fighting. (Refer to...) Figure 8 The device includes a first construction unit 310, a second construction unit 320, a determining unit 330, a third construction unit 340, a fourth construction unit 350, and an acquiring unit 360. Wherein:

[0155] The first building unit 310 is used to build a drone-assisted ISCC system scenario;

[0156] The second construction unit 320 is used to construct a drone-assisted ISCC system model based on the drone-assisted ISCC system scenario.

[0157] The determining unit 330 is used to determine the total delay and energy consumption of the UAV-assisted ISCC system model in completing the task based on the UAV-assisted ISCC system model.

[0158] The third construction unit 340 is used to construct an optimization target for the UAV-assisted ISCC system model based on the total delay and energy consumption.

[0159] The fourth construction unit 350 is used to construct an optimization problem for task scheduling and resource allocation in the unmanned aerial vehicle-assisted ISCC system model based on the optimization objective.

[0160] The acquisition unit 360 is used to model the optimization problem as a Markov decision process, and use the TD3 algorithm to optimize the task scheduling and resource allocation strategies in the UAV-assisted ISCC system model to obtain the optimal task scheduling and resource allocation scheme.

[0161] Optionally, the drone-assisted ISCC system scenario includes: a base station, an IoT device, a sensing target, and a drone;

[0162] The base station is used to receive the results of the computational tasks processed by the UAV.

[0163] The IoT device is used to collect environmental information, generate tasks, and send the tasks to the drone; the IoT device is also used to communicate with the drone and coordinate task unloading.

[0164] The sensing target is used to provide sensing tasks for the UAV;

[0165] The drone is used to collect and aggregate information from Internet of Things (IoT) devices.

[0166] Optionally, the total delay of the UAV-assisted ISCC system model in completing the task includes: local delay, transmission delay, and UAV computing delay;

[0167] The local latency is determined by the local computing power and the computing speed of the IoT device; the local computing power is the data size of the computing tasks not offloaded to the drone.

[0168] The transmission delay depends on the amount of data offloaded and the data rate transmitted; wherein the data rate transmitted is related to the available bandwidth allocated to each IoT device, the transmission power of the IoT device, the channel power gain between the IoT device and the drone, interference terms, and the power of noise.

[0169] The computational latency of the drone is determined by the amount of computation unloaded and the computational rate of the drone.

[0170] Optionally, the energy consumption of the UAV-assisted ISCC system model in completing the task includes: the energy consumption of the IoT device and the energy consumption of the UAV;

[0171] The energy consumption of IoT devices includes: the transmission energy of IoT devices and the computing energy of IoT devices;

[0172] Among them, the transmission energy of IoT devices is the product of the transmission power of IoT devices and the transmission delay of IoT devices;

[0173] Among them, the computing power of IoT devices is the product of constants related to the chip structure of IoT devices, the square of the CPU frequency of IoT devices, and the local computing latency.

[0174] The energy consumption of drones includes: energy consumption for mission computing, energy consumption for sensing fire targets, and energy consumption for propulsion.

[0175] Optionally, the optimization objectives of the UAV-assisted ISCC system model include: calculating the unloading penalty index and the perception penalty index.

[0176] Optionally, the optimization of the task scheduling and resource allocation strategy in the UAV-assisted ISCC system model using the TD3 algorithm includes:

[0177] (1) Initialize the policy network, the first Q-value network, the second Q-value network, the target network of the policy network, the target network of the first Q-value network, and the target network of the second Q-value network;

[0178] (2) Sample an action from the environment, perform the action, and obtain the next state and reward;

[0179] (3) Store the tuple of the next state into the experience replay pool;

[0180] (4) Randomly select a small batch of tuples from the experience replay pool;

[0181] (5) Calculate the target value based on the obtained mini-batch of tuples;

[0182] (6) Based on the target value, update the parameters of the first Q-value network and the second Q-value network by minimizing the loss function;

[0183] (7) Update the parameters of the policy network every few steps based on the number of times the parameters of the first Q-value network and the second Q-value network are updated;

[0184] (8) Update the parameters of the target network simultaneously according to the parameters of the updated policy network; wherein the target network includes: the target network of the policy network, the target network of the first Q-value network, and the target network of the second Q-value network;

[0185] (9) Soft update the target network based on the parameters of the target network;

[0186] (10) Repeat steps (2)-(10) until the algorithm reaches the pre-set convergence value and stops.

[0187] This invention first establishes a UAV-assisted ISCC system scenario; based on the UAV-assisted ISCC system scenario, a UAV-assisted ISCC system model is constructed; based on the UAV-assisted ISCC system model, the total latency and energy consumption for the UAV-assisted ISCC system model to complete the task are determined; secondly, based on the total latency and energy consumption, an optimization objective for the UAV-assisted ISCC system model is constructed; based on the optimization objective, an optimization problem for task scheduling and resource allocation in the UAV-assisted ISCC system model is constructed; finally, the optimization problem is modeled as a Markov decision process, and the TD3 algorithm is used to optimize the task scheduling and resource allocation strategies in the UAV-assisted ISCC system model to obtain the optimal task scheduling and resource allocation scheme.

[0188] This invention addresses the applicability of uplink ISCC in handling computationally intensive, data-diverse, and latency-sensitive forest firefighting missions by constructing an optimization model. It focuses on minimizing the weighted sum of computational offloading and perception penalty indices, while also considering various practical operational constraints, including feasible flight zones and UAV energy constraints, thus optimizing the efficiency and effectiveness of uplink ISCC. Furthermore, this invention reconstructs the task scheduling and resource allocation problem as a Markov decision process, employing the TD3 algorithm to address the joint scheduling and resource allocation of tasks in UAV-assisted ISCC systems, effectively improving task processing speed and environmental perception quality. Finally, this invention provides robust aerial support for forest firefighting, enhancing real-time fire monitoring capabilities, data processing speed, and resource optimization, thereby improving the response speed and success rate of firefighting operations.

[0189] Figure 9 This is a schematic diagram of the structure of a drone-assisted integrated sensory computing resource allocation device for forest fire fighting provided by an embodiment of the present invention, as shown below. Figure 9 As shown, the drone-assisted integrated sensor-computer resource allocation device for forest fire fighting can include the above-mentioned... Figure 8 The illustrated device is an integrated sensor-computer resource allocation system for unmanned aerial vehicle (UAV)-assisted forest firefighting. Optionally, the UAV-assisted sensor-computer resource allocation system 410 for forest firefighting may include a first processor 2001.

[0190] Optionally, the drone-assisted integrated sensory computing resource allocation device 410 for forest fire fighting may also include a memory 2002 and a transceiver 2003.

[0191] The first processor 2001, memory 2002, and transceiver 2003 can be connected via a communication bus.

[0192] The following is combined Figure 9 The following is a detailed introduction to the various components of the drone-assisted integrated sensory and computing resource allocation device 410 for forest fire fighting:

[0193] The first processor 2001 is the control center of the drone-assisted integrated sensor and computing resource allocation device 410 for forest fire fighting. It can be a single processor or a collective term for multiple processing elements. For example, the first processor 2001 can be one or more central processing units (CPUs), application-specific integrated circuits (ASICs), or one or more integrated circuits configured to implement embodiments of the present invention, such as one or more digital signal processors (DSPs), or one or more field-programmable gate arrays (FPGAs).

[0194] Optionally, the first processor 2001 can perform various functions of the drone-assisted integrated sensor-computer resource allocation device 410 for forest fire fighting by running or executing software programs stored in the memory 2002 and calling data stored in the memory 2002.

[0195] In a specific implementation, as one example, the first processor 2001 may include one or more CPUs, for example... Figure 9 CPU0 and CPU1 are shown in the diagram.

[0196] In a specific implementation, as one example, the drone-assisted integrated sensory computing resource allocation device 410 for forest fire fighting can also include multiple processors, for example... Figure 9 The first processor 2001 and the second processor 2004 are shown in the diagram. Each of these processors can be a single-core processor or a multi-core processor. Here, a processor can refer to one or more devices, circuits, and / or processing cores used to process data (such as computer program instructions).

[0197] The memory 2002 is used to store the software program that executes the present invention, and is controlled by the first processor 2001 to execute it. The specific implementation method can be referred to the above method embodiment, and will not be repeated here.

[0198] Optionally, the memory 2002 may be a read-only memory (ROM) or other type of static storage device capable of storing static information and instructions, random access memory (RAM) or other type of dynamic storage device capable of storing information and instructions, or electrically erasable programmable read-only memory (EEPROM), compact disc read-only memory (CD-ROM) or other optical disc storage, optical disc storage (including compressed optical discs, laser discs, optical discs, digital universal optical discs, Blu-ray discs, etc.), magnetic disk storage media or other magnetic storage devices, or any other medium capable of carrying or storing desired program code in the form of instructions or data structures and accessible by a computer, but not limited thereto. The memory 2002 may be integrated with the first processor 2001 or exist independently, and may be connected to the interface circuit of the integrated sensor-computer resource allocation device 410 for forest fire fighting drones. Figure 9 (Not shown in the image) is coupled to the first processor 2001, and this embodiment of the invention does not specifically limit this.

[0199] The transceiver 2003 is used to communicate with network devices or with terminal devices.

[0200] Alternatively, transceiver 2003 may include a receiver and a transmitter. Figure 9 (Not shown separately). The receiver is used to implement the receiving function, and the transmitter is used to implement the transmitting function.

[0201] Optionally, the transceiver 2003 can be integrated with the first processor 2001 or exist independently, and can assist the interface circuit of the integrated sensor-computer resource allocation device 410 through the UAV for forest fire fighting. Figure 9 (Not shown in the image) is coupled to the first processor 2001, and this embodiment of the invention does not specifically limit this.

[0202] It should be noted that, Figure 9 The structure of the drone-assisted sensor-computer integrated resource allocation device 410 for forest fire fighting shown in the figure does not constitute a limitation on the router. The actual knowledge structure recognition device may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0203] Furthermore, the technical effects of the drone-assisted integrated sensory computing resource allocation device 410 for forest fire fighting can be referred to the technical effects of the drone-assisted integrated sensory computing resource allocation method for forest fire fighting described in the above method embodiments, and will not be repeated here.

[0204] It should be understood that the first processor 2001 in the embodiments of the present invention may be a central processing unit (CPU), or it may be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor, etc.

[0205] It should also be understood that the memory in the embodiments of the present invention can be volatile memory or non-volatile memory, or may include both volatile and non-volatile memory. The non-volatile memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. The volatile memory can be random access memory (RAM), which is used as an external cache. By way of example, but not limitation, many forms of random access memory (RAM) are available, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate synchronous DRAM (DDR SDRAM), enhanced synchronous DRAM (ESDRAM), synchronous linked DRAM (SLDRAM), and direct rambus RAM (DR RAM).

[0206] The above embodiments can be implemented, in whole or in part, by software, hardware (such as circuits), firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of the present invention are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more sets of available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. A semiconductor medium can be a solid-state drive.

[0207] It should be understood that the term "and / or" in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. A and B can be singular or plural. Additionally, the character " / " in this article generally indicates an "or" relationship between the preceding and following related objects, but it can also represent an "and / or" relationship. Please refer to the context for a more accurate understanding.

[0208] In this invention, "at least one" means one or more, and "more than one" means two or more. "At least one of the following" or similar expressions refer to any combination of these items, including any combination of a single item or a plurality of items. For example, at least one of a, b, or c can represent: a, b, c, ab, ac, bc, or abc, where a, b, and c can be a single item or multiple items.

[0209] It should be understood that, in various embodiments of the present invention, the order of the above-mentioned process numbers does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0210] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0211] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the devices, apparatuses, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0212] In the several embodiments provided by this invention, it should be understood that the disclosed devices, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another device, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0213] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0214] In addition, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0215] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0216] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for resource allocation using drones assisted by integrated sensing and computing for forest fire fighting, characterized in that, The method includes: S1. Set up a scenario for a drone-assisted ISCC system; The drone-assisted ISCC system scenario includes: base station, IoT device, sensing target, and drone; S2. Construct a drone-assisted ISCC system model based on the aforementioned drone-assisted ISCC system scenario; S3. Based on the UAV-assisted ISCC system model, determine the total delay and energy consumption for the UAV-assisted ISCC system model to complete the task; S4. Based on the total delay and energy consumption, construct the optimization objective of the UAV-assisted ISCC system model; The optimization objectives of the UAV-assisted ISCC system model in S4 include: calculating the unloading penalty index and the perception penalty index; The unloading penalty index is calculated using the following formula (1): （1） in, Indicates the total delay. This represents the threshold corresponding to the maximum delay. This indicates the successful computation and unloading status of the computation task. for , This indicates taking the minimum value. Represented as from 1 to The sum of whether IoT devices unload tasks from drones during time slots. This represents the energy consumption of the computational task. This represents the threshold corresponding to the maximum energy consumption; A value of 1 indicates Successfully offloaded computational data to the drone; This represents the penalty exponent for calculating the task in time slot tn-1. This represents the maximum value of the penalty index; The perceived penalty index is expressed by the following formula (2): （2） in, This represents the perception penalty index of the fire target in time slot tn-1; Indicates the duration of the radar pulse; Indicates the radar pulse count; This represents the threshold corresponding to the maximum delay; Indicates the successful perception status of the fire target; S5. Based on the optimization objective, construct the optimization problem of task scheduling and resource allocation in the unmanned aerial vehicle-assisted ISCC system model; S6. Model the optimization problem as a Markov decision process, and use the TD3 algorithm to optimize the task scheduling and resource allocation strategies in the UAV-assisted ISCC system model to obtain the optimal task scheduling and resource allocation scheme.

2. The method for resource allocation using drones assisted by integrated sensing and computing for forest fire fighting, as described in claim 1, is characterized in that... The base station is used to receive the results of the drone's computational tasks. The IoT device is used to collect environmental information, generate tasks, and send the tasks to the drone; the IoT device is also used to communicate with the drone and coordinate task unloading. The sensing target is used to provide sensing tasks for the UAV; The drone is used to collect and aggregate information from Internet of Things (IoT) devices.

3. The method for resource allocation using drones assisted by integrated sensing and computing for forest fire fighting, as described in claim 1, is characterized in that... The total delay of the UAV-assisted ISCC system model in S3 to complete the task includes: local delay, transmission delay and UAV computing delay; The local latency is determined by the local computing power and the computing speed of the IoT device; the local computing power is the data size of the computing tasks not offloaded to the drone. The transmission delay depends on the amount of data offloaded and the data rate transmitted; the data rate transmitted is related to the available bandwidth allocated to each IoT device, the transmission power of the IoT device, the channel power gain between the IoT device and the drone, interference terms, and the power of noise. The computational latency of the drone is determined by the amount of computation unloaded and the computational rate of the drone.

4. The method for resource allocation using drones assisted by integrated sensing and computing for forest fire fighting, as described in claim 1, is characterized in that... The energy consumption of the UAV-assisted ISCC system model in S3 to complete the task includes: the energy consumption of the Internet of Things device and the energy consumption of the UAV. The energy consumption of IoT devices includes: the transmission energy of IoT devices and the computing energy of IoT devices; The transmission energy of an IoT device is the product of its transmission power and transmission delay. Among them, the computing power of IoT devices is the product of constants related to the chip structure of IoT devices, the square of the CPU frequency of IoT devices, and the local computing latency. The energy consumption of drones includes: energy consumption for mission computing, energy consumption for sensing fire targets, and energy consumption for propulsion.

5. The method for resource allocation using unmanned aerial vehicles (UAVs) assisted in forest fire fighting, as described in claim 1, is characterized in that... The S6 method employs the TD3 algorithm to optimize the task scheduling and resource allocation strategies in the UAV-assisted ISCC system model, including: S61. Initialize the policy network, the first Q-value network, the second Q-value network, the target network of the policy network, the target network of the first Q-value network, and the target network of the second Q-value network; S62. Sample an action from the environment, execute the action, and obtain the next state and reward; S63. Store the tuple of the next state into the experience replay pool. S64. Randomly select a small batch of tuples from the experience replay pool; S65. Calculate the target value based on the obtained mini-batch of tuples; S66. Based on the target value, update the parameters of the first Q-value network and the second Q-value network by minimizing the loss function; S67. Based on the number of times the parameters of the first Q-value network and the second Q-value network are updated, update the parameters of the policy network every certain number of steps. S68. Update the parameters of the target network simultaneously according to the updated policy network parameters; wherein, the target network includes: the target network of the policy network, the target network of the first Q-value network, and the target network of the second Q-value network. S69. Based on the parameters of the target network, perform a soft update of the target network; S70. Repeat steps S62-S68 until the algorithm reaches the preset convergence value and then stops.

6. A drone-assisted integrated sensory-computing resource allocation device for forest fire fighting, wherein the drone-assisted integrated sensory-computing resource allocation device for forest fire fighting is used to implement the drone-assisted integrated sensory-computing resource allocation method for forest fire fighting as described in any one of claims 1-5, characterized in that, The device includes: The first building unit is used to build a drone-assisted ISCC system scenario; The drone-assisted ISCC system scenario includes: base station, IoT device, sensing target, and drone; The second construction unit is used to construct a drone-assisted ISCC system model based on the drone-assisted ISCC system scenario. The determining unit is used to determine the total delay and energy consumption of the UAV-assisted ISCC system model in completing the task, based on the UAV-assisted ISCC system model. The third construction unit is used to construct the optimization objective of the UAV-assisted ISCC system model based on the total delay and energy consumption; The optimization objectives of the UAV-assisted ISCC system model include: calculating the unloading penalty index and the perception penalty index; The unloading penalty index is calculated using the following formula (1): （1） in, Indicates the total delay. This represents the threshold corresponding to the maximum delay. This indicates the successful computation and unloading status of the computation task. for , This indicates taking the minimum value. Represented as from 1 to The sum of whether IoT devices unload tasks from drones during time slots. This represents the energy consumption of the computational task. This represents the threshold corresponding to the maximum energy consumption; A value of 1 indicates Successfully offloaded computational data to the drone; This represents the penalty exponent for calculating the task in time slot tn-1. This represents the maximum value of the penalty index; The perceived penalty index is expressed by the following formula (2): （2） in, This represents the perception penalty index of the fire target in time slot tn-1; Indicates the duration of the radar pulse; Indicates the radar pulse count; This represents the threshold corresponding to the maximum delay; Indicates the successful perception status of the fire target; The fourth construction unit is used to construct the optimization problem of task scheduling and resource allocation in the unmanned aerial vehicle-assisted ISCC system model according to the optimization objective. The acquisition unit is used to model the optimization problem as a Markov decision process, and to optimize the task scheduling and resource allocation strategies in the UAV-assisted ISCC system model using the TD3 algorithm to obtain the optimal task scheduling and resource allocation scheme.

7. The drone-assisted integrated sensory and computational resource allocation device for forest fire fighting according to claim 6, characterized in that, The base station is used to receive the results of the drone's computational tasks. The IoT device is used to collect environmental information, generate tasks, and send the tasks to the drone; the IoT device is also used to communicate with the drone and coordinate task unloading. The sensing target is used to provide sensing tasks for the UAV; The drone is used to collect and aggregate information from Internet of Things (IoT) devices.

8. A drone-assisted integrated sensory computing resource allocation device for forest fire fighting, characterized in that, The drone-assisted integrated sensor and computing resource allocation device for forest fire fighting includes: processor; A memory storing computer-readable instructions that, when executed by the processor, implement the method as described in any one of claims 1 to 5.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium contains program code that can be invoked by a processor to execute the method as described in any one of claims 1 to 5.

Images