A dynamic control method, system, and medium for UAV formation collaboration based on HarmonyOS distributed soft bus

Abstract

This invention discloses a dynamic control method, system, and medium for UAV formation collaboration based on the HarmonyOS distributed soft bus, belonging to the technical field of UAV collaborative control. UAVs are interconnected and form a self-organizing network via the HarmonyOS distributed soft bus. Based on onboard hardware resources, UAVs are divided into high-computing-power nodes and low-computing-power nodes. The high-computing-power nodes act as the UAV master nodes, coordinating tasks and assigning reconnaissance or strike missions to the low-computing-power nodes. The UAV receiving the strike mission performs path planning based on the position information of other UAVs and the target's three-dimensional geographic coordinates. This invention dynamically constructs UAV master nodes based on UAV onboard hardware resources and rapidly distributes reconnaissance missions to lock target locations. Then, based on mission response time, it determines the most suitable UAV to perform the strike mission, achieving rapid target response and demonstrating good practicality.

Description

Technical Field

[0001] This invention belongs to the technical field of UAV collaborative control, specifically relating to a dynamic control method, system, and medium for UAV formation collaboration based on HarmonyOS distributed soft bus. Background Technology

[0002] Unmanned Aerial Vehicles (UAVs) are aircraft that achieve unmanned flight control through external remote control commands or embedded programs. Their operation modes encompass both manual remote control intervention and fully autonomous mission execution. They often utilize a high-computing-powered mission control board as the "brain" of flight mission control, decoupling the flight control system from the mission board to manage UAV control and sensor data collection, ensuring stable and reliable flight. With the increasing use of UAV swarms, multi-UAV collaborative formations have become a research hotspot. Traditional UAV networking methods often rely on point-to-point links or centralized base stations, which frequently suffer from complex deployment, poor scalability, and high communication latency.

[0003] In recent years, HarmonyOS distributed soft bus technology has provided unified distributed capabilities for cross-device communication. Its core purpose is to enable rapid discovery, connection, and efficient data transmission between devices. Through a four-stage model of "discovery-connection-networking-transmission," soft bus technology allows devices to interconnect like a unified system. Applying HarmonyOS distributed soft bus can provide UAV formations with high-bandwidth, low-latency, secure, and reliable communication capabilities. Secondly, existing UAV collaborative combat solutions mostly rely on centralized manual command or ROS systems for coordination. This approach involves complex operational procedures, significant signal transmission delays, slow system response to emergencies, and is prone to single-point-of-failure risks. If the central node fails, the entire formation may fall into a state of "mass disconnection," making it impossible to conduct real-time and accurate target tracking and strikes. Summary of the Invention

[0004] The purpose of this invention is to provide a dynamic control method, system, and medium for UAV formation collaboration based on HarmonyOS distributed soft bus, aiming to solve the above-mentioned problems.

[0005] This invention is mainly achieved through the following technical solutions:

[0006] A dynamic control method for UAV formation coordination based on HarmonyOS distributed soft bus includes the following steps:

[0007] Step S1: The drones interconnect and form a self-organizing network through the HarmonyOS distributed soft bus, sharing data among the drones to form a drone situational awareness network;

[0008] Step S2: Based on the heterogeneity of UAV onboard hardware resources, UAVs are divided into UAVs with strong computing power nodes and UAVs with weak computing power nodes.

[0009] Step S3: The drones of the high-computing-power nodes act as task coordinators, assigning reconnaissance or strike missions to the drones of the low-computing-power nodes.

[0010] Step S31: Task Issuance and Allocation: The UAV periodically broadcasts its own status through the soft bus to synchronize the resource status table; the UAV with the strong computing power node is used as the UAV master node. The UAV master node divides the formation into several task groups according to the situation information and mission objectives, and issues reconnaissance or strike missions through the soft bus.

[0011] Step S32: Target reconnaissance and calculation of target coordinates: The UAV carrying out the reconnaissance mission searches for the predetermined target in the field of view based on image recognition and analyzes the target's three-dimensional geographic coordinate information.

[0012] Step S33: Information Sharing and Mission Triggering: The UAV that obtains the reconnaissance mission broadcasts the target information to the entire UAV formation through the HarmonyOS soft bus; the UAV master node determines the most suitable UAV to perform the strike mission based on the status and mission requirements of each UAV, with the goal of minimizing the mission response time, and assigns the strike mission to that UAV.

[0013] Step S4: Obtain the position information of other drones and the target's three-dimensional geographic coordinates based on the four-dimensional spatiotemporal data of the drone tasked with the strike mission. The algorithm performs path planning.

[0014] To better implement the present invention, step S1 further includes the following steps:

[0015] Step S11: Equip the drone with the OpenEler system's mission board, flight controller, and gimbal; send control commands to the flight controller via the mission board to control the drone's flight or upload flight routes for flight missions;

[0016] Step S12: Deploy HarmonyOS distributed soft bus on the task board to establish underlying communication and data transmission paths between drones; then use HarmonyOS distributed soft bus to enable drones to discover each other automatically and seamlessly, and then establish P2P paths for authentication and networking.

[0017] Step S13: Share the data in the flight controller with the surrounding drones.

[0018] To better realize the present invention, step S2 further includes the following steps:

[0019] Step S21: First, collect resource information for each drone, including CPU model, clock speed and number of cores, RAM size, aircraft battery capacity and power consumption.

[0020] Step S22: Calculate the CPU computing power index, memory index, and energy index. Then, calculate the comprehensive computing power score = α × (CPU performance normalized value) + β × (memory normalized value) - γ × (power consumption index).

[0021] Where: α, β, and γ are weighting coefficients, respectively;

[0022] Step S23: Based on the comprehensive computing power score, the drones are divided into drones with strong computing power nodes and drones with weak computing power nodes.

[0023] To better implement the present invention, further, in step S31, if the number of drones with high computing power nodes is greater than 1, the final control decision is determined based on the decisions of several drone master nodes through a decision matrix or heuristic algorithm; or the drone with the highest computing power node is taken as the drone master node.

[0024] To better realize the present invention, in step S33, the indicators of the current payload type adaptability of the UAV, the distance to the target or flight time cost, the remaining power and the computing / performance margin are formed into a multi-indicator scoring matrix, and the UAV with the highest comprehensive score is selected as the UAV to perform the strike mission.

[0025] To better implement the present invention, step S33 further includes the following steps:

[0026] (1) First, receive the target coordinates and task type of the task, and at the same time obtain the current position, current task load, remaining power, and idle time information of each UAV;

[0027] (2) Then, based on load matching and computational load indicators, filter out drones that do not meet the basic threshold and calculate the adaptability index S_i=(1-the current proportion of the communication load).

[0028] Then perform cost calculation: take the normalized value of the distance from the drone to the target or the estimated flight time as the geometric cost C_i, and convert it into the cost value D_i=1 / (1+C_i);

[0029] Compare the remaining battery power with the flight mission consumption, and calculate the remaining energy ratio E_i = remaining battery power / flight mission consumption;

[0030] Calculate the overall score for each drone: Score_i = w1·S_i + w2·D_i + w3·E_i;

[0031] Where: w1, w2, w3 are the calculation weights;

[0032] (3) Finally, the drone with the highest comprehensive score is selected as the main response node for the strike mission, while the drone with the second highest comprehensive score is selected as the backup response node.

[0033] To better implement the present invention, further, in step S33, the comprehensive score of the UAV is Score_i = w1·S_i + w2·D_i + w3·E_i + w4·other risk coefficients; where w4 is the calculation weight.

[0034] To better realize the present invention, step S4 further includes the following steps:

[0035] Step S41: Using four-dimensional spacetime The algorithm plans the initial paths for all drones;

[0036] Step S42: Traverse all path points or line segments, check the paths of any two drones, and mark conflicting paths that intersect in space and time.

[0037] Step S43: Sort the drones that cause path conflicts according to their flight priority. Drones with higher flight priority keep their paths unchanged; drones with lower flight priority are given the opportunity to avoid the conflict.

[0038] Step S44: If the avoidance process cannot resolve the conflict, proceed to step S41 to replan the path for the low-priority drone.

[0039] This invention is mainly achieved through the following technical solutions:

[0040] A dynamic control system for UAV formation collaboration based on HarmonyOS distributed soft bus is disclosed. Based on the aforementioned dynamic control method for UAV formation collaboration based on HarmonyOS distributed soft bus, the system includes several UAVs interconnected via a self-organizing network using HarmonyOS distributed soft bus. Each UAV is equipped with a mission board, flight controller, gimbal, flight control docking module, image recognition module, data communication module, path planning module, and mission planning module. The mission board is deployed with HarmonyOS distributed soft bus, serving as a central control unit running the OpenEuler operating system to organize and manage the flight controller, gimbal, flight control docking module, image recognition module, data communication module, path planning module, and mission planning module.

[0041] The flight control docking module is used to acquire flight status, attitude, and altitude data, and to control the flight; the image recognition module is used to identify targets and calculate target positions based on image recognition algorithms; the data communication module is used to perform self-authentication networking through the HarmonyOS distributed soft bus; the task planning module is used to plan and execute tasks; and the path planning module is used to calculate the UAV path.

[0042] A computer-readable storage medium storing a computer program, characterized in that, when executed by a processor, the program implements the aforementioned dynamic control method for UAV formation collaboration based on HarmonyOS distributed soft bus.

[0043] The beneficial effects of this invention are as follows:

[0044] (1) This invention dynamically constructs a UAV master node based on the UAV's onboard hardware resources and quickly distributes reconnaissance missions to lock the target location; then, based on the mission response time, it determines the most suitable UAV to perform the strike mission, achieving rapid target response. The UAVs of this invention share status and perception information in real time, can dynamically adjust control strategies, are flexibly applicable to different environments, and have good practicality.

[0045] (2) This invention encapsulates the HarmonyOS distributed soft bus as a communication middleware, adopts a decentralized cluster design, and realizes multi-UAV collaboration. Each UAV node achieves self-organized networking, data exchange, and collaborative control based on the HarmonyOS distributed soft bus. The task of the UAV can be reversed, and the UAV can directly command another UAV. Each UAV node shares status and perception information through the data communication module, realizing real-time formation control. This invention utilizes the high bandwidth and low latency communication capabilities of the HarmonyOS soft bus to ensure real-time information synchronization, simplify the network architecture, and reduce development complexity. The system structure is flexible and scalable, and the fault recovery mode improves the reliability and robustness of the formation, meeting the actual formation flight mission requirements. Attached Figure Description

[0046] Figure 1 This is a flowchart of the dynamic control method for UAV formation collaboration based on HarmonyOS distributed soft bus according to the present invention;

[0047] Figure 2 This is a schematic diagram of the structure of the airborne software and the ground software in Example 2;

[0048] Figure 3 This is a timing diagram of the rotary-wing UAV in Example 2. Detailed Implementation

[0049] Example 1:

[0050] A dynamic control method for UAV formation coordination based on HarmonyOS distributed soft bus, such as Figure 1 As shown, it includes the following steps:

[0051] Step S1: The drones interconnect and form a self-organizing network through the HarmonyOS distributed soft bus, sharing data among the drones and building a drone situational awareness network;

[0052] Step S11: The embedded mission board of the OpenEler system is pre-installed on the UAV and connected to the onboard flight controller, gimbal and other equipment via wired connection. Flight status information, key information such as GPS, battery, attitude and heading are collected from the flight controller. At the same time, control commands can be sent to the flight controller through the mission board to control the flight of the aircraft or upload the flight path for fully automatic flight missions. High-definition video is collected from the gimbal and then hardware decoding is performed to obtain image data for target recognition and other operations.

[0053] Step S12: Deploy the HarmonyOS distributed soft bus on the taskboard of the OpenEuler system. Use the links between drones to establish underlying communication and data transmission paths. Then, use the soft bus to achieve seamless automatic discovery between drones, followed by authentication and networking to establish P2P pathways. Data collected from sensors such as flight controllers is parsed and processed, then shared with surrounding drones to build a drone situational awareness network.

[0054] Step S2: Based on the heterogeneity of UAV onboard hardware resources, the UAVs in the formation are divided into "strong computing power nodes" and "weak computing power nodes" so as to reasonably allocate computing tasks during subsequent task scheduling.

[0055] Specifically, the computing power and payload capacity of each drone are evaluated by examining metrics such as CPU performance, memory capacity, battery capacity, and power consumption. For example, high-performance computing nodes use the RK3588 chip, employing a 4×A76 + 4×A55 (up to 2.4GHz) big core + little core architecture, and integrating a tri-core NPU (6TOPS computing power); low-performance computing nodes use a Raspberry Pi 4 with 4×A72 (1.5GHz). The payload type and endurance of the drone also limit the range of tasks it can undertake; therefore, nodes should be classified based on a comprehensive consideration of computing power and energy budget.

[0056] Step S2 includes the following steps:

[0057] Step S21: First, collect resource information for each drone; collect the CPU model and clock speed, number of cores, RAM size, aircraft battery capacity and power consumption of each drone.

[0058] Step S22: Calculate the CPU computing power index, memory index, and energy index (remaining power ÷ estimated consumption rate). Specifically, benchmark the CPU or estimate the computing power value through parameters; normalize the memory capacity; and evaluate the efficiency of battery capacity / power consumption.

[0059] Then, after normalizing each indicator, the overall computing power score is calculated according to preset weights. For example, the computing power score can be designed as α × (normalized CPU performance value) + β × (normalized memory value) - γ × (power consumption index), and the weights (α, β, γ) are adjusted according to task requirements.

[0060] Step S23: Set a threshold or use a clustering algorithm to classify the scores. Nodes above the threshold are considered high-performance computing nodes, capable of handling computationally intensive or urgent tasks; high-performance computing nodes can serve as "master nodes" responsible for complex task calculations. Nodes below the threshold are considered low-performance computing nodes, mainly handling lightweight tasks or auxiliary tasks.

[0061] Step S3: Finally, high-powered computing nodes can act as task coordinators, dynamically allocating reconnaissance or strike tasks; low-powered computing nodes mainly execute the specific tasks assigned to them. This includes the following steps:

[0062] Step S31: Task Issuance and Allocation: The system presets or the ground station issues task objectives. Each UAV periodically broadcasts its own status (location, remaining battery power, payload type, etc.) via the soft bus to synchronize the resource status table. The UAV with the highest computing power acts as the master node, dividing the formation into several task groups based on situational information and task objectives, and issuing reconnaissance or strike commands via the soft bus.

[0063] Step S32: Target Reconnaissance and Calculation: The UAV tasked with the reconnaissance mission activates its image recognition module and uses the YOLOv10 deep learning algorithm to analyze camera images and locate the predetermined target within its field of view. Once the target is detected, its position on the image plane is combined with the UAV's own pose information, and the target's three-dimensional geographic coordinates are calculated using camera intrinsic parameters and altitude data.

[0064] Step S33: Information Sharing and Mission Triggering: After detecting the target and calculating its coordinates, the reconnaissance UAV broadcasts the target information (target type, 3D coordinates, detection timestamp, etc.) to the entire formation via the HarmonyOS soft bus. All attack UAVs can receive this information. Upon receiving the attack target mission, the most suitable node for execution is intelligently selected based on the status of each UAV and mission requirements to minimize mission response time and ensure mission success rate.

[0065] Specifically, the allocation process needs to comprehensively consider factors such as whether the drone's current payload type matches the mission requirements, the distance to the target or flight time cost, remaining battery power, and computational / performance margin. This results in a multi-indicator scoring matrix, selecting the drone with the highest overall score to respond to the mission, while other high-scoring drones are reserved as backups. Details are as follows:

[0066] (1) First, information is collected, including the target coordinates and task type of the task, and information such as the current position, current task load, remaining power, and computing idle time of each UAV is obtained.

[0067] (2) Then, an adaptation assessment is performed, and the adaptation index S_i (payload matching + computational load index) is calculated. The payload compatibility and computational load of each UAV are assessed, and UAVs that do not meet the basic requirements are filtered out.

[0068] Then, cost calculation is performed, taking the distance from the drone to the target or the estimated flight time as the geometric cost C_i (such as distance / speed), and converting it into a cost value D_i = 1 / (1+C_i).

[0069] Compare the remaining battery power with the flight range to calculate the remaining energy ratio E_i (remaining battery power / flight mission consumption). Assess whether the mission can be completed and the user can return or continue with subsequent missions.

[0070] Based on the above premise, a scoring model is used to calculate the comprehensive score of each drone according to the set weights: Score_i = Payload adaptability weight × (1 - Current computing / communication load ratio) + Distance weight × (Target distance normalized and inverted) + Energy margin weight × (Remaining power / Task required energy consumption).

[0071] (3) Finally, the decision is output, selecting the drone with the highest score as the primary response node, while recording the second highest-scoring node as a backup. If necessary, a multi-drone collaborative mechanism or a dynamic adjustment mechanism can be established. This process can be implemented using a decision matrix or heuristic algorithm to ensure the rapid determination of the optimal strike plan in a heterogeneous drone swarm.

[0072] Preferably, the comprehensive score Score_i = w1·S_i + w2·D_i + w3·E_i + w4·other risk coefficients. This algorithm takes into account both the performance of heterogeneous UAVs and the urgency of the mission; where w1, w2, w3, and w4 are the calculation weights.

[0073] (4) Coordinated Action: After receiving target information, the UAVs assigned to the strike mission autonomously choose the timing to take off and fly towards the target location. During the flight, the UAVs continuously exchange real-time position and situation information, providing dynamic obstacle and collision information for subsequent path planning.

[0074] Step S4: The drone receiving the attack mission, combined with the position information of other drones and the target coordinates, performs path planning. The planning process employs four-dimensional spatiotemporal collaboration. The algorithm expands the state space of the UAV into a four-dimensional (X, Y, Z, T) structure, where T is the time dimension.

[0075] Preferably, the procedure specifically includes the following steps:

[0076] (1) Environmental Modeling and Conflict Matrix: An environmental map is constructed and a conflict matrix is ​​initialized. The conflict matrix records the possible mutual interference in the future paths of each UAV. During the planning process, each path node contains the current time T. When two UAVs plan paths with the same spatial coordinates and time, they are detected as conflicting.

[0077] (2) Four-dimensional Search: Execute in the four-dimensional state space Search algorithm. The cost function for each drone is defined as:

[0078] ;

[0079] in: The actual cost from the starting point to the current node. This is the estimated heuristic distance (such as Euclidean distance) from the current node to the target. These are weighting coefficients that are dynamically adjusted over time.

[0080] Adjust in real time according to the level of environmental threat and the urgency of the mission. When the environmental threat is high, the weight is increased to make the path avoid high-risk areas; when the threat is low, the weight is decreased to seek the shortest path. Through this adaptive heuristic, the algorithm can accelerate the search while ensuring path quality.

[0081] (4) Path conflict avoidance: based on spatiotemporal four-dimensional The algorithm performs path planning for each drone. During the search process, the algorithm exchanges the planning progress of each drone in real time through a soft bus. At the same time, it detects potential spatiotemporal conflicts in the multi-drone tracks in real time and eliminates conflicts through priority adjustment and dynamic replanning, ensuring the safety and collision-free operation of drone formation flight.

[0082] Preferably, the procedure specifically includes the following steps:

[0083] Step S41: Using four-dimensional spacetime The algorithm plans the initial paths for all drones.

[0084] Step S42: Based on the conflict detection rules, check any two drone paths. If spatial-temporal overlap is found, a conflict is marked. Specifically, after generating the 4D path for each drone, all path points or line segments are traversed to determine if there is any spatiotemporal intersection. That is, if the planned trajectories of two drones occupy the same spatial region within a safe distance at a certain moment, a conflict is considered to have occurred. According to this rule, situations such as "arriving at the same waypoint simultaneously" or "path crossing" can be detected.

[0085] Step S43: Priority Adjustment Strategy; A risk assessment coefficient is assigned to each drone. Drones involved in a conflict are ranked according to their risk coefficient; those with higher risk coefficients maintain their trajectory, while those with lower risk coefficients decelerate or remain stationary. Specifically, the risk assessment coefficient is set based on factors such as mission urgency, remaining battery power, or the importance of the formation position. Drones with higher risk assessment coefficients have higher flight priority. When a conflict is detected, high-priority drones retain their original path or make minor adjustments, while low-priority drones need to avoid it. For example, waiting, deceleration, or rerouting operations can be inserted at low-priority nodes. This strategy follows the "non-interference principle," delaying collision avoidance actions until the very last moment.

[0086] Step S44: If the conflict cannot be resolved by pausing alone, dynamic replanning is performed on the drone forced to yield. For low-priority drones, path search is re-executed to generate trajectories that do not conflict with high-priority drones. Spatiotemporal four-dimensional mapping is re-executed at the current time / location. Algorithms or other avoidance algorithms generate new feasible paths. This process is similar to the conflict resolution methods in model predictive control, and continues until all conflicts are eliminated.

[0087] Step S45: Repeat the detection and adjustment until all tracks are free of spatiotemporal conflicts. This mechanism ensures that in multi-aircraft flight scenarios, track conflicts between nodes are adaptively handled through conflict rule detection and real-time scheduling, thereby achieving safe formation path planning. The generated tracks after the search are almost optimal, and simulations show that the path length deviates from the theoretical shortest path by less than 1%.

[0088] Step S46: Track Distribution and Execution: The planned safe path is distributed to the corresponding attack drone via the soft bus. The drone flies sequentially according to the path, and during flight, replanning can be periodically triggered to cope with dynamic obstacles or changes in threats.

[0089] Step S5: Each task group is equipped with a cold backup UAV. When a primary UAV experiences a communication interruption or malfunction, the backup UAV will proactively take over its mission after a predetermined time since it has not received a heartbeat from the primary node. The backup UAV can obtain the latest mission information and flight path planning from the soft bus or ground station, and quickly become the new primary node, continuing to allocate remaining tasks and update flight control. Employing a priority-based election algorithm combined with a comprehensive assessment of remaining UAV resources significantly reduces fault recovery latency, enabling rapid control switching and continuous mission execution.

[0090] Example 2:

[0091] A dynamic control system for UAV formation collaboration based on HarmonyOS distributed soft bus is presented. It employs a distributed architecture, consisting of multiple UAVs interconnected via a self-organizing network. Each UAV integrates the following core modules: an embedded RK3588 mission board, a flight controller, a gimbal (with camera), an image recognition module, a data communication module, a mission planning module, and a fault recovery module. The embedded RK3588 mission board serves as the central control unit for the UAVs, running the OpenEuler operating system and related applications, and is responsible for organizing and managing the internal modules.

[0092] Preferably, the aforementioned soft bus is encapsulated based on the HarmonyOS distributed soft bus with a Face-like architecture, which isolates internal communication from external communication to improve security. The relevant functional modules of the soft bus include a flight control docking module, an image recognition module, a data communication module, and a mission planning module; it also includes a ground service module.

[0093] The flight control docking module is used to acquire data such as flight status, attitude, and altitude, and to control the flight of the aircraft.

[0094] The image recognition module is used to perform target recognition and target position calculation through image recognition algorithms and related gimbal control.

[0095] The data communication module is used to encapsulate HarmonyOS Distributed Soft Bus, UDP, TCP, and DDS into a unified data communication middleware, which can perform self-authentication networking through the characteristics of HarmonyOS Distributed Soft Bus.

[0096] The task planning module is used by the autonomous task module and can perform task planning and execution.

[0097] The ground service module is used to receive UAV data, issue tasks, and interface with external systems.

[0098] This system comprises multiple unmanned aerial vehicles (UAVs), each leveraging the HarmonyOS distributed soft bus to achieve self-organizing networking, data exchange, and collaborative control. During system operation, the task planning module distributes collaborative tasks to each UAV, the path planning module calculates the flight formation route, and each node shares status and perception information through the data communication module, enabling real-time formation control. Utilizing the HarmonyOS soft bus, the high bandwidth and low latency communication capabilities ensure real-time information synchronization, simplifying the network architecture and reducing development complexity. The system is flexible and scalable, and the fault recovery module enhances the reliability and robustness of the formation, meeting the requirements of actual formation flight missions.

[0099] Preferably, the data communication management module is a common module that encapsulates the underlying communication. The data communication management module is designed in imitation of the FACE architecture, encapsulating communication via HarmonyOS distributed soft bus, serial port, UDP, DDS, etc., forming a unified communication interface. This provides multiple communication interfaces for upper-layer applications. Within the data communication module, the HarmonyOS distributed soft bus is compiled separately onto the OpenEuler operating system, and the RK3588 running the OpenEuler operating system serves as the onboard task board.

[0100] like Figure 2 As shown, the airborne software includes a startup management module, a flight control management module, a ground agent service module, a gimbal camera module, and a mission management module; the ground software includes a ground service module. The startup management module manages the ground agent service module, flight control management module, gimbal camera module, and mission management module, and their applications are displayed on the mission board on the aircraft. The ground service module is an application on the ground station, used for displaying information at ground nodes. Based on this software architecture, the following scenarios can be implemented:

[0101] like Figure 3 As shown, multiple conventional quadcopter drones take off for inspections. During the inspection, if one quadcopter drone detects a target, it continuously tracks and transmits target information to another quadcopter drone, directly commanding that drone to proceed to the target location to execute an attack mission. If an anomaly occurs during execution, the quadcopter drone is unable to complete the attack. It then transmits target information to another fixed-wing or quadcopter drone, transferring the attack mission to that drone, which continues to engage the target. For example, if drone 4 is conducting an inspection, drone 3 detects and tracks the target, simultaneously instructing drone 4 to return to base, while drone 1 executes the attack mission.

[0102] Preferably, the self-organizing network of the HarmonyOS distributed soft bus, based on the session authentication transmission capability, uses SessionId for data sending and receiving, realizing the interaction of business messages, streaming messages, and control commands, further improving the transmission efficiency of the HarmonyOS distributed soft bus. The built-in TLS 1.3 encryption method in the HarmonyOS soft bus is removed, as low latency is more important for communication between drone collaborative systems. Therefore, the QUIC protocol is used for encryption of communication data. A comparison is made between the modified soft bus file transfer and FTP file transfer methods, as shown in the following comparative experiment:

[0103] 1. Test environment and network topology description.

[0104] The equipment used in this test included two Raspberry Pis and one RK3588, three Minihomer aerial terminals, one Minihomer ground terminal, and one small switch. The models are shown in Table 1, and the IP configuration information is shown in Table 2.

[0105] In this latency test, IP address 192.xxx.xx.xx was used as the server, and other devices were used as clients to communicate with the server and test the latency information between them.

[0106] Test method: The client sends a message to the server and records the sending time. After receiving the message from the client, the server sends a feedback message back to the client. The client receives the feedback message from the server, records the receiving time, and estimates the latency by calculating the time difference between sending and receiving.

[0107] Average latency = ;

[0108] Where: k is the nth measurement, and n is the total number of measurements.

[0109] 2. Comparative test of file transfer based on FTP and soft bus.

[0110] Wired LAN bandwidth: 100Mbps;

[0111] Wireless LAN bandwidth (minihomer): 5Mbps.

[0112] To verify the effectiveness of the above optimizations, under the same network environment and test conditions, the test results are shown in Tables 3-5. The file transfer performance based on the soft bus and the traditional FTP method were compared. In medium file transfer scenarios of 5MB-20MB, the soft bus exhibits higher transfer rates and lower latency, making it suitable for rapid exchange of small files and command data in UAV missions. For large files (over 100MB), the performance is similar to FTP, but still maintains high stability. In environments with low wireless network bandwidth (5Mbps), the soft bus demonstrates stronger anti-interference capabilities and a higher overall transmission success rate (Table 4 shows a significantly better success rate than Table 6). Furthermore, its transmission latency for files under 20MB is superior to FTP, indicating that the soft bus's self-organizing network mechanism can maintain high stability even in complex network environments.

[0113] Table 1 Test Equipment Model Parameters

[0114]

[0115] Table 2 IP Configuration Information

[0116]

[0117] Table 3. Softbus Wired LAN Test

[0118]

[0119] Table 4. Softbus Wireless LAN Test

[0120]

[0121] Table 5 FTP Wired LAN Test

[0122]

[0123] Table 6 FTP Wireless LAN Test

[0124]

[0125] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Any simple modifications or equivalent changes made to the above embodiments based on the technical essence of the present invention shall fall within the protection scope of the present invention.

Claims

1. A dynamic control method for UAV formation collaboration based on HarmonyOS distributed soft bus, characterized in that, Includes the following steps: Step S1: The drones interconnect and form a self-organizing network through the HarmonyOS distributed soft bus, sharing data among the drones to form a drone situational awareness network; Step S2: Based on the heterogeneity of UAV onboard hardware resources, UAVs are divided into UAVs with strong computing power nodes and UAVs with weak computing power nodes. Step S3: The drones of the high-computing-power nodes act as task coordinators, assigning reconnaissance or strike missions to the drones of the low-computing-power nodes. Step S31: Task Issuance and Allocation: The UAV periodically broadcasts its own status through the soft bus to synchronize the resource status table; the UAV with the strong computing power node is used as the UAV master node. The UAV master node divides the formation into several task groups according to the situation information and mission objectives, and issues reconnaissance or strike missions through the soft bus. Step S32: Target reconnaissance and calculation of target coordinates: The UAV carrying out the reconnaissance mission searches for the predetermined target in the field of view based on image recognition and analyzes the target's three-dimensional geographic coordinate information. Step S33: Information Sharing and Mission Triggering: The drone that has obtained the reconnaissance mission broadcasts the target information to the entire drone formation via the HarmonyOS soft bus; The drone master node determines the most suitable drone to perform the strike mission based on the status and mission requirements of each drone, with the goal of achieving the highest comprehensive score, and assigns the strike mission to that drone. Step S33 includes the following steps: (1) First, receive the target coordinates and task type of the task, and at the same time obtain the current position, current task load, remaining power, and idle time information of each UAV; (2) Then, based on load matching and computational load indicators, filter out drones that do not meet the basic threshold and calculate the adaptability index S_i=(1-the current proportion of the communication load). Then perform cost calculation: take the normalized value of the distance from the drone to the target or the estimated flight time as the geometric cost C_i, and convert it into the cost value D_i=1 / (1+C_i); Compare the remaining battery power with the flight mission consumption, and calculate the remaining energy ratio E_i = remaining battery power / flight mission consumption; Calculate the overall score for each drone: Score_i = w1·S_i + w2·D_i + w3·E_i; Where: w1, w2, w3 are the calculation weights; (3) Finally, the drone with the highest comprehensive score will be used as the main response node to carry out the strike mission, while the drone with the second highest comprehensive score will be used as the backup response node. Step S4: Obtain the position information of other drones and the target's three-dimensional geographic coordinates based on the four-dimensional spatiotemporal data of the drone tasked with the strike mission. The algorithm performs path planning.

2. The dynamic control method for UAV formation collaboration based on HarmonyOS distributed soft bus according to claim 1, characterized in that, Step S1 includes the following steps: Step S11: Equip the drone with the OpenEler system's mission board, flight controller, and gimbal; send control commands to the flight controller via the mission board to control the drone's flight or upload flight routes for flight missions; Step S12: Deploy HarmonyOS distributed soft bus on the task board to establish underlying communication and data transmission paths between drones; then use HarmonyOS distributed soft bus to enable drones to discover each other automatically and seamlessly, and then establish P2P paths for authentication and networking. Step S13: Share the data in the flight controller with the surrounding drones.

3. The dynamic control method for UAV formation collaboration based on HarmonyOS distributed soft bus according to claim 1, characterized in that, Step S2 includes the following steps: Step S21: First, collect resource information for each drone, including CPU model, clock speed and number of cores, RAM size, aircraft battery capacity and power consumption. Step S22: Calculate the CPU computing power index, memory index, and energy index. Then, calculate the comprehensive computing power score = α × (CPU performance normalized value) + β × (memory normalized value) - γ × (power consumption index). Where: α, β, and γ are weighting coefficients, respectively; Step S23: Based on the comprehensive computing power score, the drones are divided into drones with strong computing power nodes and drones with weak computing power nodes.

4. The dynamic control method for UAV formation collaboration based on HarmonyOS distributed soft bus according to claim 1, characterized in that, In step S31, if the number of drones with high computing power nodes is greater than 1, the final control decision is determined based on the decisions of several drone master nodes through a decision matrix or heuristic algorithm; or the drone with the highest computing power node is taken as the drone master node.

5. The dynamic control method for UAV formation collaboration based on HarmonyOS distributed soft bus according to claim 1, characterized in that, In step S33, a multi-index scoring matrix is ​​formed by considering the drone's current payload type adaptability, distance to the target or flight time cost, remaining battery power, and computing / performance margin. The drone with the highest comprehensive score is selected as the drone to perform the strike mission.

6. The dynamic control method for UAV formation collaboration based on HarmonyOS distributed soft bus according to claim 5, characterized in that, In step S33, the comprehensive score of the UAV is Score_i = w1·S_i + w2·D_i + w3·E_i + w4·other risk coefficients; where w4 is the calculation weight.

7. The dynamic control method for UAV formation collaboration based on HarmonyOS distributed soft bus according to claim 1, characterized in that, Step S4 includes the following steps: Step S41: Using four-dimensional spacetime The algorithm plans the initial paths for all drones; Step S42: Traverse all path points or line segments, check the paths of any two drones, and mark conflicting paths that intersect in space and time. Step S43: Sort the drones that cause path conflicts according to their flight priority. Drones with higher flight priority keep their paths unchanged; drones with lower flight priority are given the opportunity to avoid the conflict. Step S44: If the avoidance process cannot resolve the conflict, proceed to step S41 to replan the path for the low-priority drone.

8. A dynamic control system for UAV formation collaboration based on HarmonyOS distributed soft bus, comprising the dynamic control method for UAV formation collaboration based on HarmonyOS distributed soft bus as described in any one of claims 1-7, characterized in that, It includes several drones, which are interconnected and form a self-organizing network via the HarmonyOS distributed soft bus. Each drone is equipped with a mission board, flight controller, gimbal, flight control docking module, image recognition module, data communication module, path planning module, and mission planning module. The mission board is deployed with the HarmonyOS distributed soft bus and serves as a central control unit, running the OpenEuler operating system to organize and manage the flight controller, gimbal, flight control docking module, image recognition module, data communication module, path planning module, and mission planning module. The flight control docking module is used to acquire flight status, attitude, and altitude data, and to control the flight; the image recognition module is used to identify targets and calculate their positions based on image recognition algorithms; the data communication module is used to perform self-authentication networking via the HarmonyOS distributed soft bus. The task planning module is used for task planning and execution; The path planning module is used to calculate the drone's path.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by the processor, the program implements the dynamic control method for UAV formation collaboration based on HarmonyOS distributed soft bus as described in any one of claims 1-7.

Images