Multi-unmanned aerial vehicle intelligent cooperation and compatibility ground control system

Abstract

The invention discloses a multi-unmanned aerial vehicle intelligent cooperation and compatibility ground control system, and the system comprises the following modules: a communication data collection module which is used for receiving an original communication data frame from an access unmanned aerial vehicle; the collaborative sparse coding modeling module is used for carrying out feature extraction on the data fragments; the protocol structure identification and semantic mapping module is used for automatically identifying the communication protocol type according to the structure feature matrix; the unified control semantic module is used for defining protocol-independent control instructions corresponding to various flight control operations; the instruction translation module is used for calling a semantic unit in a control semantic layer according to a task instruction; the instruction execution module is used for sending the translated instruction to the target unmanned aerial vehicle; and the control terminal is used for graphically displaying task issuing, instruction editing and operation of the unmanned aerial vehicle. According to the invention, a communication protocol identification and mapping algorithm based on cooperative sparse coding is fused, and a multi-unmanned aerial vehicle intelligent cooperative and compatible ground control system is constructed.

Description

Technical Field

[0001] The present invention relates to the technical field of unmanned aerial vehicle systems, and in particular to a multi-unmanned aerial vehicle intelligent coordination and compatibility ground control system. Background Art

[0002] With the widespread application of drone technology in civil, industrial, and defense fields, multi-drone collaborative systems are becoming an essential component of critical scenarios such as emergency rescue, disaster monitoring, logistics, and border patrol. Multi-drone systems typically involve the unified management and intelligent collaboration of multiple aircraft, requiring access, scheduling, control, and feedback for each drone within a ground control system. However, the current development of multi-drone control systems still faces numerous practical challenges, particularly in the areas of heterogeneous communication protocols, control interface complexity, mission coordination efficiency, system stability, and command compatibility.

[0003] In existing technologies, ground control systems typically rely on protocols and SDK interfaces provided by a single model or specific manufacturer. Their underlying communication protocols are highly coupled and closed, making them incapable of supporting unified control of heterogeneous drone clusters. Most control systems require specialized adaptation for each new type of drone, including but not limited to redeveloping driver modules, interface definitions, and field parsing logic. This requires significant engineering effort and poor scalability. Furthermore, drones from different manufacturers exhibit significant differences in protocol design, including inconsistencies in message structure, field meanings, encryption methods, and transmission frequencies, further hindering the integration and collaborative operation of multi-platform systems.

[0004] In terms of task scheduling, traditional drone control systems generally use static instruction sets or rule engines for task allocation, making it difficult to dynamically schedule drones based on real-time status (such as flight time, battery life, payload capacity, and communication link quality). This scheduling approach lacks flexibility in response to situations such as varying task urgency, task reconfiguration, or drone loss. More critically, current systems typically lack a protocol semantic abstraction layer, preventing task instructions from automatically adapting to different communication protocol formats before being sent, limiting the ground system's ability to manage heterogeneous drone swarms.

[0005] At the control semantics level, existing technologies generally use low-level instructions as the core control method, lacking the ability to uniformly abstract and reconstruct control behaviors. Due to significant differences in instruction granularity, field structure, and execution methods among different protocols, task logic needs to be manually disassembled into multiple low-level operation instructions, which not only increases the operator's burden but also makes it difficult for the system to achieve high-level intelligent task-driven control. At the same time, the return format of status data is also inconsistent, further increasing the complexity of status monitoring and task feedback. In particular, graphical control terminals often encounter problems such as data parsing and status rendering that cannot be uniformly rendered.

[0006] In response to the deficiencies of the above-mentioned prior art, the present invention proposes a multi-UAV intelligent collaborative and compatible ground control system. The system constructs a unified data processing framework from the aspects of communication protocol identification and mapping, task instruction translation, control semantic modeling, state feedback analysis, graphic interactive playback, etc. The system innovatively introduces a protocol structure recognition algorithm based on collaborative sparse coding, which can realize structured modeling and automatic protocol classification of access UAV communication data without relying on protocol documents. By establishing a unified control semantic layer, the present invention bridges the semantic gap from the upper computer task logic to the execution mechanism of different types of UAVs, and realizes protocol-independent task instruction translation.

[0007] The system also incorporates a closed-loop control structure encompassing multiple processes, including task scheduling, link transmission, state analysis, and visual rendering, ensuring consistency and reliability throughout the entire process, from task planning and control execution to status feedback. Through a graphical interactive terminal, users can view real-time information such as task execution status, command responses, flight paths, and system alarms. The system also supports tasks backtracking, log retrieval, and exception analysis.

[0008] To sum up, the existing multi-UAV ground control system still has technical shortcomings in terms of protocol compatibility, control semantic consistency, intelligent task scheduling, status display and feedback mechanism. The present invention is a systematic solution to these core problems, with obvious technological progress and practical application value in underlying control abstraction, intermediate protocol adaptation, intelligent task translation and human-computer interaction.

[0009] Therefore, how to provide a multi-UAV intelligent collaborative and compatible ground control system is a problem that technical personnel in this field urgently need to solve. Summary of the Invention

[0010] One purpose of the present invention is to propose a multi-UAV intelligent collaborative and compatible ground control system, which integrates the communication protocol identification and mapping algorithm based on collaborative sparse coding. Based on an in-depth analysis of the defects of the existing technology, it has made systematic improvements in multiple key aspects such as protocol compatibility, control unification, task automation and information interaction efficiency, significantly improving the intelligent collaborative capability of multi-UAV clusters.

[0011] A multi-UAV intelligent coordination and compatibility ground control system according to an embodiment of the present invention is characterized by including the following modules:

[0012] The communication data acquisition module is used to receive the original communication data frames from the connected drone, segment them according to a fixed time window, and output the data segments to be analyzed;

[0013] A collaborative sparse coding modeling module is used to extract features from the data segments, sparsely represent the data structure based on a preset sparse coding dictionary, and collaboratively model the communication data of multiple drones to output a structural feature matrix;

[0014] A protocol structure recognition and semantic mapping module is used to automatically identify the communication protocol type based on the structural feature matrix, extract the instruction field structure, and map the control instructions of different protocols to a unified semantic tag set;

[0015] A unified control semantics module is used to define protocol-independent control instructions corresponding to various flight control operations, namely host computer task instructions. The host computer task instructions include control semantic units such as start, navigation, hovering, return and payload control;

[0016] The instruction translation module is used to call the semantic unit in the unified control semantic module according to the task instruction of the host computer and translate the semantic unit into the native protocol instruction of the target UAV;

[0017] The command execution module is used to send the translated command to the target UAV through the corresponding communication channel and receive the status data returned by the UAV in real time;

[0018] The control terminal is used to graphically display the identification information, protocol type, mission status, and execution feedback of the connected drone. It supports mission issuance, command editing, status monitoring, and log playback operations.

[0019] A multi-UAV intelligent coordination and compatibility ground control method according to an embodiment of the present invention includes the following steps:

[0020] S1, receives the original communication data frame from the connected drone, and processes it in segments according to a fixed time window, outputting the data segments to be analyzed;

[0021] S2. Extract features from the data segments, perform sparse representation on the data structure based on a preset sparse coding dictionary, perform collaborative modeling on multiple UAV communication data, and output a structural feature matrix;

[0022] S3. Automatically identify the communication protocol type according to the structural feature matrix, extract the instruction field structure, and map the control instructions of different protocols to a unified semantic tag set;

[0023] S4. Based on the control instructions in the tag set, define various control instructions that are independent of the flight control operation protocol to form the host computer task instructions;

[0024] S5. Call the semantic unit in the unified control semantic module with the task instruction of the host computer, and translate the semantic unit into the native protocol instruction of the target UAV;

[0025] S6. Sending native protocol instructions based on the drone to the target drone through the corresponding communication channel, and receiving status data returned by the drone in real time;

[0026] S7. Graphically display the status data returned by the drone, including identification information, protocol type, mission status, and execution feedback. It supports mission issuance, command editing, status monitoring, and log playback operations.

[0027] Optionally, the S1 specifically includes:

[0028] S11. In the multi-UAV intelligent collaborative and compatible ground control system, the communication data acquisition module continuously receives the communication data stream from the connected UAV. The communication data is represented as a data sequence in the form of message frames, S = {s1, s2, ..., s t},in Represents the p-dimensional original data frame sampled at the t-th time point;

[0029] S12, set the sampling time window length of the original communication data frame to Divide the data stream S into n non-overlapping window segments in chronological order and construct a set of original data segments:

[0030] X={x1,x2,...,x n},x i ={s t |t∈[iΔt,(i+1)Δt]};

[0031] Each of these m is the combined feature dimension extracted within the window Δt;

[0032] S13, each x in the original data segment set i Perform normalization to obtain a normalized data set

[0033] S14, for each normalized Extract communication behavior feature vector f i ,in The feature vector dimensions include field change rate, bit distribution entropy, field periodicity and data jump rate.

[0034] Optionally, the S2 specifically includes:

[0035] S21. In a multi-UAV intelligent collaborative and compatible ground control system, the communication data segment set X output by step S1 is used as sparse coding input data;

[0036] S22, construct a set of basic sparse dictionaries D for offline training based on the output communication data segment set X = {d1, d2, ..., d2, ..., d k}, forming a dictionary matrix in represents the jth dictionary atom;

[0037] S23, for each input segment x of the base sparse dictionary D i , solve the sparse coding coefficients corresponding to the input segment The sparse coding coefficients are made to meet the following sparse reconstruction objectives:

[0038]

[0039] in is the sparsity regularization parameter, ‖·‖2 represents the Euclidean norm, and ‖·‖1 represents the L1 sparsity penalty term;

[0040] S24. Combine the sparse coefficient vectors of all samples into a sparse coefficient matrix Represents the sparse projection of all data segments into the dictionary space;

[0041] S25. To enhance cluster consistency, collaborative sparse constraints are applied to the data from different connected drones to construct collaborative constraint items:

[0042]

[0043] where α ij ∈{0,1} indicates whether the coding pair comes from the same drone group, and this constraint is used to optimize the similarity within the cluster;

[0044] S26, using the optimized sparse coding matrix The output is a structural feature matrix, which serves as the input of the subsequent communication protocol structure recognition and semantic mapping module.

[0045] Optionally, the S3 specifically includes:

[0046] S31, in the multi-UAV intelligent collaborative and compatible ground control system, receiving the structural feature matrix A' output from step S2 for each encoding vector a' in the structural feature matrix A' i Perform sparse pattern clustering using a similarity metric based on cosine similarity:

[0047]

[0048] Construct a similarity graph G = (V, E), where nodes V = {a'1, ..., a'n} and edges E reflect the similarity relationship between codes;

[0049] S32, dividing the graph G by using a graph clustering algorithm, classifying encoding vectors with similar structures into the same cluster, where each cluster represents a communication protocol structure template;

[0050] S33. Perform a field boundary recognition operation on the communication segments within each protocol structure template, use the maximum gradient method to analyze the bit distribution changes between fields, and determine the start and end position set of the instruction field:

[0051]

[0052] in Indicates the start and end byte index of the fth field;

[0053] S34. Extract the length, frequency and content change pattern of each field in the field sequence and establish a feature description vector of the protocol field Comparing the feature description vector with the system's built-in semantic template library;

[0054] S35. Match each identified field to a control semantic label l∈L, and obtain a mapping function from protocol to semantic label:

[0055] φ:F j →L={l1,l2,...,l q};

[0056] The output tag set serves as the basis for constructing the unified control semantic unit in step S4.

[0057] Optionally, the S4 specifically includes:

[0058] S41. In a multi-UAV intelligent collaborative and compatible ground control system, establish a control semantic description model, wherein the model defines control semantic units, each semantic unit corresponding to a typical flight control behavior;

[0059] S42, mapping the control instruction fields identified in the semantic tag set to corresponding control semantic units according to their function types, generating a semantic instruction comparison table, and establishing a correspondence between control behaviors and underlying instruction fields;

[0060] S43. Set the data structure of the control semantic unit, including the semantic number, description text, parameter template, unit format, and mapping field index, to ensure that the semantic definition has a unified standard and callable interface in the control layer;

[0061] S44. Build a semantic management module to support dynamic registration, update, and deregistration of semantic units, allowing real-time addition of new semantic definition entries based on newly identified protocol fields;

[0062] S45, standardize and encapsulate the calling logic of the semantic unit so that the semantic unit can be directly referenced by the task instruction of the host computer;

[0063] S46. Set up a semantic version control mechanism to manage the change records, compatibility verification and historical instruction comparison logs of semantic units.

[0064] Optionally, the S5 specifically includes:

[0065] S51, receiving a task instruction from a host computer, inputting the task instruction from the host computer into a semantic parsing module, identifying the control behavior category corresponding to the instruction, and calling the corresponding control semantic unit in a unified control semantic module;

[0066] S52, filling parameters of the called control semantic unit according to the control parameters contained in the host computer task instruction, and constructing a complete control semantic instruction template;

[0067] S53. Based on the communication protocol type of the target UAV, call the corresponding protocol translation mapping relationship to convert the semantic instruction template into a native protocol instruction format that can be recognized and executed by the target UAV;

[0068] S54: Store the translated native protocol instructions into a sending queue, mark the target UAV identification information, task sequence number, and sending priority, and provide a structured native instruction package for subsequent communication modules.

[0069] Optionally, the S6 specifically includes:

[0070] S61. In the multi-UAV intelligent collaborative and compatible ground control system, determine the target UAV number according to the task scheduling result, and call the communication link management module to select a currently available communication channel, wherein the communication channel includes an ad hoc network and a cellular network;

[0071] S62, encapsulating the generated native protocol instruction into a protocol, where the encapsulation content includes a frame header identifier, a control payload, a checksum, a timestamp, and a target address field, and forming a transmission frame;

[0072] S63. Send frames to the target UAV in the form of data packets via the link. The system supports link status monitoring and packet loss retransmission mechanism to ensure that each frame is confirmed to be received within the preset time window;

[0073] S64, monitoring and receiving a status data frame returned by the target UAV, wherein the status feedback frame includes flight attitude, position, speed, battery level, mission execution status, current flight control mode, and warning information;

[0074] S65. Decode and parse the received status data frame, and fill the decoding result into the status update vector as a basis for subsequent control and monitoring;

[0075] S66. Establish a task mapping relationship based on the state update vector and the original task instruction, record the task execution status, and write it into the task execution log queue for subsequent processing modules to call.

[0076] Optionally, the S7 specifically includes:

[0077] S71. In a multi-UAV intelligent collaborative and compatible ground control system, receive and organize a state data set, extract flight parameters, mission status, and feedback information of each UAV, and generate a state mapping table;

[0078] S72. Correlate and match the state mapping table with the task scheduling record to generate a multi-UAV state-task correspondence diagram for visual display. The current flight position, task execution stage, communication status, remaining power, and command response status of each UAV are displayed in the form of a real-time map and dynamic icons on the graphical control terminal;

[0079] S73. Provide visual interactive functions to support operators in issuing tasks, editing paths, switching control modes, and interrupting tasks for each drone in the terminal interface. Record all command issuance behaviors, drone status feedback, task result data, and exception information, build a task execution log, and organize it into a log linked list structure in chronological order.

[0080] S74. Set up a log playback function to allow the operator to retrieve and review the historical mission execution process based on the mission number and drone number for review, retraining or abnormal diagnosis.

[0081] The beneficial effects of the present invention are:

[0082] First, by introducing a collaborative sparse coding modeling algorithm, this invention achieves structured analysis and automatic protocol recognition of communication data from different drone models, overcoming the limitations of traditional systems that rely on manual adaptation and configuration of protocol parsing modules. This mechanism can autonomously extract instruction field structures without requiring prior access to protocol format documentation, effectively improving the rapid access capabilities of heterogeneous drones and significantly enhancing the protocol adaptability and scalability of the ground control system.

[0083] Secondly, the present invention constructs a unified control semantic layer, systematically decoupling mission logic from the underlying protocol and abstracting complex flight control behaviors into universal semantic units, such as startup, navigation, hovering, return, and payload control. Through an automatic translation mechanism from semantic instructions to native protocol instructions, the same mission instructions can be rapidly deployed and executed on multiple drones, significantly improving the efficiency and consistency of multi-drone collaborative operations and reducing the complexity of system maintenance and operations.

[0084] Furthermore, this invention introduces a closed-loop control mechanism for task execution and status interaction, supporting real-time reception and analysis of drone status data, and displaying, monitoring, and task playback in a graphical interface. Operators can not only intuitively view drone execution status, protocol types, and feedback results, but also edit, adjust, and resend tasks, enabling refined and dynamic management of the unmanned system. This integrated control approach, based on semantic modeling, protocol mapping, and task mapping, significantly enhances the system's human-machine interaction capabilities and operational flexibility.

[0085] Overall, the present invention effectively solves key problems existing in existing ground control systems, such as inconsistent protocols, incompatible semantics, inefficient task scheduling, and complex state interactions. It significantly improves the intelligent collaboration capabilities of multi-model drone clusters, and provides strong system support for air-ground linkage in complex application scenarios. It has strong technological advancement and engineering practical value. BRIEF DESCRIPTION OF THE DRAWINGS

[0086] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0087] Figure 1 This is a module structure diagram of the multi-UAV intelligent collaborative and compatible ground control system proposed by the present invention.

[0088] Figure 2 This is the flow chart of the multi-UAV intelligent collaborative control proposed in this invention. DETAILED DESCRIPTION

[0089] The present invention will now be described in further detail with reference to the accompanying drawings, which are simplified schematic diagrams that illustrate the basic structure of the present invention in a schematic manner.

[0090] refer to Figure 1 , a multi-UAV intelligent coordination and compatibility ground control system, including the following modules:

[0091] The communication data acquisition module is used to receive the original communication data frames from the connected drone, segment them according to a fixed time window, and output the data segments to be analyzed;

[0092] A collaborative sparse coding modeling module is used to extract features from the data segments, sparsely represent the data structure based on a preset sparse coding dictionary, and collaboratively model the communication data of multiple drones to output a structural feature matrix;

[0093] A protocol structure recognition and semantic mapping module is used to automatically identify the communication protocol type based on the structural feature matrix, extract the instruction field structure, and map the control instructions of different protocols to a unified semantic tag set;

[0094] A unified control semantics module is used to define protocol-independent control instructions corresponding to various flight control operations, namely host computer task instructions. The host computer task instructions include control semantic units such as start, navigation, hovering, return and payload control;

[0095] The instruction translation module is used to call the semantic unit in the unified control semantic module according to the task instruction of the host computer and translate the semantic unit into the native protocol instruction of the target UAV;

[0096] The command execution module is used to send the translated command to the target UAV through the corresponding communication channel and receive the status data returned by the UAV in real time;

[0097] The control terminal is used to graphically display the identification information, protocol type, mission status, and execution feedback of the connected drone. It supports mission issuance, command editing, status monitoring, and log playback operations.

[0098] The present invention provides a multi-UAV intelligent collaborative and compatible ground control system. Its overall architecture is oriented towards heterogeneous UAV access and intelligent task collaboration. The system modularizes key processes such as communication protocol parsing, control semantic modeling, command translation and execution, and interactive state feedback, creating an intelligent control system with high compatibility, high scalability, and high task agility. The system includes multiple core modules, including communication data acquisition, collaborative sparse modeling, protocol structure recognition and semantic mapping, unified semantic modeling, command translation, task execution, and interactive control terminals. These modules work together to form a complete closed-loop control link from "protocol-independent access" to "semantically consistent control." This creates an intelligent UAV control platform that is adaptable to a variety of communication structures. The platform is particularly suitable for the rapid deployment, task distribution, and collaborative execution of heterogeneous UAV clusters in complex scenarios, significantly improving the system's overall compatibility, intelligence, and task execution efficiency.

[0099] refer to Figure 2, a multi-UAV intelligent coordination and compatibility ground control method, comprising the following steps:

[0100] S1, receives the original communication data frame from the connected drone, and processes it in segments according to a fixed time window, outputting the data segments to be analyzed;

[0101] S2. Extract features from the data segments, perform sparse representation on the data structure based on a preset sparse coding dictionary, perform collaborative modeling on multiple UAV communication data, and output a structural feature matrix;

[0102] S3. Automatically identify the communication protocol type according to the structural feature matrix, extract the instruction field structure, and map the control instructions of different protocols to a unified semantic tag set;

[0103] S4. Based on the control instructions in the tag set, define various control instructions that are independent of the flight control operation protocol to form the host computer task instructions;

[0104] S5. Call the semantic unit in the unified control semantic module with the task instruction of the host computer, and translate the semantic unit into the native protocol instruction of the target UAV;

[0105] S6. Sending native protocol instructions based on the drone to the target drone through the corresponding communication channel, and receiving status data returned by the drone in real time;

[0106] S7. Graphically display the status data returned by the drone, including identification information, protocol type, mission status, and execution feedback. It supports mission issuance, command editing, status monitoring, and log playback operations.

[0107] The method of the present invention defines the task processing paths between functional modules in heterogeneous UAV communication and control, and constructs a complete data control link that spans perception, recognition, parsing, control, execution, and feedback based on semantic modeling and protocol mapping mechanisms. This method not only addresses the heterogeneity of communication protocols but also establishes a semantic logic closed loop from raw data input to flight control behavior output, supporting the unified scheduling and task coordination of different types of UAVs by ground systems. This not only improves interoperability between multiple UAV models, but also enhances the system's ability to ensure communication compatibility and control stability in complex mission scenarios, providing fundamental support for the rapid deployment and coordinated operations of heterogeneous UAV clusters.

[0108] In this embodiment, S1 specifically includes:

[0109] S11. In the multi-UAV intelligent collaborative and compatible ground control system, the communication data acquisition module continuously receives the communication data stream from the connected UAV. The communication data is represented as a data sequence in the form of message frames, S = {s1, s2, ..., s t},in Represents the p-dimensional original data frame sampled at the t-th time point;

[0110] S12, set the sampling time window length of the original communication data frame to Divide the data stream S into n non-overlapping window segments in chronological order and construct a set of original data segments:

[0111] X={x1,x2,...,x n},x i ={s t |t∈[iΔt,(i+1)Δt]};

[0112] Each of these m is the combined feature dimension extracted within the window Δt;

[0113] S13, each x in the original data segment set i Perform normalization to obtain a normalized data set

[0114] S14, for each normalized Extract communication behavior feature vector f i ,in The feature vector dimensions include field change rate, bit distribution entropy, field periodicity and data jump rate.

[0115] This step proposes a raw data frame processing method for heterogeneous UAV communication protocols. By introducing a window segmentation mechanism, normalization preprocessing, and a multi-dimensional feature extraction process, the unstructured communication stream is converted into a structured feature input, providing a unified time series data foundation for subsequent sparse coding modeling and automatic protocol recognition. By combining communication behavior indicators such as field change rate, bit distribution entropy, periodicity, and jump rate, the microstructure differences and behavioral pattern characteristics of different UAV communication protocols at the frame level are fully retained, effectively enhancing the model's ability to express the implicit relationship between protocol type, field distribution, and control semantics. This step plays a key role as the raw data entry point in the entire system, laying a stable data preprocessing foundation for protocol-independent control semantic extraction, and is a prerequisite for achieving highly compatible UAV access capabilities.

[0116] Optionally, the S2 specifically includes:

[0117] S21. In a multi-UAV intelligent collaborative and compatible ground control system, the communication data segment set X output by step S1 is used as sparse coding input data;

[0118] S22, construct a set of basic sparse dictionaries D for offline training based on the output communication data segment set X = {d1, d2, ..., d2, ..., d k}, forming a dictionary matrix in represents the jth dictionary atom;

[0119] S23, for each input segment x of the base sparse dictionary D i , solve the sparse coding coefficients corresponding to the input segment The sparse coding coefficients are made to meet the following sparse reconstruction objectives:

[0120]

[0121] in is the sparsity regularization parameter, ‖·‖2 represents the Euclidean norm, and ‖·‖1 represents the L1 sparsity penalty term;

[0122] S24. Combine the sparse coefficient vectors of all samples into a sparse coefficient matrix Represents the sparse projection of all data segments into the dictionary space;

[0123] S25. To enhance cluster consistency, collaborative sparse constraints are applied to the data from different connected drones to construct collaborative constraint items:

[0124]

[0125] where α ij ∈{0,1} indicates whether the coding pair comes from the same drone group, and this constraint is used to optimize the similarity within the cluster;

[0126] S26, using the optimized sparse coding matrix The output is a structural feature matrix, which serves as the input of the subsequent communication protocol structure recognition and semantic mapping module.

[0127] This step proposes a communication feature modeling method that integrates basic sparse representation and collaborative modeling mechanisms. By introducing dictionary learning and a sparse projection framework, it achieves structured representation and semantic compression modeling of communication data fragments. This step not only effectively reconstructs communication behavior characteristics based on a preset sparse dictionary, but also further introduces a collaborative sparse constraint strategy to impose consistency constraints on encoding results from the same drone group, enhancing the model's ability to express group behavior similarities. This method significantly improves the ability to identify potential structural commonalities and differences between protocols in multi-drone access scenarios and is a key support link in building a unified cross-protocol semantic model.

[0128] In this embodiment, S3 specifically includes:

[0129] S31, in the multi-UAV intelligent collaborative and compatible ground control system, receiving the structural feature matrix A' output from step S2 for each encoding vector a' in the structural feature matrix A' i Perform sparse pattern clustering using a similarity metric based on cosine similarity:

[0130]

[0131] Construct a similarity graph G = (V, E), where nodes V = {a'1, ..., a'n} and edges E reflect the similarity relationship between codes;

[0132] S32, dividing the graph G by using a graph clustering algorithm, classifying encoding vectors with similar structures into the same cluster, where each cluster represents a communication protocol structure template;

[0133] S33. Perform a field boundary recognition operation on the communication segments within each protocol structure template, use the maximum gradient method to analyze the bit distribution changes between fields, and determine the start and end position set of the instruction field:

[0134]

[0135] in Indicates the start and end byte index of the fth field;

[0136] S34. Extract the length, frequency and content change pattern of each field in the field sequence and establish a feature description vector of the protocol field Comparing the feature description vector with the system's built-in semantic template library;

[0137] S35. Match each identified field to a control semantic label l∈L, and obtain a mapping function from protocol to semantic label:

[0138] φ:F j →L={l1,l2,...,lq};

[0139] The output tag set serves as the basis for constructing the unified control semantic unit in step S4.

[0140] This step proposes an automated protocol structure parsing method based on sparse pattern clustering and field boundary identification. This method constructs a similarity graph to cluster the structure of communication data and uses the maximum gradient method to precisely delineate instruction field boundaries. This method effectively extracts structural templates for various communication protocols and maps semantic labels based on field behavior characteristics. This provides a clear and compatible structural semantic basis for unified control semantic units, and is a key step in implementing protocol-independent control strategies.

[0141] In this embodiment, the S4 specifically includes:

[0142] S41. In a multi-UAV intelligent collaborative and compatible ground control system, establish a control semantic description model, wherein the model defines control semantic units, each semantic unit corresponding to a typical flight control behavior;

[0143] S42, mapping the control instruction fields identified in the semantic tag set to corresponding control semantic units according to their function types, generating a semantic instruction comparison table, and establishing a correspondence between control behaviors and underlying instruction fields;

[0144] S43. Set the data structure of the control semantic unit, including the semantic number, description text, parameter template, unit format, and mapping field index, to ensure that the semantic definition has a unified standard and callable interface in the control layer;

[0145] S44. Build a semantic management module to support dynamic registration, update, and deregistration of semantic units, allowing real-time addition of new semantic definition entries based on newly identified protocol fields;

[0146] S45, standardize and encapsulate the calling logic of the semantic unit so that the semantic unit can be directly referenced by the task instruction of the host computer;

[0147] S46. Set up a semantic version control mechanism to manage the change records, compatibility verification and historical instruction comparison logs of semantic units.

[0148] This step proposes a unified control semantics modeling approach for heterogeneous protocol instructions. By constructing standardized control semantic units and a semantic mapping mechanism, this approach unifies the semantic abstraction and call interface between flight control behavior and underlying field instructions. The system supports dynamic expansion and version control of semantic units, providing a clearly structured, standardized, and maintainable control semantics foundation for subsequent instruction translation and task execution.

[0149] In this embodiment, the S5 specifically includes:

[0150] S51, receiving a task instruction from a host computer, inputting the task instruction from the host computer into a semantic parsing module, identifying the control behavior category corresponding to the instruction, and calling the corresponding control semantic unit in a unified control semantic module;

[0151] S52, filling parameters of the called control semantic unit according to the control parameters contained in the host computer task instruction, and constructing a complete control semantic instruction template;

[0152] S53. Based on the communication protocol type of the target UAV, call the corresponding protocol translation mapping relationship to convert the semantic instruction template into a native protocol instruction format that can be recognized and executed by the target UAV;

[0153] S54: Store the translated native protocol instructions into a sending queue, mark the target UAV identification information, task sequence number, and sending priority, and provide a structured native instruction package for subsequent communication modules.

[0154] This step proposes a semantically driven command translation mechanism. By parsing task instructions, calling semantic units, and filling in parameters, a unified control semantic template is constructed. Combined with protocol mappings, native control commands recognizable to the target drone are generated. This mechanism automatically translates high-level task logic into low-level protocol commands, improving the system's adaptability and execution consistency for control commands from multiple drone models.

[0155] In this embodiment, S6 specifically includes:

[0156] S61. In the multi-UAV intelligent collaborative and compatible ground control system, determine the target UAV number according to the task scheduling result, and call the communication link management module to select a currently available communication channel, wherein the communication channel includes an ad hoc network and a cellular network;

[0157] S62, encapsulating the generated native protocol instruction into a protocol, where the encapsulation content includes a frame header identifier, a control payload, a checksum, a timestamp, and a target address field, and forming a transmission frame;

[0158] S63. Send frames to the target UAV in the form of data packets via the link. The system supports link status monitoring and packet loss retransmission mechanism to ensure that each frame is confirmed to be received within the preset time window;

[0159] S64, monitoring and receiving a status data frame returned by the target UAV, wherein the status feedback frame includes flight attitude, position, speed, battery level, mission execution status, current flight control mode, and warning information;

[0160] S65. Decode and parse the received status data frame, and fill the decoding result into the status update vector as a basis for subsequent control and monitoring;

[0161] S66. Establish a task mapping relationship based on the state update vector and the original task instruction, record the task execution status, and write it into the task execution log queue for subsequent processing modules to call.

[0162] This step proposes a UAV communication execution mechanism that integrates communication link selection, command encapsulation, frame transmission, and state feedback analysis. This ensures the reliable issuance of control commands and closed-loop state collection in a multi-link environment. Through structured transmission frame encapsulation and state data analysis and processing, the system effectively improves communication stability and mission execution transparency in complex networking scenarios, providing a stable and efficient communication execution foundation for intelligent collaborative control of multiple UAVs.

[0163] In this embodiment, the S7 specifically includes:

[0164] S71. In a multi-UAV intelligent collaborative and compatible ground control system, receive and organize a state data set, extract flight parameters, mission status, and feedback information of each UAV, and generate a state mapping table;

[0165] S72. Correlate and match the state mapping table with the task scheduling record to generate a multi-UAV state-task correspondence diagram for visual display. The current flight position, task execution stage, communication status, remaining power, and command response status of each UAV are displayed in the form of a real-time map and dynamic icons on the graphical control terminal;

[0166] S73. Provide visual interactive functions to support operators in issuing tasks, editing paths, switching control modes, and interrupting tasks for each drone in the terminal interface. Record all command issuance behaviors, drone status feedback, task result data, and exception information, build a task execution log, and organize it into a log linked list structure in chronological order.

[0167] S74. Set up a log playback function to allow the operator to retrieve and review the historical mission execution process based on the mission number and drone number for review, retraining or abnormal diagnosis.

[0168] This step proposes a mission control interaction mechanism based on state mapping and visual management. By constructing a real-time state-task correspondence diagram, it enables dynamic display and coordinated operation of multiple UAV states on a graphical terminal. The system supports multiple interactive functions such as task scheduling, state monitoring, command recording, and log playback. This not only improves the operator's control efficiency and mission perception, but also provides systematic data support for subsequent mission review and anomaly analysis, establishing a full-process visual control closed loop for multi-UAV intelligent collaboration.

[0169] Example 1:

[0170] To verify the feasibility of this invention, it was applied to a "drone-based ground collaborative communication compatibility exercise" conducted in a certain province. The region faced complex terrain, variable communication channels, and a diverse range of drone models from three different manufacturers, using different proprietary communication protocols. These drones presented typical challenges, including protocol heterogeneity, inconsistent instructions, and difficulty in scheduling. These challenges perfectly matched the design intent and application scenarios of this invention.

[0171] The objective of the exercise mission is to complete batch image acquisition and multi-angle video coverage monitoring of areas with abnormal surface temperature differences. A total of 9 drones were deployed at three take-off points. The system was required to achieve automatic identification of drones of different models, command compatibility and synchronous task scheduling, complete multiple batches of flight missions and ensure complete visual display and backtracking of mission logs.

[0172] At the start of a mission, the ground control system continuously receives data frames from each connected drone through the communication data acquisition module. In the absence of uniform protocols, the system introduces the collaborative sparse coding communication protocol identification and mapping algorithm of the present invention, which segments the original communication data stream according to fixed time windows and extracts communication behavior characteristics such as bit distribution, field change rate, and periodic patterns to form a set of data segment feature vectors. This set is then input into the collaborative sparse modeling module, where sparse coefficients are solved using an offline constructed sparse dictionary to achieve sparse representation of different data frame structures.

[0173] During the mission, the ground control system first accessed the communication streams of nine drones. The average length of the communication protocol field of each drone was between 42 and 97 bytes, and the field order and verification mechanism were completely inconsistent. The system automatically collected the original communication frames and, through the collaborative sparse modeling algorithm of the present invention, completed the sparse structure representation extraction of all communication fragments in about 1.2 seconds, matched the protocol field distribution, and standardized all flight control instructions into a unified control semantic instruction set through the semantic mapping module. The entire protocol parsing, control template matching, and unified semantic translation process took less than 4 seconds in total, which is much better than the manual configuration method (the average configuration time for each drone is about 8-10 minutes).

[0174] During the scheduling phase, the system maps the "start-autonomous navigation-image acquisition-hovering-return" task chain into a unified command sequence based on different flight control semantic models. This sequence is then successfully translated into the corresponding protocol format for each drone, with an average latency of less than 0.6 seconds for each semantic command to be successfully translated into the native protocol. Through the graphical interface of the control terminal, operators can view the flight status, battery level, location information, and current mission progress of all drones in real time, and execute dynamic interruption or re-routing commands based on status feedback.

[0175] After the mission is complete, the system performs structured parsing of all drone return status data frames, averaging over 2,400 frames per drone. The system then enables complete trajectory reconstruction, mission result image correlation, and command response feedback log playback on the terminal interface. The total mission log storage capacity is 46.7MB, with a 100% data recovery rate. It supports retrieval and comparison based on mission and drone numbers.

[0176] Compared to traditional solutions, this invention significantly improves task scheduling time efficiency and the system's protocol adaptability. Comparative data shows that traditional methods require manual setup of adapter interfaces and configuration of protocol command mappings, resulting in long deployment cycles and high error rates. The ground control system constructed by this invention improves deployment time efficiency by 92% in similar heterogeneous UAV access scenarios, reduces scheduling response latency by approximately 84%, and achieves near-100% protocol adaptation accuracy.

[0177] The following table shows the data records at different stages of the above-mentioned task rehearsal scenario, covering indicators such as access information, protocol parsing performance, task instruction issuance efficiency, status feedback, log storage and system response performance, fully reflecting the technical effects and practical value of this invention in a real environment.

[0178]

[0179]

[0180] A comprehensive analysis of the data listed in the above table clearly shows that the "multi-UAV intelligent collaborative and compatible ground control system" of the present invention has significant advantages over traditional control systems in multiple key dimensions.

[0181] First, in terms of communication protocol recognition, the system uses collaborative sparse coding modeling and automated semantic mapping mechanisms to compress the communication protocol parsing process, which originally required manual adaptation and took over 150 seconds, to an average of 1.2 seconds, improving recognition efficiency by 98.6%. This means that the system can automatically identify and extract the structure of heterogeneous drone communication protocols in a very short time, greatly improving access efficiency.

[0182] Secondly, in terms of command semantic translation, through the design of a unified control semantic layer and automatic translation module, the average time required for native protocol translation was reduced from 12.4 seconds to 0.6 seconds, reducing the response delay by more than 95%. At the same time, the command compatibility success rate was greatly improved from 64.1% of the traditional system to 99.7%, almost achieving semantic alignment of all protocol commands, ensuring concurrent scheduling and precise execution of multiple models of drones.

[0183] In terms of mission response, the average mission response latency per drone decreased from 18.7 seconds to 2.9 seconds, an improvement of approximately 84.5%. The overall mission execution time was shortened from 45 minutes to 16.2 minutes, significantly improving scheduling efficiency. Throughout the mission cycle, the number of operators required to complete mission dispatch, monitoring, and log management has been reduced from three to one, significantly reducing human resource investment.

[0184] In terms of system feedback and visualization capabilities, the accuracy of status data decoding has been increased to 100%, ensuring highly reliable feedback on the drone's flight status and mission execution. Furthermore, the graphical terminal response latency has been reduced from 890 milliseconds to 126 milliseconds, improving real-time control. More importantly, the system provides full mission logging and backtracking capabilities, supporting the query and reproduction of over 20,000 status frames, meeting the needs of scenarios such as operation and maintenance review, anomaly analysis, and training and teaching.

[0185] The collaborative sparse coding algorithm introduced in this multi-UAV field exercise not only serves as a structural representation tool but also serves as the algorithmic foundation for addressing the core heterogeneity of multi-UAV communication protocols. Traditional protocol recognition methods rely on rule matching or field templates. However, when dealing with private, non-public UAV protocols, such methods often fail due to field offsets, encryption perturbations, or dynamic length changes. These methods are particularly prone to poor stability and high error rates in the face of communication interference or high-frequency command exchanges.

[0186] In summary, experiments have shown that the present invention not only greatly improves the deployment efficiency and control accuracy of the ground control system in heterogeneous UAV access scenarios, but also demonstrates performance superior to traditional solutions in communication analysis, task distribution, visual feedback, and system intelligent closed loop, and has good engineering feasibility and promotion prospects.

[0187] The above description is only a preferred specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any technician familiar with the technical field, within the technical scope disclosed by the present invention, who makes equivalent replacements or changes based on the technical solution and inventive concept of the present invention, should be covered by the scope of protection of the present invention.

Claims

1. A multi-UAV intelligent collaborative and compatible ground control system, characterized by: Includes the following modules: The communication data acquisition module is used to receive the original communication data frames from the connected drone, segment them according to a fixed time window, and output the data segments to be analyzed; A collaborative sparse coding modeling module is used to extract features from the data segments, sparsely represent the data structure based on a preset sparse coding dictionary, and collaboratively model the communication data of multiple drones to output a structural feature matrix; A protocol structure recognition and semantic mapping module is used to automatically identify the communication protocol type based on the structural feature matrix, extract the instruction field structure, and map the control instructions of different protocols to a unified semantic tag set; A unified control semantics module is used to define protocol-independent control instructions corresponding to various flight control operations, namely host computer task instructions. The host computer task instructions include control semantic units such as start, navigation, hovering, return and payload control; The instruction translation module is used to call the semantic unit in the unified control semantic module according to the task instruction of the host computer and translate the semantic unit into the native protocol instruction of the target UAV; The command execution module is used to send the translated command to the target UAV through the corresponding communication channel and receive the status data returned by the UAV in real time; The control terminal is used to graphically display the identification information, protocol type, mission status, and execution feedback of the connected drone. It supports mission issuance, command editing, status monitoring, and log playback operations.

2. The multi-UAV intelligent coordination and compatibility ground control system according to claim 1 is characterized in that: The modules are implemented as follows: S1, receives the original communication data frame from the connected drone, and processes it in segments according to a fixed time window, outputting the data segments to be analyzed; S2. Extract features from the data segments, perform sparse representation on the data structure based on a preset sparse coding dictionary, perform collaborative modeling on multiple UAV communication data, and output a structural feature matrix; S3. Automatically identify the communication protocol type according to the structural feature matrix, extract the instruction field structure, and map the control instructions of different protocols to a unified semantic tag set; S4. Based on the control instructions in the tag set, define various control instructions that are independent of the flight control operation protocol to form the host computer task instructions; S5. Call the semantic unit in the unified control semantic module with the task instruction of the host computer, and translate the semantic unit into the native protocol instruction of the target UAV; S6. Sending native protocol instructions based on the drone to the target drone through the corresponding communication channel, and receiving status data returned by the drone in real time; S7. Graphically display the status data returned by the drone.

3. The multi-UAV intelligent coordination and compatibility ground control system according to claim 2 is characterized in that: Said S1 specifically includes: S11. In the multi-UAV intelligent collaborative and compatible ground control system, the communication data acquisition module continuously receives the communication data stream from the connected UAV. The communication data is represented as a data sequence in the form of message frames, S = {s1, s2, ..., s t },in Represents the p-dimensional original data frame sampled at the t-th time point; S12, set the sampling time window length of the original communication data frame to Divide the data stream S into n non-overlapping window segments in chronological order and construct a set of original data segments: X={x1,x2,...,x n },x i ={s t ∣t∈[iΔt,(i+1)Δt]}; Each of these m is the combined feature dimension extracted within the window Δt; S13, each x in the original data segment set i Perform normalization to obtain a normalized data set S14, for each normalized Extract communication behavior feature vector f i ,in The feature vector dimensions include field change rate, bit distribution entropy, field periodicity and data jump rate.

4. The multi-UAV intelligent coordination and compatible ground control system according to claim 2 is characterized in that , the S2 specifically includes: S21. In a multi-UAV intelligent collaborative and compatible ground control system, the communication data segment set X output by step S1 is used as sparse coding input data; S22, construct a set of basic sparse dictionaries D = {d1, d2, ..., d j ,...,d k }, forming a dictionary matrix in represents the jth dictionary atom; S23, for each input segment x of the base sparse dictionary D i , solve the sparse coding coefficients corresponding to the input segment The sparse coding coefficients are made to meet the following sparse reconstruction objectives: in is the sparsity regularization parameter, ‖·‖2 represents the Euclidean norm, and ‖·‖1 represents the L1 sparsity penalty term; S24. Combine the sparse coefficient vectors of all samples into a sparse coefficient matrix Represents the sparse projection of all data segments into the dictionary space; S25. To enhance cluster consistency, collaborative sparse constraints are applied to the data from different connected drones to construct collaborative constraint items: where α ij ∈{0,1} indicates whether the coding pair comes from the same drone group, and this constraint is used to optimize the similarity within the cluster; S26, using the optimized sparse coding matrix The output is a structural feature matrix, which serves as the input of the subsequent communication protocol structure recognition and semantic mapping module.

5. The multi-UAV intelligent coordination and compatibility ground control system according to claim 2 is characterized in that: The S3 specifically includes: S31, in the multi-UAV intelligent collaborative and compatible ground control system, receiving the structural feature matrix A' output from step S2 for each encoding vector a' in the structural feature matrix A' i Perform sparse pattern clustering using a similarity metric based on cosine similarity: Construct a similarity graph G = (V, E), where nodes V = {a'1, ..., a'n} and edges E reflect the similarity relationship between codes; S32, dividing the graph G by using a graph clustering algorithm, classifying encoding vectors with similar structures into the same cluster, where each cluster represents a communication protocol structure template; S33. Perform a field boundary recognition operation on the communication segments within each protocol structure template, use the maximum gradient method to analyze the bit distribution changes between fields, and determine the start and end position set of the instruction field: in Indicates the start and end byte index of the fth field; S34. Extract the length, frequency and content change pattern of each field in the field sequence and establish a feature description vector of the protocol field Comparing the feature description vector with the system's built-in semantic template library; S35. Match each identified field to a control semantic label l∈L, and obtain a mapping function from protocol to semantic label: φ:F j →L={l1,l2,...,l q }; The output tag set serves as the basis for constructing the unified control semantic unit in step S4.

6. The multi-UAV intelligent coordination and compatibility ground control system according to claim 2 is characterized in that: The S4 specifically includes: S41. In a multi-UAV intelligent collaborative and compatible ground control system, establish a control semantic description model, wherein the model defines control semantic units, each semantic unit corresponding to a typical flight control behavior; S42, mapping the control instruction fields identified in the semantic tag set to corresponding control semantic units according to their function types, generating a semantic instruction comparison table, and establishing a correspondence between control behaviors and underlying instruction fields; S43. Set the data structure of the control semantic unit, including the semantic number, description text, parameter template, unit format, and mapping field index, to ensure that the semantic definition has a unified standard and callable interface in the control layer; S44. Build a semantic management module to support dynamic registration, update, and deregistration of semantic units, allowing real-time addition of new semantic definition entries based on newly identified protocol fields; S45, standardize and encapsulate the calling logic of the semantic unit so that the semantic unit can be directly referenced by the task instruction of the host computer; S46. Set up a semantic version control mechanism to manage the change records, compatibility verification and historical instruction comparison logs of semantic units.

7. The multi-UAV intelligent coordination and compatibility ground control system according to claim 2 is characterized in that: The S5 specifically includes: S51, receiving a task instruction from a host computer, inputting the task instruction from the host computer into a semantic parsing module, identifying the control behavior category corresponding to the instruction, and calling the corresponding control semantic unit in a unified control semantic module; S52, filling parameters of the called control semantic unit according to the control parameters contained in the host computer task instruction, and constructing a complete control semantic instruction template; S53. Based on the communication protocol type of the target UAV, call the corresponding protocol translation mapping relationship to convert the semantic instruction template into a native protocol instruction format that can be recognized and executed by the target UAV; S54: Store the translated native protocol instructions into a sending queue, mark the target UAV identification information, task sequence number, and sending priority, and provide a structured native instruction package for subsequent communication modules.

8. The multi-UAV intelligent coordination and compatibility ground control system according to claim 2 is characterized in that: The S6 specifically includes: S61. In the multi-UAV intelligent collaborative and compatible ground control system, determine the target UAV number according to the task scheduling result, and call the communication link management module to select a currently available communication channel, wherein the communication channel includes an ad hoc network and a cellular network; S62, encapsulating the generated native protocol instruction into a protocol, where the encapsulation content includes a frame header identifier, a control payload, a checksum, a timestamp, and a target address field, and forming a transmission frame; S63. Send frames to the target UAV in the form of data packets via the link. The system supports link status monitoring and packet loss retransmission mechanism to ensure that each frame is confirmed to be received within the preset time window; S64, monitoring and receiving a status data frame returned by the target UAV, wherein the status feedback frame includes flight attitude, position, speed, battery level, mission execution status, current flight control mode, and warning information; S65. Decode and parse the received status data frame, and fill the decoding result into the status update vector as a basis for subsequent control and monitoring; S66. Establish a task mapping relationship based on the state update vector and the original task instruction, record the task execution status, and write it into the task execution log queue for subsequent processing modules to call.

9. The multi-UAV intelligent coordination and compatibility ground control system according to claim 2 is characterized in that: The S7 specifically includes: S71. In a multi-UAV intelligent collaborative and compatible ground control system, receive and organize a state data set, extract flight parameters, mission status, and feedback information of each UAV, and generate a state mapping table; S72. Correlate and match the state mapping table with the task scheduling record to generate a multi-UAV state-task correspondence diagram for visual display. The current flight position, task execution stage, communication status, remaining power, and command response status of each UAV are displayed in the form of a real-time map and dynamic icons on the graphical control terminal; S73. Provide visual interactive functions to support operators in issuing tasks, editing paths, switching control modes, and interrupting tasks for each drone in the terminal interface. Record all command issuance behaviors, drone status feedback, task result data, and exception information, build a task execution log, and organize it into a log linked list structure in chronological order. S74. Set up a log playback function to allow the operator to retrieve and review the historical mission execution process based on the mission number and drone number for review, retraining or abnormal diagnosis.

Images