Method and system for unmanned aerial vehicle general standardized access

Abstract

The present application relates to the technical field of unmanned aerial vehicle, and discloses a method for general standardized access of unmanned aerial vehicle, comprising: obtaining protocol fingerprint identification information and multi-source capability evaluation information of the unmanned aerial vehicle to be accessed; generating a capability image comprising a plurality of capability items and a capability confidence degree representing the credibility of the capability items according to the protocol fingerprint identification information and the multi-source capability evaluation information; dynamically selecting an adapter from a preset adapter set according to the capability image and the capability confidence degree and arranging the adapter into a target adapter chain; and converting original data of the unmanned aerial vehicle to be accessed into a standardized unmanned aerial vehicle access object model through the target adapter chain, so as to realize general standardized access based on the unmanned aerial vehicle access object model. The method significantly improves the safety and efficiency of the heterogeneous unmanned aerial vehicle access unified control platform. The present application also discloses a system for general standardized access of unmanned aerial vehicle.

Description

Technical Field

[0001] This invention relates to the field of unmanned aerial vehicle (UAV) technology, and for example to a method and system for universal standardized access for UAVs. Background Technology

[0002] With the widespread application of drones in scenarios such as power line inspection, security patrol, emergency rescue, logistics and distribution, surveying and mapping, low-altitude operation supervision, urban governance, park inspection and cluster collaboration, drones from different manufacturers, models, flight control platforms, mission payloads and communication links need to be connected to a unified control platform such as a unified low-altitude operation platform, cloud control platform, mission scheduling platform, cluster control platform or third-party data fusion platform.

[0003] To achieve unified access for heterogeneous drones, several solutions have been proposed in related technologies. These solutions, based on pre-acquired drone access source information (such as device type, protocol version, and capability description), convert the raw data from the access source into a standardized format recognizable by the platform through pre-configured protocol parsing scripts or field mapping tables. For example, patent application 202610250240.X provides an adaptive secure access method for heterogeneous drone terminals oriented towards power edge computing. This method binds hardware-level identity through terminal registration and trusted anchoring, collects multimodal fingerprints, constructs a security behavior profile, performs adaptive access authentication, and establishes a dynamic behavior monitoring and baseline adjustment mechanism. Patent application 202411509111.5 provides a drone device access management and scheduling method. This method creates a standardized access interface covering data transmission protocols, control command formats, and status feedback mechanisms, allowing a cloud management platform to register, configure, and verify the functions of the access devices.

[0004] In the process of implementing the embodiments of the present invention, at least the following problems were found in the related art: When constructing security behavior profiles or verifying device functionality, relevant solutions typically rely on relatively complete protocol specifications or capability descriptions provided by the drone access source as a benchmark for judgment. However, in real-world scenarios, many drone communication protocols suffer from incomplete documentation, fragmented versions, or inconsistencies between their capability claims and actual behavior. In this situation, existing access methods struggle to reliably parse data, assess the true capabilities of drones, and establish effective connections, preventing heterogeneous drones from securely and efficiently accessing a unified control platform. Summary of the Invention

[0005] To address the shortcomings of the existing technologies, this invention provides a method and system for universal standardized access of unmanned aerial vehicles (UAVs) to improve the security and efficiency of heterogeneous UAVs accessing a unified control platform.

[0006] According to one aspect of the present invention, the present invention provides a method for universal standardized access for unmanned aerial vehicles (UAVs), comprising: obtaining protocol fingerprint identification information and multi-source capability evaluation information of the UAV to be accessed; generating a capability profile including multiple capability items and a capability confidence score characterizing the credibility of the capability items based on the protocol fingerprint identification information and the multi-source capability evaluation information; dynamically selecting adapters from a preset adapter set and arranging them into a target adapter chain based on the capability profile and the capability confidence score; and converting the raw data of the UAV to be accessed into a standardized UAV access object model through the target adapter chain, so as to achieve universal standardized access based on the UAV access object model.

[0007] Optionally, obtaining the protocol fingerprint identification information of the drone to be accessed includes: obtaining the active identification information of the drone to be accessed through active detection, and obtaining the passive identification information of the drone to be accessed through passive monitoring; and generating protocol fingerprint identification information by combining the active identification information and the passive identification information.

[0008] Optionally, based on the protocol fingerprint identification information and multi-source capability assessment information, a capability confidence score is generated, including: parsing the protocol fingerprint identification information and multi-source capability assessment information to obtain multi-source evidence corresponding to each capability item; wherein, the multi-source capability assessment information includes one or more of capability declaration information, historical access information, data sample information, sandbox verification information, and historical feedback information; determining the sub-confidence score of each type of evidence in the multi-source evidence, and assigning a weight to each type of evidence; and merging the sub-confidence scores and weights to generate the capability confidence score of the capability item.

[0009] Optionally, based on the capability profile and capability confidence, adapters are dynamically selected from a preset adapter set and arranged into a target adapter chain, including: selecting a basic adapter set from the preset adapter set according to preset arrangement rules, and constructing the basic adapter set into a basic adapter chain according to a preset execution order; dynamically adjusting the basic adapters in the basic adapter chain based on the existence and capability confidence of each capability item in the capability profile, and generating the target adapter chain.

[0010] Optionally, based on the existence and confidence level of each capability item in the capability profile, the basic adapters in the basic adapter chain are dynamically adjusted to generate a target adapter chain. This includes: traversing each capability item in the capability profile to determine the first capability item that is judged to be missing or unusable; determining the first compensation adapter associated with the first capability item, which is used to compensate for data gaps or functional deficiencies caused by the absence of the first capability item; disabling the basic adapter associated with the first capability item in the basic adapter chain and adding the first compensation adapter to the basic adapter chain to obtain an adjusted basic adapter chain; after adjusting the first capability item in the capability profile, traversing the remaining capability items in the capability profile other than the first capability item to determine the second capability item whose confidence level is lower than a preset threshold; determining the second compensation adapter associated with the second capability item, which is used to introduce reference information independent of the UAV to be accessed to compensate for the conversion result of the basic adapter corresponding to the second capability item; and adding the second compensation adapter to the adjusted basic adapter chain to generate the target adapter chain.

[0011] Optionally, the method for universal standardized access for UAVs further includes: when the target adapter chain includes multiple candidate target adapter chains, obtaining the sandbox verification pass rate, the number of times re-orchestration is triggered during operation, and the interception rate of the converted native instructions for each candidate target adapter chain within a preset statistical window; determining the historical execution effect of the candidate target adapter chain based on the sandbox verification pass rate, the number of times re-orchestration is triggered during operation, and the interception rate of the converted native instructions; and selecting one candidate target adapter chain from the multiple candidate target adapter chains as the currently effective target adapter chain based on the historical execution effect.

[0012] Optionally, the method for universal standardized access for drones further includes: after generating a drone access object model, inputting the drone access object model and / or the converted native instructions into the access sandbox for sandbox verification; and determining whether to register the drone to be accessed as a virtual standard drone endpoint based on the verification results.

[0013] Optionally, the method for universal standardized access for drones further includes: in response to receiving standardized control commands issued by the platform, performing permission verification on the standardized control commands based on capability profiles and capability confidence; after the permission verification is passed, calling the command semantic adapter in the target adapter chain to convert the standardized control commands into native commands executable by the drone to be accessed.

[0014] Optionally, the method for universal standardized access for UAVs further includes: continuously acquiring operational status data during the execution of native commands by the UAV during operation; and reducing the capability confidence of the corresponding capability item and triggering the rearrangement of the target adapter chain when the operational status data characterizes the actual capabilities of the UAV and the corresponding capability item in the capability profile are inconsistent.

[0015] According to another aspect of the present invention, the present invention also provides a system for universal standardized access for unmanned aerial vehicles (UAVs), comprising: a protocol fingerprinting module configured to obtain protocol fingerprinting information and multi-source capability assessment information of the UAV to be accessed; a capability profile generation module electrically connected to the protocol fingerprinting module, configured to generate a capability profile including multiple capability items and a capability confidence score characterizing the credibility of the capability items based on the protocol fingerprinting information and the multi-source capability assessment information; an adapter orchestration module electrically connected to the capability profile generation module, configured to dynamically select adapters from a preset adapter set and orchestrate them into a target adapter chain based on the capability profile and the capability confidence score; and a unified model conversion module electrically connected to the adapter orchestration module, configured to convert the raw data of the UAV to be accessed into a standardized UAV access object model through the target adapter chain, so as to achieve universal standardized access based on the UAV access object model.

[0016] The method and system for universal standardized access for unmanned aerial vehicles provided by this invention have at least the following beneficial technical effects: First, by obtaining the protocol fingerprint and multi-source capability assessment information of the drones to be connected, the system actively detects and analyzes their actual communication behavior characteristics. This allows for accurate identification of protocol types and field semantics even when protocol documents are missing or unreliable, laying the foundation for reliable data connection establishment. Second, based on the protocol fingerprint and multi-source capability assessment information, a capability profile and capability confidence level are generated. This comprehensively assesses the drone's true capabilities and their credibility, reducing the risk of misjudgment due to accepting false or outdated capability claims. Third, based on the capability profile and capability confidence level, adapters are dynamically selected from a pre-set adapter set and arranged into a target adapter chain. This target adapter chain then converts the raw data of the drones to be connected into a standardized drone access object model, enabling the upper-layer control platform to perceive and manage all heterogeneous drones in a consistent manner. This approach avoids the parsing errors or command execution failures caused by relying on specific protocols provided by the drones to be connected for adaptation configuration when protocol documents are missing. It also reduces the cycle of customizing adaptations for each heterogeneous drone, significantly improving the security and efficiency of heterogeneous drone access to the unified control platform. Attached Figure Description

[0017] Figure 1A flowchart illustrating a method for universal standardized access for unmanned aerial vehicles (UAVs) provided in an embodiment of the present invention; Figure 2 A flowchart illustrating another method for universal standardized access for unmanned aerial vehicles (UAVs) provided in an embodiment of the present invention; Figure 3 A flowchart illustrating another method for universal standardized access for unmanned aerial vehicles (UAVs) provided in an embodiment of the present invention; Figure 4 A flowchart illustrating another method for universal standardized access for unmanned aerial vehicles (UAVs) provided in an embodiment of the present invention; Figure 5 This is a schematic diagram of a general standardized access system for unmanned aerial vehicles (UAVs) provided in an embodiment of the present invention; Figure 6 This is a schematic diagram of a universal standardized access device for unmanned aerial vehicles (UAVs) provided in an embodiment of the present invention; Figure 7 This is a schematic diagram illustrating the connection between a unified access gateway for unmanned aerial vehicles (UAVs), a UAV, and a unified control platform, provided as an embodiment of the present invention. Detailed Implementation

[0018] To better understand the above technical solutions, a detailed description of the solutions will be provided below in conjunction with the accompanying drawings and specific embodiments. Obviously, the described embodiments are merely some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0019] The terminology used in the embodiments of this invention is for the purpose of describing particular embodiments only and is not intended to limit the invention. The singular forms “a,” “the,” and “the” as used in the embodiments of this invention and the appended claims are also intended to include the plural forms, and “multiple” generally includes at least two unless the context clearly indicates otherwise.

[0020] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that an article or device that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such an article or device. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the article or device that includes said element.

[0021] The method for universal standardized access for unmanned aerial vehicles (UAVs) according to embodiments of the present invention is implemented by an intermediate layer system deployed between the UAV and the upper-level unified control platform. This intermediate layer system could be a unified UAV access gateway or a heterogeneous UAV adaptation access system. Combined with... Figure 7 As shown, taking the unified access gateway for drones as an example, the unified access gateway for drones is connected to both the drone and the unified control platform. The unified access gateway for drones receives the raw data from the drone and sends native commands to the drone. In addition, the unified access gateway for drones receives standardized control commands from the unified control platform and provides the unified control platform with a standardized drone access object model.

[0022] It is understood that, in the embodiments of the present invention, "drone" refers to various heterogeneous access sources that need to be connected to a unified control platform, including but not limited to: drones, flight control platforms, mission payloads, remote identification modules, ground station equipment, and third-party drone management platforms or data relay platforms.

[0023] Combination Figure 1 As shown, this embodiment of the invention provides a method for universal standardized access for unmanned aerial vehicles (UAVs), comprising the following steps: S10: Obtain protocol fingerprint identification information and multi-source capability assessment information of the drone to be connected.

[0024] Here, protocol fingerprinting information is information extracted from the communication behavior of the UAV to characterize the communication protocol features; multi-source capability assessment information is assessment information other than protocol fingerprinting information, used to assist in generating capability profiles and capability confidence, such as information obtained from the UAV's own declarations, historical access processes, and operational feedback.

[0025] S20: Based on the protocol fingerprint identification information and multi-source capability assessment information, generate a capability profile including multiple capability items, as well as a capability confidence score that characterizes the credibility of the capability items.

[0026] Here, the capability profile is a collection of information describing the various capabilities and states of a UAV under current access conditions. The capability profile includes multiple capability items, each corresponding to a capability dimension of the UAV. The capability items in the capability profile include one or more of the following: flight control capability, telemetry reporting capability, payload capability, communication link capability, security authentication capability, command execution capability, and anomaly feedback capability. Specifically: flight control capability indicates whether the UAV has the ability to receive and execute flight control commands; telemetry reporting capability indicates whether the UAV has the ability to report flight status data; payload capability indicates the type of payload carried by the UAV and its data reporting capability; communication link capability indicates the type of communication link currently used by the UAV and its quality parameters such as bandwidth and latency; security authentication capability indicates the authentication methods and encryption strength supported by the UAV; command execution capability indicates the UAV's response capability and execution accuracy to various standard control commands; and anomaly feedback capability indicates the UAV's ability to proactively report faults or anomalies. Each capability item in the capability profile can be dynamically determined and updated based on protocol fingerprint identification information and multi-source capability assessment information.

[0027] Capability confidence is a quantitative indicator used to characterize the credibility of each capability item in the capability profile under the current access conditions. Specifically, capability confidence can be represented by a continuous value between 0 and 1, or by discrete levels such as high, medium, and low. A higher capability confidence indicates a higher level of credibility for that capability item; a lower capability confidence indicates a lower level of credibility for that capability item, potentially indicating false capability claims, missing capabilities, or capability degradation. Capability confidence corresponds one-to-one with each capability item in the capability profile.

[0028] S30: Based on the capability profile and capability confidence, dynamically select adapters from the preset adapter set and arrange them into a target adapter chain.

[0029] S40 converts the raw data of the UAV to be accessed into a standardized UAV access object model through the target adapter chain, so as to achieve general standardized access based on the UAV access object model.

[0030] Here, the UAV access object model is a standardized data structure used to uniformly describe and carry heterogeneous UAV data. It is the standard data representation output by the target adapter chain after transforming and processing the raw data of the UAV to be accessed. Specifically, the UAV access object model includes one or more object types, each of which carries a type of standardized data. The object types of the UAV access object model include one or more of the following: identity object, capability object, flight status object, mission status object, payload data object, link status object, and security event object. Among them: the identity object carries identity-related data such as identification information, model information, and manufacturer information of the UAV to be accessed; the capability object carries each capability item in the capability profile and its capability confidence level; the flight status object carries real-time flight status data such as flight position, altitude, speed, attitude, and heading; the mission status object carries mission execution-related data such as mission type, mission progress, and mission path; the payload data object carries payload data such as images, videos, and sensor readings collected by the mission payload; the link status object carries link quality data such as signal strength, bandwidth, latency, and packet loss rate of the communication link; and the security event object carries security-related event data such as boundary crossing alarms, link interruptions, and abnormal command interception.

[0031] For example, Table 1 below is a sample table of the UAV access object model, showing the objects included in the UAV access object model and the sample fields corresponding to each object: Table 1. Example Table of UAV Access Object Model

[0032] In some possible implementations, the UAV access object model is organized in a layered structure, including multiple layers such as the basic identity layer, flight control status layer, mission payload layer, communication link layer, mission execution layer, and security audit layer. Each layer is used to carry different types of standardized data, and the unified control platform at the upper layer can subscribe to all or some of the data from each layer according to business needs.

[0033] Understandably, the UAV Access Object Model (AEM) defines the standardized representation of UAV data on the platform side, serving as the data carrier for the standardized data subscription interface. The upper-layer unified control platform, by subscribing to the AEM, can obtain flight status, mission status, and payload data from different UAVs, enabling the fusion, display, and unified analysis of multi-source heterogeneous data. Registering a UAV as a virtual standard UAV endpoint, the data entity of this virtual standard UAV endpoint is the continuously updated AEM. All data access by the upper-layer unified control platform to the virtual standard UAV endpoint is essentially reading instances of the AEM.

[0034] The method for universal standardized access for unmanned aerial vehicles (UAVs) provided in this embodiment of the invention firstly obtains the protocol fingerprint identification information and multi-source capability assessment information of the UAV to be accessed, actively detects and analyzes its actual communication behavior characteristics, and thus can accurately identify the protocol type and field semantics even when the protocol document is missing or unreliable, laying the foundation for reliable data connection establishment; secondly, based on the protocol fingerprint information and multi-source capability assessment information, a capability profile and capability confidence level are generated, and the actual capabilities and credibility of the UAV are comprehensively evaluated, reducing the risk of misjudgment caused by accepting false or outdated capability claims; thirdly, according to the capability profile and capability confidence level, adapters are dynamically selected from a preset adapter set and arranged into a target adapter chain, and the original data of the UAV to be accessed is converted into a standardized UAV access object model through the target adapter chain, enabling the upper-layer control platform to perceive and manage all heterogeneous UAVs in a consistent manner. In this way, it avoids the parsing errors or command execution failures caused by relying on the specific protocols provided by the drones to be connected for adaptation and configuration, when the protocol documents are missing. It also reduces the cycle of customizing and adapting for each heterogeneous drone, thereby significantly improving the security and efficiency of heterogeneous drones accessing the unified control platform.

[0035] Optionally, in step S10, obtaining the protocol fingerprint identification information of the drone to be accessed includes: obtaining the active identification information of the drone to be accessed through active detection, and obtaining the passive identification information of the drone to be accessed through passive monitoring; and generating protocol fingerprint identification information by combining the active identification information and the passive identification information.

[0036] Here, active identification information includes one or more of the following: handshake probe results, version query results, capability query results, interface probe results, heartbeat request results, and parameter reading results. Passive identification information includes one or more of the following: message frame length, message header, verification field, field change period, topic name, data packet direction, port behavior, timestamp format, coordinate system type, and command confirmation method.

[0037] In practical applications, probe messages are proactively sent to the drone to be connected. Active identification information is extracted based on its response characteristics. For example: sending handshake request messages to obtain handshake probe results and determine the drone's connection establishment process and response mode; sending version query commands to obtain version query results and identify the drone's flight control firmware version and protocol version number; sending capability query commands to obtain capability query results and understand the drone's self-reported functional support range; probing access to known interface addresses to obtain interface probe results and confirm the drone's open data interface types; sending heartbeat requests to obtain heartbeat request results and determine the drone's heartbeat cycle and timeout disconnection mechanism; and sending parameter read commands to obtain parameter read results and identify the drone's parameter numbering system and value range. Without sending additional query messages, statistical analysis and feature extraction are performed on the data stream already reported by the drone to obtain passive identification information. By fusing active and passive identification information, complete protocol fingerprint identification information is generated.

[0038] By combining active detection and passive monitoring to generate protocol fingerprint information, this approach effectively solves the technical problem of traditional access methods that heavily rely on complete and accurate protocol documents provided by drones, making reliable communication impossible when protocol documents are missing, versions are fragmented, or capability claims are unreliable. Active detection, through sending standardized query commands, can quickly obtain the drone's response characteristics to specific requests; passive monitoring, through long-term, non-intrusive statistical analysis of data streams, can discover the drone's fixed communication behavior patterns. The two complement and cross-verify each other, making the protocol fingerprint information more complete and objective. This fundamentally avoids adaptation deviations caused by errors or exaggerations made unilaterally by drones, laying a reliable information foundation for the secure and efficient access of heterogeneous drones.

[0039] Optionally, combined Figure 2 As shown, in step S20, based on the protocol fingerprint identification information and multi-source capability assessment information, a capability profile including multiple capability items and a capability confidence score representing the credibility of the capability items are generated, including: S201, analyze the protocol fingerprint identification information and multi-source capability assessment information to obtain multi-source evidence corresponding to each capability item.

[0040] The multi-source capability assessment information includes one or more of the following: capability declaration information, historical access information, data sample information, sandbox verification information, and historical feedback information. Specifically, capability declaration information refers to the capability description information reported by the UAV itself through registration messages, capability lists, or configuration parameters; historical access information includes the capability performance, adaptation effects, and verification results recorded by the UAV during each access process; data sample information includes the initial telemetry data or payload data samples reported by the UAV at the initial stage of access; sandbox verification information includes the verification results generated after the access sandbox verifies the UAV's access object model and / or native commands; and historical feedback information includes the execution feedback and operational status data returned by the UAV after executing native commands during operation. It is understood that each piece of information in the multi-source capability assessment information can be used individually or in combination, and this invention does not impose any restrictions on this.

[0041] Multi-source evidence is structured evidence obtained by parsing and extracting protocol fingerprint identification information and multi-source capability assessment information. It is related to specific capability items and can be directly used for confidence assessment. Accordingly, multi-source evidence includes protocol fingerprint evidence for protocol fingerprint identification information; and multi-source evidence includes capability declaration evidence, data continuity evidence, sandbox validation evidence, and historical operation feedback evidence for multi-source capability assessment information.

[0042] In some possible implementations, the protocol fingerprint identification information and multi-source capability assessment information are parsed to obtain multi-source evidence corresponding to each capability item, including: extracting features related to each capability item from the protocol fingerprint identification information as protocol fingerprint evidence; and extracting features related to each capability item from the multi-source capability assessment information as other multi-source evidence.

[0043] A single set of multi-source data may provide evidence for multiple capabilities, and a single capability may also yield multiple pieces of evidence from the same set of multi-source data.

[0044] For example, regarding the "capability declaration information": extract descriptions such as "supports GPS / RTK positioning", "maximum ceiling 500 meters", and "supports visible light gimbal" from the capability declaration information, and use them as capability declaration evidence for "precise positioning capability", "flight control capability", and "mission payload capability", respectively.

[0045] For example, for the "precise positioning capability item": extract the coordinate system type, latitude and longitude field existence and timestamp format from the protocol fingerprint information to determine whether the positioning data reporting capability is available, and form protocol fingerprint evidence; read the various verification results in the sandbox verification report in the sandbox verification information, and extract the error statistics of positioning data in the field consistency verification and the determination result of height type in the semantic consistency verification as sandbox verification evidence.

[0046] S202, determine the sub-confidence of each type of evidence in the multi-source evidence, and assign a weight to each type of evidence.

[0047] In some possible implementations, the sub-confidence level of each type of evidence in the multi-source evidence is determined separately, and a weight is assigned to each type of evidence. This includes: determining the sub-confidence level of the evidence against the capability item based on the source characteristics and data features of each type of evidence; and assigning weights to the evidence based on the degree of objectivity of each type of evidence, wherein the evidence with a higher degree of objectivity is assigned a higher weight.

[0048] Furthermore, based on the source characteristics and data features of each type of evidence, the sub-confidence levels of the evidence for the capability items are determined, including: for protocol fingerprint data, the sub-confidence levels are determined based on the completeness and consistency of the extracted protocol features; for capability declaration evidence, the sub-confidence levels are determined based on the clarity of the capability declaration and historical verification; for data continuity evidence, the sub-confidence levels are determined based on the completeness rate, stability, and refresh frequency compliance of the relevant data fields; for sandbox validation evidence, the sub-confidence levels are determined based on the pass rate and error statistics of the relevant validation items; and for historical operational feedback evidence, the sub-confidence levels are determined based on the historical execution success rate and execution quality of the relevant instructions.

[0049] Furthermore, based on the objectivity of each type of evidence, weights are assigned to the evidence, including: allocating the weights corresponding to protocol fingerprint evidence and data continuity evidence to the first weight interval; allocating the weights corresponding to sandbox verification evidence to the second weight interval; allocating the weights corresponding to historical operation feedback evidence to the third weight interval; and allocating the weights corresponding to capability declaration evidence to the fourth weight interval; wherein, the lower limit threshold of the first weight interval is higher than the lower limit threshold of the second weight interval, the lower limit threshold of the second weight interval is higher than the lower limit threshold of the third weight interval, and the lower limit threshold of the third weight interval is higher than the lower limit threshold of the fourth weight interval.

[0050] S203, based on sub-confidence and weight, fuses to generate capability confidence of capability items.

[0051] In some possible implementations, the capability confidence of a capability item is generated by fusing sub-confidence and weights, including determining the capability confidence of a capability item according to the following formula:

[0052] Among them, C P C1 represents the confidence level of the ability item, and C1 represents the sub-confidence level corresponding to the first type of evidence. n w is the sub-confidence level corresponding to the nth type of evidence, w1 is the weight corresponding to the first type of evidence, and w n The weight corresponding to the nth type of evidence. .

[0053] For example, the capability item is "precisely located capability item". Multi-source evidence includes capability claim evidence, protocol fingerprint evidence, data continuity evidence, sandbox validation evidence, and historical operational feedback evidence; each type of evidence forms a sub-confidence level C. i The confidence scores are as follows: C1 represents the credibility of the positioning capability declared by the UAV to be connected; C2 represents the completeness and consistency of positioning fields (including latitude, longitude, altitude, timestamp, and coordinate system) in the telemetry data; C3 represents the stability of continuous trajectory data and whether there are jumps or missing values; C4 represents the consistency of errors between the positioning results and the reference trajectory or constraint boundaries; and C5 represents the positioning errors and operational stability in historical tasks. The confidence score for the precise positioning capability item is calculated using the following formula:

[0054] Among them, C Pos To accurately determine the capability confidence level of each capability item, C1 is the capability declaration sub-confidence, C2 is the protocol fingerprint sub-confidence, C3 is the data continuity sub-confidence, C4 is the sandbox verification sub-confidence, C5 is the historical operation feedback sub-confidence, w1 is the capability declaration weight, w2 is the protocol fingerprint weight, w3 is the data continuity weight, w4 is the access sandbox verification weight, and w5 is the historical operation feedback weight. For example, w1=0.15, w2=0.25, w3=0.25, w4=0.20, w5=0.10.

[0055] Protocol fingerprinting information originates from the actual communication behavior of the UAV, possessing objectivity and unforgeability. Multi-source capability assessment information provides corroboration from multiple dimensions, including historical access records, real-time data samples, sandbox verification feedback, and operational execution feedback. These two sources complement and cross-verify each other, enabling the generated capability profile to more realistically and comprehensively reflect the actual capability status of the UAV. By assigning differentiated weights to evidence from different sources and fusing them to generate capability confidence scores, the level of capability confidence objectively reflects the credibility of a capability item supported by multi-source evidence. This provides a reliable quantitative decision-making basis for the subsequent dynamic orchestration of the target adapter chain and the secure verification of command permissions.

[0056] In some possible implementations, the capability confidence of the capability item is generated by fusing sub-confidence and weight, and further includes: in the case of conflict among multiple pieces of evidence used to generate the same capability confidence, determining the evidence to be adopted first from the multiple pieces of evidence according to a preset priority rule; wherein the preset priority rule at least makes the priority of evidence obtained based on protocol fingerprint identification information higher than the priority of evidence obtained based on multi-source capability assessment information (e.g. capability declaration information).

[0057] For example, for the capability item "precise positioning capability", the following evidence is obtained simultaneously: protocol fingerprint evidence (latitude and longitude fields are continuously present in telemetry data and the value range is reasonable) and capability declaration evidence (the drone claims to support GPS positioning in the registration message). Sub-confidence is evaluated for each piece of evidence. For example, the protocol fingerprint evidence is assigned a sub-confidence of 0.9 because it directly reflects the actual data behavior; the capability declaration evidence is assigned a sub-confidence of 0.6 because it may be exaggerated or outdated. The capability confidence of the capability item is obtained by weighting and fusing the evidence according to preset weights (such as protocol fingerprint evidence weight 0.6 and capability declaration evidence weight 0.4).

[0058] During the aforementioned fusion process, if the conclusions of multiple pieces of evidence conflict, a conflict resolution mechanism is activated. For example, the capability statement evidence claims "support for RTK centimeter-level positioning," but the protocol fingerprint evidence finds that RTK-related fields never appear in the telemetry data, and the positioning accuracy is only at the meter level. In this case, the conclusions of the two pieces of evidence are clearly contradictory. Based on the preset priority rules, the evidence obtained based on the protocol fingerprint identification information (i.e., "does not support RTK") is adopted, and the final confidence level of the "precise positioning capability item" is generated accordingly.

[0059] By establishing an evidence conflict resolution mechanism and clearly prioritizing evidence obtained based on protocol fingerprinting information over evidence obtained based on multi-source capability assessment information, the technical problem of distorted capability assessment results when multi-source evidence conclusions are inconsistent is effectively solved. Protocol fingerprinting information reflects the actual protocol behavior and data characteristics of the UAV, and its objectivity and unforgeability are far superior to multi-source capability assessment information. Prioritizing the adoption of protocol fingerprint evidence can avoid misjudgments caused by exaggerated, outdated, or tampered capability claims, ensuring that the generated capability profile and capability confidence level are more realistic and reliable. This provides a credible decision-making basis for the accurate orchestration of the subsequent adapter chain and the correct verification of command permissions.

[0060] Optionally, combined Figure 3 As shown, in step S30, based on the capability profile and capability confidence level, adapters are dynamically selected from the preset adapter set and arranged into a target adapter chain, including: S301, according to the preset orchestration rules, select the basic adapter set from the preset adapter set, and construct the basic adapter set into a basic adapter chain according to the preset execution order.

[0061] Here, an adapter is a pluggable functional unit used to perform specific data processing or transformation functions. Each adapter has standardized input and output interfaces to receive input data in a specific format and output output data in a specific format. For example, a protocol parsing adapter is used to extract fields from raw data; a field mapping adapter is used to map extracted fields to standardized fields; a time synchronization adapter is used to unify timestamp references; and a coordinate transformation adapter is used to transform coordinate systems.

[0062] The preset adapter set is a pre-registered adapter library that can be dynamically selected and invoked. The preset adapter set includes various types of adapters, each with a standardized interface and attribute description, and can be dynamically selected and combined into an adapter chain according to orchestration rules.

[0063] In some possible implementations, the basic adapters constituting the basic adapter set include essential adapters and capability adapters, wherein: The essential adapters are the adapters that constitute the minimum processing capacity of the adapter chain. The essential adapters include a protocol parsing adapter and a field mapping adapter, wherein: the protocol parsing adapter is used to extract fields from the raw data; and the field mapping adapter is used to map the extracted fields to the corresponding fields in the UAV access object model.

[0064] A capability adapter is an adapter used to process the data corresponding to capability items in a capability profile. Capability adapters include one or more of the following: time synchronization adapter, coordinate transformation adapter, unit normalization adapter, payload data adapter, command semantic adapter, and security verification adapter. Specifically: the time synchronization adapter is used to unify timestamp references; the coordinate transformation adapter is used to convert coordinate system types; the unit normalization adapter is used to unify the units of measurement for physical quantities; the payload data adapter is used to parse the data format of task payloads; the command semantic adapter is used to convert standardized control commands into native commands or native command feedback into standardized responses; and the security verification adapter is used for execution command permission verification and data integrity verification.

[0065] In some possible implementations, the default execution order includes a layered order of the original protocol layer, semantic mapping layer, data aggregation layer, security verification layer, and standard interface layer.

[0066] Each adapter executes sequentially according to its assigned layer. The original protocol layer shields the differences in underlying communication; the semantic mapping layer maps heterogeneous fields to standard fields; the data aggregation layer unifies coordinate systems, timestamps, and units of measurement; the security verification layer verifies command permissions and data integrity; and the standard interface layer outputs a standardized model. Each layer has clearly defined responsibilities and boundaries. When adding or upgrading an adapter, registration or replacement is only required at the corresponding layer, without modifying adapters at other layers or the overall process. This significantly improves the maintainability and scalability of the adapter chain while ensuring the controllability and traceability of data processing.

[0067] S302, Based on the existence and confidence level of each capability item in the capability profile, dynamically adjust the basic adapters in the basic adapter chain to generate the target adapter chain.

[0068] In some possible implementations, based on the existence and confidence level of each capability item in the capability profile, the basic adapters in the basic adapter chain are dynamically adjusted to generate a target adapter chain. This includes: traversing each capability item in the capability profile to identify the first capability item that is determined to be missing or unusable; identifying the first compensation adapter associated with the first capability item, which is used to compensate for data gaps or functional deficiencies caused by the absence of the first capability item; disabling the basic adapter associated with the first capability item in the basic adapter chain and adding the first compensation adapter to the basic adapter chain to obtain an adjusted basic adapter chain; after adjusting the first capability item in the capability profile, traversing the remaining capability items in the capability profile other than the first capability item to identify the second capability item whose capability confidence level is lower than a preset threshold; identifying the second compensation adapter associated with the second capability item, which is used to introduce reference information independent of the UAV to be accessed to compensate for the conversion result of the basic adapter corresponding to the second capability item; and adding the second compensation adapter to the adjusted basic adapter chain to generate the target adapter chain.

[0069] For example, the first compensation adapter includes a field missing label adapter or a capability degradation adapter. The field missing label adapter is used to mark missing fields and set corresponding status flags. The capability degradation adapter is used to restrict operations that depend on missing confidence capabilities. The second compensation adapter includes a timestamp compensation adapter, which is used to generate platform-side time flags when timestamps are missing.

[0070] Furthermore, adding the first compensation adapter to the basic adapter chain includes: inserting the first compensation adapter into the corresponding hierarchical position in the basic adapter chain according to its type attribute. For example, the field missing marker adapter belongs to the data assimilation layer and is inserted after the protocol parsing adapter and before the field mapping adapter; the capability degradation adapter belongs to the data assimilation layer / semantic mapping layer and is inserted before the field mapping adapter.

[0071] Furthermore, adding the second compensation adapter to the adjusted base adapter chain includes: inserting the second compensation adapter into the corresponding hierarchical position in the adjusted base adapter chain according to its type attribute. For example, the timestamp compensation adapter belongs to the data aggregation layer and is inserted after the protocol parsing adapter and before the field mapping adapter.

[0072] To facilitate understanding of the specific implementation of step S302, "dynamically adjust the basic adapters in the basic adapter chain based on the existence and confidence of each capability item in the capability profile to generate the target adapter chain," the following explanation uses two drones with different capability levels as examples.

[0073] For the high-capability UAV A, which supports the standard MAVLink protocol, its protocol fingerprint information is complete and standardized, all capability items in the capability declaration are present, and preliminary verification shows that the capability confidence of each capability item is higher than the preset threshold. After generating the basic adapter chain according to the preset orchestration rules, since the capability profile shows that all capabilities exist and have high confidence, there is no need to disable any basic adapters or add compensation adapters. From the preset adapter set, the target adapter chain is selected and orchestrated in the following order: original protocol layer, semantic mapping layer, data normalization layer, security verification layer, and standard interface layer: protocol parsing adapter → field mapping adapter → time synchronization adapter → coordinate transformation adapter → payload data adapter → command semantic adapter → security verification adapter.

[0074] For UAV B, which lacks certain capabilities, it uses a proprietary protocol. Protocol fingerprinting results indicate that its telemetry data does not contain timestamp fields or positioning-related fields (such as latitude and longitude, coordinate system identifiers, etc.), yet its capability declaration claims to possess precise positioning capabilities, creating a clear contradiction. Furthermore, protocol fingerprinting results show that while its telemetry data does contain timestamp fields, the format of these fields is inconsistent across different data packets, and the values ​​exhibit significant jumps, indicating an unreliable time synchronization mechanism. During capability profiling, a conflict was found between the protocol fingerprint evidence and the capability declaration evidence regarding the assessment of "precise positioning capability." Based on a pre-defined priority rule (evidence obtained based on protocol fingerprinting information has higher priority than evidence obtained based on capability declaration information), the conflict was resolved, and the conclusion of the protocol fingerprint evidence was adopted, determining that "precise positioning capability" is missing. Regarding "time synchronization capability," based on the protocol fingerprint evidence (inconsistent timestamp format, numerical jumps) and data continuity evidence (unstable timestamp intervals), it was determined that this capability exists but its reliability is insufficient, and its capability confidence level was assessed at 0.3, lower than the pre-defined threshold of 0.5. After generating the basic adapter chain, the basic adapter chain is dynamically adjusted according to the existence and confidence level of each capability item in the capability profile, as follows: Since "precise positioning capability" is determined to be missing in the capability profile, basic adapters associated with this capability item (such as coordinate transformation adapters) are disabled in the basic adapter chain. Simultaneously, due to the missing positioning capability, a missing field marker adapter is added to set positioning-related fields in the UAV access object model to null values ​​and mark the positioning status as unavailable. A capability degradation adapter is added to reduce the confidence level of positioning-related capabilities and restrict operations that rely on positioning capabilities (such as geofence verification, waypoint flight commands, etc.). Since the confidence level of "time synchronization capability" in the capability profile is lower than a preset threshold, a timestamp compensation adapter is selected from the preset adapter set and added to the basic adapter chain. The timestamp compensation adapter introduces reference information independent of the UAV to be accessed (i.e., the system time when the platform receives the data packet) as a unified time identifier, ensuring that even if the UAV's own time information is unreliable, the data pipeline can still output a standardized model with a reliable time identifier. After the above dynamic adjustments, the target adapter chain for UAV B is as follows: Protocol parsing adapter → Field mapping adapter → Timestamp compensation adapter → Field missing marker adapter → Payload data adapter → Command semantic adapter → Capability degradation adapter → Security verification adapter.

[0075] In some possible implementations, the method for universal standardized access for UAVs also includes: when the target adapter chain includes multiple candidate target adapter chains, obtaining the sandbox verification pass rate, the number of times re-orchestration is triggered during operation, and the interception rate of the converted native instructions for each candidate target adapter chain within a preset statistical window; determining the historical execution effect of the candidate target adapter chain based on the sandbox verification pass rate, the number of times re-orchestration is triggered during operation, and the interception rate of the converted native instructions; and selecting one candidate target adapter chain from the multiple candidate target adapter chains as the currently effective target adapter chain based on the historical execution effect.

[0076] Among them, the sandbox verification pass rate is the ratio of the number of times the candidate target adapter chain passes the sandbox verification within the preset statistical window to the total number of verifications; the number of times re-orchestration is triggered during operation is the cumulative number of times the adapter chain is re-orchestrated due to the inconsistency between the actual capabilities of the UAV and the capability profile represented by the operational status data within the preset statistical window; the ratio of the intercepted native commands generated by conversion is the ratio of the number of native commands generated by conversion within the preset statistical window that are intercepted due to permission verification failure or security boundary conflict to the total number of native commands generated by conversion.

[0077] Furthermore, based on the sandbox validation pass rate, the number of times re-orchestration was triggered during runtime, and the interception rate of the converted native instructions, the historical execution performance of the candidate target adapter chain is determined. This includes: assigning weights to the sandbox validation pass rate, the number of times re-orchestration was triggered during runtime, and the interception rate of the converted native instructions; normalizing the sandbox validation pass rate, the number of times re-orchestration was triggered during runtime, and the interception rate of the converted native instructions; and weighting and summing the normalized sandbox validation pass rate, the number of times re-orchestration was triggered during runtime, and the interception rate of the converted native instructions, along with their respective weights, to obtain a comprehensive score for the candidate target adapter chain, which serves as the historical execution performance.

[0078] Furthermore, based on historical performance, one candidate target adapter chain is selected from multiple candidate target adapter chains as the currently effective target adapter chain, including selecting the candidate target adapter chain with the highest comprehensive score as the currently effective target adapter chain.

[0079] When multiple candidate target adapter chains exist, the candidate chain with the best overall performance is automatically selected as the currently effective target adapter chain based on the historical execution performance determined by the sandbox verification pass rate of each candidate chain within a preset statistical window, the number of times re-orchestration is triggered during operation, and the ratio of command interception. This enables continuous and dynamic selection of the adapter chain configuration most suitable for the current access conditions, avoiding frequent re-orchestration, command interception, or data quality degradation caused by improper adapter chain selection, and significantly improving the long-term stability and adaptability of adapter chain access.

[0080] By pre-constructing a set of preset adapters including various pluggable adapters, and dynamically selecting adapters from the set and arranging them into a target adapter chain based on capability profiles and capability confidence of each capability item, the technical problem of fixed adaptation rules and inability to adapt to the differentiated needs of drones with different capability levels in traditional static adaptation schemes is effectively solved. This enables the same access system to adaptively cope with various heterogeneous drones, from high-capability drones to lightweight payloads with partially missing capabilities, significantly improving the flexibility and applicability of adaptation, and reducing processing errors or resource waste caused by improper adapter configuration.

[0081] Optionally, in step S40, the raw data of the UAV to be accessed is converted into a standardized UAV access object model through the target adapter chain, including: extracting a first field set from the raw data through the adapter in the original protocol layer of the target adapter chain; processing the first field set through the adapter in the data normalization layer and / or semantic mapping layer of the target adapter chain to obtain a second field set; and mapping the second field set to the corresponding fields of the standardized access object model through the adapter in the standard interface layer of the target adapter chain to generate the UAV access object model.

[0082] Here, the original protocol layer is the data entry layer of the target adapter chain, responsible for the initial conversion from raw binary streams or private format messages to structured fields, so that subsequent layers do not need to be aware of the underlying protocol details. The adapters in the original protocol layer include at least a protocol parsing adapter. The data normalization layer / semantic mapping layer is the core data processing layer of the target adapter chain. It receives the structured fields output by the original protocol layer, performs normalization, compensation, and semantic mapping on them, and outputs a standardized and semantically clear set of fields for use by the standard interface layer. The adapters in the data normalization layer / semantic mapping layer include functional adapters (such as time synchronization adapters, command semantic adapters, etc.) and / or compensation adapters (such as timestamp compensation adapters, capability degradation adapters, etc.). The standard interface layer is the data exit layer of the target adapter chain. It receives the standardized set of fields output by the data normalization layer / semantic mapping layer, fills it into the corresponding fields of the UAV access object model according to the preset mapping rules, and generates the final standardized model instance. The adapters in the standard interface layer include field mapping adapters and / or security verification adapters.

[0083] In practical applications, the target adapter chain processes the raw data layer by layer in the following order: original protocol layer, data aggregation / semantic mapping layer, and standard interface layer. Each layer adapter receives the output of the upstream adapter as its input, processes it, and then passes it to the downstream adapter, forming an ordered data processing pipeline.

[0084] The following explanation uses Drone B as an example. Example of Drone B's raw data: alt_raw = 1500, alt_unit = dm, rssi = -68. The protocol parsing adapter resides at the raw protocol layer, loading the protocol fingerprint identification information corresponding to Drone B. Based on the frame structure, field offsets, data types, and other information recorded in the protocol fingerprint information, it decodes and extracts fields from the raw data stream. For Drone B's raw data, the protocol parsing adapter extracts each field and its original value from the binary data stream according to the frame structure definition of its private protocol, generating the first field set: (alt_raw, 1500, dm; rssi, -68, dBm). The data processing adapter, located in the data normalization layer, is responsible for normalizing and / or compensating the extracted raw fields. Based on the target adapter chain of UAV B, the following adapters are involved in the sequential processing: The unit normalization adapter standardizes the units of fields with physical units in the first field set. For the altitude field alt_raw = 1500 dm, the unit normalization adapter identifies its original unit as decimeters, converts it to the standard unit meters, and calculates altitude = 1500 / 10 = 150.0 m; The timestamp compensation adapter detects that a timestamp field is missing in the first field set. According to its compensation logic, it obtains the platform's current system time as the received timestamp of the data packet and generates a unified time identifier timestamp = The first field set, 2026-07-09T10:30:00.000Z, is marked as "generated by platform compensation." The missing field flag adapter detects the absence of location-related fields (latitude, longitude, coordinate system identifier, etc.) in the first field set, sets the location data to null, marks the location status as "unavailable," and generates a second field set: (altitude, 150.0, unit normalization; rssi, -68, direct pass-through; timestamp, 2026-07-09T10:30:00.000Z, generated by platform compensation; position, NULL, location unavailable). The field mapping adapter, located at the standard interface layer, is responsible for mapping the fields in the processed second field set to their corresponding fields in the standardized UAV access object model according to preset mapping rules.For `altitude = 150.0` in the second field set, the field mapping adapter maps it to the `relative_altitude` field of the flight state object in the UAV access object model; for `rssi = -68`, it maps it to the `signal_strength` field of the link state object; for `timestamp`, it maps it to the `timestamp` field of the time object with a source tag; for `position = NULL`, it maps it to the `position` field of the flight state object and sets the `position_status` field to "unavailable". Finally, the UAV access object model is: {(flight state object; relative_altitude: 150.0 m; position: NULL; position_status: unavailable); (link state object; signal_strength: -68 dBm); (time object; timestamp: 2026-07-09T10:30:00.000Z; time_source: platform compensation generated)}.

[0085] By dividing the target adapter chain into a layered sequence of original protocol layer, data normalization layer / semantic mapping layer, and standard interface layer, the conversion process of heterogeneous raw data to standardized access object model is decoupled into an ordered pipeline of "extraction-processing-mapping". This effectively solves the technical problem of the tight coupling of protocol parsing, data normalization and field mapping in traditional solutions, which makes it difficult to flexibly adapt to different UAVs.

[0086] In some embodiments, the method for universal standardized access for drones further includes: inputting the drone access object model and / or the converted native instructions into the access sandbox for sandbox verification; and determining, based on the verification result, whether to register the drone to be accessed as a virtual standard drone endpoint.

[0087] The sandbox verification includes one or more of the following: command permission verification, data replay verification, field consistency verification, semantic consistency verification, virtual command verification, security boundary verification, abnormal field injection verification, and link interruption recovery verification; the verification results include access approved, access restricted, manual review, or access denied.

[0088] Inputting the UAV access object model and / or the converted native commands into the access sandbox for sandbox verification includes: inputting the UAV access object model into the access sandbox for sandbox verification; or, inputting the converted native commands into the access sandbox for sandbox verification; or, inputting the UAV access object model and the converted native commands into the access sandbox for sandbox verification.

[0089] Here, a sandbox validation is performed on the UAV access object model to verify the accuracy of the target adapter chain in converting the UAV's raw data into a standardized model representation, i.e., to verify the correctness of the uplink data conversion. Taking data replay validation as an example, the sandbox obtains historical raw data packets reported by the UAV during the initial access phase as replay samples. The replay samples are then re-input into the target adapter chain to generate the UAV access object model in the replay state. The generated replay state is compared with the real-time UAV access object model to verify the consistency and stability of the target adapter chain's processing results. If the key fields (such as altitude, speed, and position) in the UAV access object model in the replay state deviate significantly from the output of the real-time UAV access object model under the same input conditions, the data replay validation is deemed to have failed, indicating that the target adapter chain may have state dependency issues or non-deterministic processing logic.

[0090] The generated native commands undergo sandbox validation to verify the accuracy of the target adapter chain in translating standard control commands from the platform side into native commands executable by the UAV, thus verifying the security of downlink command conversion. Taking command permission verification as an example, the sandbox constructs a series of standardized control commands for testing (including normal commands, unauthorized commands, and commands exceeding capability boundaries), calls the command semantic adapter for conversion, and checks whether the permission verification logic correctly intercepts commands that should be intercepted and correctly allows commands that should be allowed. For example, for a "fly over no-fly zone" command issued by a low-privilege operator, the sandbox verifies whether the command is correctly intercepted; for a "return to home" command issued by a high-privilege operator in normal airspace, the sandbox verifies whether the command is correctly converted and passes the verification.

[0091] A sandbox validation is performed on the UAV access object model and the converted native commands to verify the interaction logic between uplink data and downlink commands. For example, the sandbox sends a standardized query control command (such as "query current location status") to the command semantic adapter through virtual command validation, obtaining the converted native command. The sandbox simulates the normal response behavior of the UAV, generates the corresponding original response data packet, inputs it into the target adapter chain for conversion, and generates an updated UAV access object model. The sandbox checks whether the corresponding fields in the updated UAV access object model correctly reflect the content queried by the standardized control command. For example, if the standardized control command queries "location status," the location status field in the UAV access object model should be updated to "available" or "unavailable," rather than remaining empty or in another irrelevant state. Through joint validation of the UAV access object model and the converted native commands, the sandbox can discover interaction logic errors that are difficult to expose under a single validation mode.

[0092] In some possible implementations, based on the verification result, it is determined whether to register the drone to be connected as a virtual standard drone endpoint, including: registering the drone to be connected as a virtual standard drone endpoint if the verification result is successful; registering the drone to be connected as a restricted virtual standard drone endpoint (e.g., restricting the issuance of specific high-risk commands) if the verification result is subject to manual review; temporarily suspending the registration of the drone to be connected as a virtual standard drone endpoint if the verification result is denied access.

[0093] The four-level hierarchical decision-making mechanism based on the verification results—"access approved, access restricted, manual review or access denied"—further realizes the refinement and differentiation of access control. This allows drones with complete capabilities to quickly pass the verification and obtain full access permissions, drones with partially missing capabilities to access under restricted conditions, drones with ambiguous capabilities to be submitted for manual review to avoid misjudgment, and drones with serious defects to be directly denied access to ensure platform security.

[0094] By generating capability profiles and confidence levels through protocol fingerprinting and multi-source data fusion, and dynamically orchestrating target adapter chains accordingly, the raw data of heterogeneous UAVs is converted into standardized UAV access object models, thereby achieving intelligent and automated access processes. Based on this, the UAV access object models and / or the converted native commands are input into the access sandbox for pre-verification. The verification results determine whether to register the UAV to be accessed as a virtual standard UAV endpoint. This effectively solves the technical problem of traditional access methods where direct online operation after adaptation often exposes issues such as data semantic errors, command permission violations, and security boundary breaches only after formal operation, leading to high operational risks and difficult fault location. This significantly improves the security and reliability of heterogeneous UAV access to the unified control platform.

[0095] In some embodiments, the method for universal standardized access for drones further includes: in response to receiving a standardized control command issued by the platform side, performing permission verification on the standardized control command based on capability profile and capability confidence; after the permission verification is passed, calling the command semantic adapter in the target adapter chain to convert the standardized control command into native commands executable by the drone to be accessed; if the permission verification fails, intercepting the standardized control command and returning the reason for interception.

[0096] In some possible implementations, in response to receiving standardized control instructions from the platform, permission verification is performed on the standardized control instructions based on the capability profile and capability confidence level. This includes: determining the dependent capability items for executing the standardized control instructions according to the instruction type; determining the existence of dependent capability items in the capability profile and their corresponding capability confidence levels; intercepting the standardized control instructions if any dependent capability item is determined to be missing or if the capability confidence level of any dependent capability item is lower than a preset threshold; and determining that the permission verification is passed if all dependent capability items exist and the capability confidence level is higher than the preset threshold.

[0097] In some possible implementations, authorization verification for standardized control commands includes one or more of the following: operator authorization verification, command type authorization verification, flight status legitimacy verification, capability confidence verification, positioning capability verification, and geofencing verification. Specifically, operator authorization verification verifies whether the operator issuing the command has the necessary authorization to execute the control command; command type authorization verification verifies whether the UAV to be connected is permitted to execute the specified control command type under the current access conditions; flight status legitimacy verification verifies whether the current flight status of the UAV to be connected allows for the execution of the control command; positioning capability verification verifies whether the UAV to be connected currently possesses the positioning information required to execute the control command; and geofencing verification verifies whether the expected behavior after the control command is executed violates a preset airspace safety boundary.

[0098] By generating capability profiles and capability confidence levels through protocol fingerprinting and multi-source data fusion, and dynamically orchestrating the target adapter chain accordingly, the raw data of heterogeneous UAVs is transformed into a standardized UAV access object model, thereby achieving standardized access to uplink data. Based on this, when the platform issues standardized control commands, multi-dimensional permission verification is performed on the standardized control commands based on the capability profile and capability confidence level. After successful verification, the command semantic adapter in the target adapter chain is invoked to convert the standardized control commands into native commands executable by the UAV to be accessed. This effectively solves the technical problem in traditional solutions where the platform issues control commands that exceed the launch capability range, fail to execute, or even cause safety incidents due to insufficient consideration of the actual capability differences of different UAVs. This enables the unified control platform to safely and reliably schedule and manage multi-source heterogeneous UAVs with varying capabilities in a standardized manner, significantly improving the security and efficiency of large-scale UAV collaborative management.

[0099] In some embodiments, the method for universal standardized access for unmanned aerial vehicles (UAVs) further includes: continuously acquiring operational status data during the execution of native commands by the UAV during operation; and reducing the capability confidence of the corresponding capability item and triggering a re-orchestration of the target adapter chain when the operational status data characterizes the actual capabilities of the UAV and the corresponding capability item in the capability profile are inconsistent.

[0100] The operational status data includes one or more of the following: heartbeat status data, link quality data, data quality data, command execution feedback data, and abnormal event data.

[0101] In some possible implementations, based on operational status data, discrepancies between the actual capabilities of the UAV and the corresponding capability items in the capability profile are determined as follows: The operational status data is compared with the expected performance of each capability item in the capability profile; when the deviation between the operational status data and the expected performance data of the corresponding capability item in the capability profile meets a preset condition, it is determined that the actual capability of that capability item is inconsistent with the capability profile. For example, the operational status data includes data quality data; when the data integrity rate of a specific field is lower than a preset integrity rate threshold, it is determined that the actual capability of the capability item corresponding to that field is inconsistent with the capability profile. By quantitatively comparing the operational status data with the expected performance of each capability item in the capability profile, and determining inconsistencies when the deviation meets preset conditions, automated and precise detection of changes in UAV capabilities during operation is achieved, providing an objective and reliable triggering basis for the dynamic correction of subsequent capability confidence.

[0102] In some possible implementations, reducing the capability confidence of the corresponding capability item includes: determining the reduction range of capability confidence based on the degree of deviation between the operational status data and the expected performance data of the corresponding capability item in the capability profile; and generating a corrected capability confidence based on the reduction range. By determining the reduction range of capability confidence based on the degree of deviation between the operational status data and the expected performance data, a graded and gradual correction of capability confidence is achieved. This avoids overreaction to the capability profile due to instantaneous anomalies or slight fluctuations, allowing confidence adjustments to accurately reflect the severity of capability degradation and providing reliable quantitative input for the reasonable re-arrangement of the adapter chain.

[0103] In some possible implementations, triggering the reorientation of the target adapter chain includes: using the revised capability profile and capability confidence as input, dynamically selecting and reorienting adapters from a preset adapter set to obtain a new target adapter chain; inputting the UAV access object model and / or native commands generated by the new target adapter chain into the access sandbox for verification; and replacing the original target adapter chain with the new target adapter chain after the access sandbox verification passes. By triggering the reorientation of the target adapter chain after the capability confidence is corrected, and returning the output of the new target adapter chain to the access sandbox for re-verification, and only replacing it after passing the verification, it is ensured that the new target adapter chain is fully verified before formal operation, avoiding the introduction of new adaptation errors or security risks due to reorientation, and ensuring the safety of dynamic switching of adaptation logic during operation.

[0104] In some possible implementations, the method for universal standardized access for UAVs also includes: adjusting the data reporting priority when the link quality of operational status data deteriorates, or when the data refresh frequency of a specific object in the UAV access object model falls below its corresponding preset threshold. This adjustment includes prioritizing the reporting of data from flight status objects and / or safety event objects, while delaying or reducing the reporting frequency of payload data objects. By prioritizing the reporting of flight status data and / or safety event data, and delaying or reducing the reporting frequency of payload data when link quality deteriorates or the data refresh frequency falls below the threshold, priority protection of critical data under limited communication bandwidth resources is achieved. This ensures that the platform can still monitor the flight safety status of UAVs in real time under adverse communication conditions, avoiding operational risks caused by the loss or delay of critical safety information due to link congestion.

[0105] By generating capability profiles and capability confidence levels through protocol fingerprinting and multi-source data fusion, and dynamically orchestrating the target adapter chain accordingly, the raw data of heterogeneous UAVs is transformed into a standardized UAV access object model, thereby achieving intelligent and automated access for heterogeneous UAVs. Furthermore, by continuously acquiring operational status data during UAV execution of native commands, including heartbeat status, link quality, data quality, command execution feedback, and abnormal events, and automatically reducing the capability confidence level of the corresponding capability item and triggering a re-orchestration of the target adapter chain when the operational status data indicates a discrepancy between the actual capabilities of the UAV and the corresponding capability items in the capability profile, this approach effectively solves the technical problems of statically fixed adaptation rules and inability to respond to dynamic changes in UAV capabilities during operation in traditional access schemes. This significantly improves the access stability and adaptability of heterogeneous UAVs operating in long-term, complex environments.

[0106] Combination Figure 4 As shown, the method for universal standardized access for unmanned aerial vehicles includes the following steps: S10: Obtain protocol fingerprint identification information and multi-source capability assessment information of the drone to be connected.

[0107] S20: Based on the protocol fingerprint identification information and multi-source capability assessment information, generate a capability profile including multiple capability items, as well as a capability confidence score that characterizes the credibility of the capability items.

[0108] S30: Based on the capability profile and capability confidence, dynamically select adapters from the preset adapter set and arrange them into a target adapter chain.

[0109] S40 converts the raw data of the UAV to be accessed into a standardized UAV access object model through the target adapter chain, so as to achieve general standardized access based on the UAV access object model.

[0110] S50 inputs the UAV access object model and / or the native commands generated by the conversion into the access sandbox for sandbox verification.

[0111] S60 determines, based on the verification results, whether to register the drone to be connected as a virtual standard drone endpoint.

[0112] S70, in response to receiving standardized control commands from the platform, performs permission verification on the standardized control commands based on capability profiles and capability confidence levels.

[0113] After the permission verification is passed, the S80 calls the command semantic adapter in the target adapter chain to convert the standardized control commands into native commands that can be executed by the drone to be connected.

[0114] The S90 continuously acquires operational status data during the drone's execution of native commands.

[0115] S100: If the actual capabilities of the UAV as represented by the operational status data are inconsistent with the corresponding capability items in the capability profile, reduce the capability confidence of the corresponding capability items and trigger the rearrangement of the target adapter chain.

[0116] Taking a specific application scenario as an example, this disclosure provides a detailed explanation of the method for universal standardized access for unmanned aerial vehicles.

[0117] A low-altitude operation platform needs to simultaneously connect to multiple heterogeneous inspection drones, each employing different flight control platforms, communication links, and mission payloads. The unified drone access gateway first receives access requests from each drone. Through passive identification, it acquires drone identifiers, communication link types (e.g., cellular or private network), flight control types, payload types, and protocol versions. Then, through active identification methods such as handshake detection and version lookup, it acquires handshake detection results, version lookup results, and capability lookup results. Combining the active and passive identification information, it generates protocol fingerprint information for each drone. Based on the protocol fingerprint information, combined with multi-source capability assessment information such as the capability declaration information reported by the drones and historical access records, it generates a capability profile for each drone and a capability confidence level for each capability item.

[0118] Based on capability profiles and capability confidence levels, the unified UAV access gateway dynamically selects adapters from a pre-defined adapter set and arranges them into a target adapter chain. For example, for UAVs that support reporting position, attitude, battery level, and visible light images, a protocol parsing adapter, field mapping adapter, time synchronization adapter, coordinate transformation adapter, and payload data adapter are selected sequentially to form its target adapter chain. For UAVs that also need to support flight control command issuance, command semantic adapters and security verification adapters are further added to the target adapter chain. Through the target adapter chain, the UAV's raw telemetry data (including flight status, position, attitude, battery level, and image payload data, etc.) is converted into a standardized UAV access object model, achieving a unified expression of data from different UAV models at the model level.

[0119] Before officially connecting UAVs to the low-altitude operation platform, the unified UAV access gateway inputs the UAV access object model and the generated native commands into the access sandbox for verification. Specifically, it verifies the field refresh frequency and coordinate system consistency through historical data playback, verifies command reachability through virtual command verification, and confirms whether the corresponding commands are allowed to be issued during the current flight phase through command permission verification. After the access sandbox verification is passed, the UAV is registered as a virtual standard UAV endpoint. After registration, using the UAV access object model, the low-altitude operation platform obtains the flight status, mission status, and payload data of each virtual standard UAV endpoint through a standardized data subscription interface. When the low-altitude operation platform issues standardized control commands such as "return to home," "hover," "adjust flight path," and "payload photography" through the standardized command call interface, the unified UAV access gateway first performs permission verification based on the corresponding UAV's capability profile and capability confidence. After the permission verification is passed, it calls the command semantic adapter in the target adapter chain to convert the standardized control commands into native commands that the UAV can execute and issues them.

[0120] During drone operation, the unified access gateway continuously acquires operational status data during the execution of native commands, including heartbeat status, link quality, data quality, command execution feedback, and abnormal events. When standard control commands issued by the low-altitude operation platform are converted and executed, if the native command execution feedback indicates that the drone failed to respond correctly and does not match the expected "command execution capability" in the capability profile, the capability confidence of the corresponding command execution capability is reduced, triggering the reselection and reorganization of command semantic adapters in the target adapter chain. If necessary, the modified target adapter chain and conversion results are returned to the access sandbox for re-verification.

[0121] The method for universal standardized access for unmanned aerial vehicles (UAVs) provided in this invention generates protocol fingerprint information by combining active detection and passive monitoring, eliminating the reliance on complete protocol documents. Based on the protocol fingerprint information, a capability profile and capability confidence level are generated, ensuring that the assessment of UAV capabilities is based on objective evidence and effectively overcoming the risk of misjudgment caused by unreliable capability claims. The method dynamically selects and orchestrates adapter chains from a preset adapter set according to the capability profile and capability confidence level, achieving high-precision matching between adaptation logic and actual protocol behavior, avoiding data parsing errors or command execution failures caused by static adaptation. By using an access sandbox to perform pre-conversion multi-dimensional verification on the UAV access object model and native commands, adaptation errors or security risks are intercepted before formal access, reducing the failure rate during operation. By continuously monitoring the consistency between command execution feedback and the capability profile during operation, dynamic correction of capability confidence and closed-loop re-orchestration of the adapter chain are achieved, enabling adaptation to real-time changes in UAV capabilities and maintaining long-term stable access quality. This solves the problems of long adaptation cycles, poor reliability, and lack of dynamic response capabilities when heterogeneous drones are connected to a unified control platform, and significantly improves the security and efficiency of large-scale heterogeneous drone access.

[0122] In some embodiments, combined with Figure 5As shown, the system 80 for universal standardized access for unmanned aerial vehicles (UAVs) includes: a protocol fingerprinting module 801, a capability profile generation module 802, an adapter orchestration module 803, and a unified model conversion module 804. Specifically: the protocol fingerprinting module 801 is configured to obtain protocol fingerprinting information and multi-source capability assessment information of the UAV to be accessed; the capability profile generation module 802 is electrically connected to the protocol fingerprinting module 801 and is configured to generate a capability profile including multiple capability items and a capability confidence score characterizing the credibility of the capability items, based at least on the protocol fingerprinting information and the multi-source capability assessment information; the adapter orchestration module 803 is electrically connected to the capability profile generation module 802 and is configured to dynamically select adapters from a preset adapter set and orchestrate them into a target adapter chain based on the capability profile and capability confidence score; the unified model conversion module 804 is electrically connected to the adapter orchestration module 803 and is configured to convert the raw data of the UAV to be accessed into a standardized UAV access object model through the target adapter chain, thereby achieving universal standardized access based on the UAV access object model.

[0123] In some embodiments, continue to combine Figure 5 As shown, the system 80 for universal standardized access for drones also includes an access sandbox verification module 805, which is configured to perform access sandbox verification on the input drone access object model and / or the converted native instructions; and determine whether to register the drone to be accessed as a virtual standard drone endpoint based on the verification result.

[0124] Continue to combine Figure 5 As shown, the system 80 for universal standardized access of UAVs also includes an authorization verification module 806. The authorization verification module 806 is configured to, in response to receiving standardized control commands issued by the platform side, perform authorization verification on the standardized control commands based on capability profile and capability confidence; after the authorization verification is passed, it calls the command semantic adapter in the target adapter chain to convert the standardized control commands into native commands that can be executed by the UAV to be accessed.

[0125] Continue to combine Figure 5 As shown, the system for universal standardized access for UAVs also includes an operational self-calibration module 807. The operational self-calibration module 807 is configured to continuously acquire operational status data during the execution of native commands by the UAV during operation; if the operational status data characterizes the actual capabilities of the UAV and the corresponding capability items in the capability profile are inconsistent, the capability confidence of the corresponding capability item is reduced, and the target adapter chain is re-arranged.

[0126] Combination Figure 6As shown, this disclosure provides an apparatus 100 for universal standardized access for unmanned aerial vehicles (UAVs), including a processor 101 and a memory 102, and may also include a communication interface 103 and a bus 104. The processor 101, communication interface 103, and memory 102 can communicate with each other via the bus 104. The communication interface 103 can be used for information transmission. The processor 101 can call logical instructions in the memory 102 to execute the method for universal standardized access for UAVs described in the above embodiment.

[0127] Furthermore, the logical instructions in the aforementioned memory 102 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium.

[0128] The memory 102, as a computer-readable storage medium, can be used to store software programs and computer-executable programs, such as program instructions / modules corresponding to the methods in the embodiments of this disclosure. The processor 101 executes functional applications and data processing by running the program instructions / modules stored in the memory 102, that is, it implements the method for universal standardized access for UAVs in the above method embodiments.

[0129] The memory 102 may include a program storage area and a data storage area. The program storage area may store the operating system and applications required for at least one function; the data storage area may store data created based on the use of the terminal device. Furthermore, the memory 102 may include high-speed random access memory and may also include non-volatile memory.

[0130] Processor 101 is a circuit with signal processing capabilities. In one implementation, processor 101 can be a circuit with instruction read and execute capabilities, such as a central processing unit (CPU), microprocessor, graphics processing unit (GPU) (which can be understood as a type of microprocessor), or digital signal processor (DSP). In another implementation, processor 101 can implement certain functions through the logical relationships of hardware circuits. These logical relationships of hardware circuits are fixed or reconfigurable. For example, processor 101 can be a hardware circuit implemented as an ASIC or a programmable logic device (PLD), such as an FPGA. In a reconfigurable hardware circuit, the process of the processor loading a configuration document and configuring the hardware circuit can be understood as the process of the processor loading instructions to implement the functions of some or all of the above modules. Furthermore, it can also be a hardware circuit designed for artificial intelligence, which can be understood as a type of ASIC, such as a neural network processing unit (NPU), tensor processing unit (TPU), or deep learning processing unit (DPU). The processor 101 is used to execute related programs to implement the functions required by the units in the apparatus for universal standardized access for unmanned aerial vehicles in the embodiments of this application, or to execute the method for universal standardized access for unmanned aerial vehicles in the embodiments of this application.

[0131] As can be seen, each module in the above device can be one or more processors (or processing circuits) configured to implement the above methods, such as: CPU, GPU, NPU, TPU, DPU, microprocessor, DSP, ASIC, FPGA, or a combination of at least two of these processor types.

[0132] Furthermore, the modules in the above devices can be integrated in whole or in part, or they can be implemented independently. In one implementation, these modules are integrated together as a system-on-a-chip (SOC). The SOC may include at least one processor for implementing any of the above methods or for implementing the functions of the modules of the device. The at least one processor may be of different types, such as CPU and FPGA, CPU and artificial intelligence processor, CPU and GPU, etc.

[0133] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including both the preferred embodiments and all changes and modifications falling within the scope of the invention. Clearly, those skilled in the art can make various alterations and modifications to the invention without departing from its spirit and scope. Thus, if these modifications and modifications of the invention fall within the scope of the claims and their equivalents, the invention is also intended to include these modifications and modifications.

Claims

1. A method for universal standardized access for unmanned aerial vehicles (UAVs), characterized in that, include: Obtain the protocol fingerprint identification information and multi-source capability assessment information of the drone to be connected; Based on protocol fingerprinting information and multi-source capability assessment information, a capability profile including multiple capability items and capability confidence scores representing the credibility of capability items are generated. This includes: parsing protocol fingerprinting information and multi-source capability assessment information to obtain multi-source evidence corresponding to each capability item; wherein, multi-source capability assessment information includes one or more of capability declaration information, historical access information, data sample information, sandbox verification information, and historical feedback information; determining the sub-confidence score of each type of evidence in the multi-source evidence and assigning weights to each type of evidence; and fusing the sub-confidence scores and weights to generate the capability confidence score of the capability item. Based on the capability profile and capability confidence level, adapters are dynamically selected from the preset adapter set and arranged into a target adapter chain; The target adapter chain converts the raw data of the drone to be connected into a standardized drone access object model, so as to achieve general standardized access based on the drone access object model.

2. The method according to claim 1, characterized in that, Obtain the protocol fingerprint information of the drone to be connected, including: Active identification information of the drone to be connected is obtained through active detection, and passive identification information of the drone to be connected is obtained through passive monitoring. By combining active and passive identification information, protocol fingerprint identification information is generated.

3. The method according to claim 1, characterized in that, Determine the sub-confidence of each type of evidence in the multi-source evidence, and assign weights to each type of evidence, including: Based on the source characteristics and data features of each type of evidence, determine the sub-confidence level of the evidence against the capability item; Each type of evidence is assigned a weight based on its degree of objectivity, with higher-level evidence receiving a higher weight.

4. The method according to claim 1, characterized in that, Based on capability profiles and capability confidence levels, adapters are dynamically selected from a pre-defined adapter set and arranged into a target adapter chain, including: Based on preset orchestration rules, a set of basic adapters is selected from a preset set of adapters, and the set of basic adapters is constructed into a basic adapter chain according to a preset execution order; Based on the existence and confidence level of each capability item in the capability profile, the basic adapters in the basic adapter chain are dynamically adjusted to generate the target adapter chain.

5. The method according to claim 4, characterized in that, Based on the existence and confidence level of each capability item in the capability profile, the basic adapters in the basic adapter chain are dynamically adjusted to generate the target adapter chain, including: Iterate through each ability item in the ability profile and determine the first ability item that is judged to be missing or unusable in the ability profile. A first compensation adapter is identified as associated with the first capability item. The first compensation adapter is used to compensate for data gaps or functional deficiencies caused by the absence of the first capability item. Disable the base adapter associated with the first capability item in the base adapter chain, and add the first compensation adapter to the base adapter chain to obtain the adjusted base adapter chain; After adjusting the first ability item in the ability profile, iterate through the remaining ability items in the ability profile except for the first ability item, and determine the second ability item whose ability confidence is lower than the preset threshold. A second compensation adapter is identified as associated with the second capability item. The second compensation adapter is used to introduce reference information independent of the UAV to be accessed in order to compensate for the conversion result of the base adapter corresponding to the second capability item. The second compensation adapter is added to the adjusted base adapter chain to generate the target adapter chain.

6. The method according to claim 4, characterized in that, Also includes: When the target adapter chain includes multiple candidate target adapter chains, obtain the sandbox verification pass rate, the number of times re-orchestration is triggered during operation, and the rate at which the generated native instructions are intercepted for each candidate target adapter chain within a preset statistical window; Based on the sandbox validation pass rate, the number of times re-orchestration was triggered during runtime, and the rate at which the generated native instructions were intercepted, the historical execution performance of the candidate target adapter chain was determined. Based on historical performance, select one candidate target adapter chain from multiple candidate target adapter chains as the currently active target adapter chain.

7. The method according to any one of claims 1 to 6, characterized in that, After generating the drone access object model, the following is also included: Input the drone access object model and / or the converted native commands into the access sandbox for sandbox verification; Based on the verification results, determine whether to register the drone to be connected as a virtual standard drone endpoint.

8. The method according to any one of claims 1 to 6, characterized in that, Also includes: In response to receiving standardized control commands from the platform, the system performs permission verification on the standardized control commands based on the capability profile and capability confidence level. After the permission verification is passed, the command semantic adapter in the target adapter chain is invoked to convert the standardized control commands into native commands that can be executed by the drone to be connected.

9. The method according to any one of claims 1 to 6, characterized in that, Also includes: During drone operation, continuously acquire operational status data of the drone as it executes native commands; When the actual capabilities of a drone as represented by operational status data are inconsistent with the corresponding capability items in its capability profile, the capability confidence of the corresponding capability item is reduced, and the target adapter chain is reorganized.

10. A system for universal standardized access for unmanned aerial vehicles (UAVs), characterized in that, include: The protocol fingerprinting module is configured to obtain the protocol fingerprinting information and multi-source capability assessment information of the drone to be connected; The capability profile generation module is electrically connected to the protocol fingerprint recognition module and is configured to generate a capability profile including multiple capability items and a capability confidence level representing the credibility of the capability items based on the protocol fingerprint recognition information and multi-source capability assessment information. The adapter orchestration module is electrically connected to the capability profile generation module and is configured to dynamically select adapters from a preset adapter set and orchestrate them into a target adapter chain based on the capability profile and capability confidence. The unified model conversion module, electrically connected to the adapter orchestration module, is configured to convert the raw data of the UAV to be accessed into a standardized UAV access object model through the target adapter chain, so as to achieve universal standardized access based on the UAV access object model.

Images