A robot quick-change load interface self-learning protocol modeling method and system

Abstract

The embodiment of the application discloses a kind of robot quick-change load interface self-learning protocol modeling method and system, it is related to robot quick-change interface and load communication configuration technical field.The method comprises the following steps: after detecting target load access robot quick-change interface, obtaining quick-change locking state, interface contact stability parameter and initial communication state;Generation hierarchical safety detection instruction set and send to the target load;Receive corresponding detection response data, and synchronously collect load working current data and robot end motion physical state parameters;Extract response frame structure features and response timing features;Generation field action correlation;Based on the field action correlation, the protocol semantic state model of target load is constructed;According to the protocol semantic state model, the protocol configuration parameters corresponding to the target load are generated.The application improves the degree of automation and safety and reliability of robot quick-change load access.

Description

Technical Field

[0001] This application relates to the field of robot quick-change interface and payload communication configuration technology, and in particular to a self-learning protocol modeling method and system for robot quick-change payload interface. Background Technology

[0002] With the increasing application of industrial robots in flexible manufacturing, intelligent assembly, and material handling and inspection, robot end effectors need to frequently change different loads such as grippers, suction cups, vision modules, and force control tools. Quick-change interfaces are typically used to achieve mechanical locking, power connection, and communication connection of loads, enabling robots to quickly switch between different tools. However, loads from different manufacturers, models, or versions often use different communication protocols and control fields. When connecting to unknown loads, robots usually need to rely on preset protocol templates or manually configure communication parameters, resulting in low load connection efficiency and potential status feedback errors or motion control anomalies when protocol fields are not accurately identified.

[0003] Existing protocol identification methods mostly rely on analyzing the communication messages themselves, such as identifying frame headers, frame trailers, field lengths, checksums, and response timing. However, these methods struggle to determine whether communication fields truly correspond to the locking, enabling, action execution, or anomaly protection states of the robot's quick-change payload. In robot quick-change scenarios, the semantics of protocol fields should not only be reflected in the response data but also match the interface contact stability, payload operating current, and the physical state of the robot's end effector.

[0004] Therefore, there is an urgent need for a protocol configuration scheme that can combine communication response and fast load change electromechanical response for self-learning modeling, so as to improve the automation and safety reliability of robot fast load change access. Summary of the Invention

[0005] This application provides a self-learning protocol modeling method and system for robot quick-load-changing interfaces, which improves the automation level and safety and reliability of robot quick-load-changing access.

[0006] This application provides the following solution: According to the first aspect, a self-learning protocol modeling method for a robot quick-change payload interface is provided. The method includes: detecting that after a target payload is connected to the robot's quick-change interface, acquiring the quick-change locking state, interface contact stability parameters, and initial communication state; when the quick-change locking state and the interface contact stability parameters meet preset access conditions, generating a hierarchical safety detection instruction set and sending it to the target payload; receiving corresponding detection response data and simultaneously collecting payload operating current data and robot end-effector motion physical state parameters; extracting response frame structure features and response timing features based on the detection response data; generating field action association relationships based on the response frame structure features, response timing features, payload operating current data, and robot end-effector motion physical state parameters; constructing a protocol semantic state model of the target payload based on the field action association relationships; and generating protocol configuration parameters corresponding to the target payload based on the protocol semantic state model.

[0007] According to one achievable method in this application embodiment, the interface contact stability parameter is generated by: collecting the locking position signal, locking holding force, terminal contact impedance, power supply voltage fluctuation, and communication handshake success rate of the robot quick-change interface; generating a mechanical locking stability value based on the locking position signal and locking holding force; generating an electrical connection stability value based on the terminal contact impedance, power supply voltage fluctuation, and communication handshake success rate; and generating the interface contact stability parameter based on the mechanical locking stability value and the electrical connection stability value.

[0008] According to one achievable method in an embodiment of this application, the generation of a hierarchical security detection instruction set includes: determining the current detection permission level based on the quick-switch locking state and the interface contact stability parameters, wherein the detection permission level includes a first detection permission level, a second detection permission level, and a third detection permission level; generating a status query type detection instruction under the first detection permission level; generating an enable handshake type detection instruction under the second detection permission level; generating a low-amplitude action verification type detection instruction under the third detection permission level; and configuring corresponding allowable current upper limit, response timeout time, action amplitude upper limit, and disabled field identifier for each detection instruction.

[0009] According to one achievable method in an embodiment of this application, the generation of the low-amplitude motion verification detection command includes: determining a target communication channel from the communication channels supported by the robot quick-change interface based on the initial communication state of the target payload, and determining the basic frame transmission format and response waiting window corresponding to the target communication channel; obtaining low-amplitude verification constraint parameters corresponding to the third detection permission level, the low-amplitude verification constraint parameters including the upper limit of allowable current, the upper limit of motion amplitude, the upper limit of end-effector pose offset, and the upper limit of duration; selecting candidate verification templates from a preset safe motion template library that do not change the working mode of the target payload and do not trigger continuous motion; encapsulating the candidate verification templates according to the basic frame transmission format, response waiting window, and low-amplitude verification constraint parameters to generate a low-amplitude motion verification detection command; and configuring abnormal stop conditions for the low-amplitude motion verification detection command.

[0010] According to one achievable method in an embodiment of this application, the synchronous acquisition of load operating current data and robot end-effector motion physical state parameters includes: adding a unified timestamp to the detection response data, load operating current data, and robot end-effector motion physical state parameters; aligning the detection response data and the robot end-effector motion physical state parameters in time according to the communication response delay and the robot control cycle; and constructing the time-aligned detection response data, load operating current data, and robot end-effector motion physical state parameters into a multi-source response sequence within the same detection cycle.

[0011] According to one achievable method in this application embodiment, the step of extracting response frame structure features and response timing features from the probe response data includes: determining the response frame boundary, field distribution position, field length, and check field position based on frame header changes, frame tail changes, frame length changes, and check field changes in continuous probe response data, and generating the response frame structure features; dividing the probe response data into field units based on the response frame structure features to obtain multiple candidate field units; determining candidate command fields, candidate status fields, and candidate interlock fields from the multiple candidate field units based on the value changes, position stability, and field length changes of each candidate field unit corresponding to different probe commands; and generating the response timing features based on the response delay, change order, and duration of each candidate command field, candidate status field, and candidate interlock field relative to the probe command sending time.

[0012] According to one achievable method in this application embodiment, the step of generating field action associations based on the response frame structure features, response timing features, load operating current data, and robot end-effector motion physical state parameters includes: determining candidate command fields, candidate state fields, and candidate interlock fields in the detection response data based on the response frame structure features; determining the field change time, field change order, and field holding time of each candidate command field, candidate state field, and candidate interlock field relative to the detection command sending time based on the response timing features; determining the current mutation time, current duration, and current recovery slope of the target load within the corresponding detection period based on the load operating current data; determining the end-effector pose change, end-effector velocity change, and end-effector torque change of the target load within the corresponding detection period based on the robot end-effector motion physical state parameters; and generating the field action associations based on the temporal correspondence and direction consistency between the field change time, field change order, and field holding time and the current mutation time, current duration, current recovery slope, end-effector pose change, end-effector velocity change, and end-effector torque change.

[0013] According to one achievable method in the embodiments of this application, the construction of the protocol semantic state model of the target payload based on the field action association includes: determining the lock confirmation state, power supply stability state, enable state, action execution state, and abnormal protection state of the target payload according to the field action association, and taking each state as a protocol semantic state node; generating state transition edges according to the state change sequence triggered by different detection commands; and writing the field that triggers the state transition edge, the corresponding load operating current change condition, and the robot end effector motion physical state change condition into the protocol semantic state model.

[0014] According to one achievable method in this application embodiment, generating the protocol configuration parameters corresponding to the target payload based on the protocol semantic state model includes: determining the frame parsing rules, command mapping rules, state feedback mapping rules, and safety interlock rules corresponding to the target payload based on the protocol semantic state model; when a change in a certain field does not cause a change in the corresponding payload operating current or a change in the physical state of the robot end effector, the field is marked as a field to be confirmed, and it is prohibited from being written as an action control field into the command mapping rules; generating the protocol configuration parameters based on the frame parsing rules, command mapping rules, state feedback mapping rules, and safety interlock rules.

[0015] According to the second aspect, a self-learning protocol modeling system for a robot quick-change payload interface is provided. The system includes: an access state acquisition module, used to detect when a target payload is connected to the robot quick-change interface, and acquire the quick-change locking state, interface contact stability parameters, and initial communication state; a detection command generation module, used to generate a hierarchical safety detection command set and send it to the target payload when the quick-change locking state and the interface contact stability parameters meet preset access conditions; a data acquisition module, used to receive corresponding detection response data and simultaneously acquire payload operating current data and robot end-effector motion physical state parameters; a feature extraction module, used to extract response frame structure features and response timing features based on the detection response data; an association modeling module, used to generate field action association relationships based on the response frame structure features, response timing features, payload operating current data, and robot end-effector motion physical state parameters; a model construction module, used to construct a protocol semantic state model of the target payload based on the field action association relationships; and a configuration output module, used to generate protocol configuration parameters corresponding to the target payload based on the protocol semantic state model.

[0016] According to the specific embodiments provided in this application, the following technical effects are disclosed: This application reduces the risk of false triggering during the unknown load access phase by comprehensively judging the quick-switch locking state, interface contact stability parameters, and initial communication state after the target load is connected, and performing hierarchical safety detection after the access conditions are met. At the same time, by synchronously collecting detection response data, load operating current data, and robot end-effector motion physical state parameters, the changes in communication fields are combined with changes in load current and end-effector motion response to generate field action correlations and construct a protocol semantic state model. This makes the protocol configuration parameters no longer rely solely on message structure recognition, improving the accuracy, adaptation efficiency, and quick-switch operation safety of the load protocol self-learning model.

[0017] Of course, any product implementing this application does not necessarily need to achieve all of the advantages described above at the same time. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 A flowchart illustrating the robot quick-change payload interface self-learning protocol modeling method provided in this application embodiment; Figure 2This is a schematic diagram illustrating the generation of a hierarchical security detection instruction set provided in an embodiment of this application. Figure 3 This is a schematic diagram illustrating the generation of protocol configuration parameters provided in an embodiment of this application. Figure 4 This is a structural block diagram of the robot quick-change payload interface self-learning protocol modeling system provided in the embodiments of this application. Detailed Implementation

[0020] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of this application are within the scope of protection of this application.

[0021] Figure 1 This is a flowchart illustrating the robot quick-change payload interface self-learning protocol modeling method provided in an embodiment of this application. Figure 1 As shown, the method may include the following steps: Step 101: After the target load is connected to the robot's quick-change interface, obtain the quick-change locking status, interface contact stability parameters, and initial communication status.

[0022] Step 102: When the quick-change locking state and the interface contact stability parameter meet the preset access conditions, generate a hierarchical security detection instruction set and send it to the target payload.

[0023] Step 103: Receive the corresponding detection response data and simultaneously collect the load working current data and the physical state parameters of the robot end effector motion.

[0024] Step 104: Extract response frame structure features and response timing features based on the probe response data.

[0025] Step 105: Generate field action associations based on the response frame structure features, response timing features, load operating current data, and robot end effector motion physical state parameters.

[0026] Step 106: Construct a protocol semantic state model of the target payload based on the field action association relationship.

[0027] Step 107: Generate protocol configuration parameters corresponding to the target payload based on the protocol semantic state model.

[0028] As can be seen from the above process, this application reduces the risk of false triggering during the unknown load access phase by comprehensively judging the quick-switch locking state, interface contact stability parameters, and initial communication state after the target load is accessed, and performing hierarchical safety detection after the access conditions are met. At the same time, by synchronously collecting detection response data, load operating current data, and robot end-effector motion physical state parameters, the changes in communication fields are combined with changes in load current and end-effector motion response to generate field action correlations and construct a protocol semantic state model. This makes the protocol configuration parameters no longer solely dependent on message structure recognition, improving the accuracy, adaptation efficiency, and quick-switch operation safety of the load protocol self-learning model.

[0029] The following describes in detail each step of the above process and the effects that can be further produced, with reference to the embodiments.

[0030] Step 101 specifically involves: "After the target load is connected to the robot's quick-change interface, the quick-change locking status, interface contact stability parameters, and initial communication status are obtained."

[0031] After the target load is connected to the robot's quick-change interface, the system first obtains the initial connection status through the locking detection unit, electrical connection detection unit, and communication detection unit on the quick-change interface. The quick-change locking status can be obtained from the locking position switch, locking pin position sensor, or pressure detection element within the locking mechanism, used to determine whether the target load has completed mechanical insertion and entered the locking position. The initial communication status can be obtained by the robot controller sending a basic handshake signal to the target load or listening to the target load's power-on response signal, used to determine whether the target load has communication response capability. The interface contact stability parameter comprehensively reflects whether the mechanical and electrical connections are stable after the target load is connected, avoiding direct entry into the protocol detection process when the load is not locked or the terminal contact is poor.

[0032] When generating interface contact stability parameters, the system can collect the locking position signal, locking holding force, terminal contact impedance, power supply voltage fluctuation, and communication handshake success rate of the robot quick-change interface. The locking position signal indicates whether the quick-change mechanism is in the correct locking position, and the locking holding force indicates whether the mechanical connection between the target load and the robot end effector has sufficient holding capacity. The terminal contact impedance reflects the contact quality of the power supply terminal or communication terminal, the power supply voltage fluctuation reflects the stability of the power supply after the target load is connected, and the communication handshake success rate reflects the link reliability of the target load during the initial communication phase. The above data can be continuously sampled within a short detection window after the target load is powered on to reduce the impact of single detection errors on stability judgment.

[0033] The system can generate a mechanical locking stability value based on the locking position signal and the locking holding force. For example, when the locking position signal is continuously valid and the locking holding force is within the preset holding force range, it indicates that the target load has been reliably locked, and the mechanical locking stability value is high. When the locking position signal changes intermittently, the locking holding force is lower than the preset threshold, or the holding force fluctuates greatly, it indicates that the target load may not be fully engaged, the locking may be loose, or there may be abnormal force, and the mechanical locking stability value decreases.

[0034] The system can also generate electrical connection stability values ​​based on terminal contact impedance, power supply voltage fluctuations, and communication handshake success rates. For example, when the terminal contact impedance is low and changes smoothly, the power supply voltage fluctuation is less than a preset fluctuation threshold, and the communication handshake success rate is higher than a preset success rate threshold, it indicates that the power supply and communication links of the quick-switch interface are in a stable state, and the electrical connection stability value is high. When the terminal contact impedance suddenly increases, the power supply voltage drops significantly, or the communication handshake fails frequently, it indicates that there may be poor contact, momentary disconnection, or unstable communication link at the interface terminals, and the electrical connection stability value decreases.

[0035] After obtaining the mechanical locking stability value and the electrical connection stability value, the system can normalize them and fuse them according to preset weights to generate interface contact stability parameters. For example, for loads with large bearing mass or high operational risk, the weight of the mechanical locking stability value can be increased; for loads with high communication frequency or high power supply, the weight of the electrical connection stability value can be increased. The generated interface contact stability parameters are used to subsequently determine whether preset access conditions are met. Only when the quick-change locking state is valid and the interface contact stability parameters reach the preset threshold will the system allow the generation of a hierarchical safety detection instruction set and enter the protocol self-learning modeling process; otherwise, the system can wait, re-detect, or output an access anomaly prompt, thereby improving the safety and reliability of the robot's quick-change load access process.

[0036] Step 102 specifically involves: "When the quick-change locking state and the interface contact stability parameter meet the preset access conditions, a hierarchical security detection instruction set is generated and sent to the target payload."

[0037] The system only enters the probe command generation phase after the quick-switch locking state and interface contact stability parameters meet the preset access conditions. The preset access conditions may include the quick-switch locking state being in place, the interface contact stability parameters reaching a preset stability threshold, and the initial communication state not exhibiting continuous handshake failures or power supply anomalies. This judgment ensures that the target load has basic mechanical, electrical, and communication response conditions, avoiding the sending of probe commands before the load is stably connected, thereby reducing the risk of false triggering, communication anomalies, and interface damage.

[0038] Figure 2This diagram illustrates the generation of a hierarchical security detection instruction set provided in this embodiment. The hierarchical security detection instruction set can be generated progressively based on the current detection permission level. The system determines the current detection permission level based on the fast-switch locking state and interface contact stability parameters. When the locking state only confirms the target payload's arrival but the interface contact stability parameters are within a low confidence range, it is determined to be at the first detection permission level. When the locking state is stable and the interface contact stability parameters reach a high confidence range, but the target payload's safe enabling has not yet been confirmed, it is determined to be at the second detection permission level. When the locking state is stable, the interface contact stability parameters are higher than a preset security threshold, and no anomalies are found in the preceding state query or enable handshake results, it is determined to be at the third detection permission level. This permission level division allows the protocol's self-learning process to gradually transition from low-risk communication queries to restricted action verification, rather than directly executing action-based detection after unknown payload access.

[0039] Under the first detection permission level, the system generates status query detection commands. These commands are mainly used to read the basic status information of the target payload, such as online status, basic response code, device presence identifier, lockout confirmation field, or fault status field. Status query detection commands do not change the operating mode of the target payload, do not trigger payload actions, and do not grant payload enable permissions. Their purpose is to confirm whether the target payload can stably return response data under the current communication link and to provide basic samples for subsequent extraction of response frame structure features and response timing features.

[0040] Under the second detection permission level, the system generates enable handshake detection commands. These commands are used to confirm whether the target payload is allowed to enter a controlled enabled state. For example, they determine whether the target payload has the conditions for further action verification through methods such as safety handshake, soft interlock confirmation, capability query, and operating mode confirmation. Enable handshake detection commands can put the target payload into a responsive but restricted ready state, but do not directly trigger continuous action or high-energy output. The system can determine whether the target payload has conditions such as being prohibited from enabling, interlock not being met, or abnormal protection based on the response fields returned at this stage.

[0041] At the third detection permission level, the system generates low-amplitude motion verification detection commands. These commands are used to verify the correspondence between the target load's motion response and communication fields under strict constraints. For example, they perform short-term, small-amplitude opening and closing verification on gripper loads, short-term on / off verification on adsorption loads, and low-output verification on force control tools. The motion amplitude, duration, and output energy of low-amplitude motion verification detection commands are all limited, and they are typically only allowed to be generated after the first two levels of detection have not detected any anomalies and the interface contact stability parameters meet safety requirements. This ensures that motion verification will not cause erroneous movement of the target load or abnormal displacement of the robot's end effector.

[0042] The low-amplitude motion verification detection commands are generated as follows: Based on the initial communication state of the target payload, the target communication channel is determined from the communication channels supported by the robot's quick-switch interface, and the basic frame transmission format and response waiting window corresponding to the target communication channel are determined; the low-amplitude verification constraint parameters corresponding to the third detection permission level are obtained, including the upper limit of allowable current, the upper limit of motion amplitude, the upper limit of end-effector pose offset, and the upper limit of duration; candidate verification templates that do not change the working mode of the target payload and do not trigger continuous motion are selected from the preset safe motion template library; the candidate verification templates are encapsulated according to the basic frame transmission format, response waiting window, and low-amplitude verification constraint parameters to generate low-amplitude motion verification detection commands; and abnormal stop conditions are configured for the low-amplitude motion verification detection commands.

[0043] Specifically, when generating such instructions, the system first determines the target communication channel based on the initial communication state of the target payload. The initial communication state of the target payload can include the handshake response status of different communication channels, the basic response code, response latency, number of communication errors, and initial frame format characteristics. If the robot's quick-switch interface simultaneously supports Ethernet, serial bus, USB, or other industrial communication channels, the system can select a communication channel with a higher handshake success rate, more stable response latency, and fewer errors as the target communication channel. After determining the target communication channel, the system further determines the basic frame transmission format and response waiting window corresponding to that communication channel. For example, for a serial communication channel, the transmission format of the start identifier, address field, function field, data field, and checksum field can be determined; for an Ethernet communication channel, the message encapsulation method, port or session identifier, and response waiting time can be determined. The response waiting window is used to limit the time range for waiting for the target payload to return a valid response after sending the probe instruction.

[0044] Low-amplitude verification constraint parameters are used to limit the output intensity and duration that low-amplitude motion verification probe commands may cause. Specifically, these can include upper limits on allowable current, motion amplitude, end-effector pose offset, and duration. The upper limit on allowable current limits the maximum operating current of the target load during verification, preventing overcurrent due to accidental triggering by unknown protocols. The upper limit on motion amplitude limits the gripper opening and closing stroke, suction on / off duration, tool output, or other load motion amplitude. The upper limit on end-effector pose offset limits the allowed position or orientation changes of the robot's end-effector during verification, preventing excessive end-effector offset caused by target load motion. The upper limit on duration limits the execution time of the verification motion, ensuring that the verification motion remains a short-duration rather than a continuous motion.

[0045] After obtaining the low-amplitude verification constraint parameters, the system selects candidate verification templates from a preset safety action template library. This library can pre-set multiple template types according to load type, communication channel type, output method, and safety level. Examples include small-amplitude opening and closing templates for gripper-type loads, short-time on / off templates for adsorption-type loads, low-output templates for force control tools, and short-time trigger acquisition templates for visual loads. The selected candidate verification template should not change the target load's operating mode, should not write permanent parameters, should not perform high-risk operations such as reset, zeroing, continuous operation, or forced output, and should be able to generate a detectable slight response within a short time. If the target load type cannot be clearly identified, the system can select a general low-risk verification template, such as a short-time state switching template or a capability response confirmation template.

[0046] The system encapsulates candidate verification templates according to the base frame transmission format, response waiting window, and low-amplitude verification constraint parameters corresponding to the target communication channel, generating low-amplitude action verification probe commands. Encapsulation refers to writing the verification action identifier, action limitation parameters, response waiting requirements, and safety constraint information from the candidate verification template into a command frame that conforms to the transmission requirements of the target communication channel. For example, the system can write the upper limit of action amplitude, upper limit of duration, and upper limit of allowable current as data fields or control constraint fields into the command frame, and supplement address fields, function fields, and verification fields according to the base frame transmission format. The resulting probe commands can be recognized by the target payload communication interface and are also subject to the limitations of the low-amplitude verification constraint parameters, avoiding the formation of high-risk action commands.

[0047] Finally, the system configures abnormal stop conditions for low-amplitude motion verification detection commands. Abnormal stop conditions may include load operating current exceeding the allowable current limit, robot end-effector pose deviation exceeding the limit, end-effector torque mutation exceeding the safety threshold, target load failing to return a valid response within the response waiting window, returning an abnormal status code, communication interruption, or a change in the quick-change lock state. When any abnormal stop condition is triggered, the system immediately stops the continued execution of the current low-amplitude motion verification detection command and may switch to a safe holding state, lower the detection permission level again, or output an access anomaly prompt.

[0048] When generating various probe commands, the system configures an allowable current limit, response timeout, action amplitude limit, and disabled field flags for each probe command. The allowable current limit restricts the maximum operating current of the target load during probe testing; if the current exceeds the limit, an abnormal stop or degradation process is immediately triggered. The response timeout limits the time range within which the target load should return a probe response; if no valid response is returned within this time, the current probe is considered a failure or a communication link anomaly. The action amplitude limit restricts the displacement, angle, clamping stroke, or output that action-type probe commands can cause, preventing excessive actions under unknown protocols. The disabled field flags mask high-risk fields such as reset, parameter writing, continuous operation, and forced output, ensuring that only probe commands conforming to safety constraints are sent during the self-learning protocol modeling phase.

[0049] In this way, the graded safety detection instruction set does not simply send fixed test frames to the target payload, but dynamically determines the detection permission level based on the locking state and contact stability of the robot's quick-change interface, and gradually opens different types of detection instructions under different permission levels.

[0050] Step 103 specifically involves: "receiving the corresponding detection response data and simultaneously collecting the load operating current data and the physical state parameters of the robot's end effector motion."

[0051] After sending a graded safety detection command to the target payload, the system receives the detection response data returned by the target payload and simultaneously initiates the synchronous acquisition of payload operating current data and robot end-effector motion physical state parameters. The detection response data reflects the target payload's response at the communication layer, the payload operating current data reflects whether the target payload generates an actual power consumption response during the execution of the detection command, and the robot end-effector motion physical state parameters reflect whether the target payload's actions cause changes in end-effector pose, velocity, or torque.

[0052] To ensure that data from different sources can be analyzed under the same time reference, the system adds a unified timestamp to the detection response data, load operating current data, and robot end-effector motion physical state parameters. This unified timestamp can be provided by the robot controller, quick-switch interface control unit, or upper-level control system, and is used to mark the time of detection command transmission, response data reception, current sampling, and end-effector physical state sampling. By using the unified timestamp, it can be determined whether changes in a certain field, current, and end-effector motion occur within the same detection cycle, providing a temporal basis for subsequently generating field action correlations.

[0053] Since communication response and robot motion control typically have different sampling periods and transmission delays, the system needs to perform time alignment based on the communication response delay and the robot control cycle. The communication response delay can be determined based on the time difference between the time the probe command is sent and the time the probe response data is received, while the robot control cycle can be determined based on the robot controller's control refresh cycle. The system can compensate the probe response data according to the communication response delay and map the compensated response data to the corresponding robot control cycle, enabling the probe response data and the robot's end effector motion physical state parameters to be compared within the same time window.

[0054] After time alignment, the system constructs a multi-source response sequence within the same detection cycle from the probe response data, load operating current data, and robot end-effector motion physical state parameters. This multi-source response sequence can be organized according to the probe command number, transmission time, response time, current change data, end-effector pose change data, end-effector velocity change data, and end-effector torque change data. Each detection cycle corresponds to a complete set of communication and physical response records, enabling the system to determine whether the target load produces a consistent response in the communication fields, current changes, and end-effector physical state after a probe command is triggered.

[0055] Step 104 specifically involves: "Extracting response frame structure features and response timing features based on the probe response data."

[0056] After receiving the probe response data returned by the target payload, the system first performs structural analysis on the response data over multiple consecutive probe cycles. Since different target payloads may use different communication frame formats, the system does not pre-assume a fixed protocol template. Instead, it determines the structural boundaries of the response frame based on recurring or regularly changing data segments in the continuous probe response data. For example, if a data segment consistently appears at the beginning of multiple response data sessions, it can be considered a candidate region for the frame header; if a data segment consistently appears at the end of multiple response data sessions, it can be considered a candidate region for the frame tail; if the length of the response data corresponding to different probe commands shows a regular change, the frame length field or data area length can be determined accordingly; if a field changes accordingly with the content of the preceding field, it can be considered a candidate region for the check field. By comprehensively judging changes in the frame header, frame tail, frame length, and check field, the system determines the response frame boundaries, field distribution locations, field lengths, and check field locations, and generates the structural features of the response frame.

[0057] After generating the structural features of the response frame, the system divides the probe response data into field units based on the response frame boundaries and field distribution locations. A field unit can be understood as a data segment in the response frame with a relatively stable position, length, or variation pattern. The system can divide a complete response frame into multiple candidate field units according to field length, field start position, field end position, and verification coverage. Through this processing, the original continuous response data is transformed into a structured set of fields that facilitates subsequent analysis, providing a foundation for identifying command fields, status fields, and interlock fields.

[0058] Subsequently, the system compares and analyzes the candidate field units corresponding to different detection commands. For each candidate field unit, the system can statistically analyze its value changes, position stability, and field length changes under different detection commands. When the value of a candidate field unit changes with the type of detection command, and its position and length are relatively stable, it can be identified as a candidate command field; when a candidate field unit changes when the target load state changes, but remains stable during the detection cycle without state changes, it can be identified as a candidate state field; when a candidate field unit corresponds to locking, power supply, enabling, or abnormal protection conditions, and affects whether subsequent commands are allowed to be executed, it can be identified as a candidate interlock field. Through the above methods, the system can extract candidate fields with semantic potential from the response data in unknown protocol scenarios.

[0059] After identifying candidate command fields, candidate status fields, and candidate interlock fields, the system further analyzes the temporal variation patterns of these fields relative to the time of probe command transmission. Specifically, the system records the time of probe command transmission, the time of response data reception, and the time when each candidate field changes value, calculating the response latency of each candidate field. Simultaneously, the system determines the order of changes based on the sequence of changes across multiple fields and the duration of a field's state at a given value. For example, after sending an enable handshake probe command, if an interlock field changes first, followed by a status field change, a temporal correlation can be considered between the two. After sending a low-amplitude action verification probe command, if a status field changes rapidly and persists until the action ends, this field may be related to the action's execution state.

[0060] Finally, the system generates response timing features based on the response delays, change order, and durations of each candidate command field, candidate state field, and candidate interlock field. These response timing features characterize the temporal response relationships of different fields after the probe command is triggered, enabling the system not only to identify which fields are in the response frame but also to determine the sequential effects and durations of these fields during protocol interaction. These response timing features can be combined with subsequently acquired load operating current data and robot end-effector motion physical state parameters to generate field action correlations, thereby further verifying whether the candidate fields truly correspond to the target load's actions, state feedback, or safety interlock semantics.

[0061] Step 105 specifically involves: "Generating field action associations based on the response frame structure features, response timing features, load operating current data, and robot end-effector motion physical state parameters."

[0062] After obtaining the structural and temporal characteristics of the response frame, the system further performs correlation analysis between the communication fields and the actual electromechanical response of the target load. The purpose of this step is to determine whether candidate fields in the response data truly correspond to load actions, load states, or safety interlocks, rather than relying solely on changes in field values ​​for protocol semantic judgment. For robot fast-change loads, fields truly related to actions or states should typically exhibit a stable temporal correspondence and consistent direction of change with variations in load operating current, robot end-effector pose, end-effector velocity, or end-effector torque.

[0063] The system first determines the candidate command fields, candidate status fields, and candidate interlock fields in the probe response data based on the structural features of the response frame. The structural features of the response frame include information such as the field distribution location, field length, field boundaries, and check field location, which the system uses to determine the position and structural range of each candidate field in the response frame. Candidate command fields typically reflect the target payload's recognition or execution of a certain type of probe command; candidate status fields typically reflect the current state of the target payload; and candidate interlock fields typically reflect safety conditions such as locking, power supply, enabling, or anomaly protection. Through this processing, the system transforms the raw response data into a set of fields with candidate semantics.

[0064] The system determines the field change time, field change order, and field retention time of each candidate command field, candidate status field, and candidate interlock field relative to the time the probe command was sent, based on the response timing characteristics. The field change time indicates when a field's value changes after the probe command is sent; the field change order indicates the sequential relationship between multiple fields; and the field retention time indicates the duration for which a field remains in a specific value state. For example, after sending an enable handshake probe command, the interlock field may change first, followed by the status field entering an enabled state; after sending a low-amplitude action verification probe command, the command field or status field may change within a short period and retain its corresponding value during the action verification period.

[0065] The system determines the moment of current abrupt change, duration of current, and current recovery slope of the target load within the corresponding detection period based on the load's operating current data. The moment of current abrupt change indicates the time when the target load begins to perform an action or switch states; the duration of current indicates the length of time the target load maintains its action or state; and the current recovery slope indicates how quickly the operating current recovers to a stable level after the action ends. For example, when a gripper performs low-amplitude clamping verification, the operating current typically experiences a short-term rise at the start of the action and gradually decreases after the action ends; if an adsorption-type load performs short-term on / off verification, the current change will also correspond to the on / off action.

[0066] The system also determines the changes in end-effector pose, velocity, and torque of the target load within the corresponding detection period based on the robot's end-effector motion physical state parameters. The end-effector pose change reflects whether the target load's motion causes a change in the robot's end-effector position or attitude; the end-effector velocity change reflects whether the end-effector's motion state changes during the motion; and the end-effector torque change reflects whether the target load's motion, gripping, contact, or output resistance has a mechanical impact on the robot's end-effector. For example, when the gripper performs a slight closing motion and contacts the test object, the end-effector torque may change in the corresponding direction; when the target load performs a small displacement motion, the end-effector pose or velocity may change accordingly.

[0067] After obtaining information on field changes, current changes, and end-effector physical changes, the system generates field action correlations based on their temporal correspondence and consistency of change direction. Temporal correspondence can be understood as whether the moment of field change falls within the same detection period as the moment of current abrupt change, end-effector pose change, or end-effector torque change, and whether this meets a preset time deviation range. Consistency of change direction can be understood as whether the direction of action or state change represented by the field value change matches the current change trend, end-effector pose change direction, end-effector velocity change direction, or end-effector torque change direction. If a field change is always accompanied by a corresponding current abrupt change and end-effector physical response, and the two are close in time and consistent in direction, the system can establish a field action correlation between that field and the corresponding action or state.

[0068] For example, after a low-amplitude motion verification probe command is sent, if a candidate state field switches from a first value to a second value, and simultaneously the target load's operating current experiences a sudden increase after a short delay and remains so for a period of time, and the robot's end effector torque also changes in the same direction as clamping, then the system can associate this candidate state field as a clamping motion feedback field. Similarly, if an interlock field only allows the motion field to take effect after the locking state is stable, the power supply is stable, and the enable handshake is successful, then the system can associate this candidate interlock field as a safety interlock field. Conversely, if a field changes value but there is no corresponding current change or end effector physical state change within the same probe cycle, or if its change direction is inconsistent with the end effector physical response direction, then this field can be marked as a weakly associated field or a field to be confirmed, and will not be used as the basis for motion control or critical state feedback.

[0069] The field-action relationships generated in the above manner can include field identifiers, field types, associated action types, associated state types, corresponding detection commands, field change times, current change times, end-effector physical response parameters, and association reliability. These field-action relationships provide the foundation for subsequent construction of the protocol semantic state model, enabling the protocol modeling results to reflect not only the communication frame structure but also the correspondence between communication fields and the actual electromechanical responses of the robot's fast-change payloads, thereby improving the accuracy and security of self-learning modeling for unknown payload protocols.

[0070] Step 106 specifically involves: "Constructing a protocol semantic state model of the target payload based on the field action association relationship".

[0071] After generating the field action relationships, the system can further abstract the communication status, power supply status, enable status, action status, and abnormal status of the target payload during the quick-switch access and detection process into a protocol semantic state model. This model does not simply describe the communication frame format, but uses changes in communication fields, changes in payload operating current, and the physical response of the robot end effector as the basis for state identification, thereby enabling the model to reflect the true working semantics of the target payload under the robot's quick-switch interface.

[0072] The system first determines multiple protocol semantic states of the target load based on the field action relationships. The latching confirmation state indicates that the target load has completed mechanical latching and has a reliable connection foundation with the fast-switch interface; the power supply stability state indicates that the power supply voltage, current, and terminal contact status are within a stable range after the target load is connected; the enable state indicates that the target load has passed security handshake or interlock confirmation and is allowed to enter the restricted action verification stage; the action execution state indicates that the target load has generated a corresponding action response triggered by a low-amplitude action verification probe command; and the anomaly protection state indicates risk states such as communication anomalies, overcurrent, latching anomalies, terminal physical response anomalies, or load feedback anomalies. The system uses these states as protocol semantic state nodes and records the corresponding field conditions, current conditions, and terminal physical state conditions for each state node.

[0073] When generating state transition edges, the system models the state changes in the order triggered by different probe commands. State transition edges represent the triggering conditions for the target payload to switch from one protocol semantic state to another. For example, after a state query probe command is sent, if the target payload returns a lock confirmation field and the quick-change lock state is valid, the system can generate a state transition edge from the initial access state to the lock confirmation state. After an enable handshake probe command is sent, if a candidate interlock field changes to an allowed state and the payload operating current remains stable, the system can generate a state transition edge from a stable power supply state to an enabled / allowed state. After a low-amplitude motion verification probe command is sent, if a candidate motion field or state field undergoes an expected change, and the payload operating current experiences a corresponding abrupt change, resulting in a matching change in the robot's end-effector pose, velocity, or torque, the system can generate a state transition edge from an enabled / allowed state to an motion execution state.

[0074] In one specific implementation, the system establishes a sequence of detection events according to the order in which the detection commands are sent. Each detection event includes at least the detection command type, the time when the detection command is sent, the corresponding response field change, the load operating current change, and the change in the physical state of the robot's end effector. The system first takes the initial state after the target load is connected as the starting state node, and then analyzes the state changes caused by state query detection commands, enable handshake detection commands, and low-amplitude motion verification detection commands in sequence.

[0075] When a status query probe command is sent, if the probe response data contains a lock confirmation field, and the fast-change lock status is valid and the interface contact stability parameters meet the preset access conditions, the system determines that the target payload has changed from the initial access state to the lock confirmation state, and generates a state transition edge from the initial access state to the lock confirmation state. This state transition edge records the status query probe command that triggered the transition, the lock confirmation field, the time of field change, and the corresponding lock-in conditions.

[0076] When an enable handshake probe command is sent, if the candidate interlock field changes from a prohibited state to an enabled state, and the load operating current remains within a stable range without any abnormal current surges, the system determines that the target load has changed from a latch-up confirmation state or a power supply stable state to an enable-enabled state, and generates a corresponding state transition edge. This state transition edge records the enable handshake field that triggered the transition, the field holding time, the current stabilization condition, and the interlock condition that allows subsequent action verification.

[0077] When a low-amplitude motion verification probe command is sent, if the candidate command field or candidate state field changes within a preset response time, and the load operating current exhibits a short-term change matching the motion duration, and the robot end-effector pose change, end-effector velocity change, or end-effector torque change is consistent with the motion direction, then the system determines that the target load has changed from the enabled state to the motion execution state, and generates a state transition edge from the enabled state to the motion execution state. This state transition edge records the fields that triggered the motion, the motion feedback field, the time of the current change, the current duration, the end-effector physical response conditions, and the state retention conditions after the motion is completed.

[0078] If, after any detection command is sent, the detection response data returns an anomaly field, or the load operating current exceeds the allowable current limit, or the change in robot end-effector torque or end-effector pose offset exceeds the safety threshold, the system determines that the target load has changed from its current state to an anomaly protection state and generates a state transition edge from the current state to the anomaly protection state. This state transition edge records the anomaly trigger field, the time of anomaly occurrence, the anomaly current condition, the anomaly end-effector physical state condition, and the corresponding stop or degradation processing condition.

[0079] The generation of state transition edges depends not only on whether the communication fields change, but also on the sequential relationship between field changes and physical responses. For example, after a probe command is sent, the system can record changes in the command field, interlock field, state field, current surge, end torque change, and field hold time in chronological order. If these changes occur sequentially according to preset logic, and the time intervals between them are within the allowable range, it indicates that the probe command has indeed triggered a state transition of the target load. If a field change occurs, but there is no corresponding current change or end physical state change, the system may not generate a state transition edge related to the action execution, or may mark the transition as a low-confidence transition.

[0080] For abnormal protection states, the system can generate corresponding state transition edges based on the abnormal response field, current anomaly, and end-effector physical anomaly. For example, when the target load returns an abnormal code, the load operating current exceeds the allowable current limit, the robot end-effector torque suddenly exceeds the safety threshold, or the quick-change locking state fails during detection, the system can generate a state transition edge from the current state to the abnormal protection state. This state transition edge can record the abnormal trigger field, abnormal current condition, abnormal end-effector physical state condition, and corresponding safety handling strategy, thereby enabling the protocol semantic state model to have safe interlocking and abnormal rollback capabilities.

[0081] When constructing the protocol semantic state model, the system also writes the fields that trigger each state transition edge, the corresponding load operating current change conditions, and the robot end effector motion physical state change conditions into the model. For example, a state transition edge from an enable state to an action execution state can record the candidate command field that triggers the transition, the action state feedback field, the time of current change, the duration of current change, the current recovery slope, the end effector pose change, and the end effector torque change. In this way, when the robot control system uses the protocol model, it can not only know how a certain field should be parsed, but also know what kind of current change and end effector physical response should accompany the field's activation.

[0082] The protocol semantic state model constructed in the above manner can unify the communication protocol fields of the target payload, the security state of the quick-switch interface, and the electromechanical response of the payload into the same state model.

[0083] Step 107 specifically involves: "Generating protocol configuration parameters corresponding to the target payload based on the protocol semantic state model".

[0084] After obtaining the protocol semantic state model of the target payload, the system can generate the corresponding protocol configuration parameters based on the model. The protocol semantic state model already records each protocol semantic state node, state transition edge, field that triggers state transition, corresponding load operating current change conditions, and robot end effector motion physical state change conditions. Therefore, the system can extract a set of rules that can be used for actual control configuration from the model, enabling the robot control system to perform communication parsing, command sending, state reading, and safety interlock control of the target payload according to these rules.

[0085] Figure 3 This diagram illustrates the generation of protocol configuration parameters provided in this embodiment. In specific implementation, the system first determines frame parsing rules based on the protocol semantic state model. Frame parsing rules may include response frame boundaries, frame header position, frame tail position, field distribution position, field length, check field position, and check method, etc., to guide the robot quick-change interface or robot controller in correctly splitting and reading the response data returned by the target payload. Through these frame parsing rules, the system can identify different fields from the raw communication data and provide basic data for subsequent command mapping, state feedback mapping, and interlock judgment.

[0086] The system also determines command mapping rules based on the protocol semantic state model. Command mapping rules describe the correspondence between robot control commands and target payload protocol fields. For example, they specify which fields in the target payload communication frame, their values, and the transmission order of enable commands, low-amplitude motion verification commands, gripping commands, release commands, or stop commands in the robot control system should be mapped to. The generation of command mapping rules is not simply based on whether a field changes, but rather on whether the field is confirmed as a valid trigger field in the state transition edge and whether it is accompanied by a change in the corresponding payload operating current or a change in the physical state of the robot's end effector. Only fields that meet these conditions can be written into the command mapping rules as motion control-related fields.

[0087] The system further determines state feedback mapping rules based on the protocol semantic state model. These rules describe the correspondence between the state fields returned by the target load and the internal state variables of the robot control system. For example, a field can be mapped to a lock confirmation state, a power supply stability state, an enable state, an action execution state, or an anomaly protection state. Since these state fields have already been verified in the protocol semantic state model through field-action relationships, their state feedback results can be corroborated by changes in load operating current or changes in the physical state of the robot's end effector, thereby improving the reliability of the state feedback.

[0088] Safety interlock rules are used to restrict which commands the target load is allowed or prohibited from executing under different protocol semantic states. For example, when the lockout confirmation state is not established, enabling handshake probe commands are prohibited from being sent; when the power supply stability state is not established, low-amplitude action verification probe commands are prohibited from being sent; when the enable state is not established, writing to the action control field is prohibited; when entering an abnormal protection state, continuing to send action commands is prohibited, and stop, de-enable, or degradation processing is triggered. Safety interlock rules can be generated based on the interlock field, abnormal field, current limit conditions, and end-point physical state limit conditions in the state transition edge.

[0089] During the generation of protocol configuration parameters, the system also performs a safety screening of fields. If a change in a field does not cause a corresponding change in the load's operating current or the physical state of the robot's end effector, it indicates that although the field changes in the communication response data, it lacks actual electromechanical response support, and a genuine correlation between it and the target load's action cannot be confirmed. In this case, the system marks the field as a field to be confirmed and prohibits it from being written into the command mapping rules as an action control field. Fields to be confirmed can be retained in the auxiliary parsing section of the protocol configuration parameters for further observation or manual confirmation, but cannot be directly used to trigger the target load's action.

[0090] For example, in low-amplitude motion verification detection, if a candidate field changes from zero to one in the response frame, but the load operating current does not show a corresponding abrupt change, and the robot's end-effector pose, velocity, or torque does not show a corresponding change, the system will not configure this field as a gripping motion field or a releasing motion field, but will instead mark it as a field to be confirmed. Conversely, if a change in a field has a stable correspondence with the time of the current abrupt change, the duration of the motion, and the direction of the end-effector torque change, then this field can be written into the command mapping rule or the state feedback mapping rule.

[0091] Finally, the system generates protocol configuration parameters corresponding to the target payload based on frame parsing rules, command mapping rules, status feedback mapping rules, and safety interlock rules. These protocol configuration parameters can include communication frame parsing configuration, command field configuration, status field configuration, interlock field configuration, response timeout configuration, current limit configuration, motion amplitude limit configuration, and anomaly handling configuration. The robot control system or quick-switch interface control unit can load these protocol configuration parameters for subsequent automatic identification, automatic configuration, status monitoring, and safety control of the target payload.

[0092] To further illustrate the technical effects of the present invention, a specific embodiment and its test results are provided.

[0093] The method of this application is applied to a six-axis industrial robot and its end effector quick-change interface. The target load includes an electric gripper, a vacuum adsorption module, and a vision inspection module. The robot quick-change interface is equipped with a locking position detection switch, a locking holding force detection unit, a terminal contact impedance detection unit, a voltage sampling unit, a current sampling unit, and a communication interface control unit. The communication interface supports Ethernet, RS485, and USB communication methods. After the target load is connected, the system first collects the locking position signal, locking holding force, terminal contact impedance, power supply voltage fluctuation, and communication handshake success rate, and generates interface contact stability parameters. When the locking position signal is continuously valid, the locking holding force reaches the set holding range, the terminal contact impedance and power supply voltage fluctuation are both within the allowable range, and the communication handshake success rate meets the requirements, the system determines that the target load meets the access conditions and enters the graded safety detection process.

[0094] During the graded safety detection process, the system first generates status query detection commands to obtain the basic response and status fields of the target payload. If no anomalies are found in the status query, an enable handshake detection command is generated to confirm the enable and interlock states of the target payload. Once the target payload enters the third detection permission level, the system determines the target communication channel based on the initial communication state and selects the corresponding low-amplitude action verification template from a pre-set safety action template library. For example, for an electric gripper, the system selects a low-amplitude opening and closing verification template and sets the upper limits for allowable current, action amplitude, end-effector pose offset, and duration. This template is then encapsulated into a low-amplitude action verification detection command according to the basic frame transmission format of the target communication channel. After sending the detection command, the system synchronously collects detection response data, payload operating current data, and robot end-effector pose, velocity, and torque data, and constructs a multi-source response sequence within the same detection cycle using a unified timestamp.

[0095] The system then determines the response frame boundaries, field distribution locations, field lengths, and check field locations based on changes in the frame header, frame tail, frame length, and check field in the continuous detection response data, generating response frame structural features. Based on these features, the detection response data is further divided into field units, yielding multiple candidate field units. The system then determines candidate command fields, candidate state fields, and candidate interlock fields based on the value changes, positional stability, and field length changes of these candidate field units under different detection commands. Furthermore, it generates response timing features based on the response delay, change order, and duration of these fields relative to the detection command transmission time. Finally, the system performs correlation analysis between the field change time, field change order, and field hold time, and the current abrupt change time, current duration, and current recovery slope in the load operating current data, as well as the end-effector pose change, end-effector velocity change, and end-effector torque change in the robot's end-effector motion physical state parameters, generating field action correlation relationships.

[0096] Based on the field-action relationship, the system constructs a protocol semantic state model for the target payload. Taking an electric gripper as an example, the system uses the initial access state, lock confirmation state, power supply stability state, enable state, action execution state, and anomaly protection state as protocol semantic state nodes. When a state query probe command triggers a change in the lock confirmation field and the quick-change lock state is valid, the system generates a state transition edge from the initial access state to the lock confirmation state. When an enable handshake probe command triggers an interlock field change from a prohibited state to an enabled state, and the payload operating current remains stable, the system generates a state transition edge from the power supply stability state to the enable state. When a low-amplitude action verification probe command triggers a change in the action field, and the payload operating current experiences a short-term increase while the robot's end effector torque changes in the corresponding direction, the system generates a state transition edge from the enable state to the action execution state. The system writes the fields that trigger the state transition edge, the corresponding current change conditions, and the end effector physical state change conditions into the protocol semantic state model, and generates frame parsing rules, command mapping rules, state feedback mapping rules, and safety interlock rules based on this model as protocol configuration parameters for the target payload.

[0097] To verify the effectiveness of the proposed solution, three different target payloads were selected for testing. Thirty fast-switch access and protocol self-learning modeling tests were conducted on each payload type. During the testing, the manual configuration method, the protocol modeling method based solely on communication frame structure recognition, and the electromechanical response coupling self-learning modeling method of this application were compared. Test results show that the proposed solution was able to automatically generate protocol configuration parameters after stable access of the target payload in all 90 tests, with an average modeling time of 16.8 seconds, significantly shorter than the manual configuration method. For the electric gripper and vacuum adsorption module, the proposed solution accurately identified action execution-related fields, status feedback fields, and safety interlock fields, with a field semantic recognition accuracy rate exceeding 96%.

[0098] According to another embodiment, a self-learning protocol modeling system for robot quick-change payload interface is provided. Figure 4 A schematic block diagram of a robot quick-load-change interface self-learning protocol modeling system according to one embodiment is shown. Figure 4 As shown, the system includes: The access status acquisition module 401 is used to detect when the target load is connected to the robot quick-change interface and to acquire the quick-change locking status, interface contact stability parameters and initial communication status.

[0099] The detection instruction generation module 402 is used to generate a hierarchical security detection instruction set and send it to the target payload when the quick-change locking state and the interface contact stability parameter meet the preset access conditions.

[0100] The data acquisition module 403 is used to receive the corresponding detection response data and simultaneously acquire the load operating current data and the physical state parameters of the robot end effector.

[0101] The feature extraction module 404 is used to extract response frame structure features and response timing features based on the detection response data.

[0102] The association modeling module 405 is used to generate field action associations based on the response frame structure features, response timing features, load working current data and robot end motion physical state parameters.

[0103] The model building module 406 is used to build a protocol semantic state model of the target payload based on the field action association relationship.

[0104] The configuration output module 407 is used to generate protocol configuration parameters corresponding to the target payload based on the protocol semantic state model.

[0105] As one possible implementation, the interface contact stability parameters of the access status acquisition module 401 are generated in the following way: The locking position signal, locking holding force, terminal contact impedance, power supply voltage fluctuation, and communication handshake success rate of the robot's quick-change interface are collected; a mechanical locking stability value is generated based on the locking position signal and locking holding force; an electrical connection stability value is generated based on the terminal contact impedance, power supply voltage fluctuation, and communication handshake success rate; and the interface contact stability parameters are generated based on the mechanical locking stability value and the electrical connection stability value.

[0106] As one possible implementation, the detection instruction generation module 402 generates a hierarchical security detection instruction set, including: determining the current detection permission level based on the quick-change locking state and the interface contact stability parameters, wherein the detection permission level includes a first detection permission level, a second detection permission level, and a third detection permission level; generating a status query type detection instruction under the first detection permission level; generating an enable handshake type detection instruction under the second detection permission level; and generating a low-amplitude action verification type detection instruction under the third detection permission level. Configure corresponding allowable current limit, response timeout, action amplitude limit, and disabled field flags for each detection command.

[0107] As one possible implementation, the generation of the low-amplitude motion verification detection command in the detection command generation module 402 includes: determining the target communication channel from the communication channels supported by the robot quick-change interface according to the initial communication state of the target payload, and determining the basic frame transmission format and response waiting window corresponding to the target communication channel; obtaining the low-amplitude verification constraint parameters corresponding to the third detection permission level, the low-amplitude verification constraint parameters including the upper limit of allowable current, the upper limit of motion amplitude, the upper limit of end pose offset, and the upper limit of duration; selecting candidate verification templates from the preset safe action template library that do not change the working mode of the target payload and do not trigger continuous actions; encapsulating the candidate verification templates according to the basic frame transmission format, response waiting window, and low-amplitude verification constraint parameters to generate a low-amplitude motion verification detection command; and configuring abnormal stop conditions for the low-amplitude motion verification detection command.

[0108] As one possible implementation, the data acquisition module 403 synchronously acquires load operating current data and robot end-effector motion physical state parameters, including: adding a unified timestamp to the detection response data, load operating current data, and robot end-effector motion physical state parameters; aligning the detection response data and robot end-effector motion physical state parameters in time according to the communication response delay and robot control cycle; and constructing the time-aligned detection response data, load operating current data, and robot end-effector motion physical state parameters into a multi-source response sequence within the same detection cycle.

[0109] As one possible implementation, the feature extraction module 404 extracts response frame structure features and response timing features based on the probe response data, including: determining the response frame boundary, field distribution position, field length, and check field position based on frame header changes, frame tail changes, frame length changes, and check field changes in the continuous probe response data, and generating the response frame structure features; dividing the probe response data into field units based on the response frame structure features to obtain multiple candidate field units; determining candidate command fields, candidate status fields, and candidate interlock fields from the multiple candidate field units based on the value changes, position stability, and field length changes of each candidate field unit corresponding to different probe commands; and generating the response timing features based on the response delay, change order, and duration of each candidate command field, candidate status field, and candidate interlock field relative to the probe command sending time.

[0110] As one possible implementation, the association modeling module 405 generates field action associations based on the response frame structure features, response timing features, load operating current data, and robot end-effector motion physical state parameters. This includes: determining candidate command fields, candidate state fields, and candidate interlock fields in the detection response data based on the response frame structure features; determining the field change time, field change order, and field holding time of each candidate command field, candidate state field, and candidate interlock field relative to the detection command sending time based on the response timing features; determining the current mutation time, current duration, and current recovery slope of the target load within the corresponding detection period based on the load operating current data; determining the end-effector pose change, end-effector velocity change, and end-effector torque change of the target load within the corresponding detection period based on the robot end-effector motion physical state parameters; and generating the field action associations based on the temporal correspondence and direction consistency between the field change time, field change order, and field holding time and the current mutation time, current duration, current recovery slope, end-effector pose change, end-effector velocity change, and end-effector torque change.

[0111] As one possible implementation, the model building module 406 constructs a protocol semantic state model of the target payload based on the field action association relationship, including: determining the lock confirmation state, power supply stability state, enable state, action execution state, and abnormal protection state of the target payload according to the field action association relationship, and taking each state as a protocol semantic state node; generating state transition edges according to the state change sequence triggered by different detection commands; and writing the fields that trigger the state transition edges, the corresponding load operating current change conditions, and the robot end effector motion physical state change conditions into the protocol semantic state model.

[0112] As one possible implementation, the configuration output module 407 generates protocol configuration parameters corresponding to the target payload based on the protocol semantic state model, including: determining the frame parsing rules, command mapping rules, state feedback mapping rules, and safety interlock rules corresponding to the target payload based on the protocol semantic state model; when a change in a certain field does not cause a change in the corresponding payload operating current or a change in the physical state of the robot end effector, the field is marked as a field to be confirmed, and it is prohibited from being written as an action control field into the command mapping rules; and the protocol configuration parameters are generated based on the frame parsing rules, command mapping rules, state feedback mapping rules, and safety interlock rules.

[0113] In addition, embodiments of this application also provide a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the steps of the method described in any of the foregoing method embodiments.

[0114] And an electronic device comprising: one or more processors; and a memory associated with the one or more processors, the memory being used to store program instructions that, when read and executed by the one or more processors, perform the steps of the method described in any of the foregoing method embodiments.

[0115] The technical solutions provided in this application have been described in detail above. Specific examples have been used to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. Furthermore, those skilled in the art will recognize that, based on the ideas of this application, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A method for modeling a self-learning protocol for a robot's quick-change payload interface, characterized in that, The method includes: After the target load is connected to the robot's quick-change interface, the quick-change locking status, interface contact stability parameters, and initial communication status are obtained. When the quick-change locking state and the interface contact stability parameter meet the preset access conditions, a hierarchical security detection instruction set is generated and sent to the target payload. Receive the corresponding detection response data and simultaneously collect the load operating current data and the physical state parameters of the robot end effector motion; Extract response frame structure features and response timing features from the probe response data; Based on the structural features of the response frame, candidate command fields, candidate status fields, and candidate interlock fields are determined in the probe response data. Based on the response timing characteristics, determine the field change time, field change order, and field retention time of each candidate command field, candidate status field, and candidate interlock field relative to the probe command sending time; Based on the load operating current data, determine the time of current abrupt change, the duration of current, and the current recovery slope of the target load within the corresponding detection period. The changes in end-effector pose, end-effector velocity, and end-effector torque of the target load within the corresponding detection period are determined based on the physical state parameters of the robot's end-effector motion. Based on the time correspondence and direction of change between the field change time, field change order and field holding time and the current change time, current duration, current recovery slope, end pose change, end velocity change and end torque change, the field action association relationship is generated. Based on the aforementioned field action relationships, a protocol semantic state model of the target payload is constructed; The protocol configuration parameters corresponding to the target payload are generated based on the protocol semantic state model.

2. The method according to claim 1, characterized in that, The interface contact stability parameters are generated in the following manner: Collect the locking position signal, locking holding force, terminal contact impedance, power supply voltage fluctuation, and communication handshake success rate of the robot's quick-change interface; A mechanical locking stability value is generated based on the locking position signal and the locking holding force; An electrical connection stability value is generated based on the terminal contact impedance, power supply voltage fluctuation, and communication handshake success rate. The interface contact stability parameters are generated based on the mechanical locking stability value and the electrical connection stability value.

3. The method according to claim 1, characterized in that, The generation of the hierarchical security detection instruction set includes: The current detection permission level is determined based on the quick-change locking state and the interface contact stability parameters. The detection permission level includes a first detection permission level, a second detection permission level, and a third detection permission level. Generate status query probe commands under the first probe permission level; Generate handshake-type probe commands under the second probe permission level; Generate low-amplitude motion verification probe commands under the third probe permission level; Configure corresponding allowable current limit, response timeout, action amplitude limit, and disabled field flags for each detection command.

4. The method according to claim 3, characterized in that, The generation of the low-amplitude motion verification detection command includes: Based on the initial communication state of the target payload, the target communication channel is determined from the communication channels supported by the robot quick-change interface, and the basic frame transmission format and response waiting window corresponding to the target communication channel are determined. Obtain the low-amplitude verification constraint parameters corresponding to the third detection permission level. The low-amplitude verification constraint parameters include the upper limit of allowable current, the upper limit of action amplitude, the upper limit of end pose offset, and the upper limit of duration. Select a candidate verification template from the preset safety action template library that does not change the target load working mode and does not trigger continuous actions; According to the basic frame transmission format, response waiting window and low amplitude verification constraint parameters, the candidate verification template is encapsulated to generate low amplitude action verification detection instructions; Configure abnormal stop conditions for the low-amplitude motion verification detection command.

5. The method according to claim 1, characterized in that, The synchronous acquisition of load operating current data and robot end-effector motion physical state parameters includes: A unified timestamp is added to the detection response data, load operating current data, and robot end effector motion physical state parameters; Based on the communication response delay and the robot control cycle, the detection response data is time-aligned with the physical state parameters of the robot end effector. The time-aligned detection response data, load operating current data, and robot end effector motion physical state parameters are constructed into a multi-source response sequence within the same detection period.

6. The method according to claim 3, characterized in that, The step of extracting response frame structure features and response timing features based on the probe response data includes: Based on the changes in frame header, frame tail, frame length, and check field in the continuous detection response data, the response frame boundary, field distribution location, field length, and check field location are determined, and the response frame structure features are generated. Based on the structural features of the response frame, the probe response data is divided into field units to obtain multiple candidate field units; Based on the value changes, position stability, and field length changes of each candidate field unit corresponding to different detection commands, candidate command fields, candidate status fields, and candidate interlock fields are determined from the multiple candidate field units. The response timing features are generated based on the response delay, change order, and duration of each candidate command field, candidate status field, and candidate interlock field relative to the time of probe command transmission.

7. The method according to claim 1, characterized in that, The construction of the protocol semantic state model of the target payload based on the field action association includes: Based on the field action association relationship, determine the target load's lockout confirmation state, power supply stability state, enable state, action execution state, and abnormal protection state, and use each state as a protocol semantic state node; State transition edges are generated according to the sequence of state changes triggered by different detection commands; The fields that trigger the state transition edge, the corresponding load operating current change conditions, and the robot end effector motion physical state change conditions are written into the protocol semantic state model.

8. The method according to claim 7, characterized in that, The step of generating protocol configuration parameters corresponding to the target payload based on the protocol semantic state model includes: The frame parsing rules, command mapping rules, state feedback mapping rules, and security interlocking rules corresponding to the target payload are determined based on the protocol semantic state model. When a change in a certain field does not cause a change in the corresponding load operating current or a change in the physical state of the robot end effector, the field is marked as a field to be confirmed and is prohibited from being written into the command mapping rule as a motion control field. The protocol configuration parameters are generated based on the frame parsing rules, command mapping rules, status feedback mapping rules, and security interlock rules.

9. A self-learning protocol modeling system for a robot quick-change payload interface, characterized in that, The system includes: The access status acquisition module is used to detect when the target load is connected to the robot's quick-change interface and to acquire the quick-change locking status, interface contact stability parameters, and initial communication status. The detection instruction generation module is used to generate a hierarchical security detection instruction set and send it to the target payload when the quick-change locking state and the interface contact stability parameter meet the preset access conditions; The data acquisition module is used to receive the corresponding detection response data and simultaneously acquire the load operating current data and the physical state parameters of the robot end effector. The feature extraction module is used to extract response frame structure features and response timing features based on the detection response data; The association modeling module is used to determine candidate command fields, candidate state fields, and candidate interlock fields in the detection response data based on the response frame structure characteristics; determine the field change time, field change order, and field holding time of each candidate command field, candidate state field, and candidate interlock field relative to the detection command sending time based on the response timing characteristics; determine the current mutation time, current duration, and current recovery slope of the target load in the corresponding detection period based on the load operating current data; determine the end-effector pose change, end-effector velocity change, and end-effector torque change of the target load in the corresponding detection period based on the robot end-effector motion physical state parameters; and generate the field action association relationship based on the temporal correspondence and change direction consistency between the field change time, field change order, and field holding time and the current mutation time, current duration, current recovery slope, end-effector pose change, end-effector velocity change, and end-effector torque change. The model building module is used to construct a protocol semantic state model of the target payload based on the field action association relationship; The configuration output module is used to generate protocol configuration parameters corresponding to the target payload based on the protocol semantic state model.

Images