Intelligent lock remote control method based on intelligent interaction

Abstract

The application relates to the technical field of intelligent lock remote authentication and control, and discloses a smart lock remote control method based on intelligent interaction, which comprises the following steps: firstly, completing lock end sensing, security digest, key storage and terminal interactive communication configuration, and calibrating lock bolt coordinates, action categories, permissions and device identification; collecting conversation interaction time and action categories, constructing a remote interaction event matrix, generating interaction intent concentration, candidate control actions and action unique determination quantities; collecting lock bolt travel, door magnetic gap, motor current and lock body acceleration, generating a door lock physical state vector, action physical feasibility, conversation time sequence continuity determination quantity, state binding consistency determination quantity and action access determination quantity; after access is passed, the lock bolt is driven in segments, execution residual error and final in-place error are calculated, and a remote control execution result is output.

Description

Technical Field

[0001] This invention relates to the field of remote authentication and control technology for smart locks, specifically a remote control method for smart locks based on intelligent interaction. Background Technology

[0002] Existing smart lock remote control technologies typically rely on account verification, password verification, temporary authorization, or single command issuance as the primary control basis. In scenarios such as home entrance doors, long-term rental apartments, office access control, and remote opening and closing of unattended warehouses, there is a lack of synchronous constraints between the terminal-side interactive behavior and the actual physical state of the lock. This leads to the user's control intentions formed under conditions of accidental touch, repeated clicks, network jitter, or multiple logins being easily simplified into a single unlock or lock command. At the same time, whether the door is closed, whether the bolt is in the operable range, whether the motor load is abnormal, and whether there is vibration disturbance in the lock body are often provided as post-event status prompts, making it difficult to establish pre-action constraints during the command access stage. Existing remote sessions often rely on fixed timeouts or server forwarding times for judgment, failing to adequately differentiate between network transmission latency, terminal processing time, and the continuity of lock challenge sequence numbers. This can easily lead to unclear instruction timing boundaries in scenarios with weak networks, delayed receipts, or replayed messages. Other solutions, while employing challenge-response or digest methods to improve communication security, often verify interaction intent, lock status, user permissions, and control commands separately, making it difficult to guarantee consistency between pre-control intent and the final control message. In addition, existing smart locks mostly rely on motor power-on time, Hall trigger or lock tongue position signal to judge the control result during the execution phase. For situations such as lock tongue jamming, free spinning, half closing, rebounding or abnormal motor load, there is no execution residual judgment mechanism to cross-check the lock tongue displacement and current load integration, which may lead to a deviation between the execution result received by the remote end and the actual mechanical state of the lock body. Therefore, there is an urgent need for a remote control method for smart locks based on intelligent interaction to solve the problems in the existing technologies mentioned above, such as the difficulty in accurately identifying the remote interaction intent, the difficulty in constraining the on-site state of the lock, the insufficient binding of session timing and state, and the difficulty in verifying the execution result of the lock tongue. Summary of the Invention

[0003] This invention provides a remote control method for smart locks based on intelligent interaction, which helps to solve the problems mentioned in the background art.

[0004] This invention provides the following technical solution: A remote control method for a smart lock based on intelligent interaction includes: Configure sensing, security summary and key storage functions on the lock end, and configure interaction, login, challenge and communication links on the terminal to complete the calibration of lock tongue coordinates, action category, permissions, device identification and pairing key; Collect the session interaction time and action categories of logged-in users, construct a remote interaction event matrix, and extract the proportion of action categories; Based on the interval between adjacent interactions and the proportion of action categories, a concentration of interaction intent is generated to determine candidate control actions and unique action determination quantities. Collect bolt travel, door magnetic gap, motor current and lock body acceleration to generate door lock physical state vector and resolution mapping vector; Based on the candidate action's preceding interval and the door lock's physical state vector, generate the action's physical feasibility quantity; Based on the session handshake message, processing verification receipt, terminal monotonic sequence number and lock-end challenge sequence number, generate the session timing continuity determination quantity; Based on the pre-control intent field group, remote control instruction field group, lock end status commitment quantity and pairing key, generate a status binding consistency judgment quantity, and combine it with user permissions to generate an action access judgment quantity; After the access is approved, the target latch position is determined. The maximum allowable number of segments and the motor drive duty cycle are generated by combining the latch stroke resolution. The latch is driven segment by segment. The conversion factor from current integral to latch displacement is called to calculate the execution residual, generate the final positioning error, and output the remote control execution result.

[0005] Optionally, the steps for completing the calibration of the latch coordinates, action category, permissions, device identifier, and pairing key specifically include: The lock end is equipped with functions for acquiring bolt travel, door magnetic gap, motor current, lock body acceleration, lock end counting, lock end timing, security summary and key storage; Configure the terminal with interactive event collection, user login identifier collection, terminal counting, terminal timing, challenge value generation and remote communication functions; With the fully open position of the latch as the reference zero point and the locking direction of the latch as the positive direction, record the latch position, door clearance, motor no-load state, average current of the complete stroke, load integral of the calibration current of the complete stroke, complete stroke time, sensor resolution, clock resolution and motor duty cycle boundary. The registration includes an action category table, action prerequisite range, user permission set, action required permission set, device identifier, and pairing key.

[0006] Optionally, the step of constructing the remote interaction event matrix and extracting the proportion of action categories specifically includes: Collect the time and type of interaction events generated by the currently logged-in user within a single remote control session; Based on the time difference between the collection of adjacent interaction events and the time span of the first and last interaction events of this session, the adjacent interaction interval is normalized to obtain the normalized interaction interval. Based on the registered action category table, each interaction event is converted into an action category tag vector; The action category marker vectors are combined according to the order in which the interaction events occur to form a remote interaction event matrix; Perform mean statistics on the remote interaction event matrix according to action category to obtain the proportion of each action category in the current session.

[0007] Optionally, the steps of generating the interaction intent set quantity and determining the candidate control actions and the action uniqueness determination quantity specifically include: The mean and dispersion of the normalized interaction interval are calculated to obtain the interaction rhythm characteristics. Normalized entropy is calculated for the proportion of action categories, and zero-probability logarithmic correction is applied to action categories that do not appear to obtain the features of the action set. By combining interaction rhythm features with action concentration features, an interaction intent concentration quantity is generated; Select the action category with the largest proportion as the candidate control action; Based on whether the proportion of action categories corresponding to candidate control actions meets the strict majority condition, a unique decision quantifier for the action is generated.

[0008] Optionally, the step of generating the door lock physical state vector and resolution mapping vector specifically includes: Collect the current latch travel, current door magnetic gap, and current motor current; Collect the lock body's triaxial acceleration within a fixed sampling window before the current session challenge is issued, and calculate the root mean square of the current lock body's triaxial acceleration; Based on the fully open position, fully closed position, and bolt travel resolution, the current bolt travel is normalized and saturated. Based on the door closing gap, the door opening reference gap, and the door magnetic resolution, the current door magnetic gap is normalized and saturated. Based on the motor no-load current, the average current during the entire stroke, and the current sampling resolution, the current of the current motor is normalized and saturated. Based on the standard gravitational acceleration, the root mean square of the current lock body's three-axis acceleration, and the acceleration sampling resolution, the vibration state of the lock body is normalized and processed by resolution propagation to generate the door lock's physical state vector and resolution mapping vector.

[0009] Optionally, the step of generating the physical feasibility quantity of the action based on the candidate action pre-interval and the door lock physical state vector specifically includes: Based on the candidate control actions, extract the corresponding action pre-interval from the action category table; The action pre-interval includes the lower and upper bounds of each physical state component. Based on the resolution mapping vector, the lower and upper bounds of the action pre-interval are expanded. Each physical state component in the door lock's physical state vector is included in the corresponding expanded interval; The inclusion determination results of each physical state component are synthesized under necessary conditions to generate the physical feasible quantity of the action.

[0010] Optionally, the step of generating the session timing continuity determination quantity specifically includes: Based on the set of handshake messages in the current session, record the internal time difference between when the terminal sends a handshake message and when it receives a handshake return message. Record the internal processing time of the lock from receiving the handshake message to sending the handshake return message; Based on the internal time difference of the terminal and the internal processing time of the lock, extract the one-way transmission time estimate; Perform nonnegation processing on each one-way transmission time estimate, select the maximum nonnegative one-way transmission time estimate in the current session, and superimpose the device clock resolution to generate a session response time window; Based on the processing verification messages and their receipts in the current session, record the local time difference between the lock end sending the processing verification message and receiving the processing verification receipt. Subtract the round-trip transmission compensation amount determined by the session response time window and the device clock resolution from the local time difference of the lock end; The deduction result is non-negative, and the maximum non-negative processing compensation amount in the current session is selected to generate the terminal processing compensation duration. Among them, the terminal processing load for processing the verification receipt is no less than the terminal status binding token calculation and remote control instruction message encapsulation load. The session timing continuity determination quantity is generated by jointly determining the lock challenge issuance time, lock command reception time, session response time window, terminal processing compensation duration, terminal monotonic sequence number, and lock challenge sequence number.

[0011] Optionally, the step of generating a consistency determination quantity for state binding and generating an action access determination quantity in conjunction with user permissions specifically includes: The interactive terminal sends a pre-control intent field group containing user identifier, candidate control action, interaction intent concentration, action uniqueness determination quantity, terminal device identifier, and terminal challenge value; The lock terminal generates a lock terminal state commitment quantity based on the pre-control intent field group, door lock physical state vector, action physical feasibility quantity, lock terminal device identifier, lock terminal challenge value, lock terminal challenge sequence number, terminal processing compensation time and pairing key; The interactive terminal generates a terminal status binding token based on the remote control instruction field group, the lock state commitment quantity, the terminal monotonic sequence number, and the pairing key; The interactive terminal sends a remote control instruction field group containing fields with the same name, the terminal monotonic sequence number, and the terminal status binding token; The lock terminal recalculates the lock terminal status binding token based on the remote control instruction field group, the lock terminal status commitment quantity, the terminal monotonic sequence number, and the pairing key; Perform consistency checks on the terminal state binding token, the lock state binding token, the pre-control intent field group, and the remote control instruction field group to generate a state binding consistency check quantity. Generate the permissible permissions based on the current logged-in user's permission set and the permission set required for candidate control actions; The necessary conditions are combined to generate the action admission criteria: accessibility criteria, action uniqueness criteria, session sequence continuity criteria, action physical feasibility criteria, state binding consistency criteria, and interaction intent concentration criteria.

[0012] Optionally, the steps of driving the locking tongue segment by segment, calculating the execution residual, generating the final positioning error, and outputting the remote control execution result specifically include: When the action admission judgment quantity meets the admission conditions, the motor execution flag and target latch position are determined according to the candidate control action; The maximum number of segments is generated based on the target latch position, the actual latch position before execution, and the latch travel resolution. Based on the remaining displacement between the target latch position and the current segment start position, and the initial total displacement between the target latch position and the actual latch position before execution, the initial total displacement is compensated in combination with the latch stroke resolution, and the motor drive duty cycle of the corresponding segment is generated by modulating between the minimum start duty cycle and the maximum allowable duty cycle according to the compensated displacement ratio. Based on the complete stroke displacement, the complete stroke calibration current load integral, the current sampling resolution, and the complete stroke time, a conversion factor from the current integral to the latch displacement is generated; Drive the bolt segment by segment within the maximum allowed number of segments, and stop within the resolution range of reaching the target position; Record the actual number of segments executed and the actual position of the locking tongue after stopping; Based on the relative relationship between the target latch position and the current segment start position, the movement direction of the corresponding actual segment is generated; Based on the start position, end position, motor current load integral, direction of motion, conversion factor from current integral to latch displacement, and latch stroke resolution of each actual execution segment, the execution residual is generated. The final positioning error is generated based on the target bolt position and the actual bolt position after stopping. Based on the final positioning error and each execution residual, the remote control execution result is output.

[0013] The present invention has the following beneficial effects: 1. By configuring functions such as bolt travel, door magnetic gap, motor current, lock body acceleration, counting and timing, security summary, and key storage at the lock end, and configuring interactive data acquisition, login recognition, challenge generation, and remote communication functions at the terminal, the smart lock remote control has the foundation of mechanical state perception, identity source recognition, time sequence recording, and key binding before execution. By using the fully open position of the bolt as the reference zero point and recording the complete travel current load integral, complete travel time, sensor resolution, and duty cycle boundary, subsequent displacement judgment, current load conversion, clock compensation, and motor control can all be carried out based on the installation calibration results, thereby reducing control misjudgments caused by differences in initial equipment installation, door gap, and motor load. 2. By collecting the interaction event times and categories of the currently logged-in user within a single remote control session, and normalizing the time differences between adjacent interactions according to the time span of the beginning and end of the session, remote sessions of different durations can form interaction rhythm data under a unified scale. By converting each interaction event into an action category label vector and forming a remote interaction event matrix according to the order of occurrence, and then statistically analyzing the proportion of action categories according to the action category direction, continuous clicks, repeated selections, accidental touch switching, and mixed operations on the terminal side can be compressed into a comparable category distribution. This provides structured input for subsequent candidate control action judgment and reduces the risk of malfunction caused by a single trigger command directly driving the lock body. 3. By performing mean and dispersion calculations on the normalized interaction intervals and combining them with normalized entropy analysis of the proportion of action categories, the remote control intent is simultaneously constrained by the stability of the interaction rhythm and the concentration of action categories. By performing zero-probability logarithmic correction on action categories that do not appear, the concentrated action features can still maintain computability even when categories are missing. By selecting the category with the largest proportion of action categories as candidate control actions and generating a unique decision quantity for the action with strict majority conditions, the candidate actions not only come from the most frequent interactions but also need to meet the intent dominance condition of exceeding the sum of other actions, thereby suppressing the interference of accidental touches, hesitant switching, and concurrent input of multiple actions on remote locking or unlocking actions. 4. By collecting the current bolt travel, door magnetic gap, motor current, and triaxial acceleration within a fixed sampling window, and uniformly normalizing the bolt position, door closure status, motor load status, and lock vibration status, the on-site state of the lock end can be expressed in the same physical state vector. By mapping the sensor resolution to the dimensionless resolution of each state component, subsequent interval judgment can absorb hardware quantization errors. This processing enables remote control to no longer rely solely on terminal commands or cloud authorization for entry action judgment, but instead incorporates whether the door is closed, whether the bolt is in a reasonable range, whether the motor has an abnormal load, and whether the lock body is disturbed into the pre-access state characterization. 5. By extracting the corresponding action pre-interval based on candidate control actions and expanding the interval boundary using a resolution mapping vector, the physical state determination of the lock end can simultaneously reflect the necessary mechanical conditions of the action and the minimum resolvable error of the sensor. By determining the inclusion of the bolt state, door closing state, motor load state, and lock body vibration state with the expanded interval, and synthesizing all determination results with necessary conditions, the subsequent access chain cannot be entered if any key physical state does not meet the corresponding action conditions. This process can avoid local risks such as locking when the door is not closed, unlocking when the bolt is not in the expected position, and continuing to drive when the lock body vibrates abnormally. 6. By recording the terminal's internal round-trip time and the lock's internal processing time based on the session handshake messages, and dividing the round-trip transmission process into two equal parts to obtain a one-way transmission time estimate, the session response time window can reflect the actual latency of the current communication link. By processing the verification message and its receipt to calculate the terminal processing compensation time, the time required for the terminal to generate tokens and encapsulate instructions is included in the timing determination. Furthermore, by combining the continuous judgment of the terminal's monotonic sequence number and the lock's challenge sequence number, the remote control instructions must simultaneously meet the requirements of time range, terminal sequence number progression, and lock challenge progression, thereby reducing the possibility of delayed messages, replayed messages, and out-of-order messages being mistakenly accepted. 7. By pre-loading the user identifier, candidate control actions, interaction intent set, action uniqueness determination, terminal device identifier, and terminal challenge value into the pre-control intent field group, the lock end can form a state commitment quantity related to the current physical state of the door lock and the feasibility of the action before the formal control command arrives. The terminal state binding token and the lock end state binding token are generated by the remote control command field group, the lock end state commitment quantity, the terminal monotonic sequence number, and the pairing key, respectively. At the same time, the token consistency and the field group consistency are verified, so that the user intent, device identity, lock end state, command content, and timing sequence number are bound to the same access basis. Then, the action access determination quantity is generated by combining the permission reachability quantity and each necessary determination quantity, which can reduce the false execution when the permission is valid but the on-site state or session state does not match. 8. After the action is admitted, the motor execution flag and target latch position are determined based on the candidate control actions, so that remote locking, remote unlocking and status query can enter the corresponding execution paths respectively; by generating the maximum allowable number of segments according to the target displacement and latch stroke resolution, and modulating the segment drive intensity according to the remaining displacement, initial total displacement and duty cycle boundary, the latch movement process can be controlled segment by segment according to the degree of position proximity; by using the full stroke to calibrate the current load integral to generate the conversion factor from current integral to latch displacement, and then comparing the actual displacement with the current integral to generate the execution residual, and combining the final positioning error to output the execution result, the latch jamming, free spin, rebound and half-closed state can be identified before the result is output at the remote end; 9. This solution is designed for remote interactive control scenarios of smart locks. Addressing the issues of terminal intent being susceptible to accidental touches and weak network latency, and the potential disconnect between the lock's mechanical state and remote commands, the solution first completes sensing, timing, challenge, key, and action permission calibration at both the lock and terminal. Then, it extracts the action category ratio, interaction rhythm, and intent concentration state from session interaction events, while simultaneously collecting bolt travel, door magnetic gap, motor current, and lock body acceleration to form the lock's physical state for pre-action judgment. Based on this, session handshake time, processing verification receipts, terminal monotonic sequence numbers, and lock challenge sequence numbers jointly define the command timing. The pre-control intent field group, remote control command field group, lock state commitment quantity, and pairing key are further... By binding user intent, lock state, and command messages, candidate actions are only executed when the interaction intent is clear, the on-site state is feasible, the session sequence is continuous, the field bindings are consistent, and the permissions are accessible. In this technical solution, interaction recognition no longer determines unlocking and closing in isolation, physical state no longer provides only post-event prompts, security authentication no longer only verifies the identity field, and execution feedback no longer simply relies on the position signal. Instead, through segmented duty cycle modulation, current integration to lock tongue displacement conversion, and linkage verification of execution residuals and final position errors, remote intent confirmation, on-site mechanical constraints, session security binding, and lock tongue execution status are incorporated into the same control chain, thereby improving the consistency between remote control results and the actual action state of the lock body. Attached Figure Description

[0014] Figure 1 This is a flowchart of the overall process for the remote control method of the smart lock of the present invention.

[0015] Figure 2 This is a schematic diagram of the system functional architecture of the lock and terminal of the present invention.

[0016] Figure 3 This is a schematic diagram of the remote interactive event matrix and candidate control action generation of the present invention.

[0017] Figure 4 This is a schematic diagram illustrating the construction of the physical state vector and resolution mapping vector of the door lock according to the present invention.

[0018] Figure 5 This is a schematic diagram illustrating the determination of the pre-action interval and the physical feasibility of the action in this invention. Detailed Implementation

[0019] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] Example, refer to Figure 1 A remote control method for smart locks based on intelligent interaction includes the following steps: Step 1: Perform initial configuration and deterministic calibration of the lock and terminal; like Figure 2 As shown, the lock end is equipped with functions for acquiring lock tongue travel, door magnetic gap, motor current, lock body acceleration, lock end counting, lock end timing, security summary, and key storage. Configure the terminal with interactive event collection function, user login identifier collection function, terminal counting function, terminal timing function, challenge value generation function and remote communication function; The fully open position of the bolt is used as the reference zero point of the bolt travel coordinate; the direction of the bolt from the fully open position to the fully closed position is used as the locking direction; Record the following parameters: fully open lock tongue position, fully closed lock tongue position, door closing gap, door opening reference gap, motor no-load state, average current during the entire stroke, load integral of the calibration current during the entire stroke, time during the entire stroke, lock tongue stroke sensor resolution, door magnetic sensor resolution, current sampling resolution, acceleration sampling resolution, terminal clock resolution, lock end clock resolution, minimum starting duty cycle, and maximum allowable duty cycle. During the installation and calibration process, the fully open position of the latch is first determined as the reference zero position. Then, the complete stroke displacement corresponding to the latch moving from the fully open position to the fully closed position is measured. Subsequently, the motor current is continuously collected during the execution of the complete stroke. The absolute value of the difference between the instantaneous motor current and the motor no-load current during this period is processed. Then, the current load amplitude during the entire complete stroke time is accumulated over time to obtain the complete stroke calibration current load integral. At the same time, the terminal clock resolution and the lock end clock resolution are read, and the larger of the two is determined as the equipment clock resolution compensation amount. During system registration, a mapping relationship is established between user identifier, terminal device identifier, lock device identifier, user permission set, action-required permission set, action pre-registration interval, and pairing key; the action category table includes at least remote locking action, remote unlocking action, and status query action; each action category corresponds to an action pre-registration interval; the action pre-registration interval includes the lower and upper bounds of each physical state component in the door lock physical state vector; Among them, the fully open position of the latch, the fully closed position of the latch, the door closing gap, the door opening reference gap, the motor no-load current, the average current of the complete stroke, the load integral of the calibration current of the complete stroke, the complete stroke time, the resolution of each sensor, the clock resolution, and the motor duty cycle boundary are all known quantities that are directly measured or read by hardware during installation; the user identifier, device identifier, permission set, action category table, action pre-interval, and pairing key are all known data that are written or mapped during system registration. Parameter attribute analysis: The resolution of the bolt travel sensor, door magnetic sensor, current sampling resolution, acceleration sampling resolution, and clock resolution are all derived from hardware specifications or installation calibration; the minimum starting duty cycle is preferably determined as the minimum duty cycle when the bolt just overcomes static friction and produces continuous displacement; the maximum allowable duty cycle is preferably determined as the upper limit duty cycle when the complete stroke can be executed normally without triggering overcurrent protection; a minimum starting duty cycle that is too small will cause the motor to not start stably, while a minimum starting duty cycle that is too large will cause increased initial mechanical shock; a maximum allowable duty cycle that is too small will cause excessive execution time, while a maximum allowable duty cycle that is too large will cause increased current load and increased impact on the lock body; By configuring functions such as bolt travel, door magnetic gap, motor current, lock body acceleration, counting and timing, security summary, and key storage at the lock end, and configuring interactive data acquisition, login recognition, challenge generation, and remote communication functions at the terminal, the smart lock remote control has the foundation of mechanical state perception, identity source recognition, timing recording, and key binding before execution. By using the fully open bolt position as the reference zero point and recording the complete travel current load integral, complete travel time, sensor resolution, and duty cycle boundary, subsequent displacement judgment, current load conversion, clock compensation, and motor control can all be carried out based on the installation calibration results, thereby reducing control misjudgments caused by differences in initial equipment installation, door gap, and motor load. Step 2: Collect session interaction events and construct a remote interaction event matrix; like Figure 3As shown, the system collects the time and category of each interaction event formed by the currently logged-in user in a single remote control session; based on the time difference between two adjacent interaction events and the time span between the first and last interaction events in the current session, it performs normalization processing on the adjacent interaction intervals; during the normalization process, the time difference between adjacent interaction events is used as the processed quantity, and the sum of the time span of the first and last interaction events in the current session and the terminal clock resolution is used as the normalization benchmark to obtain the dimensionless normalized interaction interval; Based on the registered action category table, each interaction event is converted into an action category tag vector. In the action category tag vector, the position corresponding to the action category to which the current interaction event belongs is set as a valid tag, and the positions corresponding to other action categories are set as invalid tags. All action category tag vectors are arranged in order of occurrence of the interaction events to form a remote interaction event matrix. Then, the remote interaction event matrix is ​​averaged according to the action category direction to obtain the proportion of each action category in the current session. Among them, the interaction event time is a known quantity directly measured by the terminal interaction event acquisition function; the interaction event category is a known quantity directly acquired by the terminal interaction event acquisition function and mapped through the action category table; the normalized interaction interval is an intermediate quantity calculated in this step; the action category label vector is an intermediate quantity generated in this step based on the interaction event category and the action category table; the remote interaction event matrix is ​​an intermediate quantity obtained in this step by combining all action category label vectors in chronological order; and the action category ratio is a target quantity obtained by statistically analyzing the remote interaction event matrix in this step. Parameter attribute analysis: This step requires that a single remote control session contain at least two interaction events, based on the requirement of at least one adjacent interaction interval to generate interaction rhythm features. If the number of interaction events is insufficient, stable interaction rhythm data cannot be formed, and subsequent interaction intent calculations should be stopped. The total number of action categories is determined by the registered action category table, which should at least cover remote locking actions, remote unlocking actions, and status query actions. If there are too few action categories, it will be impossible to distinguish different remote control intents. If there are too many action categories and they include categories without actual control meaning, the proportion of action categories will be dispersed, and the stability of candidate control action determination will be reduced. By collecting the interaction event times and categories of the currently logged-in user within a remote control session, and normalizing the time differences between adjacent interactions according to the time span of the beginning and end of the session, remote sessions of different durations can form interaction rhythm data under a unified scale. By converting each interaction event into an action category label vector and forming a remote interaction event matrix according to the order of occurrence, and then statistically analyzing the proportion of action categories according to the action category direction, continuous clicks, repeated selections, accidental touch switching, and mixed operations on the terminal side can be compressed into a comparable category distribution. This provides structured input for subsequent candidate control action judgment and reduces the risk of malfunction caused by a single trigger command directly driving the lock body. Step 3: Generate a concentrated set of interactive intents and determine candidate control actions; The mean of all normalized interaction intervals obtained in step two is calculated to obtain the mean of normalized interaction intervals; then the dispersion of each normalized interaction interval relative to the mean is calculated to obtain the interaction rhythm characteristics. Normalized entropy calculation is performed on the proportion of each action category to obtain the action set features; When the proportion of action categories is zero, zero-probability logarithmic correction is used to ensure that no undefined logarithmic operations are generated when no action category appears. The action concentration features are then combined with the interaction rhythm features to obtain the interaction intent concentration quantity; During the process of determining candidate control actions, the proportions of each action category are compared; the action category with the largest proportion is determined as the candidate control action; when multiple action categories have the same largest proportion, the action category with the highest code in the stable coding order in the registered action category table is selected as the candidate control action, but this selection is only used to generate stable candidate action codes and does not mean that the action has passed the admission test. Then, it is determined whether the proportion of the action category corresponding to the candidate control action exceeds half of all interaction events; if it does, a valid action unique decision value is generated; if it does not, an invalid action unique decision value is generated. Wherein, the normalized interaction interval is the intermediate quantity calculated in step two; the action category ratio is the target quantity calculated in step two; the mean of the normalized interaction interval is the intermediate quantity calculated in this step; the interaction rhythm feature is the intermediate quantity calculated in this step based on the dispersion of the normalized interaction interval; the action set feature is the intermediate quantity calculated in this step based on the normalized entropy of the action category ratio; the zero probability logarithmic correction term is the intermediate quantity generated in this step when performing logarithmic correction on action categories that do not appear; the interaction intent set quantity is the target quantity calculated in this step; the candidate control action is the target quantity selected in this step based on the largest action category ratio; and the action uniqueness determination quantity is the target quantity generated in this step based on the strict majority condition. Parameter attribute analysis: The strict majority boundary is preferably one where the number of candidate control action events exceeds the sum of the number of other action events, based on the mathematical definition of majority decision. A higher boundary will reduce the probability of a legitimate session entering the subsequent admission chain and increase the possibility of false rejection when the number of interactions is small. A lower boundary will allow candidate control actions to still form when multiple action intentions are mixed, reducing the action uniqueness constraint. The basis for zero probability logarithm correction is that the zero probability term in the information entropy calculation should not produce undefined logarithmic values, and its processing result should ensure that the absence of action categories does not contribute additionally to the features of the action set. By performing mean and dispersion calculations on the normalized interaction intervals and combining them with normalized entropy analysis of the proportion of action categories, the remote control intent is simultaneously constrained by the stability of the interaction rhythm and the concentration of action categories. By applying zero-probability logarithmic correction to action categories that do not appear, the concentrated action features can still maintain computability even when categories are missing. By selecting the category with the largest proportion of action categories as candidate control actions and generating a unique action decision quantity with strict majority conditions, the candidate actions not only come from the most frequent interactions but also need to meet the intent dominance condition of exceeding the sum of other actions, thereby suppressing the interference of accidental touches, hesitant switching, and concurrent input of multiple actions on remote locking or unlocking actions. Step 4: Generate the door lock physical state vector and resolution mapping vector; like Figure 4 As shown, the current latch travel, current door magnetic gap, and current motor current are collected; the three-axis acceleration of the lock body is collected within a fixed sampling window before the current session challenge is issued; the acceleration sample values ​​in the three axes within the fixed sampling window are squared respectively; the squared values ​​of the three directions at the same sampling point are summed; then the mean of the summation values ​​of all sampling points within the fixed sampling window is calculated, and the mean is square rooted to obtain the root mean square of the current lock body's three-axis acceleration. Based on the fully open position, fully closed position, and travel resolution of the latch, the current latch travel is normalized, and the result is limited to the range of zero to one through saturation processing; based on the door closing gap, door opening reference gap, and door magnetic resolution, the current door magnetic gap is normalized, and the result is limited to the range of zero to one through saturation processing. Based on the motor's no-load current, the average current over the entire stroke, and the current sampling resolution, the current of the current motor is normalized, and the result is limited to the range of zero to one through saturation processing; based on the standard gravitational acceleration and the root mean square of the current lock body's triaxial acceleration, the vibration state of the lock body is normalized. The physical state vector of the door lock is composed of the above-mentioned bolt state component, door closing state component, motor load state component, and lock body vibration state component; Furthermore, the resolution of the latch travel sensor is mapped to the dimensionless resolution of the latch state component; the resolution of the door magnetic sensor is mapped to the dimensionless resolution of the door closing state component; the current sampling resolution is mapped to the dimensionless resolution of the motor load state component; and the acceleration sampling resolution is mapped to the dimensionless resolution of the lock body vibration state component through the resolution propagation relationship of the lock body vibration state normalization function; the resolution mapping vector is composed of the above dimensionless resolutions. Among them, the current latch travel is a known quantity directly acquired by the latch travel acquisition function; the current door magnetic gap is a known quantity directly acquired by the door magnetic gap acquisition function; the current motor current is a known quantity directly acquired by the motor current acquisition function; the lock body triaxial acceleration sampling value is a known quantity directly acquired by the lock body acceleration acquisition function within a fixed sampling window; the current lock body triaxial acceleration root mean square is an intermediate quantity calculated in this step; the door lock physical state vector is the target quantity calculated in this step; the resolution mapping vector is the target quantity calculated in this step; the latch fully open position, latch fully closed position, door closing gap, door opening reference gap, motor no-load current, average current of the complete travel, and the resolution of each sensor are all known quantities recorded or calibrated in step one. Parameter attribute analysis: The preferred length of the fixed sampling window should be no less than several times the triaxial acceleration sampling period and should be located before the current session challenge is issued; a fixed sampling window that is too short will cause instantaneous jitter to have an excessive impact on the root mean square result; a fixed sampling window that is too long will cause the sampling result to be mixed with earlier states, reducing the accuracy of the representation of the current lock body state; the boundary of saturation processing is preferably zero to one, based on the fact that each physical state component is mapped to the same dimensionless scale; a boundary that is too small will truncate normal state changes; a boundary that is too large will destroy the uniform scale of each component in the door lock physical state vector; By collecting data on the current latch travel, door magnetic gap, motor current, and triaxial acceleration within a fixed sampling window, and normalizing the latch position, door closure status, motor load status, and lock vibration status, the on-site state of the lock end can be expressed in the same physical state vector. By mapping the sensor resolution to the dimensionless resolution of each state component, subsequent interval judgments can absorb hardware quantization errors. This processing enables remote control to no longer rely solely on terminal commands or cloud authorization for entry actions, but instead incorporates whether the door is closed, whether the latch is within a reasonable range, whether the motor has an abnormal load, and whether the lock body is experiencing disturbances into the pre-access state characterization. Step 5: Generate the physical feasibility of the action based on the action pre-interval; like Figure 5As shown, based on the candidate control actions obtained in step three, the corresponding action pre-interval is extracted from the action category table; the action pre-interval includes the lower and upper bounds of the interval corresponding to each physical state component in the door lock physical state vector; then, based on the resolution mapping vector obtained in step four, the lower bounds of each interval of the action pre-interval are extended downwards, and the upper bounds of each interval are extended upwards; the extended interval is used as the inclusion determination interval of the current lock end physical state. The bolt state component, door closing state component, motor load state component, and lock vibration state component in the door lock physical state vector are each included in the corresponding extended action pre-interval. If a physical state component falls within the corresponding extended interval, the inclusion judgment result for that component is valid; if it does not fall within the corresponding extended interval, the inclusion judgment result for that component is invalid. Subsequently, the inclusion judgment results of all physical state components are synthesized under necessary conditions. Only when all inclusion judgment results are valid is the physical feasible quantity of the action valid; if any physical state component does not satisfy the corresponding interval, the physical feasible quantity of the action is invalid. In this process, the candidate control action is the intermediate quantity calculated in step three; the action pre-interval is the known data determined in step one; the lower and upper bounds of the interval are the known boundary data in the action pre-interval; the door lock physical state vector is the intermediate quantity calculated in step four; the resolution mapping vector is the intermediate quantity calculated in step four; the expanded action pre-interval is the intermediate data generated in this step based on the action pre-interval and the resolution mapping vector; the inclusion determination result of each physical state component is the intermediate quantity calculated in this step; and the action physical feasibility quantity is the target quantity calculated in this step. Parameter attribute analysis: The lower and upper bounds of the action pre-interval are preferably determined by the necessary mechanical conditions and normalized boundaries of the action; the remote locking action requires at least that the door body closed state component is in the closed interval; the remote unlocking action requires at least that the bolt locked state component is in the locked interval; no motion pre-restriction is imposed on the state query action; a narrow interval boundary may lead to unnecessary rejection due to sensor quantization error; a wide interval boundary may allow actions that do not meet the mechanical pre-restriction conditions to enter the subsequent admission chain; the resolution mapping vector is used to expand the interval boundary, and its basis is the minimum resolvable value of the sensor; By extracting the corresponding action pre-interval based on candidate control actions and expanding the interval boundary using a resolution mapping vector, the physical state determination of the lock end can simultaneously reflect the necessary mechanical conditions of the action and the minimum resolvable error of the sensor. By determining the inclusion of the bolt state, door closing state, motor load state, and lock body vibration state with the expanded interval, and synthesizing all determination results with necessary conditions, the subsequent access chain cannot be entered if any key physical state does not meet the corresponding action condition. This process can avoid local risks such as locking when the door is not closed, unlocking when the bolt is not in the expected position, and continuing to drive when the lock body vibrates abnormally. Step 6: Generate session temporal continuity determination variables; Based on the set of handshake messages in the current session, record the internal time difference between the terminal sending a handshake message and receiving a handshake return message; record the internal processing time of the lock terminal from receiving a handshake message to sending a handshake return message; subtract the lock terminal's internal processing time from the internal time difference, and then divide the process into two equal parts according to the round-trip transmission process to obtain the one-way transmission time estimate corresponding to each handshake; perform non-negation processing on each one-way transmission time estimate; select the largest non-negative one-way transmission time estimate in the current session; and then superimpose the device clock resolution compensation amount to generate a session response time window. Based on the processing verification messages and their receipts within the current session, record the local time difference between the lock end sending the processing verification message and receiving the processing verification receipt; subtract the round-trip transmission compensation amount determined by the session response time window from this local time difference, and further subtract the device clock resolution compensation amount; perform non-negation processing on the subtraction result; select the largest non-negative processing compensation amount within the current session to generate the terminal processing compensation duration; wherein, the terminal processing load corresponding to the processing verification receipt is not less than the terminal state binding token calculation and remote control command message encapsulation load; The time difference between the time the lock-end challenge is issued and the time the lock-end command is received is taken as the current session command response time. It is determined whether the command response time is within a non-negative range and whether it does not exceed the response upper limit formed by the session response time window, the device clock resolution compensation amount, and the terminal processing compensation duration. At the same time, it is determined whether the terminal monotonic sequence number carried in the remote control command message is equal to the next order of the terminal monotonic sequence number of the previous confirmed interaction command. It is also determined whether the current lock-end challenge sequence number is equal to the next order of the previous confirmed challenge sequence number. The above timing range determination, terminal monotonic sequence number continuity determination, and lock-end challenge sequence number continuity determination are combined with necessary conditions to generate the session timing continuity determination quantity. The handshake message set is a known set formed by the remote communication link within the current session; the time when the terminal sends the handshake message and the time when the terminal receives the handshake return message are known quantities directly measured by the terminal timing function; the time when the lock receives the handshake message and the time when the lock sends the handshake return message are known quantities directly measured by the lock timing function; the one-way transmission time estimate is an intermediate quantity calculated in this step; the session response time window is an intermediate quantity calculated in this step; the processing verification message and its receipt set is a known set formed by the remote communication link within the current session; the time when the lock sends the processing verification message and the time when the lock receives the processing verification receipt are known quantities directly measured by the lock timing function; the terminal processing compensation time is an intermediate quantity calculated in this step; the lock challenge issuance time, lock instruction reception time, terminal monotonic sequence number, the previous confirmed terminal monotonic sequence number, the current lock challenge sequence number, and the previous confirmed challenge sequence number are all known quantities directly obtained by the lock or terminal from the current session and historical confirmation records; the session timing continuity determination quantity is the target quantity calculated in this step. Parameter attribute analysis: The session response time window is generated by the actual handshake data of the current session and the device clock resolution; a larger session response time window will broaden the time acceptance range of remote control commands; a smaller session response time window will increase the probability of false rejection under legitimate link latency; the terminal processing load corresponding to the processing verification receipt should preferably be set to no less than the terminal state binding token calculation and remote control command message encapsulation load, based on the fact that the terminal processing compensation time should cover the subsequent actual command processing time; a smaller processing load will lead to insufficient terminal processing compensation; a processing load much larger than the actual command load will excessively broaden the timing window; By recording the terminal's internal round-trip time and the lock's internal processing time based on the session handshake messages, and dividing the round-trip transmission process into two equal parts to obtain a one-way transmission time estimate, the session response time window can reflect the actual latency of the current communication link. By processing the verification message and its acknowledgment to calculate the terminal processing compensation time, the time required for the terminal to generate tokens and encapsulate instructions is included in the timing determination. Furthermore, by combining the continuous judgment of the terminal's monotonic sequence number and the lock's challenge sequence number, the remote control instructions must simultaneously meet the requirements of time range, terminal sequence number progression, and lock challenge progression, thereby reducing the possibility of delayed messages, replayed messages, and out-of-order messages being mistakenly accepted. Step 7: Generate state binding consistency criteria and action admission criteria; The interactive terminal sends a pre-control intent field group; this pre-control intent field group includes user identifier, candidate control action, interaction intent concentration quantity, action uniqueness determination quantity, terminal device identifier, and terminal challenge value; after receiving the pre-control intent field group, the lock terminal performs deterministic digest authentication processing based on the pre-control intent field group, the door lock physical state vector obtained in step four, the action physical feasibility quantity obtained in step five, the lock terminal device identifier, the lock terminal challenge value, the lock terminal challenge sequence number, the terminal processing compensation time and pairing key obtained in step six, and generates the lock terminal state commitment quantity; The interactive terminal constructs a remote control instruction field group; this remote control instruction field group includes a user identifier, candidate control actions, interaction intent concentration, action uniqueness determination quantity, terminal device identifier, and terminal challenge value corresponding to the pre-control intent field group; based on the remote control instruction field group, lock state commitment quantity, terminal monotonic sequence number, and pairing key, the interactive terminal performs deterministic digest authentication processing to generate a terminal state binding token; subsequently, the interactive terminal sends the remote control instruction field group, terminal monotonic sequence number, and terminal state binding token; The lock terminal performs the same deterministic digest authentication process based on the received remote control instruction field group, the locally stored lock terminal state commitment quantity, the terminal monotonic sequence number carried in the remote control instruction message, and the pairing key; it then recalculates the lock terminal state binding token; it performs a consistency check between the terminal state binding token and the lock terminal state binding token; and it performs a consistency check between the pre-control intent field group and the remote control instruction field group. If both the token and the field group are consistent, a valid state binding consistency determination quantity is generated; otherwise, an invalid state binding consistency determination quantity is generated. Based on the current logged-in user's permission set and the permission set required for candidate control actions, it is determined whether the current logged-in user's permission set covers the permission set required for candidate control actions, generating a permission reachability quantity. Subsequently, the permission reachability quantity, action uniqueness determination quantity, session temporal continuity determination quantity, action physical feasibility quantity, state binding consistency determination quantity, and interaction intent set quantity are synthesized into a necessary condition to generate an action admission determination quantity. When all necessary conditions are met and the interaction intent set quantity forms a valid input, the action admission determination quantity is used to proceed to subsequent segment execution; when any necessary condition is not met, the action admission determination quantity is used to refuse to execute the candidate control action. Among them, the pre-control intent field group is the set of pre-control data for the current session sent by the interactive terminal to the lock terminal; the user identifier is a known quantity obtained by the terminal login identifier collection function and confirmed by registration mapping; the candidate control action is the intermediate quantity calculated in step three; the interaction intent set quantity is the intermediate quantity calculated in step three; the action unique determination quantity is the intermediate quantity calculated in step three; the terminal device identifier is the known quantity obtained by registration in step one; the terminal challenge value is the known quantity generated by the terminal challenge value generation function in the current session; the door lock physical state vector is the intermediate quantity calculated in step four; the action physical feasibility quantity is the intermediate quantity calculated in step five; the lock terminal device identifier is the known quantity obtained by registration in step one; the lock terminal challenge value is the known quantity generated by the lock terminal challenge value generation function in the current session; and the lock terminal challenge sequence number. The known quantities used in step six; the terminal processing compensation time is the intermediate quantity calculated in step six; the pairing key is the known quantity obtained from registration in step one; the lock state commitment quantity is the intermediate quantity calculated in this step; the remote control instruction field group is the data set carried by the terminal when issuing remote control instructions; the terminal monotonic sequence number is the known quantity generated by the terminal counting function and sent with the remote control instruction; the terminal state binding token is the intermediate quantity calculated by the interactive terminal; the lock state binding token is the intermediate quantity obtained from lock recalculation; the state binding consistency determination quantity is the intermediate quantity calculated in this step; the user permission set and the permission set required for the action are both known sets obtained from registration or authorization update in step one; the permission reachability quantity is the intermediate quantity calculated in this step; the action admission determination quantity is the target quantity calculated in this step. Parameter attribute analysis: The pairing key is secure data written during the registration phase; the deterministic digest authentication process should have the attributes of same input and same output and key participation; if the granularity of the pairing key management is too coarse, it will reduce the isolation between devices; if the pairing key update frequency is too high, it will increase the synchronization cost between the terminal and the lock; the admission boundary of the action admission judgment quantity is preferably when all necessary conditions are met and the interaction intent concentration forms a valid input, based on the composite structure of multiple necessary conditions; if the admission boundary is too low, the necessary conditions will lose their constraints; if the admission boundary is too high, it will additionally exclude legitimate sessions that have already met the necessary conditions. By pre-loading the user identifier, candidate control actions, interaction intent set, action uniqueness determination, terminal device identifier, and terminal challenge value into the pre-control intent field group, the lock end can form a state commitment quantity related to the current physical state of the door lock and the feasibility of the action before the formal control command arrives. The terminal state binding token and the lock end state binding token are generated by the remote control command field group, the lock end state commitment quantity, the terminal monotonic sequence number, and the pairing key, respectively, and the token consistency and the field group consistency are verified at the same time, so that the user intent, device identity, lock end state, command content, and timing sequence number are bound to the same access basis. Then, the action access determination quantity is generated by combining the permission reachability quantity and each necessary determination quantity, which can reduce the erroneous execution when the permission is valid but the on-site state or session state does not match. Step 8: Execute segmented locking bolt control and output the remote control execution result; When the action access determination quantity meets the access conditions, the motor execution flag and target latch position are determined according to the candidate control action; when the candidate control action is a remote locking action or a remote unlocking action, the motor execution flag is valid; when the candidate control action is a status query action, the motor execution flag is invalid; when the candidate control action is a remote locking action, the target latch position is determined to be the fully closed position; when the candidate control action is a remote unlocking action, the target latch position is determined to be the fully open position; when the candidate control action is a status query action, the target latch position is determined to be the actual position of the latch before execution. Based on the distance between the target latch position and the actual latch position before execution, and the latch travel resolution, the maximum allowable number of segments is generated. Specifically, the distance between the target latch position and the actual latch position before execution is segmented and converted according to the latch travel resolution, and the conversion result is rounded up. If the motor execution flag is invalid, the maximum allowable number of segments is determined to be no-motion segments. When generating the motor drive duty cycle for each segment, the remaining displacement between the target latch position and the starting position of the current segment is first calculated; then the initial total displacement between the target latch position and the actual latch position before execution is calculated; subsequently, the initial total displacement is compensated using the latch travel resolution to prevent the proportional calculation from failing due to the initial total displacement being zero; then, according to the compensated displacement ratio, the motor drive duty cycle of the corresponding segment is modulated between the minimum starting duty cycle and the maximum allowable duty cycle to generate the corresponding segment. Based on the full stroke displacement, the full stroke calibration current load integral, the current sampling resolution, and the full stroke time, a conversion factor from the current integral to the latch displacement is generated. Specifically, the full stroke displacement is used as the displacement reference; the full stroke calibration current load integral and the compensation amount formed by the current sampling resolution and the full stroke time are combined to form the current integral reference; and the ratio of the displacement reference to the current integral reference is used as the conversion factor from the current integral to the latch displacement. Drive the bolt segment by segment within the maximum allowable number of segments, and stop within the bolt travel resolution range to reach the target bolt position; record the actual number of segments executed and the actual position of the bolt after stopping; for each actual segment executed, generate the movement direction based on the relative relationship between the target bolt position and the current segment start position; if the target bolt position is in the locking direction of the current segment start position, the movement direction is the locking direction; if the target bolt position is in the unlocking direction of the current segment start position, the movement direction is the unlocking direction; if the two are the same, the movement direction is the no-displacement direction. For each actual execution segment, calculate the actual displacement between the end position and the beginning position of the segment; collect the instantaneous motor current during the execution of the segment, and accumulate the absolute value of the difference between the instantaneous motor current and the motor no-load current over time to obtain the motor current load integral of the segment; multiply the motor current load integral with the conversion factor from the current integral to the latch displacement, and combine it with the direction of motion to obtain the current integral estimated displacement of the segment; calculate the difference between the actual displacement and the current integral estimated displacement, and normalize it with the latch stroke resolution to generate the execution residual of the segment; Based on the difference between the actual position of the latch bolt after stopping and the target position of the latch bolt, the final positioning error is generated; when the motor execution flag is invalid, the corresponding remote control execution result is output as a status query; when the motor execution flag is valid, the final positioning error is compared with the latch bolt travel resolution, and the execution residuals of all actual execution segments are compared with the normalized boundary corresponding to a latch bolt travel resolution; if the final positioning error does not exceed the latch bolt travel resolution, and the execution residuals of all actual execution segments do not exceed the corresponding boundary, the execution is output as successful; otherwise, the execution is output as abnormal. Among them, the action admission judgment quantity is the target quantity calculated in step seven; the candidate control action is the intermediate quantity calculated in step three; the motor execution flag is the intermediate quantity generated in this step based on the candidate control action; the target latch position is the intermediate quantity generated in this step based on the candidate control action; the actual latch position before execution is the known quantity directly measured by the latch stroke acquisition function before execution; the latch stroke resolution is the known quantity recorded in step one; the maximum allowable number of segments is the intermediate quantity calculated in this step; the minimum starting duty cycle and the maximum allowable duty cycle are both known quantities calibrated in step one; the motor drive duty cycle is the control quantity calculated in this step; the complete stroke displacement, complete... The integral of the calibration current load, the total stroke time, and the current sampling resolution are all known quantities obtained in step one; the conversion factor from current integral to latch displacement is an intermediate quantity calculated in this step; the actual number of segments executed is an intermediate quantity obtained by directly counting the lock end during segment execution; the actual position of the latch after stopping is a known quantity obtained by directly measuring the latch stroke acquisition function; the direction of movement is an intermediate quantity calculated in this step based on the relative relationship between the target latch position and the current segment start position; the execution residual is an intermediate quantity calculated in this step; the final positioning error is an intermediate quantity calculated in this step; and the remote control execution result is the target quantity output in this step. Parameter attribute analysis: The maximum allowable number of segments is determined by the target displacement and the lock tongue travel resolution; too few segments will increase the single-segment control displacement and reduce the execution residual positioning accuracy; too many segments will lengthen the execution link and increase sampling and control overhead; the minimum start duty cycle and the maximum allowable duty cycle are both calibrated in step one; the execution residual judgment boundary is preferably one lock tongue travel resolution, based on the minimum resolvable position change of the lock tongue travel sensor; if this boundary is too large, mechanical jamming or idling may be detected with delay; if this boundary is too small, normal quantization error may be misjudged as execution abnormality; the final positioning error boundary is preferably one lock tongue travel resolution, based on the fact that the stopping position cannot reliably distinguish errors smaller than this resolution; if this boundary is too large, the positioning judgment accuracy decreases; if this boundary is too small, the probability of false rejection under normal positioning conditions increases; After the action is admitted, the motor execution flag and target latch position are determined based on the candidate control actions, so that remote locking, remote unlocking and status query can enter the corresponding execution paths respectively; by generating the maximum allowable number of segments according to the target displacement and latch stroke resolution, and modulating the segment drive intensity according to the remaining displacement, initial total displacement and duty cycle boundary, the latch movement process can be controlled segment by segment according to the degree of position proximity; by using the full stroke to calibrate the current load integral to generate the conversion factor from current integral to latch displacement, and then comparing the actual displacement with the current integral to generate the execution residual, and combining the final positioning error to output the execution result, the latch jamming, free spin, rebound and half-closed state can be identified before the result is output at the remote end; It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0021] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the technical principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A remote control method for a smart lock based on intelligent interaction, characterized in that, include: Configure sensing, security summary and key storage functions on the lock end, and configure interaction, login, challenge and communication links on the terminal to complete the calibration of lock tongue coordinates, action category, permissions, device identification and pairing key; Collect the session interaction time and action categories of logged-in users, construct a remote interaction event matrix, and extract the proportion of action categories; Based on the interval between adjacent interactions and the proportion of action categories, a concentration of interaction intent is generated to determine candidate control actions and unique action determination quantities. Collect bolt travel, door magnetic gap, motor current and lock body acceleration to generate door lock physical state vector and resolution mapping vector; Based on the candidate action's preceding interval and the door lock's physical state vector, generate the action's physical feasibility quantity; Based on the session handshake message, processing verification receipt, terminal monotonic sequence number and lock-end challenge sequence number, generate the session timing continuity determination quantity; Based on the pre-control intent field group, remote control instruction field group, lock end status commitment quantity and pairing key, generate a status binding consistency judgment quantity, and combine it with user permissions to generate an action access judgment quantity; After the access is approved, the target latch position is determined. The maximum allowable number of segments and the motor drive duty cycle are generated by combining the latch stroke resolution. The latch is driven segment by segment. The conversion factor from current integral to latch displacement is called to calculate the execution residual, generate the final positioning error, and output the remote control execution result.

2. The method for remote control of a smart lock based on intelligent interaction according to claim 1, characterized in that, The steps for completing the calibration of the latch coordinates, action category, permissions, device identifier, and pairing key specifically include: The lock end is equipped with functions for acquiring bolt travel, door magnetic gap, motor current, lock body acceleration, lock end counting, lock end timing, security summary and key storage; Configure the terminal with interactive event collection, user login identifier collection, terminal counting, terminal timing, challenge value generation and remote communication functions; With the fully open position of the latch as the reference zero point and the locking direction of the latch as the positive direction, record the latch position, door clearance, motor no-load state, average current of the complete stroke, load integral of the calibration current of the complete stroke, complete stroke time, sensor resolution, clock resolution and motor duty cycle boundary. The registration includes an action category table, action prerequisite range, user permission set, action required permission set, device identifier, and pairing key.

3. The method for remote control of a smart lock based on intelligent interaction according to claim 1, characterized in that, The steps of constructing the remote interactive event matrix and extracting the proportion of action categories specifically include: Collect the time and type of interaction events generated by the currently logged-in user within a single remote control session; Based on the time difference between the collection of adjacent interaction events and the time span of the first and last interaction events of this session, the adjacent interaction interval is normalized to obtain the normalized interaction interval. Based on the registered action category table, each interaction event is converted into an action category tag vector; The action category marker vectors are combined according to the order in which the interaction events occur to form a remote interaction event matrix; Perform mean statistics on the remote interaction event matrix according to action category to obtain the proportion of each action category in the current session.

4. The method for remote control of a smart lock based on intelligent interaction according to claim 1, characterized in that, The steps of generating a set of interactive intent quantities and determining candidate control actions and unique action determination quantities specifically include: The mean and dispersion of the normalized interaction interval are calculated to obtain the interaction rhythm characteristics. Normalized entropy is calculated for the proportion of action categories, and zero-probability logarithmic correction is applied to action categories that do not appear to obtain the features of the action set. By combining interaction rhythm features with action concentration features, an interaction intent concentration quantity is generated; Select the action category with the largest proportion as the candidate control action; Based on whether the proportion of action categories corresponding to candidate control actions meets the strict majority condition, a unique decision quantifier for the action is generated.

5. The method for remote control of a smart lock based on intelligent interaction according to claim 1, characterized in that, The steps for generating the physical state vector and resolution mapping vector of the door lock specifically include: Collect the current latch travel, current door magnetic gap, and current motor current; Collect the lock body's triaxial acceleration within a fixed sampling window before the current session challenge is issued, and calculate the root mean square of the current lock body's triaxial acceleration; Based on the fully open position, fully closed position, and bolt travel resolution, the current bolt travel is normalized and saturated. Based on the door closing gap, the door opening reference gap, and the door magnetic resolution, the current door magnetic gap is normalized and saturated. Based on the motor no-load current, the average current during the entire stroke, and the current sampling resolution, the current of the current motor is normalized and saturated. Based on the standard gravitational acceleration, the root mean square of the current lock body's three-axis acceleration, and the acceleration sampling resolution, the vibration state of the lock body is normalized and processed by resolution propagation to generate the door lock's physical state vector and resolution mapping vector.

6. The method for remote control of a smart lock based on intelligent interaction according to claim 1, characterized in that, The step of generating the physical feasibility of an action based on the candidate action pre-interval and the door lock physical state vector specifically includes: Based on the candidate control actions, extract the corresponding action pre-interval from the action category table; The action pre-interval includes the lower and upper bounds of each physical state component. Based on the resolution mapping vector, the lower and upper bounds of the action pre-interval are expanded. Each physical state component in the door lock's physical state vector is included in the corresponding expanded interval; The inclusion determination results of each physical state component are synthesized under necessary conditions to generate the physical feasible quantity of the action.

7. The method for remote control of a smart lock based on intelligent interaction according to claim 1, characterized in that, The step of generating the session timing continuity determination quantity specifically includes: Based on the set of handshake messages in the current session, record the internal time difference between when the terminal sends a handshake message and when it receives a handshake return message. Record the internal processing time of the lock from receiving the handshake message to sending the handshake return message; Based on the internal time difference of the terminal and the internal processing time of the lock, extract the one-way transmission time estimate; Perform nonnegation processing on each one-way transmission time estimate, select the maximum nonnegative one-way transmission time estimate in the current session, and superimpose the device clock resolution to generate a session response time window; Based on the processing verification messages and their receipts in the current session, record the local time difference between the lock end sending the processing verification message and receiving the processing verification receipt. Subtract the round-trip transmission compensation amount determined by the session response time window and the device clock resolution from the local time difference of the lock end; The deduction result is non-negative, and the maximum non-negative processing compensation amount in the current session is selected to generate the terminal processing compensation duration. Among them, the terminal processing load for processing the verification receipt is no less than the terminal status binding token calculation and remote control instruction message encapsulation load. The session timing continuity determination quantity is generated by jointly determining the lock challenge issuance time, lock command reception time, session response time window, terminal processing compensation time, terminal monotonic sequence number, and lock challenge sequence number.

8. The method for remote control of a smart lock based on intelligent interaction according to claim 1, characterized in that, The step of generating a consistent state binding determination quantity and generating an action access determination quantity in conjunction with user permissions specifically includes: The interactive terminal sends a pre-control intent field group containing user identifier, candidate control action, interaction intent concentration, action uniqueness determination quantity, terminal device identifier, and terminal challenge value; The lock terminal generates a lock terminal state commitment quantity based on the pre-control intent field group, door lock physical state vector, action physical feasibility quantity, lock terminal device identifier, lock terminal challenge value, lock terminal challenge sequence number, terminal processing compensation time and pairing key; The interactive terminal generates a terminal status binding token based on the remote control instruction field group, the lock state commitment quantity, the terminal monotonic sequence number, and the pairing key; The interactive terminal sends a remote control instruction field group containing fields with the same name, the terminal monotonic sequence number, and the terminal status binding token; The lock terminal recalculates the lock terminal status binding token based on the remote control instruction field group, the lock terminal status commitment quantity, the terminal monotonic sequence number, and the pairing key; Perform consistency checks on the terminal state binding token, the lock state binding token, the pre-control intent field group, and the remote control instruction field group to generate a state binding consistency check quantity. Generate the permissible permissions based on the current logged-in user's permission set and the permission set required for candidate control actions; The necessary conditions are combined to generate the action admission criteria: accessibility criteria, action uniqueness criteria, session sequence continuity criteria, action physical feasibility criteria, state binding consistency criteria, and interaction intent concentration criteria.

9. The method for remote control of a smart lock based on intelligent interaction according to claim 1, characterized in that, The steps of driving the locking tongue segment by segment, calculating the execution residual, generating the final positioning error, and outputting the remote control execution result specifically include: When the action admission judgment quantity meets the admission conditions, the motor execution flag and target latch position are determined according to the candidate control action; The maximum number of segments is generated based on the target latch position, the actual latch position before execution, and the latch travel resolution. Based on the remaining displacement between the target latch position and the current segment start position, and the initial total displacement between the target latch position and the actual latch position before execution, the initial total displacement is compensated in combination with the latch stroke resolution, and the motor drive duty cycle of the corresponding segment is generated by modulating between the minimum start duty cycle and the maximum allowable duty cycle according to the compensated displacement ratio. Based on the complete stroke displacement, the complete stroke calibration current load integral, the current sampling resolution, and the complete stroke time, a conversion factor from the current integral to the latch displacement is generated; Drive the bolt segment by segment within the maximum allowed number of segments, and stop within the resolution range of reaching the target position; Record the actual number of segments executed and the actual position of the locking tongue after stopping; Based on the relative relationship between the target latch position and the current segment start position, the movement direction of the corresponding actual segment is generated; Based on the start position, end position, motor current load integral, direction of motion, conversion factor from current integral to latch displacement, and latch stroke resolution of each actual execution segment, the execution residual is generated. The final positioning error is generated based on the target bolt position and the actual bolt position after stopping. Based on the final positioning error and each execution residual, the remote control execution result is output.

Images