Intelligent door control method and system based on artificial intelligence

Abstract

The application provides an artificial intelligence-based multi-scene access control electric door linkage control method and system, comprising: acquiring the position information of each electric door entrance in the building through an electric door entrance coordinate acquisition unit, combining the multi-scene access control electric door linkage layout, extracting the actual walking distance between two electric door entrances according to the spatial distribution between the electric door entrances, and obtaining an electric door entrance walking distance table; extracting the identification time of the same voucher at different electric door entrances from a voucher time sequence recording unit, comparing the time interval between two adjacent recognitions with the minimum passing time of the corresponding electric door entrance pair, and marking as a preliminary potential return suspect when the time interval is less than the minimum passing time; analyzing the results of the behavior legality label, and adding a preset buffer time to the minimum passing time of the electric door entrance pair for the situation of recognized legal door changing behavior to obtain an extended time window.

Description

Technical Field

[0001] This invention relates to the field of information technology, and in particular to a multi-scenario access control method and system for electric doors based on artificial intelligence. Background Technology

[0002] In the field of AI-based multi-scenario access control and electric door linkage control, access control systems manage personnel entry and exit through credential recognition. Their core objective is to ensure building security while maintaining an efficient passage experience. This area is crucial because large building complexes, office parks, hospitals, schools, and other locations experience frequent personnel flow. Credential sharing or tailgating can directly threaten the security and order of these areas. Therefore, access control systems must possess both strong anti-cheating capabilities and be user-friendly for normal passage. Currently, most access control systems use fixed anti-followback time thresholds to determine if credential sharing has occurred. That is, if the same credential appears at different access points within a short period, it is considered a follow-back and passage is denied. This method works well when entrances are far apart; however, when multiple entrances are adjacent and the walking distance is short, the problems with fixed thresholds become apparent.

[0003] Authorized personnel, due to various practical needs, such as finding long queues at a certain entrance or the inconvenience of carrying items, often temporarily change their passage routes, turning from one entrance to another that is very close by. This normal door-switching behavior can easily fall within a fixed threshold in terms of time, causing the system to mistakenly identify legitimate passage as a sneakback attempt, thus triggering a rejection response at the second entrance. The root cause of this misjudgment is that there is no effective correlation between the time window on which the anti-sneakback judgment relies and the spatial distance in the actual passage scenario. The physical walking distance between entrances varies, and the time required to complete a legitimate movement from one door to another varies significantly. If the time threshold is set too short, normal door-switching between adjacent entrances will be frequently blocked, causing passage interruptions, and users will have to repeatedly swipe their cards or seek manual intervention; if the threshold is set too wide, it will be difficult to effectively intercept genuine credential sharing behavior, such as one person holding a card and others following quickly through multiple adjacent entrances.

[0004] Therefore, the system is constantly caught in a dilemma between preventing credential sharing and ensuring authorized personnel can freely change doors. How to accurately distinguish between legitimate door-changing behavior and malicious attempts to sneak back, based on the actual walking distance between different entrances and the actual movement speed of personnel, has become a critical issue that urgently needs to be addressed in the current multi-scenario access control system for electric doors. Summary of the Invention

[0005] This invention provides a multi-scenario access control method for electric doors based on artificial intelligence, including: The location information of each electric door entrance in the building is obtained by the electric door entrance coordinate acquisition unit. Combined with the multi-scenario access control electric door linkage layout, the actual walking distance between two electric door entrances is extracted according to the spatial distribution between the electric door entrances, and an electric door entrance walking distance table is generated. Based on the walking distance table of the electric gate entrances and the preset normal walking speed of people, a walking time model is constructed. The walking time model outputs the estimated passage time between each pair of electric gate entrances, and the minimum passage time between each pair of electric gate entrances is determined based on the estimated passage time. Extract the identification time of the same voucher at different electric gate entrances from the voucher timing record unit, compare the time interval between two adjacent identifications with the minimum passage time of the corresponding electric gate entrance pair, and mark it as a preliminary suspected sneakback when the time interval is less than the minimum passage time. For credentials marked as suspected of initial backtracking, obtain the authorized personnel's historical access path data, and use a path matching algorithm to evaluate the consistency between the current electric gate entrance switching behavior and the historical path to obtain a behavior legality label; Based on the legality label of the behavior, for cases where the door-changing behavior is determined to be legal, a preset buffer time is added to the minimum passage time of the electric door entrance pair to generate an extended time window; The current identification interval is evaluated based on the extended time window. For identification events that exceed the extended time window, a rejection response is triggered and recorded as a passage interruption event. For identification events that are within the extended time window, passage is allowed and the current passage path is added to the historical path data.

[0006] Furthermore, the location information of each electric door entrance within the building is obtained through the electric door entrance coordinate acquisition unit. Combined with the multi-scenario access control electric door linkage layout, the actual walking distance between two electric door entrances is extracted based on the spatial distribution between them, generating an electric door entrance walking distance table, including: The three-dimensional coordinate data of each electric door entrance in the building is read by the electric door entrance coordinate acquisition unit. The three-dimensional coordinate data includes the horizontal position, vertical position and floor identification. The electric door entrance location distribution map is formed according to the multi-scenario access control electric door linkage layout. Based on the distribution map of the electric gate entrance locations, the area occupied by walls and fixed facilities is identified from the building structure data and marked as an obstacle area. For the connection relationship between two electric gate entrances, the turning points that can be bypassed are determined along the boundary of the obstacle area as pedestrian path nodes, thus obtaining the pedestrian path node sequence between each pair of electric gate entrances. The straight-line distance between adjacent path nodes is calculated and accumulated using the walking path node sequence. If the path passes through different floors, the vertical movement distance is superimposed according to the floor height corresponding to the floor identifier to obtain the actual walking distance value between the two electric door entrances. The electric door entrance pairs are indexed and arranged to generate an electric door entrance walking distance table.

[0007] Furthermore, based on the walking distance table for the electric gate entrances and a preset normal walking speed, a walking time model is constructed. This model outputs the estimated passage time between each pair of electric gate entrances, and the minimum passage time between each pair of electric gate entrances is determined based on the estimated passage time. This includes: Read the actual walking distance values ​​between each pair of electric gate entrances from the electric gate entrance walking distance table, obtain the preset lower limit value of normal walking speed for personnel, and construct a walking time model based on the correspondence between the actual walking distance values ​​and the lower limit value of walking speed. The walking time model is used to calculate the time for each electric gate entrance. The time conversion method is to divide the actual walking distance by the lower limit of the walking speed to obtain the estimated passage time. If different floors are involved, the vertical movement time is calculated and summed according to the floor height and vertical movement speed. The estimated passage time value corresponding to each electric gate entrance is determined as the minimum passage duration, and it is stored in association with the corresponding electric gate entrance pair to generate an index table of minimum passage duration for electric gate entrance pairs.

[0008] Furthermore, the step of extracting the identification times of the same voucher at different electric gate entrances from the voucher timing record unit, comparing the time interval between two adjacent identifications with the minimum passage time for the corresponding electric gate entrance pair, and marking it as a preliminary suspected backtracking when the time interval is less than the minimum passage time includes: Extract the identification time of the same voucher from the voucher time sequence record unit and arrange them in chronological order to obtain the electric gate entrance number corresponding to two adjacent identification times; Based on the electric gate entrance number, retrieve the corresponding minimum passage time from the electric gate entrance minimum passage time index table; The time interval is obtained by subtracting two adjacent identification times. The time interval is then compared with the minimum passage duration obtained from the query. If the time interval is less than the minimum passage duration, the current identification event is marked as a preliminary suspected backtracking event.

[0009] Furthermore, for credentials marked as initially suspected of re-entry, the historical access path data of the authorized personnel is obtained. A path matching algorithm is used to evaluate the consistency between the current electric gate entry switching behavior and the historical path, resulting in a behavior legitimacy label, including: For credentials marked as suspected of initial backtracking, the historical access path data of the authorized personnel corresponding to the credentials is obtained from the access record storage unit. The historical access path data includes the sequence of electric gate entrances passed in the past time period and the corresponding identification time sequence, which are arranged in chronological order to form a historical path sequence set. Extract the electric gate entrance switching behavior that triggers the initial backtracking suspicion. The entrance switching behavior consists of an entrance pair consisting of the previous identified entrance and the next identified entrance. Retrieve historical path segments containing the same entrance pair from the historical path sequence set to obtain a set of historical path segments that match the current entrance pair. For the set of historical path segments, the edit distance is used to measure the difference between the entry sequence of the current entry switching behavior and the entry sequence of each historical path segment. The edit distance is the number of insertion, deletion or replacement operations required to convert the current entry sequence into a historical entry sequence. The path matching degree value is obtained by taking the reciprocal of the edit distance and normalizing it. The path matching score is compared with a preset matching threshold. If the path matching score is greater than or equal to the matching threshold, the current entry switching behavior is determined to be a legitimate door-switching behavior and is assigned a legitimate label. If the path matching score is less than the matching threshold, the current entry switching behavior is determined to be a genuine backdoor attempt and is assigned an illegal label, thus obtaining a behavior legitimacy label.

[0010] Furthermore, based on the behavior legality label, for cases determined to be legal door-swapping behavior, an extended time window is generated by adding a preset buffer time to the minimum passage time of the electric door entrance pair, including: Obtain the determination result of the behavior legality label. If the behavior legality label is a legal label, then read the minimum passage time corresponding to the current electric gate entrance from the minimum passage time index table of electric gate entrances. The minimum passage time is superimposed with the preset buffer time to obtain an extended time window, and the extended time window is associated with and stored with the current electric gate entrance.

[0011] Furthermore, the step of evaluating the current recognition interval based on the extended time window, triggering a rejection response and recording recognition events exceeding the extended time window as passage interruption events, and allowing passage and adding the current passage path to the historical path data for recognition events within the extended time window, includes: The time interval of the current recognition event is obtained, the extended time window corresponding to the current electric gate entrance is read from the associated storage, and the time interval is compared with the extended time window to obtain the recognition interval evaluation result. If the time interval exceeds the extended time window, a rejection response command is sent to the multi-scenario access control electric door linkage control unit, and the credential number, electric door entrance pair, recognition time and over-window mark of this identification event are recorded as an access interruption event. If the time interval is within the extended time window, a passage permission command is sent to the multi-scenario access control electric door linkage control unit, and the electric door entrance pair involved in this identification event is added to the historical path data as a new path segment.

[0012] On the other hand, the present invention provides a multi-scenario access control electric door linkage control system based on artificial intelligence, the system comprising: The electric door entrance coordinate acquisition module is used to obtain the location information of each electric door entrance in the building through the electric door entrance coordinate acquisition unit. Combined with the multi-scenario access control electric door linkage layout, the actual walking distance between two electric door entrances is extracted according to the spatial distribution between the electric door entrances, and an electric door entrance walking distance table is generated. The walking time model construction module is used to construct a walking time model based on the walking distance table of the electric gate entrance and the preset normal walking speed of people. The walking time model outputs the estimated passage time between each pair of electric gate entrances, and determines the minimum passage time between each pair of electric gate entrances based on the estimated passage time. The preliminary suspected backtracking module is used to extract the identification time of the same voucher at different electric gate entrances from the voucher time sequence recording unit, compare the time interval between two adjacent identifications with the minimum passage time of the corresponding electric gate entrance pair, and mark it as a preliminary suspected backtracking when the time interval is less than the minimum passage time. The behavior legitimacy identification module is used to obtain the historical access path data of the authorized personnel for credentials marked as initially suspected of sneaking back, and to evaluate the consistency between the current electric gate entrance switching behavior and the historical path through a path matching algorithm to obtain a behavior legitimacy label; An extended time window generation module is used to generate an extended time window by adding a preset buffer time to the minimum passage time of the electric gate entrance pair for cases where the behavior is determined to be a legitimate door-changing behavior, based on the behavior legality label. The access control and path update module is used to evaluate the current identification interval according to the extended time window, trigger a rejection response and record the identification event that exceeds the extended time window as an access interruption event, and allow the identification event that is within the extended time window and add the current access path to the historical path data.

[0013] The technical solutions provided by the embodiments of the present invention may include the following beneficial effects: This invention discloses a multi-scenario access control method for electric doors based on artificial intelligence. By collecting the actual walking distances at each electric door entrance within a building and combining this with normal walking speeds, a walking time model is constructed to accurately extract the minimum passage time required for normal passage between each pair of electric door entrances. Based on this, the recognition time interval of the same credential at adjacent electric door entrances is compared with the minimum passage time. When the interval is less than the minimum passage time, it is marked as a preliminary suspected backtracking attempt. Subsequently, for suspected credentials, historical passage path data is introduced, and a path matching algorithm is used to determine whether the current access control switching behavior is a legitimate door switching or a genuine backtracking attempt, achieving preliminary identification of backtracking behavior. For cases determined to be legitimate door switching, this invention automatically adds a buffer time to the minimum passage time of the electric door entrance pair to form an extended time window, and dynamically evaluates the recognition interval accordingly: within the extended time window, normal passage is allowed and historical path data is updated; outside the window, a multi-scenario linkage rejection response is triggered, and the passage interruption event is recorded. By further integrating the time, location, and frequency information of interruption events, and analyzing interruption trends through anomaly detection algorithms, the system can optimize walking speed and minimum passage duration parameters. This effectively prevents security risks such as credential sharing and backtracking, while maximizing the user's normal passage experience and achieving intelligent balance control between anti-cheating capabilities and passage convenience. Attached Figure Description

[0014] Figure 1 This is a flowchart of the multi-scenario access control electric door linkage control method based on artificial intelligence according to the present invention.

[0015] Figure 2 This is a schematic diagram of the multi-scenario access control electric door linkage control method based on artificial intelligence according to the present invention.

[0016] Figure 3 This is a schematic diagram of the structure of the multi-scenario access control electric door linkage control system based on artificial intelligence of the present invention. Detailed Implementation

[0017] To enable those skilled in the art to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this specification, and not all embodiments. Based on the embodiments in this specification, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of this specification.

[0018] like Figures 1-3 The multi-scenario access control electric door linkage control method and system based on artificial intelligence in this embodiment may specifically include: S101. Obtain the location information of each electric door entrance in the building through the electric door entrance coordinate acquisition unit. Combine the multi-scenario access control electric door linkage layout, extract the actual walking distance between two electric door entrances according to the spatial distribution between the electric door entrances, and obtain the electric door entrance walking distance table.

[0019] The three-dimensional coordinate data of each electric door entrance within the building is read by the electric door entrance coordinate acquisition unit. This data includes the horizontal and vertical positions of the entrances, as well as floor markings. Based on the multi-scenario access control electric door linkage layout, the distribution of each electric door entrance on the building plan is extracted, forming an electric door entrance location distribution map. According to this map, the occupied areas of walls and fixed facilities are identified from the building structure data and marked as obstacle areas. For the connection relationship between two electric door entrances, detourable turning points are sequentially determined along the boundary of the obstacle area as pedestrian path nodes, resulting in a pedestrian path node sequence between each pair of electric door entrances. Using this pedestrian path node sequence, the straight-line distance between adjacent path nodes is calculated. The straight-line distances of all adjacent nodes within the same electric door entrance pair are accumulated. If the path passes through different floors, the vertical movement distance is superimposed based on the floor height corresponding to the floor markings, yielding the actual walking distance value between the two electric door entrances. These actual walking distance values ​​are then indexed and arranged by electric door entrance pair to generate an electric door entrance walking distance table.

[0020] In the actual deployment of multi-scenario access control and electric door linkage control, the electric door entrance coordinate acquisition unit typically uses building information modeling or on-site measurement equipment to obtain the three-dimensional coordinate data of each electric door entrance. The lateral and longitudinal positions in the three-dimensional coordinate data reflect the two-dimensional distribution of the entrances on the building plane, while the floor markings are used to distinguish the entrances belonging to different floors. Together, these three elements constitute the spatial positioning basis of the entrances.

[0021] For example, in an office building with multiple floors, the ground floor has a main entrance on the east side and a freight entrance on the west side, and the second floor has a connecting corridor entrance. The three-dimensional coordinate data of each entrance are recorded in the building coordinate system and the floor number to which it belongs, thus forming a distribution map of the electric door entrance locations.

[0022] In one possible implementation, when identifying the occupancy range of walls and fixed facilities from building structural data, the building structural data comes from architectural design drawings or spatial models generated by on-site scanning. The wall occupancy range refers to the boundary area covered by the wall entity in a plane, while the fixed facility occupancy range refers to the planar projection area of ​​permanent structures such as elevator shafts, stairwells, and columns. After uniformly marking these occupancy ranges as obstacle areas, if the line connecting two electric gate entrances crosses an obstacle area, this line cannot be used as an actual walking path; a detour point must be found along the boundary of the obstacle area.

[0023] Specifically, the process of determining the nodes of the walking path includes: using the starting entrance as the origin, performing global path planning using the A algorithm. The evaluation function of the A algorithm is f(n) = g(n) + h(n), where g(n) is the cumulative actual walking distance from the starting entrance to the current node n (in meters), and h(n) is the heuristically estimated distance from the current node n to the target entrance, calculated using Euclidean distance. , where (x n ,y n (x) represents the coordinates of the current node. goal ,y goal Let f(n) be the coordinates of the target entrance. The algorithm starts from the initial entrance and expands by selecting the node with the smallest f(n) value each time until the target entrance is reached, thus obtaining the shortest walking path node sequence between each pair of electric gate entrances.

[0024] It should be noted that when two electric door entrances are located on different floors, the pedestrian path node sequence must include the location nodes of the stairwell or elevator lobby, and the vertical movement distance must be added when calculating the actual walking distance. The vertical movement distance is determined based on the floor height corresponding to the floor markings. The floor height refers to the vertical distance between the surfaces of two adjacent floor slabs, and this value is extracted from the building structure data.

[0025] In one embodiment, when calculating the actual walking distance using the walking path node sequence, the straight-line distance between adjacent path nodes is measured sequentially and accumulated. Finally, the actual walking distance values ​​of each electric gate entrance pair are indexed and arranged according to the entrance number to generate an electric gate entrance walking distance table. This table stores the walking distance between any two entrances in matrix form for subsequent passage time determination.

[0026] S102. Based on the walking distance table of the electric gate entrances and the preset normal walking speed of people, construct a walking time model, output the estimated passage time between each pair of electric gate entrances through the walking time model, and identify the minimum passage time required for each pair of electric gate entrances to complete normal walking based on the estimated passage time.

[0027] The actual walking distance between each pair of electric gate entrances is read from the walking distance table. A preset lower limit for normal walking speed is obtained. A walking time model is constructed based on the correspondence between the actual walking distance and the lower limit for walking speed. The input of the walking time model is the actual walking distance, and the output is the estimated passage time. The walking time model is used to perform time conversion on each pair of electric gate entrances. The time conversion method is to divide the actual walking distance by the lower limit for walking speed to obtain the estimated passage time between each pair of electric gate entrances. If the pair of electric gate entrances involves different floors, the vertical movement time is calculated based on the floor height and vertical movement speed corresponding to the floor identification and added to the estimated passage time. Based on the estimated passage time, the estimated passage time value corresponding to each electric gate entrance is determined as the minimum passage time required to complete normal walking. Each pair of electric gate entrances is associated with and stored with its corresponding minimum passage time to obtain an index table of minimum passage time for electric gate entrance pairs.

[0028] In the practical application of multi-scenario access control and electric door linkage control, the electric door entrance walking distance table records the actual walking distance between each pair of electric door entrances in the building after passing through the obstacle area. This value reflects the actual path length of people moving from one entrance to another.

[0029] Specifically, the walking time model is constructed based on the fundamental conversion relationship between distance and speed. Its input is the actual walking distance, and its output is the estimated travel time. The lower limit of walking speed refers to the relatively slow walking speed of an adult under normal conditions. Using this lower limit for time conversion can cover the travel situations of people of different ages and physical conditions, avoiding misjudgment of legitimate passage due to overestimation of speed.

[0030] In one possible implementation, the time conversion process is as follows: read the actual walking distance of a pair of electric gate entrances, divide the value by the lower limit of walking speed, and obtain the estimated passage time between the pair of entrances.

[0031] For example, if the actual walking distance between the main entrance on the east side and the freight entrance on the west side is 120 meters and the minimum walking speed is 1 meter per second, then the estimated passage time is 120 seconds.

[0032] It should be noted that when the electric gate entrances involve different floors, the personnel movement path includes vertical movement segments via stairs or elevators. In this case, calculating only the horizontal walking time cannot reflect the actual passage time. For such cross-floor scenarios, the corresponding floor height values ​​are extracted based on the floor markings, and the vertical movement time is calculated by combining this with the vertical movement speed, which refers to the average movement rate of personnel climbing stairs or ascending and descending in elevators. The vertical movement time is then superimposed on the estimated horizontal passage time to obtain the complete estimated passage time for the cross-floor entrance pair.

[0033] In one embodiment, the height between the ground floor entrance and the second-floor connecting corridor entrance is 4 meters. The vertical movement speed is referenced to the normal stair climbing speed of an adult, taken as 0.5 meters per second, resulting in a vertical movement time of 8 seconds. This time is added to the horizontal walking time to form the estimated passage time for this entrance pair. The normal stair climbing speed of an adult is based on the average vertical displacement speed measured in ergonomics. Based on the estimated passage time, the estimated passage time value corresponding to each electric gate entrance is directly determined as the minimum passage time required to complete normal walking. This is then stored in association using the entrance pair number as an index, forming an index table of minimum passage time for electric gate entrance pairs. This index table records the lower limit of the time required for authorized personnel to complete normal walking between any two entrances in matrix form.

[0034] S103. Extract the identification time of the same voucher at different electric gate entrances from the voucher timing record unit, compare the time interval between two adjacent identifications with the minimum passage time of the corresponding electric gate entrance pair, and mark it as a preliminary suspected backtracking when the time interval is less than the minimum passage time.

[0035] The identification times of the same credential at different automatic gate entrances are extracted from the credential timing record unit. These identification times are arranged chronologically, and the automatic gate entrance numbers corresponding to two adjacent identification times are obtained. The corresponding minimum passage time is then retrieved from the minimum passage time index table based on the automatic gate entrance number. The time interval is obtained by subtracting two adjacent identification times. This time interval is compared with the retrieved minimum passage time. If the time interval is less than the minimum passage time, the identification event corresponding to that credential is marked as a preliminary suspected backtracking event.

[0036] The credential timing record unit stores the identification records of each credential at each automatic gate entrance. Each record includes the credential number, the automatic gate entrance number, and the identification time. In actual access control scenarios, when the same credential appears at multiple automatic gate entrances within a short period of time, it is necessary to determine whether its passage behavior is reasonable.

[0037] Specifically, for a given credential, all recognition times at different automated gate entrances are extracted and arranged chronologically. Adjacent records correspond to the previous and subsequent recognition entrances, respectively, thus identifying the automated gate entrance pair. Based on the combination of numbers for this entrance pair, the corresponding minimum passage time value is retrieved from the minimum passage time index table for automated gate entrance pairs.

[0038] In one embodiment, the time interval is obtained by subtracting the previous identification time from the subsequent identification time. If the time interval is less than the minimum passage time obtained by query, it indicates that the cardholder is physically unable to complete the normal walking movement from the previous entrance to the next entrance. At this time, the identification event corresponding to the credential is marked as a preliminary suspected sneakback, for further verification in the subsequent behavior determination stage.

[0039] S104. For credentials marked as suspected preliminary backdoor entry, obtain the authorized person's historical access path data, evaluate the consistency between the current electric gate entry switching behavior and the historical path through the path matching algorithm, identify whether it is a legitimate door switching behavior or a real backdoor entry attempt, and obtain the behavior legality label.

[0040] For credentials marked as suspected initial backtracking, the historical access path data of the authorized personnel corresponding to the credential is retrieved from the access record storage unit. This historical access path data includes the sequence of electric gate entrances traversed by the authorized personnel in past time periods and the corresponding identification time sequence, arranged chronologically to form a historical path sequence set. The electric gate entrance switching behavior that currently triggers the initial backtracking suspicion is extracted. This entrance switching behavior consists of an entrance pair formed by the previous and subsequent identification entrances. Historical path segments containing the same entrance pair are retrieved from the historical path sequence set to obtain a set of historical path segments matching the current entrance pair. For this set of historical path segments, edit distance is used to measure the difference between the entrance sequence of the current entrance switching behavior and the entrance sequences of each historical path segment. The edit distance is the number of insertion, deletion, or replacement operations required to convert the current entrance sequence into a historical entrance sequence. The reciprocal of the edit distance is taken and normalized to obtain the path matching degree value. The path matching score is compared with a preset matching threshold. If the path matching score is greater than or equal to the matching threshold, the current entry switching behavior is determined to be a legitimate door-switching behavior and is assigned a legitimate label. If the path matching score is less than the matching threshold, the current entry switching behavior is determined to be a genuine backdoor attempt and is assigned an illegal label, thus obtaining a behavior legitimacy label.

[0041] In the actual operation of multi-scenario access control and electric door linkage control, when a certain credential is marked as a preliminary suspected trespass, it is necessary to further determine whether the credential holder's entry switching behavior is consistent with their daily access habits. The access record storage unit stores all identification records of each authorized person in the past time period, and these records are indexed and stored by credential number.

[0042] Specifically, the process of acquiring historical access route data is as follows: Based on the credential number, all identification records of the authorized person within a preset time range are retrieved. Each record includes the electric gate entrance number and the identification time. These records are arranged in chronological order of identification time, and the entrance numbers of adjacent records are connected sequentially to form an entrance sequence, thus constituting the authorized person's historical route sequence. If the person has multiple access routes on different dates, each route independently forms a historical route sequence, and all sequences are collected to form a historical route sequence set.

[0043] In one possible implementation, the entry switching behavior that triggers the initial suspected backtracking consists of an entry pair formed by the previous and subsequent identified entry points, such as an authorized person switching from the east main entry point to the west freight entry point. When retrieving from the historical path sequence set, each historical path sequence is traversed one by one to locate path segments containing the same entry pair. A path segment refers to a local section in a historical path sequence where the entry pair is a continuous subsequence. If a historical path sequence contains continuous records from the east main entry point to the west freight entry point, that section is extracted as the historical path segment matching the current entry pair.

[0044] It's important to note that edit distance is a method for measuring the degree of difference between two sequences. Its basic principle is to calculate the minimum number of operations required to transform one sequence into another. Operation types include inserting an element, deleting an element, and replacing one element with another. In the scenario of comparing electric gate entrance sequences, each element corresponds to an electric gate entrance number. Edit distance reflects the structural difference between the entrance sequence of the current entrance switching behavior and the entrance sequences of historical path segments.

[0045] For example, if the current entrance sequence is from the east main entrance to the west freight entrance to the second-floor connecting corridor entrance, and the entrance sequence of a certain historical path segment is from the east main entrance to the west freight entrance, then converting the current sequence to a historical sequence requires deleting the element of the second-floor connecting corridor entrance, with an edit distance of one. The formula for calculating the path matching degree is: M = 1 / (ED + 1), where M is the path matching degree (value range (0,1]), and ED is the edit distance. This formula ensures that the smaller the edit distance, the closer the matching degree is to 1.

[0046] In one embodiment, if there are three segments in the historical path segment set with edit distances of zero, one, and three to the current entry sequence, the corresponding path matching degrees are 1 / (0+1)=1, 1 / (1+1)=0.5, and 1 / (3+1)=0.25, respectively. The maximum value of 1 is taken as the final path matching degree value.

[0047] Understandably, the preset matching threshold is used to delineate the boundary between legitimate door-switching behavior and genuine backdooring attempts. When the path matching score exceeds this threshold, it indicates that the current entrance switching behavior has a high frequency or similarity in historical access records, conforming to the authorized person's daily access habits, and is therefore judged as a legitimate door-switching behavior and assigned a legitimate label. When the path matching score is below this threshold, it indicates that the current entrance switching behavior deviates from the historical access pattern, and there is a possibility that the credential has been misused or shared by others, and is therefore judged as a genuine backdooring attempt and assigned an illegal label. The behavior legitimacy label, as the output of this step, is marked in the current identification event record of the credential. This label is valid for 24 hours on the same day and automatically expires and is re-evaluated upon the first access the following day. If the same credential again exhibits a path matching score below the preset threshold within the validity period, the preset threshold is usually set to 0.75, then the original legitimate label is overwritten and updated to an illegal label.

[0048] S105. Analyze the results of the behavior legality label. For cases identified as legal door-changing behavior, add a preset buffer time to the minimum passage time of the electric door entrance pair to obtain an extended time window.

[0049] The determination result of the behavior legality label is obtained. If the behavior legality label is a legal label, the minimum passage time corresponding to the current electric gate entrance is read from the minimum passage time index table of electric gate entrances, and a preset buffer time is obtained at the same time. The minimum passage time and the preset buffer time are superimposed to obtain an extended time window, and the extended time window is associated with and stored with the current electric gate entrance pair.

[0050] In the actual operation of multi-scenario access control and electric door linkage control, when the behavior legality label is determined to be a legal label, it indicates that the current credential holder's entry switching behavior conforms to their historical access habits and is a normal door switching behavior rather than a credential sharing attempt.

[0051] Specifically, after retrieving the minimum passage time corresponding to the current electric gate entrance from the minimum passage time index table, a preset buffer time needs to be added to form an extended time window. The preset buffer time is used to accommodate the additional time spent by authorized personnel during actual passage due to queuing, short stops, or fluctuations in walking speed.

[0052] In one embodiment, if the minimum passage time between the main entrance on the east side and the freight entrance on the west side is 90 seconds, and the preset buffer time is 30 seconds, then the superimposed extended time window is 120 seconds. After associating and storing this extended time window with the current electric gate entrance pair, the time interval determination of subsequent identification events will use the extended time window as the comparison benchmark, thereby avoiding misjudgment of legitimate door-swapping behavior.

[0053] S106. Evaluate the current recognition interval based on the extended time window to obtain the recognition interval evaluation result. For recognition events that exceed the extended time window, trigger the multi-scenario access control electric door linkage to refuse response and record it as a passage interruption event. For recognition events that are within the extended time window, allow passage and add the current passage path to the historical path data.

[0054] The time interval of the current identification event is obtained, and the extended time window corresponding to the current electric gate entrance is read from the associated storage. The time interval is compared with the extended time window to obtain the identification interval evaluation result. According to the identification interval evaluation result, if the time interval exceeds the extended time window, a rejection response command is sent to the multi-scenario access control electric gate linkage control unit, and the credential number, electric gate entrance pair, identification time, and out-of-window marker of this identification event are recorded as an access interruption event; if the time interval is within the extended time window, an access permission command is sent to the multi-scenario access control electric gate linkage control unit, and the electric gate entrance pair involved in this identification event is added as a new path segment to the historical path data of the authorized personnel corresponding to the credential.

[0055] In the actual operation of multi-scenario access control and electric door linkage control, the identification interval evaluation result is obtained based on the comparison between the time interval of the current identification event and the value of the extended time window. The time interval refers to the difference between the current identification time and the previous identification time, and the extended time window is formed by adding a preset buffer time to the minimum passage duration. In the identification interval evaluation, the time interval of the current identification event is defined as the difference Δt between the current identification time and the previous identification time, and the extended time window Texp is the minimum passage duration Tmin plus the preset buffer time Tbuf (Texp=Tmin+Tbuf). The system determines whether they belong to the same passage event by comparing the magnitudes of Δt and Texp: if Δt≤Texp, it is considered the same passage; if Δt>Texp, it is considered a new passage event.

[0056] Specifically, when the time interval exceeds the extended time window, it indicates that the time taken for the credential holder to move between the two identifications is significantly longer than the time required for normal walking. However, since the preceding steps have already marked the credential as a legitimate door-changing behavior, such timeouts are usually due to temporary waiting or other non-malicious factors. Passage should be allowed and additional monitoring should be triggered to maintain access control order.

[0057] In one possible implementation, after receiving a rejection response command, the multi-scenario access control motorized door linkage control unit sends a closing signal to the access control actuator at the corresponding motorized door entrance, preventing the credential holder from passing through that entrance. Simultaneously, the credential number of this identification event, the pair of motorized door entrances involved, and the time of identification are recorded as passage interruption events for subsequent statistical analysis and backtracking.

[0058] It should be noted that when the time interval is within the extended time window, it indicates that the credential holder's movement speed is consistent with historical passage habits, conforming to the time characteristics of normal door-changing behavior. In this case, a passage permission command is sent to the multi-scenario access control electric door linkage control unit. After the access control actuator responds, it opens the electric door, allowing the authorized personnel to pass smoothly. Furthermore, after passage is permitted, the electric door entrance pair involved in this identified event should be added as a new path segment to the historical path data of the authorized personnel corresponding to this credential.

[0059] For example, if an authorized person switches from the main entrance on the east side to the freight entrance on the west side, that entrance will be added to their historical path sequence set, so that subsequent path matching assessments can refer to richer historical passage records, thereby improving the accuracy of identifying legitimate door-switching behavior.

[0060] In another embodiment of this application, it further includes S107, integrating the occurrence time of the passage interruption event, the data involving the electric gate entrance location and the interruption frequency, combining the identification interval evaluation results, analyzing the evolution trend of the number of interruptions through an anomaly detection algorithm, evaluating the accuracy of the current walking time model, and feeding back the optimized walking speed to the minimum passage time extraction stage, thereby completing the balance control between preventing credential sharing and the user's normal passage experience.

[0061] Interruption event data for each electric gate entrance pair is extracted from the traffic interruption event record. This data includes the occurrence time and the location of the involved electric gate entrance. The data is then grouped and statistically analyzed by electric gate entrance pair to obtain the interruption frequency of each pair within a preset time period. Based on this interruption frequency and the proportion of events deemed to exceed the extended time window in the identification interval evaluation results, an isolation forest algorithm is used to detect anomalies in the interruption frequency sequence of each electric gate entrance pair. The isolation forest algorithm takes the interruption frequency sequence as input and outputs electric gate entrance pairs whose interruption frequencies significantly deviate from the normal fluctuation range, thus obtaining anomaly entrance pair identifiers. Based on these anomaly entrance pair identifiers, the actual time interval and corresponding actual walking distance in the historical traffic events of that entrance pair are extracted. The actual walking distance is divided by the actual time interval to obtain the corrected walking speed. The corrected walking speed is then subtracted from the original walking speed lower limit to obtain the speed deviation value. If the speed deviation value exceeds the preset speed adjustment threshold, the original walking speed lower limit value is added to the speed deviation value to obtain the adjusted walking speed lower limit value. The adjusted walking speed lower limit value is fed back to the time conversion stage of the walking time model, and the minimum passage time of the electric gate entrance pair is recalculated and updated to the minimum passage time index table of the electric gate entrance pair, thus completing the balance control between preventing credential sharing and the normal passage experience of users.

[0062] During the long-term operation of multi-scenario access control electric door linkage control, the cumulative record of access interruption events reflects the frequency of rejection response of each electric door entrance pair in actual use. When extracting the interruption event data of each electric door entrance pair from the access interruption event record, the interruption event data is arranged chronologically according to the occurrence time, and statistical analysis is performed based on the electric door entrance pair as the grouping basis to obtain the interruption frequency of each electric door entrance pair within a preset time period.

[0063] Specifically, the preset time period can be set to one week or one month, during which the number of passage interruption events for each electric gate entrance pair is counted. If the interruption frequency of a certain electric gate entrance pair is significantly higher than that of other entrance pairs, it indicates that the minimum passage duration setting for that entrance pair may deviate from the actual movement of people, causing legitimate passage to be frequently misjudged as sneaking back.

[0064] It should be noted that the Isolation Forest algorithm is an unsupervised anomaly detection method based on the idea of ​​Random Forest. Its core principle is to construct multiple isolation trees by randomly selecting features and randomly partitioning data points. Anomaly data points, because their feature values ​​differ significantly from normal data, are more easily identified early in the isolation trees. The anomaly score is calculated using the formula: s(x,n)=2 (-E(h(x)) / c(n))Where s(x,n) is the anomaly score of sample x in a dataset with a sample size of n (range (0,1)), E(h(x)) is the average path length of sample x in all isolation trees, and c(n) is the average path length of the binary search tree with a sample size of n, calculated as c(n) = 2H(n-1) - 2(n-1) / n, where H(i) is the harmonic number. The closer the anomaly score is to 1, the more likely the sample is to be an anomaly. In this scheme, an anomaly score threshold is set, and entry pairs with anomaly scores exceeding this threshold are marked as anomaly entry pairs.

[0065] In one possible implementation, the interruption frequency sequence is weighted by combining the proportion of events deemed to exceed the extended time window in the evaluation results of the identification interval. The extended time window is a standard time window T calculated based on standard walking speed plus a buffer threshold, such as 1.2T. The proportion of events exceeding the window is the number of events exceeding this window divided by the total number of events. If a certain electric gate entrance pair has a high proportion of timeout events and a high interruption frequency, it is more likely to be identified as an anomaly by the Isolation Forest algorithm, and an anomaly entrance pair identifier is output. Further, for the electric gate entrance pair corresponding to the anomaly entrance pair identifier, actual passage data needs to be extracted from historical passage events to calculate the corrected walking speed. The formula for calculating the corrected walking speed is: Va dj =(1 / N)×Σ(D i / T i ), where V adj To correct for walking speed (unit: m / s), N is the number of historical allowed passage events, and D... i T represents the actual walking distance (in meters, derived from the walking distance table for the electric gate entrance) for the i-th passage. i Let be the actual time interval of the i-th passage (in seconds, representing the difference between two adjacent recognition times). This formula takes the arithmetic mean of the instantaneous walking speeds of multiple passages to obtain the corrected walking speed that reflects the actual mobility efficiency of the entrance.

[0066] For example, if the actual walking distance between the main entrance on the east side and the freight entrance on the west side is 120 meters, and the average actual time interval between ten permitted passage events in history is 100 seconds, then the corrected walking speed is 1.2 meters per second. This value reflects the actual movement efficiency of authorized personnel between this pair of entrances.

[0067] Understandably, the speed deviation value is used to measure the difference between the corrected walking speed and the original lower limit of walking speed. The speed deviation value is obtained by subtracting the original lower limit of walking speed from the corrected walking speed. If the speed deviation value is positive and exceeds the preset speed adjustment threshold, it indicates that the original lower limit of walking speed is set too low and needs to be adjusted upward; if the speed deviation value is negative and the absolute value exceeds the preset speed adjustment threshold, it indicates that the original lower limit of walking speed is set too high and needs to be adjusted downward.

[0068] In one embodiment, the preset speed adjustment threshold is 0.1 meters per second. If the speed deviation is 0.2 meters per second, the original lower limit of walking speed is added to the speed deviation to obtain the adjusted lower limit of walking speed. The adjusted lower limit of walking speed is fed back to the time conversion stage of the walking time model, and the minimum passage time of the electric gate entrance is recalculated based on the adjusted lower limit of walking speed.

[0069] Preferably, after the recalculated minimum passage time is updated in the minimum passage time index table for electric gate entrance pairs, subsequent passage determinations for that entrance pair will be based on the updated minimum passage time. Through the above closed-loop adjustment mechanism, time constraints are relaxed for entrance pairs with abnormal interruption frequency, reducing misjudgments of legitimate gate-switching behavior. At the same time, the original constraints are maintained for entrance pairs with normal interruption frequency, continuously preventing credential sharing behavior and achieving a balance between preventing credential sharing and ensuring a normal user passage experience.

[0070] This invention provides a multi-scenario access control electric door linkage control system based on artificial intelligence, the system comprising: The electric door entrance coordinate acquisition module is used to obtain the location information of each electric door entrance in the building through the electric door entrance coordinate acquisition unit. Combined with the multi-scenario access control electric door linkage layout, the actual walking distance between two electric door entrances is extracted according to the spatial distribution between the electric door entrances, and an electric door entrance walking distance table is generated. The walking time model construction module is used to construct a walking time model based on the walking distance table of the electric gate entrance and the preset normal walking speed of people. The walking time model outputs the estimated passage time between each pair of electric gate entrances, and determines the minimum passage time between each pair of electric gate entrances based on the estimated passage time. The preliminary suspected backtracking module is used to extract the identification time of the same voucher at different electric gate entrances from the voucher time sequence recording unit, compare the time interval between two adjacent identifications with the minimum passage time of the corresponding electric gate entrance pair, and mark it as a preliminary suspected backtracking when the time interval is less than the minimum passage time. The behavior legitimacy identification module is used to obtain the historical access path data of the authorized personnel for credentials marked as initially suspected of sneaking back, and to evaluate the consistency between the current electric gate entrance switching behavior and the historical path through a path matching algorithm to obtain a behavior legitimacy label; An extended time window generation module is used to generate an extended time window by adding a preset buffer time to the minimum passage time of the electric gate entrance pair for cases where the behavior is determined to be a legitimate door-changing behavior, based on the behavior legality label. The access control and path update module is used to evaluate the current identification interval according to the extended time window, trigger a rejection response and record the identification event that exceeds the extended time window as an access interruption event, and allow the identification event that is within the extended time window and add the current access path to the historical path data. In another embodiment of this application, a model optimization and balance control module is also included, which integrates the occurrence time of the passage interruption event, the location of the electric gate entrance and the interruption frequency data, combines the identification interval evaluation results, analyzes the evolution trend of the number of interruptions through an anomaly detection algorithm, evaluates the accuracy of the current walking time model, and feeds back the optimized walking speed to the calculation stage of the minimum passage time.

[0071] If the technical solution of this application involves the processing of personal information, the relevant products have established a sound user authorization mechanism: before collecting, using, or sharing personal information, the obligation to inform is fulfilled in accordance with the law, and the individual's voluntary and explicit consent is obtained; if sensitive personal information is involved, the user's separate and explicit consent is further obtained. Specific measures include, but are not limited to: setting up prominent prompts in the information collection area, or clearly displaying the processing rules (including the processor, purpose, method, information type, etc.) through electronic interfaces such as pop-ups, checkboxes, and active submissions, to ensure that users voluntarily authorize based on their knowledge. All personal information processing activities strictly comply with national laws and regulations, especially the relevant provisions of the "Personal Information Protection Law of the People's Republic of China," to effectively safeguard the legitimate rights and interests of personal information subjects.

[0072] The preferred embodiments of the present invention disclosed above are merely illustrative of the invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the invention to any specific implementation. Clearly, many modifications and variations can be made based on the content of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention. The invention is limited only by the claims and their full scope and equivalents.

Claims

1. A multi-scenario access control method for electric doors based on artificial intelligence, characterized in that, include: The location information of each electric door entrance in the building is obtained by the electric door entrance coordinate acquisition unit. Combined with the multi-scenario access control electric door linkage layout, the actual walking distance between two electric door entrances is extracted according to the spatial distribution between the electric door entrances, and an electric door entrance walking distance table is generated. Based on the walking distance table of the electric gate entrances and the preset normal walking speed of people, a walking time model is constructed. The walking time model outputs the estimated passage time between each pair of electric gate entrances, and the minimum passage time between each pair of electric gate entrances is determined based on the estimated passage time. Extract the identification time of the same voucher at different electric gate entrances from the voucher timing record unit, compare the time interval between two adjacent identifications with the minimum passage time of the corresponding electric gate entrance pair, and mark it as a preliminary suspected sneakback when the time interval is less than the minimum passage time. For credentials marked as suspected of initial backtracking, obtain the authorized personnel's historical access path data, and use a path matching algorithm to evaluate the consistency between the current electric gate entrance switching behavior and the historical path to obtain a behavior legality label; Based on the legality label of the behavior, for cases where the door-changing behavior is determined to be legal, a preset buffer time is added to the minimum passage time of the electric door entrance pair to generate an extended time window; The current identification interval is evaluated based on the extended time window. For identification events that exceed the extended time window, a rejection response is triggered and recorded as a passage interruption event. For identification events that are within the extended time window, passage is allowed and the current passage path is added to the historical path data.

2. The multi-scenario access control electric door linkage control method based on artificial intelligence according to claim 1, characterized in that, The method involves acquiring the location information of each electric door entrance within the building through the electric door entrance coordinate acquisition unit, combining it with the multi-scenario access control electric door linkage layout, extracting the actual walking distance between two electric door entrances based on the spatial distribution between them, and generating an electric door entrance walking distance table, including: The three-dimensional coordinate data of each electric door entrance in the building is read by the electric door entrance coordinate acquisition unit. The three-dimensional coordinate data includes the horizontal position, vertical position and floor identification. The electric door entrance location distribution map is formed according to the multi-scenario access control electric door linkage layout. Based on the distribution map of the electric gate entrance locations, the area occupied by walls and fixed facilities is identified from the building structure data and marked as an obstacle area. For the connection relationship between two electric gate entrances, the turning points that can be bypassed are determined along the boundary of the obstacle area as pedestrian path nodes, thus obtaining the pedestrian path node sequence between each pair of electric gate entrances. The straight-line distance between adjacent path nodes is calculated and accumulated using the walking path node sequence. If the path passes through different floors, the vertical movement distance is superimposed according to the floor height corresponding to the floor identifier to obtain the actual walking distance value between the two electric door entrances. The electric door entrance pairs are indexed and arranged to generate an electric door entrance walking distance table.

3. The multi-scenario access control electric door linkage control method based on artificial intelligence according to claim 1, characterized in that, The process involves constructing a walking time model based on the walking distance table for each electric gate entrance and a preset normal walking speed. This model outputs the estimated passage time between each pair of electric gate entrances. Based on the estimated passage time, the minimum passage time between each pair of electric gate entrances is determined, including: Read the actual walking distance values ​​between each pair of electric gate entrances from the electric gate entrance walking distance table, obtain the preset lower limit value of normal walking speed for personnel, and construct a walking time model based on the correspondence between the actual walking distance values ​​and the lower limit value of walking speed. The walking time model is used to calculate the time for each electric gate entrance. The time conversion method is to divide the actual walking distance by the lower limit of the walking speed to obtain the estimated passage time. If different floors are involved, the vertical movement time is calculated and summed according to the floor height and vertical movement speed. The estimated passage time value corresponding to each electric gate entrance is determined as the minimum passage duration, and it is stored in association with the corresponding electric gate entrance pair to generate an index table of minimum passage duration for electric gate entrance pairs.

4. The multi-scenario access control electric door linkage control method based on artificial intelligence according to claim 1, characterized in that, The step of extracting the identification times of the same voucher at different electric gate entrances from the voucher timing record unit, comparing the time interval between two adjacent identifications with the minimum passage time for the corresponding electric gate entrance pair, and marking it as a preliminary suspected backtracking when the time interval is less than the minimum passage time includes: Extract the identification time of the same voucher from the voucher time sequence record unit and arrange them in chronological order to obtain the electric gate entrance number corresponding to two adjacent identification times; Based on the electric gate entrance number, retrieve the corresponding minimum passage time from the electric gate entrance minimum passage time index table; The time interval is obtained by subtracting two adjacent identification times. The time interval is then compared with the minimum passage duration obtained from the query. If the time interval is less than the minimum passage duration, the current identification event is marked as a preliminary suspected backtracking event.

5. The multi-scenario access control electric door linkage control method based on artificial intelligence according to claim 1, characterized in that, For credentials marked as initially suspected of re-entry, the authorized personnel's historical access path data is obtained. A path matching algorithm is used to evaluate the consistency between the current automatic gate entry switching behavior and the historical path, resulting in a behavior legitimacy label, including: For credentials marked as suspected of initial backtracking, the historical access path data of the authorized personnel corresponding to the credentials is obtained from the access record storage unit. The historical access path data includes the sequence of electric gate entrances passed in the past time period and the corresponding identification time sequence, which are arranged in chronological order to form a historical path sequence set. Extract the electric gate entrance switching behavior that triggers the initial backtracking suspicion. The entrance switching behavior consists of an entrance pair consisting of the previous identified entrance and the next identified entrance. Retrieve historical path segments containing the same entrance pair from the historical path sequence set to obtain a set of historical path segments that match the current entrance pair. For the set of historical path segments, the edit distance is used to measure the difference between the entry sequence of the current entry switching behavior and the entry sequence of each historical path segment. The edit distance is the number of insertion, deletion or replacement operations required to convert the current entry sequence into a historical entry sequence. The path matching degree value is obtained by taking the reciprocal of the edit distance and normalizing it. The path matching score is compared with a preset matching threshold. If the path matching score is greater than or equal to the matching threshold, the current entry switching behavior is determined to be a legitimate door-switching behavior and is assigned a legitimate label. If the path matching score is less than the matching threshold, the current entry switching behavior is determined to be a genuine backdoor attempt and is assigned an illegal label, thus obtaining a behavior legitimacy label.

6. The multi-scenario access control electric door linkage control method based on artificial intelligence according to claim 1, characterized in that, Based on the legality label of the behavior, for cases where the door-changing behavior is determined to be legal, an extended time window is generated by adding a preset buffer time to the minimum passage time of the electric door entrance pair, including: Obtain the determination result of the behavior legality label. If the behavior legality label is a legal label, then read the minimum passage time corresponding to the current electric gate entrance from the minimum passage time index table of electric gate entrances. The minimum passage time is superimposed with the preset buffer time to obtain an extended time window, and the extended time window is associated with and stored with the current electric gate entrance.

7. The multi-scenario access control electric door linkage control method based on artificial intelligence according to claim 1, characterized in that, The process of evaluating the current recognition interval based on the extended time window, triggering a rejection response and recording the recognition event as a passage interruption event for recognition events exceeding the extended time window, and allowing passage for recognition events within the extended time window and adding the current passage path to the historical path data includes: The time interval of the current recognition event is obtained, the extended time window corresponding to the current electric gate entrance is read from the associated storage, and the time interval is compared with the extended time window to obtain the recognition interval evaluation result. If the time interval exceeds the extended time window, a rejection response command is sent to the multi-scenario access control electric door linkage control unit, and the credential number, electric door entrance pair, recognition time and over-window mark of this identification event are recorded as an access interruption event. If the time interval is within the extended time window, a passage permission command is sent to the multi-scenario access control electric door linkage control unit, and the electric door entrance pair involved in this identification event is added to the historical path data as a new path segment.

8. A multi-scenario access control and electric door linkage control system based on artificial intelligence, characterized in that: The system includes: The electric door entrance coordinate acquisition module is used to obtain the location information of each electric door entrance in the building through the electric door entrance coordinate acquisition unit. Combined with the multi-scenario access control electric door linkage layout, the actual walking distance between two electric door entrances is extracted according to the spatial distribution between the electric door entrances, and an electric door entrance walking distance table is generated. The walking time model construction module is used to construct a walking time model based on the walking distance table of the electric gate entrance and the preset normal walking speed of people. The walking time model outputs the estimated passage time between each pair of electric gate entrances, and determines the minimum passage time between each pair of electric gate entrances based on the estimated passage time. The preliminary suspected backtracking module is used to extract the identification time of the same voucher at different electric gate entrances from the voucher time sequence recording unit, compare the time interval between two adjacent identifications with the minimum passage time of the corresponding electric gate entrance pair, and mark it as a preliminary suspected backtracking when the time interval is less than the minimum passage time. The behavior legitimacy identification module is used to obtain the historical access path data of the authorized personnel for credentials marked as initially suspected of sneaking back, and to evaluate the consistency between the current electric gate entrance switching behavior and the historical path through a path matching algorithm to obtain a behavior legitimacy label; An extended time window generation module is used to generate an extended time window by adding a preset buffer time to the minimum passage time of the electric gate entrance pair for cases where the behavior is determined to be a legitimate door-changing behavior, based on the behavior legality label. The access control and path update module is used to evaluate the current identification interval according to the extended time window, trigger a rejection response and record the identification event that exceeds the extended time window as an access interruption event, and allow the identification event that is within the extended time window and add the current access path to the historical path data.

Images