An AI tour guide management system and method for a tourist attraction

Abstract

The application discloses a kind of tourism-oriented AI guide management system and management method, it is related to guide management technical field, including: coordinate construction module, according to the structure data of ancient building doorway, establishes doorway coordinate system;Based on the tourist position data and doorway coordinate system collected by tourist terminal equipment, obtain local threshold crossing segment, and the lateral position of each collection time in local threshold crossing segment is classified, to obtain lateral attribution state;Let threshold evidence module, according to lateral attribution state, generates corner tending state, let threshold round angle state and regression queue state, and determines let threshold evidence intensity;Off the queue counter evidence module, the lateral attribution state of tourist in doorway inside segment is continuously identified to generate first off the queue counter evidence item;The application improves the accuracy of tourist state identification in ancient building doorway area.

Description

Technical Field

[0001] This invention relates to the field of tour guide management technology, and in particular to an AI tour guide management system and method for tourist attractions. Background Technology

[0002] With the development of digital management in tourist attractions, more and more scenic spots are deploying AI tour guide management devices in group tour scenarios. These devices collect tourist location trajectories through tour guide devices, tourist devices, and location markers to achieve integrity management of tourist groups, tour guide route management, and early warning management of abnormal departures from the group. In ancient architectural scenic spots such as temples, ancestral halls, old houses, and guild halls, courtyards are usually connected by narrow doorways. The doorway areas generally have wooden threshold structures, and the edges of the thresholds are also equipped with small copper corner protectors to protect the wooden structure. When tourists pass through the doorways, especially elderly tourists or tourists carrying personal belongings, they are prone to short-term lateral deviations, partial pauses, and edge-keeping behaviors due to avoiding the copper corner protectors, slowing down the threshold crossing action, or turning sideways around the threshold. Although the tourists are still moving in the direction of the tour guide, their local trajectory will form a clear lateral detour process in the threshold area, which will cause the tourists to show a spatial change state similar to leaving the group in the trajectory data.

[0003] Existing technologies for processing visitor trajectories in the archway areas of ancient buildings typically rely on the distance between visitors and tour guides, formation deviations, or continuous deviations from the trajectory to determine whether visitors have strayed from the group. However, they lack the ability to recognize localized detours caused by the copper corner protectors at the threshold edges. Therefore, when visitors briefly pause or detour to avoid the threshold corners, existing technologies can easily misinterpret this normal passage as genuine departure from the group, resulting in incorrect management instructions being issued to the tour guide. This leads to tour guides receiving frequent and invalid reminders, reducing tour guide management efficiency and disrupting the normal group passage order in the archway areas of ancient buildings. Summary of the Invention

[0004] The purpose of this invention is to address the shortcomings of existing technologies that disrupt the normal group passage order in the archway area of ​​ancient buildings, and to propose an AI tour guide management system and method for tourist attractions.

[0005] To address the problems existing in the prior art, the present invention adopts the following technical solution: An AI-powered tour guide management system for tourist attractions includes: The coordinate construction module establishes a doorway coordinate system based on the structural data of the ancient building's doorway; based on the visitor location data collected by the visitor terminal device and the doorway coordinate system, it obtains local threshold segments and classifies the lateral positions at each collection time within the local threshold segments to obtain the lateral attribution status. The threshold evidence collection module generates corner-attaching trend state, threshold-around-corner state, and return queue state based on the horizontal attribution state, and determines the threshold evidence strength. The module for refuting the departure from the group continuously identifies the lateral state of tourists in the segment inside the doorway to generate the first item for refuting the departure from the group; based on the guide's position data and local threshold segments collected by the guide terminal device, it identifies the direction of tourists' movement inside the doorway to generate the second item for refuting the departure from the group; and based on the first and second items for refuting the departure from the group, it determines the strength of the refuting the departure from the group. The strategy optimization module generates an initial particle swarm based on the set of tour guide collaborative strategies corresponding to the doorways of ancient buildings; it calculates the fitness value of each particle based on the strength of evidence for yielding the threshold and the strength of evidence for refuting leaving the group in the set of historical threshold event samples; and iteratively updates the initial particle swarm based on the fitness value to obtain the recommended tour guide collaborative strategy. The instruction generation module matches the recommended tour guide collaboration strategy with the instruction type by considering the threshold evidence strength and the disproving evidence strength of leaving the queue, thus obtaining tour guide management instructions.

[0006] Preferably, establishing a coordinate system for the doorway includes: Obtain the structural data of the doorway of the ancient building; the structural data includes the coordinates of the center point of the doorway, the vector of the passage direction of the doorway, the position data of the threshold line, the coordinates of the left and right copper corner brackets; A coordinate system for the doorway is established based on the coordinates of the center point of the doorway and the direction vector of passage through the doorway in the structural data.

[0007] Preferably, obtaining the lateral attribution status includes: Acquire tourist location data collected by tourist terminal devices; Based on tourist location data and the doorway coordinate system, generate axial and lateral position sequences for tourists; Based on the axial position sequence of tourists, continuous trajectory segments are extracted from the planar position coordinates of tourists to obtain local threshold segments; Based on the tourists' lateral position sequence, the coordinates of the left and right copper corners, the lateral positions at each collection time within the local threshold segment are classified to obtain the lateral attribution status.

[0008] Preferably, determining the strength of evidence for a threshold includes: Based on the lateral attribution status, the local threshold segment is segmented to obtain the segment outside the doorway, the segment near the threshold, and the segment inside the doorway. Determine the corner-attaching tendency state based on the lateral attribution state of the segment outside the doorway; Determine the threshold-around-angle state based on the lateral attribution state of the segments adjacent to the threshold; The regression queue status is determined based on the lateral attribution status of the segment inside the doorway; Based on the corner-touching tendency state, the threshold-around-corner state, and the return queue state, generate a set of threshold-yielding evidence items for tourists; The strength of the threshold evidence is determined based on the number of evidence items in the threshold evidence set.

[0009] Preferably, generating the first off-line counter-evidence entry includes: Based on the local threshold segments and lateral attribution status, extract the data on the changes in the lateral attribution status of tourists in the segment inside the doorway. Based on the data on changes in lateral attribution status, continuous lateral attribution status identification is performed on the tourists' lateral attribution status in the segment inside the doorway to obtain the lateral non-regression status. Based on the lateral non-regression state, generate the first off-line counter-evidence entry.

[0010] Preferably, determining the strength of the disproving evidence includes: Acquire tour guide location data collected by the tour guide terminal device; Extract the planar position coordinates of the tour guides corresponding to the local threshold segments from the tour guide position data; Based on the tour guide's planar position coordinates, generate the tour guide's travel direction data; Based on local threshold segments, generate tourist movement direction data; Based on the data of tourists' walking direction and the data of the tour guide's walking direction, the deviation state of tourists' walking direction inside the doorway is identified, and the deviation state is obtained. Based on the direction divergence state, generate a second disjointed counter-evidence entry; Perform item association processing on the first and second off-listed counter-evidence items to obtain the set of off-listed counter-evidence items; The strength of an off-staff counter-evidence is determined by the number of counter-evidence entries in the off-staff counter-evidence set.

[0011] Preferably, generating the initial particle swarm includes: Obtain the set of tour guide collaboration strategies corresponding to the doorways of ancient buildings; Each tour guide collaboration strategy in the set of tour guide collaboration strategies is numbered to obtain the strategy number; Use the strategy number as the initial particle position; Determine the adjacent strategy numbers of each particle based on the order of the strategy numbers; The initial particle velocity is determined based on the initial particle position and the adjacent strategy numbers. An initial particle swarm is generated based on the initial particle position and initial particle velocity of each particle.

[0012] Preferably, the recommended tour guide collaboration strategy includes: Obtain a sample set of historical threshold events corresponding to the doorways of ancient buildings; By labeling the strength of evidence for yielding the threshold and the strength of evidence for refusal for leaving the queue in the historical threshold event sample set, a sample set for AI strategy optimization is obtained. Based on the AI ​​strategy, the sample set is optimized, and the historical samples of the guide cooperative strategy corresponding to each initial particle position in the initial particle swarm are replayed to obtain the mismatch count. Calculate the fitness value of the particle based on the mismatch count; Based on fitness values, determine the global optimal particle position and the individual optimal particle position; Based on the global optimal particle position and the individual optimal particle position, the initial particle swarm is iteratively updated to obtain the particle swarm after the iteration. The recommended guide cooperative strategy is obtained by selecting guide cooperative strategies from the set of guide cooperative strategies after the iteration is completed by using the particle swarm optimization.

[0013] Preferably, the guide management instructions include: The strength of the evidence for yielding the threshold for each tourist in the group is summed to obtain the total strength of the evidence for yielding the threshold. The strength of the disproving evidence for each tourist leaving the group is summed to obtain the total strength of the disproving evidence for leaving the group. Based on the sum of the evidence strength of thresholding and the sum of the evidence strength of leaving the group, the instruction type of the recommended tour guide collaboration strategy is matched to obtain tour guide management instructions.

[0014] To address the aforementioned problems, this invention also provides an AI-powered tour guide management method for tourist attractions, the method comprising: Based on the structural data of the ancient building's doorway, a doorway coordinate system is established; based on the tourist location data collected by the tourist terminal device, local threshold segments are obtained, and the lateral positions at each collection time within the local threshold segments are classified to obtain the lateral attribution status. Based on the lateral attribution status, the corner-attaching tendency status, the threshold-around-corner status, and the regression queue status are generated, and the threshold-attaching evidence strength is determined. The lateral state of tourists in the segment inside the doorway is continuously identified to generate the first counter-evidence item for leaving the group; based on the guide position data and local threshold segments collected by the guide terminal device, the direction of tourists' movement inside the doorway is identified to generate the second counter-evidence item for leaving the group; based on the first counter-evidence item for leaving the group and the second counter-evidence item for leaving the group, the strength of the counter-evidence item for leaving the group is determined. An initial particle swarm is generated based on the set of tour guide collaborative strategies corresponding to the doorways of ancient buildings; the fitness value of each particle is calculated based on the strength of evidence for yielding the threshold and the strength of evidence for refuting leaving the queue in the set of historical threshold event samples; the initial particle swarm is iteratively updated based on the fitness value to obtain the recommended tour guide collaborative strategy. By matching the strength of threshold evidence and the strength of disproving evidence for leaving the group, the recommended tour guide collaboration strategy is analyzed to obtain tour guide management instructions.

[0015] Compared with the prior art, the beneficial effects of the present invention are: 1. This invention establishes a coordinate system for the doorway of ancient buildings and combines the position of the threshold line, the coordinates of the copper-clad corners, and the local threshold crossing trajectory of tourists to conduct a detailed analysis of the lateral position changes of tourists in the doorway area. It can identify the continuous actions of tourists, such as sideways detours, close-to-the-edge passages, and return to the main passageway, caused by avoiding the threshold corners. This avoids misjudging normal threshold crossing behavior as real separation behavior and improves the accuracy of tourist status identification in the doorway area of ​​ancient buildings.

[0016] 2. This invention distinguishes and analyzes tourists' yielding and leaving behaviors in doorway areas by separately constructing evidence strength for yielding and evidence strength for leaving the group. It also identifies whether tourists have truly deviated from the main direction of the group by combining the directional deviation state inside the doorway and the lateral non-return state. This can reduce false alarms caused by tourists' short-term pauses or detours around thresholds and reduce the probability of tour guides receiving invalid management reminders in complex doorway areas.

[0017] 3. This invention uses particle swarm optimization to iteratively analyze the adaptation of different tour guide collaboration strategies in historical threshold event samples. Based on the sum of threshold evidence strength and the sum of disproving evidence strength, it outputs corresponding tour guide management instructions, which can match tour guide management actions with the current passage status of the doorway. This improves the efficiency of tour guide organization during the passage of groups in ancient building scenic areas and enhances the stability of tourist passage order in narrow doorway areas. Attached Figure Description

[0018] The accompanying drawings, which are included to provide a further understanding of the invention and form part of this application, illustrate exemplary embodiments of the invention and, together with their description, serve to explain the invention and do not constitute an undue limitation thereof. In the drawings: Figure 1 This is a functional module diagram of an AI tour guide management system for tourist attractions, provided as an embodiment of the present invention. Detailed Implementation

[0019] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0020] Example 1: This example provides an AI tour guide management system for tourist attractions. See [link to example]. Figure 1 Specifically, including: The coordinate construction module establishes a doorway coordinate system based on the structural data of the ancient building's doorway; based on the visitor location data collected by the visitor terminal device and the doorway coordinate system, it obtains local threshold segments and classifies the lateral positions at each collection time within the local threshold segments to obtain the lateral attribution status. In an embodiment of the present invention, establishing a doorway coordinate system includes: Obtain the structural data of the doorway of the ancient building; the structural data includes the coordinates of the center point of the doorway, the vector of the passage direction of the doorway, the position data of the threshold line, the coordinates of the left and right copper corner brackets; The doorway of an ancient building refers to the passage area for people set in the wall, courtyard enclosure structure or hall partition structure of an ancient building. When tourists pass through the doorway of an ancient building, their direction of movement is restricted by the width of the doorway, the boundary of the door frame and the position of the threshold, thus forming a passage state that moves along the length of the doorway. Structural data refers to the dataset used to characterize the spatial structural relationships of ancient building doorways. It describes the spatial distribution of various physical components within the doorway area and the positional relationship of visitors relative to these components. Doorway center point coordinates refer to the planar coordinate data of the geometric center of the doorway's passage area, representing the central reference position for visitors crossing the doorway. Doorway passage direction vector refers to the directional data established along the normal passage direction of visitors, indicating the main direction of movement when visitors enter the doorway from the outside. Threshold line position data refers to the extension position of the threshold on the ground, representing the spatial boundary when visitors cross the threshold. Left copper corner coordinates refer to the planar coordinate data corresponding to the location of the left copper corner of the threshold, a metal cladding structure fixed to the left edge of the threshold, with its surface higher than or close to the threshold edge. Right copper corner coordinates refer to the planar coordinate data corresponding to the location of the right copper corner of the threshold, a metal cladding structure fixed to the right edge of the threshold, its spatial position located on either side of the threshold. Data was collected on the plan view of the area where the ancient building's doorway was located, including the left and right boundaries of the door frame, the front and rear edges of the threshold, the left and right ends of the threshold, and the edges of the left and right brass corner brackets. The midpoint between the left and right boundaries of the door frame was determined as the center point of the doorway. The direction from the center point of the doorway to the middle of the rear edge of the threshold, when a visitor enters the doorway from the outside, was determined as the passage direction vector of the doorway. The threshold line position data is determined by combining the front edge position data and the rear edge position data. The left copper corner edge position data is determined by the point located at the left end of the threshold and close to the tourist passage area. The right copper corner edge position data is determined by the point located at the right end of the threshold and close to the tourist passage area. The coordinates of the center point of the doorway, the passage direction vector of the doorway, the threshold line position data, the left copper corner coordinates, and the right copper corner coordinates are unified according to the same plane coordinate reference to obtain the structural data of the ancient building doorway.

[0021] A coordinate system for the doorway is established based on the coordinates of the center point of the doorway and the direction vector of passage through the doorway in the structural data.

[0022] The doorway coordinate system refers to a spatial reference system established with the coordinates of the center point of the doorway as the origin and the vector of the passage direction of the doorway as the axis. The axis in the doorway coordinate system is used to represent the movement position of tourists along the length of the doorway, and the horizontal axis in the doorway coordinate system is used to represent the offset position of tourists relative to the center area of ​​the doorway towards the left or right corner.

[0023] Using the center point coordinates of the doorway as the origin of the doorway coordinate system, the passage direction vector of the doorway is normalized to obtain the axial unit vector. The axial unit vector is then rotated 90 degrees in the same plane to obtain the lateral unit vector. The origin, axial unit vector, and lateral unit vector are combined to form the doorway coordinate system. The threshold line position data, left brass corner coordinates, and right brass corner coordinates in the structural data are transformed into the doorway coordinate system to obtain the axial position of the threshold line, the lateral position of the left brass corner, and the lateral position of the right brass corner. The doorway coordinate system is then used as the spatial reference for converting the visitor position data into axial and lateral position sequences.

[0024] In embodiments of the present invention, obtaining the lateral attribution state includes: Acquire tourist location data collected by tourist terminal devices; As tourists move with the group, the tourist terminal device generates tourist location data according to the collection time. The tourist location data includes tourist identification, collection time, and planar position coordinates. The planar position coordinates are represented using the same planar coordinate reference as the doorway coordinate system. Multiple collection times corresponding to the same tourist identification are arranged in chronological order, and collection records lacking planar position coordinates are removed. The planar position coordinates corresponding to adjacent collection times are connected in chronological order to form continuous tourist movement trajectory data.

[0025] Based on tourist location data and the doorway coordinate system, generate axial and lateral position sequences for tourists; Tourist terminal equipment refers to data acquisition devices carried by tourists and used to collect tourists' spatial positions; tourist position data refers to the planar position coordinate data of tourists at multiple collection times; axial position sequence refers to the position arrangement data formed by the change of tourists along the passage direction of the doorway at continuous collection times, and the axial position sequence is used to represent the change of tourists' forward position as they approach the doorway, cross the threshold, and leave the doorway; lateral position sequence refers to the position arrangement data formed by the shift of tourists to the left or right relative to the center area of ​​the doorway at continuous collection times, and the lateral position sequence is used to represent the lateral movement state of tourists as they approach the left brass corner area, approach the right brass corner area, or are located in the center area of ​​the doorway. Using the origin of the doorway coordinate system as the position transformation reference, the axial unit vector of the doorway coordinate system as the axial projection direction, and the lateral unit vector of the doorway coordinate system as the lateral projection direction, the planar position coordinates of each tourist at each collection time are projected onto the axial unit vector and the lateral unit vector respectively, thus obtaining the axial position and lateral position of the tourist at each collection time. The axial positions of the same tourist are arranged in the order of collection time to form an axial position sequence, and the lateral positions of the same tourist are arranged in the order of collection time to form a lateral position sequence.

[0026] Based on the axial position sequence of tourists, continuous trajectory segments are extracted from the planar position coordinates of tourists to obtain local threshold segments; Planar position coordinates refer to the position data of tourists in a two-dimensional plane. Planar position coordinates include position data along the passage direction of the doorway and position data perpendicular to the passage direction of the doorway. Continuous trajectory segment extraction refers to extracting the continuous position change process of tourists entering the doorway and crossing the threshold from the continuous movement trajectory of tourists. Local threshold crossing segment refers to the continuous position trajectory data corresponding to the process of tourists entering the doorway from the outside to the inside of the doorway. Based on the threshold line position data in the doorway coordinate system, the axial position corresponding to the threshold is determined. The axial position sequence of each tourist is traversed in the order of collection time to determine the continuous collection time range from the outside to the inside of the threshold. This continuous collection time range is taken as the tourist's threshold crossing period. The planar position coordinates of the tourist during the threshold crossing period are extracted from the tourist position data. The extracted planar position coordinates are arranged in the order of collection time to obtain the tourist's local threshold crossing segment.

[0027] Based on the tourists' lateral position sequence, the coordinates of the left and right copper corners, the lateral positions at each collection time within the local threshold segment are classified to obtain the lateral attribution status.

[0028] Horizontal position classification refers to the process of dividing the horizontal position of tourists into areas close to the left corner, the center of the doorway, or the right corner, based on their position in the horizontal direction of the doorway. Horizontal affiliation status refers to the horizontal position category data of tourists at each collection time.

[0029] Transform the coordinates of the left and right braced corners to the doorway coordinate system to obtain the lateral positions of the left and right braced corners. Determine the lateral origin of the doorway coordinate system as the lateral position of the doorway center. Define the lateral range between the left braced corner lateral position and the doorway center lateral position as the left transition range, and the lateral range between the right braced corner lateral position and the doorway center lateral position as the right transition range. Use the doorway center lateral position as the boundary between the left and right transition ranges. Define the side of the left transition range closest to the left braced corner lateral position as the left braced corner side. Within the lateral transition range, the side closest to the right brass corner is defined as the right brass corner side. The portion of the left and right transition ranges that is close to the center of the doorway is defined as the center of the doorway side. The lateral position of the tourist at each collection moment within the local threshold segment is compared with the lateral positions of the left brass corner, the center of the doorway, and the right brass corner. Based on the comparison results, the tourist's lateral position is assigned to the left brass corner side, the center of the doorway side, or the right brass corner side. The classification results obtained at each collection moment are arranged in chronological order to obtain the tourist's lateral affiliation status.

[0030] The threshold evidence collection module generates corner-attaching trend state, threshold-around-corner state, and return queue state based on the horizontal attribution state, and determines the threshold evidence strength. In embodiments of the present invention, determining the strength of threshold evidence includes: Based on the lateral attribution status, the local threshold segment is segmented to obtain the segment outside the doorway, the segment near the threshold, and the segment inside the doorway. Segmented processing refers to the process of dividing a local threshold crossing segment of the same tourist into different continuous trajectory parts according to the tourist's position relative to the threshold line. This is used to separately represent the tourist's movement process when approaching the threshold, being near the threshold, and leaving the threshold. The segment outside the doorway refers to the continuous position data formed when the tourist has not yet crossed the threshold line and is still in the area outside the doorway. The segment near the threshold refers to the continuous position data formed when the tourist reaches the area where the threshold line is located and performs the action of crossing the threshold. The segment inside the doorway refers to the continuous position data formed when the tourist crosses the threshold line and enters the area inside the doorway of the ancient building. The lateral attribution status corresponding to each collection time within a local threshold segment is arranged in chronological order of collection time. The threshold line position data is converted to the doorway coordinate system, and the axial position corresponding to the threshold line is obtained. The continuous collection time range where the axial position in the local threshold segment is outside the threshold line is determined as the doorway outer time period. The continuous collection time range where the axial position in the local threshold segment coincides with or crosses the threshold line is determined as the threshold adjacent time period. The continuous collection time range where the axial position in the local threshold segment is inside the threshold line is determined as the doorway inner time period. The tourist plane position coordinates and lateral attribution status corresponding to the doorway outer time period are extracted from the local threshold segment to obtain the doorway outer segment. The tourist plane position coordinates and lateral attribution status corresponding to the threshold adjacent time period are extracted from the local threshold segment to obtain the threshold adjacent segment. The tourist plane position coordinates and lateral attribution status corresponding to the doorway inner time period are extracted from the local threshold segment to obtain the doorway inner segment.

[0031] Determine the corner-attaching tendency state based on the lateral attribution state of the segment outside the doorway; The lateral attribution states corresponding to each collection time within the segment outside the doorway are arranged in chronological order. The arranged lateral attribution states are then associated with their corresponding lateral positions at the collection times. The continuous lateral attribution states near the starting position of the segment adjacent to the threshold are taken as the end state sequence outside the doorway. The change order of the lateral attribution states in the end state sequence outside the doorway is compared moment by moment. When the lateral attribution state changes from the center side of the doorway to the left copper corner side in the continuous collection time, it is determined that the tourist is moving towards the left copper corner in the area outside the doorway. When the lateral attribution state changes from the center side of the doorway to the right copper corner side in the continuous collection time, it is determined that the tourist is moving towards the right copper corner in the area outside the doorway. The lateral state change process corresponding to the tourist moving towards the left copper corner or the right copper corner in the area outside the doorway is determined as the corner-adhering tendency state. The start collection time, end collection time, and corresponding lateral attribution state change order corresponding to the corner-adhering tendency state are recorded.

[0032] Determine the threshold-around-angle state based on the lateral attribution state of the segments adjacent to the threshold; Arrange the lateral attribution states corresponding to each collection time within the segment adjacent to the threshold in chronological order of collection time. Associate the arranged lateral attribution states with the lateral positions corresponding to each collection time within the segment adjacent to the threshold. Perform continuous state traversal on the lateral attribution state change process in the segment adjacent to the threshold. When the lateral attribution state changes from the left copper corner side to the center of the doorway and continues to change back to the left copper corner side in the continuous collection time, it is determined that the tourist has a left corner movement process in the area adjacent to the threshold. When the lateral attribution state changes from the right copper corner side to the center of the doorway and continues to change back to the right copper corner side in the continuous collection time, it is determined that the tourist has a right corner movement process in the area adjacent to the threshold. The lateral state change process corresponding to the left or right corner movement process is determined as the threshold corner movement state, and the continuous collection time range, lateral attribution state change order, and corresponding lateral position change process corresponding to the threshold corner movement state are recorded.

[0033] The regression queue status is determined based on the lateral attribution status of the segment inside the doorway; The "corner-approaching state" refers to the state in which a tourist's lateral belonging state in the segment outside the doorway moves towards the left or right brass corner; the "leaning around the corner state" refers to the state in which a tourist's lateral belonging state in the segment near the threshold moves from the side near the brass corner to the center of the doorway and continues to complete the threshold movement; the "returning to the queue state" refers to the state in which a tourist's lateral belonging state in the segment inside the doorway returns from the left or right brass corner to the center of the doorway. The lateral attribution states corresponding to each collection time within the inner segment of the doorway are arranged in chronological order of collection time. The arranged lateral attribution states are then associated with the lateral positions corresponding to each collection time within the inner segment of the doorway. Continuous state comparisons are performed on the changes in lateral attribution states within the inner segment of the doorway. When the lateral attribution state changes from the left copper corner side to the center of the doorway in continuous collection time, it is determined that the tourist is returning from the left side of the doorway to the center area. When the lateral attribution state changes from the right copper corner side to the center of the doorway in continuous collection time, it is determined that the tourist is returning from the right side of the doorway to the center area. The lateral state change process corresponding to the tourist returning from the left side of the doorway to the center area or from the right side of the doorway to the center area is determined as the return queue state. The range of continuous collection times, the order of lateral attribution state changes, and the corresponding lateral position change process corresponding to the return queue state are recorded.

[0034] Based on the corner-touching tendency state, the threshold-around-corner state, and the return queue state, generate a set of threshold-yielding evidence items for tourists; The threshold evidence item set refers to the data set generated from the corner-approaching state, the threshold-around-the-corner state, and the regression queue state; the evidence item refers to the data item in the threshold evidence item set used to record the existence of a single state. When the corner-hugging tendency state has been determined, extract the tourist identifier, start collection time, end collection time, outer segment identifier of the doorway, lateral attribution state change sequence, and lateral position change process corresponding to the corner-hugging tendency state to generate corner-hugging tendency evidence entries. When the threshold-around-corner state has been determined, extract the tourist identifier, continuous collection time range, threshold-adjacent segment identifier, lateral attribution state change sequence, and lateral position change process corresponding to the threshold-around-corner state to generate threshold-around-corner evidence entries. When the regression queue state has been determined, extract the tourist identifier, continuous collection time range, inner segment identifier of the doorway, lateral attribution state change sequence, and lateral position change process corresponding to the regression queue state to generate regression queue evidence entries. Group the corner-hugging tendency evidence entries, threshold-around-corner evidence entries, and regression queue evidence entries into the same data set according to the tourist identifier to obtain the threshold-around-corner evidence entry set for that tourist.

[0035] The strength of the threshold evidence is determined based on the number of evidence items in the threshold evidence set.

[0036] The strength of evidence for yielding the threshold refers to the data obtained from the number of evidence items in the set of evidence items for yielding the threshold, which is used to represent the number of states corresponding to the behavior of yielding the threshold during the tourist's crossing of the threshold.

[0037] Traverse the set of threshold-crossing evidence entries corresponding to the same tourist identifier, count the corner-touching evidence entries, threshold-crossing-around-corner evidence entries, and regression queue evidence entries as valid evidence entries, and obtain the number of evidence entries. Determine the number of evidence entries as the threshold-crossing evidence strength of the tourist. Establish a correspondence between the tourist identifier, the local threshold-crossing segment identifier, the threshold-crossing evidence entry set, the number of evidence entries, and the threshold-crossing evidence strength to obtain the threshold-crossing evidence strength data corresponding to the local threshold-crossing segment of the tourist.

[0038] It should be noted that when tourists cross the threshold of ancient buildings, their physical movement is affected by the threshold height, the position of the copper corner, and the lateral passage space of the doorway. To avoid contact between the soles of their shoes, toes, or sides of their bodies and the copper corner, tourists will undergo a continuous spatial movement process, moving towards the copper corner, going around the edge of the copper corner, and returning to the center of the doorway. This continuous spatial movement process corresponds to the obstacle avoidance movement pattern of the human body in a restricted passage space. Since the human body will continuously undergo lateral position shift, lateral position rotation, and lateral position return during the obstacle avoidance movement, the corner-touching state, the threshold-yielding and corner-circling state, and the return queue state correspond to the approaching stage, the bypassing stage, and the recovery stage of the human body's obstacle avoidance movement process, respectively. When the number of states in the three stages increases, it indicates that tourists have undergone more spatial position changes related to copper corner avoidance within the threshold area, indicating that tourists have a more obvious shoulder-tilting behavior to yield the threshold. Therefore, determining the strength of the threshold yielding evidence based on the number of evidence items in the threshold yielding evidence set can reflect the degree of continuous obstacle avoidance movement of tourists within the threshold area due to copper corner avoidance.

[0039] The module for refuting the departure from the group continuously identifies the lateral state of tourists in the segment inside the doorway to generate the first item for refuting the departure from the group; based on the guide's position data and local threshold segments collected by the guide terminal device, it identifies the direction of tourists' movement inside the doorway to generate the second item for refuting the departure from the group; and based on the first and second items for refuting the departure from the group, it determines the strength of the refuting the departure from the group. In an embodiment of the present invention, generating a first off-line counter-evidence entry includes: Based on the local threshold segments and lateral attribution status, extract the data on the changes in the lateral attribution status of tourists in the segment inside the doorway. Lateral attribution status change data refers to the change record of the lateral attribution status of tourists at each collection time in the segment inside the doorway, formed in chronological order. It is used to represent the lateral position change process of tourists after entering the inside of the doorway between the left brass corner side, the center side of the doorway, and the right brass corner side. Based on the segment inside the doorway that has been divided in the local threshold segment, the start time, end time, and corresponding plane position coordinates of tourists are obtained from the segment inside the doorway. The lateral attribution status records between the start time and the end time are extracted from the lateral attribution status. Each lateral attribution status record is mapped to the plane position coordinates of tourists at the same time. The mapped data are arranged in chronological order of the time of collection to form the lateral attribution status change data of tourists in the segment inside the doorway. The lateral attribution status change data includes the time of collection, plane position coordinates of tourists, lateral position, lateral attribution status, and the order of status arrangement.

[0040] Based on the data on changes in lateral attribution status, continuous lateral attribution status identification is performed on the tourists' lateral attribution status in the segment inside the doorway to obtain the lateral non-regression status. Continuous lateral state recognition refers to the process of continuously comparing the lateral attribution state change data in the segment inside the doorway to determine whether the tourist continues to stay on the left or right brass corner side; lateral non-return state refers to the state in which the tourist, after entering the inside of the doorway, still remains on the left or right brass corner side and does not return to the center side of the doorway. The lateral attribution status change data is traversed item by item according to the status arrangement order. Adjacent lateral attribution statuses are compared. When both the previous and next lateral attribution statuses are on the left copper corner side, the corresponding continuous status record is assigned to the left continuous lateral record. When the center side of the doorway does not appear in the continuous status record, a left lateral non-returning status is generated. When both the previous and next lateral attribution statuses are on the right copper corner side, the corresponding continuous status record is assigned to the right continuous lateral record. When the center side of the doorway does not appear in the continuous status record, a right lateral non-returning status is generated. The left or right lateral non-returning status is taken as the tourist's lateral non-returning status, and the corresponding continuous status record, start collection time, end collection time, and lateral attribution direction are retained.

[0041] It should be noted that the left brass corner side refers to the area in the horizontal region of the doorway that is close to the left brass corner. This area is located between the center of the doorway and the left brass corner. When a visitor is on the left brass corner side, their horizontal position is offset from the center of the doorway towards the left brass corner. Furthermore, the spatial distance between the visitor and the left brass corner during the process of passing through the threshold is less than the spatial distance between the visitor and the center of the doorway.

[0042] Based on the lateral non-regression state, generate the first off-line counter-evidence entry.

[0043] The first departure counter-evidence entry refers to a data entry generated based on the lateral non-regression state, used to record the spatial state of tourists who remain in the lateral area of ​​the doorway after passing through the threshold.

[0044] After generating the lateral non-return state, the tourist identifier, local threshold segment identifier, doorway inner segment identifier, start collection time, end collection time, lateral attribution direction, continuous state record, lateral attribution state change data, and corresponding tourist plane position coordinates corresponding to the lateral non-return state are written into the same data entry. This data entry is marked as the first departure counter-evidence entry, and the first departure counter-evidence entry is associated with and saved with the tourist's local threshold segment.

[0045] In embodiments of the present invention, determining the strength of off-line counter-evidence includes: Acquire tour guide location data collected by the tour guide terminal device; The tour guide terminal device refers to the data acquisition device carried by the tour guide and moving with the tour guide. This device can form a data record of the tour guide's plane position at multiple acquisition moments. The tour guide position data refers to the set of position data generated by the tour guide terminal device at continuous acquisition moments. This set is used to represent the changes in the tour guide's position as he moves near the doorway. The tour guide terminal device generates tour guide location data according to the collection time. The tour guide location data includes tour guide identification, collection time, and tour guide plane position coordinates. The tour guide plane position coordinates are represented using the same plane coordinate reference as the local threshold segment. The tour guide plane position coordinates with the same tour guide identification are arranged in chronological order of collection time to obtain continuous tour guide location data. The continuous tour guide location data is used as the data object for tour guide direction extraction.

[0046] Extract the planar position coordinates of the tour guides corresponding to the local threshold segments from the tour guide position data; The guide position records corresponding to the local threshold segments are arranged in chronological order of collection time. The coordinates of two guide plane positions corresponding to adjacent collection times are connected to form guide position change line segments. The directional change relationship of the guide position change line segments relative to the axial direction of the doorway coordinate system is calculated. The directional change result corresponding to each guide position change line segment is taken as a guide direction record. All guide direction records are arranged in chronological order of collection time to obtain guide travel direction data. The guide travel direction data includes collection time, guide position change line segments, directional change results, and corresponding guide plane position coordinates.

[0047] Based on the tour guide's planar position coordinates, generate the tour guide's travel direction data; The tour guide's planar position coordinates refer to the tour guide's position coordinates under the same planar coordinate reference. These coordinates can be spatially correlated with the tourist's planar position coordinates, the center point coordinates of the doorway, and the position data of the threshold line. The tour guide's travel direction data refers to the direction data formed by the change of the tour guide's planar position coordinates at adjacent collection times. This direction data is used to indicate the direction of the tour guide's movement from one position to the next. The guide direction data is continuously traversed, and the guide direction record corresponding to the current collection time is compared with the guide direction record corresponding to the next collection time. When the direction change results of adjacent collection times remain continuous, the corresponding guide direction records are grouped into the same guide continuous travel segment. All guide continuous travel segments are arranged in the order of collection time to obtain guide continuous travel direction data corresponding to local threshold segments. The guide continuous travel direction data is then associated and saved with the corresponding local threshold segments.

[0048] Based on local threshold segments, generate tourist movement direction data; Tourist movement direction data refers to the direction data formed by the change of the tourist's planar position coordinates at adjacent collection times within a local threshold segment. This direction data is used to indicate the direction of the tourist's movement as they pass through the threshold and enter the inside of the doorway. Based on the planar position coordinates of tourists at each collection time in the local threshold segment, the planar position coordinates of the same tourist at adjacent collection times are paired according to the order of collection time. The coordinates of the previous collection time in each pair of adjacent tourist planar position coordinates are taken as the starting point of the line segment and the coordinates of the next collection time are taken as the ending point of the line segment to generate a tourist position change line segment. The tourist position change line segment is transformed into the doorway coordinate system, and the axial component of the tourist position change line segment in the axial direction and the lateral component in the lateral direction of the doorway coordinate system are obtained. The travel direction of the tourist in the corresponding collection time interval is determined according to the axial component and the lateral component. The tourist position change line segment, axial component, lateral component and travel direction corresponding to each collection time interval are associated to obtain the tourist direction record. All tourist direction records are arranged according to the order of collection time to form tourist travel direction data.

[0049] Based on the data of tourists' walking direction and the data of the tour guide's walking direction, the deviation state of tourists' walking direction inside the doorway is identified, and the deviation state is obtained. The state of direction deviation refers to the situation where, after entering the inside of the doorway, the direction in which the tourists move deviates from the direction in which the tour guide is moving, and this deviation is manifested in the state in which the tourists no longer continue to move in the direction led by the tour guide. The tourist direction records in the tourist movement direction data are matched with the tour guide direction records in the tour guide movement direction data according to the collection time. The matched tourist position change segments and tour guide position change segments are transformed to the same doorway coordinate system. The axial components of the tourist position change segments and tour guide position change segments are compared for directional similarity. The lateral components of the tourist position change segments and tour guide position change segments are compared for lateral separation. The component changes of the tourist position change segments relative to the passage direction of the doorway center are compared for distance. When the tourist position change segments and tour guide position change segments are not in the same direction, or the tourist position change segments show a lateral separation relationship relative to the tour guide position change segments, or the tourist position change segments show a distance relationship relative to the passage direction of the doorway center, the tourist direction records in the corresponding collection time interval are identified as direction deviation records. The direction deviation records are collected in chronological order of collection time, and the tourist identifier, local threshold segment identifier, tourist position change segment, and tour guide position change segment corresponding to the direction deviation records are associated to obtain the direction deviation state.

[0050] Based on the direction divergence state, generate a second disjointed counter-evidence entry; The second counter-evidence entry for leaving the group refers to the data entry generated after the direction deviation state is determined. This data entry records tourist identification, local threshold segment identification, doorway inner segment identification, tourist travel direction data, tour guide travel direction data, and direction deviation state. Once the direction deviation state has been determined, extract the tourist identifier, local threshold segment identifier, doorway inner segment identifier, direction deviation record, tourist position change line segment, tour guide position change line segment, tourist travel direction data, and tour guide travel direction data corresponding to the direction deviation state. Write the extracted data into the same data entry to obtain the second departure counter-evidence entry, and associate and save the second departure counter-evidence entry with the corresponding tourist's local threshold segment.

[0051] Perform item association processing on the first and second off-listed counter-evidence items to obtain the set of off-listed counter-evidence items; Item association processing refers to the process of grouping the first and second off-line counter-evidence items into the same data set corresponding to the same tourist based on the tourist identifier, the local threshold segment identifier, and the inner side of the doorway segment identifier; the off-line counter-evidence item set refers to the set of the first and second off-line counter-evidence items formed by the same tourist in the same local threshold segment; Extract the tourist identifier, partial threshold segment identifier, and doorway inner segment identifier from the first departure counter-evidence entry. Extract the tourist identifier, partial threshold segment identifier, and doorway inner segment identifier from the second departure counter-evidence entry. Group the first departure counter-evidence entries and the second departure counter-evidence entries with the same tourist identifier, the same partial threshold segment identifier, and the same doorway inner segment identifier into the same data set to obtain the departure counter-evidence entry set corresponding to the tourist.

[0052] The strength of an off-staff counter-evidence is determined by the number of counter-evidence entries in the off-staff counter-evidence set.

[0053] A counter-evidence item refers to a single data item in the set of counter-evidence items for leaving the group, used to record the spatial state related to tourists leaving the group. The counter-evidence strength for leaving the group refers to the degree of spatial deviation exhibited by tourists after passing through the archway of the ancient building and entering the inner area of ​​the archway, relative to the direction the tour guide is leading and the central passage area of ​​the archway. This degree of spatial deviation is used to represent the cumulative state of tourists continuously staying in the lateral area, not returning to the central passage area of ​​the archway, and deviating from the direction of movement of the tour guide in the inner area of ​​the archway. The higher the counter-evidence strength for leaving the group, the more the movement state of tourists in the inner area of ​​the archway deviates from the movement state of the tour guide leading the group, and the weaker the spatial consistency between tourists and the main passage direction of the group.

[0054] Iterate through each counter-evidence entry in the set of exit-from-the-group counter-evidence entries, count the first and second exit-from-the-group counter-evidence entries as counter-evidence entries respectively, and obtain the number of counter-evidence entries. Determine the number of counter-evidence entries as the exit-from-the-group counter-evidence intensity corresponding to the tourist. Establish a correspondence between the tourist identifier, the local threshold segment identifier, the doorway inner segment identifier, the set of exit-from-the-group counter-evidence entries, the number of counter-evidence entries, and the exit-from-the-group counter-evidence intensity, and obtain the exit-from-the-group counter-evidence intensity data corresponding to the local threshold segment of the tourist.

[0055] It should be noted that if tourists continue to linger in the lateral area of ​​the archway or their movement direction continues to deviate from the guide's direction after passing through the archway, it indicates that the tourists' spatial movement is no longer in accordance with the overall movement trajectory of the group. Since tourists usually return to the central passage area of ​​the archway after passing through the threshold and continue to move along the guide's direction during normal group passage, the lateral non-return state and the directional deviation state correspond to the tourists' deviation behavior in lateral spatial position and direction of movement, respectively. When the number of counter-evidence items for leaving the group increases, it indicates that the tourists' spatial deviation behavior after passing through the threshold continues to accumulate, and the consistency of movement between the tourists and the overall movement direction of the group gradually weakens, indicating that the tourists are more likely to be in a state of deviating from the normal passage trajectory of the group. Therefore, determining the strength of the counter-evidence item for leaving the group based on the number of counter-evidence items in the set of counter-evidence items for leaving the group can reflect the degree to which tourists deviate from the main passage behavior of the group within the archway area.

[0056] The strategy optimization module generates an initial particle swarm based on the set of tour guide collaborative strategies corresponding to the doorways of ancient buildings; it calculates the fitness value of each particle based on the strength of evidence for yielding the threshold and the strength of evidence for refuting leaving the group in the set of historical threshold event samples; and iteratively updates the initial particle swarm based on the fitness value to obtain the recommended tour guide collaborative strategy. In an embodiment of the present invention, generating an initial particle swarm includes: Obtain the set of tour guide collaboration strategies corresponding to the doorways of ancient buildings; The guide coordination strategy set refers to the data set of guide organizational behavior established for the process of tourists passing through the threshold of the same ancient building. Different guide coordination strategies in the guide coordination strategy set correspond to different positions of the guide in the threshold area, different prompting actions, different directions of group guidance, and different ways of organizing tourists to pass through. The guide coordination strategies in the guide coordination strategy set are arranged according to the structure of the threshold, the position of the threshold, and the spatial distribution of the brass corners. The guide coordination strategy refers to the specific organizational action data performed by the guide during the process of tourists passing through the ancient building threshold. This organizational action data includes the guide's movement direction, stopping area, prompting direction, and group guidance method in the threshold area. Based on the structure of the doorway, the position of the threshold, the position of the left and right brass corners, and the direction of tourist passage of the same ancient building, the possible passage organization actions that tour guides can perform at the doorway of the ancient building are compiled. The tour guides standing on the outside of the doorway to guide tourists through, the tour guides standing on the inside of the doorway to receive tourists through, the tour guides prompting tourists to pass through the center of the doorway, the tour guides prompting tourists to avoid the left and right brass corners, the tour guides organizing tourists to pass through in a single line, and the tour guides organizing tourists to pass through in batches are respectively regarded as tour guide coordination strategies. Each tour guide coordination strategy is categorized into the data set corresponding to the same ancient building doorway, resulting in the tour guide coordination strategy set corresponding to the ancient building doorway.

[0057] Each tour guide collaboration strategy in the set of tour guide collaboration strategies is numbered to obtain the strategy number; The strategy number refers to the sequential identifier data generated according to the order of the tour guide collaboration strategy in the tour guide collaboration strategy set. Different strategy numbers correspond to different tour guide collaboration strategies. According to the order of the guide coordination strategies in the guide coordination strategy set, each guide coordination strategy is assigned a different strategy number. The strategy number is then associated with the corresponding guide coordination strategy to form a strategy number record. The strategy number record includes the ancient building doorway sign, strategy number, guide positioning content, guide prompts content, and team organization content. All strategy number records are saved in the order of the strategy numbers to obtain the strategy number corresponding to each guide coordination strategy in the guide coordination strategy set.

[0058] Use the strategy number as the initial particle position; A particle refers to a data object in the particle swarm optimization algorithm that represents a candidate guide cooperative strategy. Each particle corresponds to a strategy number and the position change state associated with that strategy number. The initial particle position refers to the position of the strategy number corresponding to the particle before the start of the particle swarm optimization algorithm iteration. This position is used to indicate the position of the guide cooperative strategy currently represented by the particle in the set of guide cooperative strategies. The strategy number corresponding to each tour guide collaboration strategy in the tour guide collaboration strategy set is used as the initial particle position of a particle. The ancient building doorway identifier, strategy number, tour guide position content, tour guide prompt content, and team organization content recorded in the strategy number are used as the initial particle position association data of the particle. The correspondence between each particle and the corresponding tour guide collaboration strategy is established to obtain the initial particle position data represented by the strategy number.

[0059] Determine the adjacent strategy numbers of each particle based on the order of the strategy numbers; Adjacent strategy numbers refer to the strategy numbers that are located before or after the current strategy number in the strategy number arrangement order. Adjacent strategy numbers are used to indicate the movable direction of a particle in the guide cooperative strategy set. Sort the strategy numbers according to the order of arrangement. Perform position retrieval on the strategy number corresponding to each particle to obtain the strategy number before and after the strategy number in the arrangement order. Determine the strategy number before and after the strategy number as the adjacent strategy number of the particle. When the strategy number is located at the beginning of the arrangement order, the strategy number after the strategy number is determined as the adjacent strategy number of the particle. When the strategy number is located at the end of the arrangement order, the strategy number before the strategy number is determined as the adjacent strategy number of the particle.

[0060] The initial particle velocity is determined based on the initial particle position and the adjacent strategy numbers. Initial particle velocity refers to the position change data corresponding to the particle's change from the current strategy number to the adjacent strategy number. This position change data is used to represent the particle's movement direction and movement range in the guide cooperative strategy set. Based on the strategy number corresponding to the initial particle position of each particle and the adjacent strategy numbers of the particle, the difference between the adjacent strategy numbers and the strategy number corresponding to the initial particle position is calculated. The difference is used as the number change of the particle. When the adjacent strategy number is greater than the strategy number corresponding to the initial particle position, the number change is determined as a positive change. When the adjacent strategy number is less than the strategy number corresponding to the initial particle position, the number change is determined as a negative change. The number change is associated with the corresponding number change direction to obtain the initial particle velocity, and the initial particle velocity is written into the particle data record of the corresponding particle.

[0061] An initial particle swarm is generated based on the initial particle position and initial particle velocity of each particle.

[0062] The initial particle swarm refers to a data set consisting of multiple particles according to their respective initial particle positions and initial particle velocities. This data set is used to represent the overall distribution state of multiple candidate guide cooperative strategies in the initial stage of the particle swarm algorithm.

[0063] Extract the initial particle position, initial particle velocity, corresponding strategy number, corresponding guide cooperation strategy, adjacent strategy number, and ancient building doorway identifier for each particle. Assign the initial particle position to the initial particle velocity. Combine the particle data records with different initial particle positions and different initial particle velocities according to the order of strategy numbers. Put all the combined particle data records into the same data set to obtain the initial particle swarm.

[0064] In embodiments of the present invention, a recommended tour guide collaboration strategy is obtained, including: Obtain a sample set of historical threshold events corresponding to the doorways of ancient buildings; The historical threshold event sample set refers to the data set formed by the same ancient building doorway during the passage of previous groups. This data set includes the strength of evidence for letting the threshold pass each time the group passes the threshold, the strength of evidence for leaving the group to refute it, the tour guide's coordination strategy, and the passing result. Based on the ancient building doorway markings, data on threshold events formed when past groups passed through the ancient building doorway are collected. Each threshold event data includes the ancient building doorway marking, group marking, tourist marking, partial threshold crossing segment marking, strength of evidence of yielding, strength of evidence of leaving the group, implemented tour guide coordination strategies, and tour guide management results. Threshold event data with the same ancient building doorway marking are grouped into the same data set to obtain a historical threshold event sample set corresponding to the ancient building doorway.

[0065] By labeling the strength of evidence for yielding the threshold and the strength of evidence for refusal for leaving the queue in the historical threshold event sample set, a sample set for AI strategy optimization is obtained. Sample status labeling refers to the process of classifying and recording the pass status corresponding to historical threshold event samples based on the strength of threshold evidence and the strength of disproving evidence; the AI ​​strategy optimization sample set refers to the historical threshold event sample set after sample status labeling is completed, which is used to provide sample basis for the optimization selection of tour guide collaborative strategies. After receiving the historical threshold event sample set, the AI ​​strategy optimization unit aggregates the historical threshold event samples according to the ancient building doorway identifier and the time period of team passage. It extracts the yielding evidence strength and the departure counter-evidence strength in pairs for each historical threshold event sample, forming sample evidence pairs. The AI ​​strategy optimization unit identifies the dominant relationship of the execution evidence in the sample evidence pairs. When the yielding evidence strength is greater than or equal to the departure counter-evidence strength, the historical threshold event sample is marked as a yielding state sample; when the yielding evidence strength is less than the departure counter-evidence strength, it is marked as a departure state sample. The AI ​​strategy optimization unit writes the sample state labeling result, yielding evidence strength, departure counter-evidence strength, the tour guide collaboration strategy corresponding to the historical threshold event sample, the tour guide handling result corresponding to the historical threshold event sample, and the ancient building doorway identifier into the same sample record. The completed sample record is used as the AI ​​strategy optimization sample, and all AI strategy optimization samples corresponding to the same ancient building doorway are aggregated to obtain the AI ​​strategy optimization sample set.

[0066] It should be noted that the actual passage status of tourists in the archways of ancient buildings is usually characterized by a relative dominance between yielding behavior and leaving the group behavior. Since the strength of evidence for yielding indicates the degree of continuous obstacle avoidance movement caused by tourists avoiding the corner, and the strength of evidence for leaving the group indicates the degree of continuous deviation of tourists from the guide's direction and the central passage area of ​​the archway, when the strength of evidence for yielding is greater than or equal to the strength of evidence for leaving the group, it indicates that the tourists' spatial movement is mainly influenced by the corner avoidance behavior, and their movement trajectory still follows the overall direction of the group. In this case, labeling the threshold event data as yielding reflects that the tourists are engaging in normal obstacle avoidance behavior. When the strength of evidence for leaving the group is greater than the strength of evidence for yielding, the situation changes. According to the intensity, it indicates that tourists continuously deviate from the overall movement trajectory of the group after passing through the doorway. Their spatial movement is no longer mainly dominated by threshold avoidance behavior. At this time, the threshold event data is marked as a departure state, which can reflect that tourists have a tendency to deviate from the group's movement direction. By associating the sample state labeling results with the ancient building doorway markers, group markers, tourist markers, local threshold crossing segment markers, threshold avoidance evidence strength, departure disproving evidence strength, and implemented tour guide coordination strategies, a data correspondence relationship can be formed that simultaneously includes the doorway spatial structure, tourist movement state, and tour guide organizational actions. This provides a historical sample basis with spatial behavior correspondence for the subsequent optimization selection of tour guide coordination strategies.

[0067] Based on the AI ​​strategy, the sample set is optimized, and the historical samples of the guide cooperative strategy corresponding to each initial particle position in the initial particle swarm are replayed to obtain the mismatch count. Historical sample playback refers to the process of applying a certain tour guide collaboration strategy to historical threshold event samples and forming an evaluation result based on the correspondence between the treatment type corresponding to the tour guide collaboration strategy and the sample status labeling result; mismatch count refers to the number of times the treatment type corresponding to the tour guide collaboration strategy does not match the sample status labeling result of the historical threshold event sample. For each initial particle position in the initial particle swarm, the guide cooperative strategy corresponding to that initial particle position is extracted. The guide cooperative strategy is mapped to each AI strategy optimization sample in the AI ​​strategy optimization sample set. The disposal type corresponding to the guide cooperative strategy is compared with the sample state labeling result in the AI ​​strategy optimization sample. When the disposal type corresponding to the guide cooperative strategy is threshold yield disposal and the sample state labeling result is a sample in the out-of-line state, a mismatch is recorded. When the disposal type corresponding to the guide cooperative strategy is out-of-line disposal and the sample state labeling result is a threshold yield state sample, a mismatch is recorded. The number of mismatches formed by the same guide cooperative strategy in the AI ​​strategy optimization sample set is accumulated to obtain the mismatch count corresponding to the initial particle position.

[0068] Calculate the fitness value of the particle based on the mismatch count; Specifically, extract the mismatch count corresponding to the particle, count the total number of samples in the AI ​​strategy optimization sample set, calculate the matching ratio, matching ratio = (total number of samples - mismatch count) / total number of samples; determine the matching ratio as the fitness value of the particle, and write the particle position, the guide cooperative strategy corresponding to the particle, the mismatch count, the matching count, the total number of samples, the matching ratio and the fitness value into the same particle fitness record.

[0069] It should be noted that the fitness value corresponding to a particle is used to represent the degree of matching between the particle's corresponding guide cooperative strategy and historical threshold event samples. In the same ancient building doorway, the total number of historical threshold event samples represents the total number of historical passage environments corresponding to the guide cooperative strategy. The mismatch count represents the number of inconsistencies between the guide cooperative strategy and the historical passage results. According to the frequency relationship in statistics, when the proportion of the matching count to the total number of samples is higher, it indicates that the guide cooperative strategy has a wider range of applicability in the historical passage environment and is closer to the actual passage behavior in historical threshold events. Therefore, dividing the matching count by the total number of samples to obtain the matching ratio can represent the effective fit of the particle's corresponding guide cooperative strategy in all historical threshold event samples. The higher the matching ratio, the stronger the actual adaptability of the particle's corresponding guide cooperative strategy in the ancient building doorway. Therefore, the matching ratio is determined as the fitness value corresponding to the particle.

[0070] Based on fitness values, determine the global optimal particle position and the individual optimal particle position; Fitness value refers to the data used to represent the quality of the guide cooperative strategy corresponding to a particle. Fitness value corresponds to mismatch count. The lower the mismatch count, the more suitable the guide cooperative strategy corresponding to the particle is for the historical passage of the ancient building's doorway. Global optimal particle position refers to the strategy number position corresponding to the particle with the best fitness value in the current particle swarm. Individual optimal particle position refers to the strategy number position with the best fitness value that a single particle has reached during the iteration process. Based on the fitness records of each particle, the fitness value obtained by the same particle in the current replay process is compared with the fitness value obtained by the particle in history. The particle position corresponding to the highest fitness value obtained by the particle in history is determined as the individual optimal particle position. The fitness values ​​of all particles in the initial particle swarm are compared, and the particle position corresponding to the highest fitness value among all particles is determined as the global optimal particle position. The individual optimal particle position and the global optimal particle position are used as the basis for particle iterative update.

[0071] Based on the global optimal particle position and the individual optimal particle position, the initial particle swarm is iteratively updated to obtain the particle swarm after the iteration. Iterative update refers to the process of adjusting particle position and particle velocity based on the global optimal particle position and the individual optimal particle position; the particle swarm after iteration refers to the set of particles formed after the particle swarm completes the iterative update. For each particle in each iteration, extract its current particle position, current particle velocity, individual optimal particle position, and global optimal particle position. Use the difference in index between the individual optimal particle position and the current particle position as the individual guiding change, and the difference in index between the global optimal particle position and the current particle position as the global guiding change. Sum the current particle velocity, individual guiding change, and global guiding change to obtain the updated particle velocity. Sum the current particle position and the updated particle velocity to obtain the updated particle position. Map the updated particle position to the strategy number in the guide coordination strategy set to obtain the particle data record after this iteration. Use the updated particle data record of each iteration as the data object for the next iteration until all particles have completed the iteration update, resulting in the particle swarm after the iteration ends.

[0072] The recommended guide cooperative strategy is obtained by selecting guide cooperative strategies from the set of guide cooperative strategies after the iteration is completed by using the particle swarm optimization.

[0073] Recommended tour guide collaboration strategies refer to the tour guide organization action data selected from the set of tour guide collaboration strategies for the current ancient building archway.

[0074] After the iteration, extract the particle position and fitness value of each particle in the particle swarm. Arrange the particles in descending order of fitness value and determine the position of the particle with the highest fitness value as the recommended particle position. Based on the strategy number corresponding to the recommended particle position, extract the corresponding guide cooperative strategy from the guide cooperative strategy set and determine the extracted guide cooperative strategy as the recommended guide cooperative strategy.

[0075] The instruction generation module matches the recommended tour guide collaboration strategy with the instruction type by considering the threshold evidence strength and the disproving evidence strength of leaving the queue, thus obtaining tour guide management instructions.

[0076] In an embodiment of the present invention, obtaining tour guide management instructions includes: The strength of the evidence for yielding the threshold for each tourist in the group is summed to obtain the total strength of the evidence for yielding the threshold. The total strength of evidence of yielding to the threshold refers to the cumulative data of the strength of evidence of yielding to the threshold formed by all tourists in the group when passing through the doorway of the ancient building. This data is used to indicate the degree of passage deviation of the group as a whole in the threshold area due to avoiding the copper corner, getting close to the side of the doorway, or going around the threshold. Extract the tourist identifier and threshold evidence strength for each tourist in the group. Treat the threshold evidence strengths with the same tourist identifier as the same tourist for summation. Iterate through the threshold evidence strengths of all tourists in the group according to their tourist identifiers. Add each threshold evidence strength encountered to the same summation value until the threshold evidence strengths of all tourists in the group have been summed to obtain the total threshold evidence strength. Then, associate and save the total threshold evidence strength with the group identifier, the ancient building doorway identifier, and the current threshold event.

[0077] The strength of the disproving evidence for each tourist leaving the group is summed to obtain the total strength of the disproving evidence for leaving the group. The total strength of evidence of departure refers to the cumulative data of the strength of evidence of departure formed by all tourists in the group after passing through the gate of the ancient building. This data is used to represent the degree of continuous lateral deviation of the group as a whole in the area inside the gate and the degree of overall deviation from the direction led by the tour guide. Extract the tourist identifier and the strength of the evidence of leaving the group for each tourist in the group. Treat the evidence of leaving the group with the same tourist identifier as the summation object for the same tourist. Iterate through the evidence of leaving the group for all tourists in the group according to the tourist identifier. Add each evidence of leaving the group that has been traversed to the same summation value until the evidence of leaving the group for all tourists in the group has been accumulated to obtain the total evidence of leaving the group. Then associate and save the total evidence of leaving the group with the group identifier, the ancient building doorway identifier and the current threshold event.

[0078] Based on the sum of the evidence strength of thresholding and the sum of the evidence strength of leaving the group, the instruction type of the recommended tour guide collaboration strategy is matched to obtain tour guide management instructions.

[0079] Instruction type matching refers to the process of mapping recommended tour guide collaboration strategies to different tour guide management action categories based on the relationship between the sum of threshold evidence strength and the sum of dissenting evidence strength; tour guide management instructions refer to data instructions used to guide tour guides in performing tourist organization, tourist reminders, group adjustments, or passage control in the archway area of ​​ancient buildings.

[0080] The total strength of evidence for yielding the threshold is compared with the total strength of evidence for rebuttal due to leaving the group. When the total strength of evidence for yielding the threshold is greater than or equal to the total strength of evidence for rebuttal due to leaving the group, the result is determined as a threshold-yielding management type. When the total strength of evidence for yielding the threshold is less than the total strength of evidence for rebuttal due to leaving the group, the result is determined as a leaving-the-group management type. The guide positioning content, guide prompts, and team organization content are extracted from the recommended guide coordination strategy. When the result is a threshold-yielding management type, the recommended guide coordination strategy is matched as a threshold yielding guide management instruction. When the result is a leaving-the-group management type, the recommended guide coordination strategy is matched as a leaving-the-group handling guide management instruction. The matched instruction type, guide positioning content, guide prompts, team organization content, team identifier, ancient building doorway identifier, and current threshold event are written into the same instruction record to obtain the guide management instruction.

[0081] The "Threshold Avoidance Management" category refers to the tour guide management category corresponding to the overall behavior of tourists in the group when passing through the archway of an ancient building, such as avoiding the copper corner, going around the threshold, and returning to the queue after passing through the archway. This management category is used to guide tour guides to implement threshold avoidance prompts, guide the group to move slowly, and maintain traffic order. The "Leaving the Group Management" category refers to the tour guide management category corresponding to the overall behavior of tourists in the group after passing through the archway of an ancient building, such as continuously deviating from the guide's direction of travel and failing to return to the main passage area of ​​the group. This management category is used to guide tour guides to implement tourist confirmation, group recovery, and handling of leaving the group.

[0082] It should be noted that the total strength of evidence for yielding to the threshold represents the overall degree of yielding behavior among tourists in the group when passing through the archway of an ancient building, due to avoiding the copper corner, slowing down to go around the threshold, and rejoining the group. The total strength of evidence for leaving the group represents the overall degree of leaving the group when tourists in the group continue to deviate from the guide's direction of travel after passing through the archway, continue to stay in the side area of ​​the archway, and fail to return to the main passage area of ​​the group. When the total strength of evidence for yielding to the threshold is greater than or equal to the total strength of evidence for leaving the group, it indicates that the overall behavior of the group is still mainly about avoiding the threshold and rejoining the group after avoiding the copper corner, and the overall movement of the tourists is still in accordance with the direction led by the guide. Therefore, the corresponding management focus should be on threshold yielding management. When the total strength of evidence for yielding to the threshold is less than the total strength of evidence for leaving the group, it indicates that the behavior of continuously deviating from the group direction and failing to return to the main passage area is dominant in the overall behavior of the group, and the overall movement of the tourists has deviated from the guide's trajectory. Therefore, the corresponding management focus should be on managing leaving the group.

[0083] The "Threshold Yield" guide management instruction refers to the guide's handling instructions generated when the overall passage of the group is mainly about avoiding the copper corner, going around the threshold, and rejoining the queue. These instructions include adjusting the guide's position, prompting visitors to avoid the threshold, guiding visitors to slow down, and controlling the order of the group's passage, all to maintain the safe passage order of visitors in the archway area of ​​the ancient building. The "Off-Group Handling" guide management instruction refers to the guide's handling instructions generated when the overall passage of the group is mainly about continuous lateral deviation, continuous deviation from the guide's direction of travel, and failure to rejoin the main passage area. These instructions include visitor confirmation, group regrouping, guide direction correction, and reminders to visitors who have left the group, all to restore the consistency of movement between visitors and the main passage direction of the group.

[0084] Upon receiving the tour guide management instructions, the corresponding tour guide positioning, prompts, and group organization details are sent to the tour guide's terminal device. The terminal device then outputs corresponding voice prompts, movement guidance information, or group adjustment information, enabling the tour guide to intervene on-site in the passage of tourists through the ancient building's archway according to the tour guide management instructions. After the tour guide completes the corresponding management actions, the system continues to collect tourist and tour guide location data, re-identifying the tourists' movement within the archway area to generate new evidence of yielding and evidence of leaving the group, which is used to continuously update subsequent tour guide management instructions.

[0085] It should be noted that in scenarios where ancient building doorways have copper-clad corners, tourists may simply be avoiding the corners by turning sideways, pausing briefly, and returning to the main passageway. However, they may continue to veer to the side or away from the guide's direction after passing through the doorway, creating a genuine risk of leaving the group. Therefore, it is difficult to determine whether the guide should slow down or leave the group based solely on the instantaneous position of a single tourist. The sum of evidence strength for yielding the threshold is used to represent the degree of normal passage adjustment made by the group as a whole due to avoiding the corners, while the sum of evidence strength for leaving the group is used to represent the degree of risk of the group as a whole deviating from the main passageway and the guide's direction. By comparing the two, it can be determined whether the current group status is closer to yielding the threshold or the risk of leaving the group. Then, the comparison result is matched with the recommended guide coordination strategy to convert the recommended strategy into a clear guide management instruction for yielding the threshold or leaving the group, thereby avoiding misreporting normal actions of avoiding the corners as leaving the group and providing the guide with a unique and executable management instruction.

[0086] Example 2: Based on Example 1, an AI tour guide management method for tourist attractions is also provided. Specifically, the steps of this method include: Based on the structural data of the ancient building's doorway, a doorway coordinate system is established; based on the tourist location data collected by the tourist terminal device, local threshold segments are obtained, and the lateral positions at each collection time within the local threshold segments are classified to obtain the lateral attribution status. Based on the lateral attribution status, the corner-attaching tendency status, the threshold-around-corner status, and the regression queue status are generated, and the threshold-attaching evidence strength is determined. The lateral state of tourists in the segment inside the doorway is continuously identified to generate the first counter-evidence item for leaving the group; based on the guide position data and local threshold segments collected by the guide terminal device, the direction of tourists' movement inside the doorway is identified to generate the second counter-evidence item for leaving the group; based on the first counter-evidence item for leaving the group and the second counter-evidence item for leaving the group, the strength of the counter-evidence item for leaving the group is determined. An initial particle swarm is generated based on the set of tour guide collaborative strategies corresponding to the doorways of ancient buildings; the fitness value of each particle is calculated based on the strength of evidence for yielding the threshold and the strength of evidence for refuting leaving the queue in the set of historical threshold event samples; the initial particle swarm is iteratively updated based on the fitness value to obtain the recommended tour guide collaborative strategy. By matching the strength of threshold evidence and the strength of disproving evidence for leaving the group, the recommended tour guide collaboration strategy is analyzed to obtain tour guide management instructions.

[0087] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. An AI-powered tour guide management system for tourist attractions, characterized in that: include: The coordinate construction module establishes a coordinate system for the doorway based on the structural data of the ancient building's doorway. Based on the tourist location data and the doorway coordinate system collected by the tourist terminal device, a local threshold segment is obtained, and the lateral position at each collection time within the local threshold segment is classified to obtain the lateral attribution status. The threshold evidence collection module generates corner-attaching trend state, threshold-around-corner state, and return queue state based on the horizontal attribution state, and determines the threshold evidence strength. The module for refuting the departure from the group continuously identifies the lateral state of tourists in the segment inside the doorway to generate the first item for refuting the departure from the group; based on the guide's position data and local threshold segments collected by the guide terminal device, it identifies the direction of tourists' movement inside the doorway to generate the second item for refuting the departure from the group; and based on the first and second items for refuting the departure from the group, it determines the strength of the refuting the departure from the group. The strategy optimization module generates an initial particle swarm based on the set of tour guide collaborative strategies corresponding to the doorways of ancient buildings; it calculates the fitness value of each particle based on the strength of evidence for yielding the threshold and the strength of evidence for refuting leaving the group in the set of historical threshold event samples; and iteratively updates the initial particle swarm based on the fitness value to obtain the recommended tour guide collaborative strategy. The instruction generation module matches the recommended tour guide collaboration strategy with the instruction type by considering the threshold evidence strength and the disproving evidence strength of leaving the queue, thus obtaining tour guide management instructions.

2. The AI ​​tour guide management system for tourist attractions according to claim 1, characterized in that, Establishing the coordinate system for the doorway includes: Obtain the structural data of the doorway of the ancient building; the structural data includes the coordinates of the center point of the doorway, the vector of the passage direction of the doorway, the position data of the threshold line, the coordinates of the left and right copper corner brackets; A coordinate system for the doorway is established based on the coordinates of the center point of the doorway and the direction vector of passage through the doorway in the structural data.

3. The AI ​​tour guide management system for tourist attractions according to claim 1, characterized in that, The horizontal attribution status is obtained, including: Acquire tourist location data collected by tourist terminal devices; Based on tourist location data and the doorway coordinate system, generate axial and lateral position sequences for tourists; Based on the axial position sequence of tourists, continuous trajectory segments are extracted from the planar position coordinates of tourists to obtain local threshold segments; Based on the tourists' lateral position sequence, the coordinates of the left and right copper corners, the lateral positions at each collection time within the local threshold segment are classified to obtain the lateral attribution status.

4. The AI ​​tour guide management system for tourist attractions according to claim 1, characterized in that, Determining the strength of evidence for a threshold includes: Based on the lateral attribution status, the local threshold segment is segmented to obtain the segment outside the doorway, the segment near the threshold, and the segment inside the doorway. Determine the corner-attaching tendency state based on the lateral attribution state of the segment outside the doorway; Determine the threshold-around-angle state based on the lateral attribution state of the segments adjacent to the threshold; The regression queue status is determined based on the lateral attribution status of the segment inside the doorway; Based on the corner-touching tendency state, the threshold-around-corner state, and the return queue state, generate a set of threshold-yielding evidence items for tourists; The strength of the threshold evidence is determined based on the number of evidence items in the threshold evidence set.

5. The AI ​​tour guide management system for tourist attractions according to claim 1, characterized in that, Generate the first off-queue counter-evidence entry, including: Based on the local threshold segments and lateral attribution status, extract the data on the changes in the lateral attribution status of tourists in the segment inside the doorway. Based on the data on changes in lateral attribution status, continuous lateral attribution status identification is performed on the tourists' lateral attribution status in the segment inside the doorway to obtain the lateral non-regression status. Based on the lateral non-regression state, generate the first off-line counter-evidence entry.

6. The AI ​​tour guide management system for tourist attractions according to claim 1, characterized in that, Determine the strength of the disproving evidence by leaving the group, including: Acquire tour guide location data collected by the tour guide terminal device; Extract the planar position coordinates of the tour guides corresponding to the local threshold segments from the tour guide position data; Based on the tour guide's planar position coordinates, generate the tour guide's travel direction data; Based on local threshold segments, generate tourist movement direction data; Based on the data of tourists' walking direction and the data of the tour guide's walking direction, the deviation state of tourists' walking direction inside the doorway is identified, and the deviation state is obtained. Based on the direction divergence state, generate a second disjointed counter-evidence entry; Perform item association processing on the first and second off-listed counter-evidence items to obtain the set of off-listed counter-evidence items; The strength of an off-staff counter-evidence is determined by the number of counter-evidence entries in the off-staff counter-evidence set.

7. The AI ​​tour guide management system for tourist attractions according to claim 1, characterized in that, Generate the initial particle swarm, including: Obtain the set of tour guide collaboration strategies corresponding to the doorways of ancient buildings; Each tour guide collaboration strategy in the set of tour guide collaboration strategies is numbered to obtain the strategy number; Use the strategy number as the initial particle position; Determine the adjacent strategy numbers of each particle based on the order of the strategy numbers; The initial particle velocity is determined based on the initial particle position and the adjacent strategy numbers. An initial particle swarm is generated based on the initial particle position and initial particle velocity of each particle.

8. The AI ​​tour guide management system for tourist attractions according to claim 1, characterized in that, Recommended tour guide collaboration strategies include: Obtain a sample set of historical threshold events corresponding to the doorways of ancient buildings; By labeling the strength of evidence for yielding the threshold and the strength of evidence for refusal for leaving the queue in the historical threshold event sample set, a sample set for AI strategy optimization is obtained. Based on the AI ​​strategy, the sample set is optimized, and the historical samples of the guide cooperative strategy corresponding to each initial particle position in the initial particle swarm are replayed to obtain the mismatch count. Calculate the fitness value of the particle based on the mismatch count; Based on fitness values, determine the global optimal particle position and the individual optimal particle position; Based on the global optimal particle position and the individual optimal particle position, the initial particle swarm is iteratively updated to obtain the particle swarm after the iteration. The recommended guide cooperative strategy is obtained by selecting guide cooperative strategies from the set of guide cooperative strategies after the iteration is completed by using the particle swarm optimization.

9. The AI ​​tour guide management system for tourist attractions according to claim 1, characterized in that, Receive tour guide management instructions, including: The strength of the evidence for yielding the threshold for each tourist in the group is summed to obtain the total strength of the evidence for yielding the threshold. The strength of the disproving evidence for each tourist leaving the group is summed to obtain the total strength of the disproving evidence for leaving the group. Based on the sum of the evidence strength of thresholding and the sum of the evidence strength of leaving the group, the instruction type of the recommended tour guide collaboration strategy is matched to obtain tour guide management instructions.

10. A method for managing AI-powered tour guides in tourist attractions, characterized in that: The method includes: Based on the structural data of the ancient building's doorway, a doorway coordinate system is established; based on the tourist location data collected by the tourist terminal device, local threshold segments are obtained, and the lateral positions at each collection time within the local threshold segments are classified to obtain the lateral attribution status. Based on the lateral attribution status, the corner-attaching tendency status, the threshold-around-corner status, and the regression queue status are generated, and the threshold-attaching evidence strength is determined. The lateral state of tourists in the segment inside the doorway is continuously identified to generate the first counter-evidence item for leaving the group; based on the guide position data and local threshold segments collected by the guide terminal device, the direction of tourists' movement inside the doorway is identified to generate the second counter-evidence item for leaving the group; based on the first counter-evidence item for leaving the group and the second counter-evidence item for leaving the group, the strength of the counter-evidence item for leaving the group is determined. An initial particle swarm is generated based on the set of tour guide collaborative strategies corresponding to the doorways of ancient buildings; the fitness value of each particle is calculated based on the strength of evidence for yielding the threshold and the strength of evidence for refuting leaving the queue in the set of historical threshold event samples; the initial particle swarm is iteratively updated based on the fitness value to obtain the recommended tour guide collaborative strategy. By matching the strength of threshold evidence and the strength of disproving evidence for leaving the group, the recommended tour guide collaboration strategy is analyzed to obtain tour guide management instructions.

Images