Virtual hotel guide customization system based on intelligent optimization and multi-modal interaction

Abstract

The invention discloses a virtual hotel guide customization system based on intelligent optimization and multi-modal interaction, which relates to the technical field of virtual reality and augmented reality application and comprises a multi-source data acquisition module, a VR guide module, an AR customization module, a user interaction interface and a data storage management module. The multi-source data acquisition module acquires hotel 3D data and user behavior data, the VR guide module plans a user moving path, the AR customization module adjusts room layout according to user behaviors and selects an optimal scheme, and the user interaction interface provides touch voice interaction. Through the intelligent optimization and multi-mode interaction technology, personalized navigation and room layout customization are realized, the user experience is improved, the optimal path is planned, the layout is intelligently adjusted according to the user behavior, the touch voice interaction interface is friendly, the data security is guaranteed through the encryption storage technology, and the user experience is improved. And an innovative solution is provided for the technical field of virtual reality and augmented reality application.

Description

Technical Field

[0001] The present invention relates to the technical field of virtual reality and augmented reality applications, and particularly relates to a virtual hotel tour customization system based on intelligent optimization and multimodal interaction. Background Art

[0002] Virtual hotel tour is a tourism guiding service that combines virtual reality and augmented reality technologies. It uses devices such as head-mounted displays and sensors to bring users into a computer-generated three-dimensional hotel environment, achieving an immersive guiding experience. Through virtual hotel tours, guests can intuitively understand the internal structure of the hotel, room layout, facility locations, and surrounding scenic spot information before check-in. This guiding method not only provides a more vivid and specific display effect but also makes guests feel as if they are on the spot, experiencing the hotel atmosphere in advance, choosing their favorite rooms, or exploring the activity areas they will participate in. In addition, virtual hotel tours can also integrate online booking functions, facilitating guests to directly arrange their activity schedules, greatly simplifying the travel planning process. With the popularization and integration of related technologies, virtual hotel tours will provide more personalized and interactive tourism information services for customers and become an indispensable part of smart hotels.

[0003] To solve the problems of weak immersion and poor interactivity in traditional virtual hotel tours, the existing technology uses basic 3D modeling and simple interaction interfaces for processing. Although this method can provide a certain degree of visual display, allowing users to browse through various corners of the hotel in a virtual environment, there are still situations where scene switching is rigid and interactive elements are limited. Users often can only browse according to a preset route during the tour, unable to freely explore or interact deeply with the environment, lacking a real sense of immersion and participation. As a result, the user experience is poor, and they cannot fully feel the unique charm and service details of the hotel, which also affects the effect of the hotel attracting potential customers and enhancing its brand image through virtual tours. Therefore, a virtual hotel tour customization system based on intelligent optimization and multimodal interaction is proposed. Summary of the Invention

[0004] The purpose of the present invention is to provide a virtual hotel tour customization system based on intelligent optimization and multimodal interaction to solve the problems raised in the above background art.

[0005] To solve the above technical problems, the technical solution adopted by the present invention is: a virtual hotel tour customization system based on intelligent optimization and multimodal interaction, including a multi-source data acquisition module, a VR tour module, an AR customization module, a user interaction interface, and a data storage and management module; The multi-source data acquisition module collects 3D spatial point cloud data and user customization behavior data of the hotel, and performs preprocessing and feature extraction on them; The VR navigation module calculates the positional relationship between the user's virtual character position and each bounding box according to the 3D spatial point cloud data, and plans the optimal movement path by using a collision detection algorithm based on the bounding box; The AR customization module selects and adjusts the room layout elements based on the user's customized behavior data, calculates the layout scores of each scheme, and selects the optimal layout scheme; The user interaction interface generates a touch and voice interaction interface for displaying the VR navigation results and AR customization effects; The data storage and management module encrypts the hotel VR navigation data and AR customization data, and stores the encrypted data by using big data storage and management technologies.

[0006] A further improvement of the technical solution of the present invention lies in that: in the multi-source data acquisition module, the acquisition processes of the 3D spatial point cloud data and the user's customized behavior data of the hotel include: Deploy lidar sensors at the central positions of the tops of the hotel lobby, corridors, and guest rooms, covering a 360-degree horizontal field of view downward at a fixed inclination angle, with a vertical field of view angle covering 60 degrees. Deploy depth cameras at the four corners of the guest rooms and the columns in the public areas, 2.5 meters above the ground, with a vertical view angle covering the height of the room and a horizontal field of view angle of 90 degrees. The lidar measures the time difference between the emission and reflection reception of the light beam by emitting pulsed laser beams, calculates the distance to the object surface, and generates 3D spatial point cloud data of a space where each point contains three-dimensional coordinate and reflection intensity information. The depth camera projects a structured light pattern, calculates the depth value by capturing the pattern offset after the deformation of the object surface and combining the triangulation principle, and generates dense 3D spatial point cloud data with a resolution of 0.5 cm to supplement the details of the furniture edges; Embed a touch terminal into the hotel official website and mobile application interface, support touch screen operation, collect the RGB values of the bedding colors selected by the user and the furniture translation instructions, integrate the voice acquisition module into the mobile device microphone and the VR headset, receive voice commands in real time, and distinguish the user's identity through voiceprint recognition. When the user touches the screen to select a color block, the virtual hotel navigation customization system extracts the RGB value corresponding to the screen pixel coordinates to collect the user's color selection data. When the user drags the virtual furniture model, the virtual hotel navigation customization system records the coordinate difference between the starting point and the ending point to obtain the layout adjustment instruction.

[0007] A further improvement of the technical solution of the present invention lies in that: in the multi-source data acquisition module, the preprocessing process of the collected 3D spatial point cloud data and the user's customized behavior data includes: Based on the statistical filtering principle, calculate the average distance and standard deviation between each point in the point cloud and its neighboring points. If there is a point whose average distance from its neighboring points exceeds the average distance plus three times the standard deviation, it is determined as an outlier noise point and removed. Using the adjacent complete wall point cloud data, complete the missing wall area by fitting a plane equation. In the preprocessed three-dimensional point cloud, mark the path nodes in the passable area and assign a cost value to each node. Merge the denoised point cloud data with the completed wall data to generate a three-dimensional space model of the virtual hotel, and input it into the collision detection and path planning algorithms of the VR navigation module. The cost value table of the path nodes is transmitted to the path planning algorithm as the weight basis for calculating the moving path; Convert the RGB values selected by the user to numerical values in the range of [0, 1] through normalization processing. Combine the room coordinate system to convert the coordinate difference between the starting point and the ending point into absolute coordinates. Associate the normalized color data and the absolute coordinate layout instructions to the user account and store them encrypted, and input them into the AR customization module.

[0008] A further improvement of the technical solution of the present invention is that in the multi-source data acquisition module, the process of feature extraction from the preprocessed 3D spatial point cloud data and user customization behavior data includes: Project the preprocessed three-dimensional point cloud data onto the horizontal plane, divide it into grids with a side length of 0.1 meter, count the number of grids covering the wall point cloud of the virtual hotel, multiply by the area of a single grid, and accumulate to obtain the total area of the room. For the vertex coordinates of the furniture bounding box, calculate the planar distance between the center points of each piece of furniture. Based on the vertex coordinates of the obstacle bounding box, calculate the length, width, and height. In the point cloud of the passable area, extract the ground center point as a path node, with a distance of 0.5 meter between adjacent nodes. Assign a corridor node cost value of 1 and a staircase node cost value of 2 according to the area attributes where the path nodes are located, and use them as the weight parameters of the path planning algorithm; Statistically analyze the translation vectors in the user's historical adjustment data, calculate the high-frequency adjustment direction and amplitude, and classify the user's adjustment data through the K-means clustering algorithm to extract the translation vectors corresponding to the cluster centers; Input the total area of the room, the distance between the furniture center points, and the obstacle size into the collision detection and path planning algorithms of the VR navigation module, and input the layout adjustment features into the AR customization module.

[0009] A further improvement of the technical solution of the present invention is that in the VR navigation module, the process of using the bounding box-based collision detection algorithm to calculate the position relationship between the user's virtual character position and each bounding box includes: Extract the vertex coordinates of the walls and furniture from the preprocessed 3D spatial point cloud data, calculate the minimum and maximum values of their three-dimensional coordinates respectively, and form an axial bounding box; The user moves through the VR device, and the virtual hotel tour customization system obtains the three-dimensional coordinates of the virtual character in real time based on the positioning system of the VR device, traverses the bounding boxes in the scene, and determines one by one whether the current position of the virtual character overlaps with the spatial range of the bounding box. If the x-axis, y-axis, and z-axis of the three-dimensional coordinates of the virtual character are all within the minimum and maximum values of the three-dimensional coordinates of the bounding box, it is determined that the virtual character collides with the bounding box; If a collision is detected, the virtual hotel tour customization system enforces constraints according to the bounding box type, prohibits the virtual character from entering the wall bounding box area, corrects the virtual character coordinates to the nearest edge outside the bounding box, allows the virtual character to approach the furniture bounding box area but restricts penetration, and triggers a red light flashing prompt. The bounding box types include walls and furniture.

[0010] A further improvement of the technical solution of the present invention lies in: in the VR tour module, the process of planning the best movement path from the user's starting point to the search end point includes: Receiving path node coordinates, cost value tables, user starting point, and end point coordinate data through the VR tour module; Accumulate the cost values of the path nodes to obtain the actual cost from the starting point to the current node. Calculate the straight-line distance from the current node to the end point using the Euclidean distance to obtain the estimated cost. Use the A* algorithm to define the total cost function as the sum of the actual cost and the estimated cost; Add the starting point to the open list, record its total cost function value, and perform iterative search. The iterative search process includes selecting the node with the smallest total cost function value from the open list as the current node. If the current node is the end point, backtrack the path and output it. If the current node is not the end point, traverse its adjacent nodes, calculate the actual cost and the estimated cost for each adjacent node, and update the total cost function value. If the adjacent node is not in the open list and there is a better path, update the current node information and add it to the open list until there is no feasible path to terminate the iteration, output the optimal path node sequence, transmit the optimal path node sequence to the VR scene rendering engine to display it in real time with green lines, and dynamically adjust the node cost values according to the hotel traffic flow.

[0011] A further improvement of the technical solution of the present invention lies in: in the AR customization module, the process of using a color recognition algorithm based on the RGB color space to match the bedding color selected by the user includes: Match the normalized RGB values selected by the user with the bedding color library of the virtual hotel, calculate the Euclidean distance between the user color and the colors in the bedding color library, select the color with the smallest distance as the matching result, record its RGB value as the user preference feature, associate the RGB value of the optimal matching color with the user account, encrypt and store it in the data storage management module, and overlay the matched bedding color on the virtual guest room model in the AR scene in real time.

[0012] A further improvement of the technical solution of the present invention lies in that: in the AR customization module, the simulated annealing algorithm is used to calculate the scores of each furniture layout plan, and the process of selecting the best furniture layout plan includes: According to the preprocessed three-dimensional room model, furniture size data and user's historical layout preference characteristics, an initial furniture layout plan is generated, and the furniture positions are defined using three-dimensional coordinates and orientation angles. The furniture spacing and the distance from the wall are comprehensively analyzed to calculate the score of the furniture layout plan; Apply position perturbations to random furniture in the current layout and calculate the score of the new furniture layout plan If the score of the new furniture layout plan is lower than the score of the current furniture layout plan then accept the new furniture layout plan, otherwise accept the new plan with probability P, where T is the current temperature parameter, and the initial temperature is set to 100. When the score of the furniture layout plan has not been improved after 10 consecutive iterations, stop the search, output the three-dimensional coordinates and orientation parameters of the current optimal layout plan, and transmit them to the AR scene rendering engine to be superimposed on the virtual guest room model for the user to preview.

[0013] A further improvement of the technical solution of the present invention lies in that: in the user interaction interface, the process of generating a touch and voice interaction interface for displaying the VR tour result and the AR customization effect includes: According to the user's historical operation data, dynamically adjust the display priority of the interface module, generate a touch and voice interaction interface through a template engine, support gesture sliding and voice command to switch the tour and customization functions. The touch and voice interaction interface includes a tour control area, a customization operation area and an information display area. The tour control area displays VR scene switching buttons and path planning status. The customization operation area embeds a touch color palette to display the RGB value in real time, embeds a furniture layout adjustment panel to display the translation vector, and the information display area dynamically displays the basic information of the hotel, the RGB value of the matching bed color and the scores of each furniture layout plan; When the user clicks on the color palette to select a color, the virtual hotel tour customization system captures the screen pixel coordinates, extracts the corresponding RGB value, and normalizes it to a numerical value in the range of [0,1]. When the user drags the furniture model, record the coordinate difference between the starting point and the ending point and convert it into an absolute displacement amount; Collect voice signals through the mobile device microphone and the VR headset, use voiceprint recognition technology to distinguish the user's identity, set voice command keywords, and trigger corresponding operations; Transmit the optimal path node sequence output by the VR tour guide module to the interface tour guide control area, and dynamically display the moving path with green connection lines. Overlay the RGB values of the bed product colors and the furniture layout plan of the AR customization module onto the information display area in real time, support perspective rotation and zoom preview. When the furniture layout score is lower than the set score threshold, prompt with red light flashing and push it to the information display area.

[0014] A further improvement of the technical solution of the present invention lies in: in the data storage and management module, the process of storing hotel VR tour guide data and AR customization data using encryption technology includes: The data storage and management module receives VR tour guide and AR customization data. The VR tour guide data includes a virtual hotel three-dimensional model, path node coordinates, a cost value table, and a user movement trajectory. The AR customization data includes the RGB values of the bed product colors selected by the user, the three-dimensional coordinates and orientation angles of the furniture layout plan, and the translation vector clustering center of the historical adjustment record. Perform field-level encryption on the structured data in the VR tour guide and AR customization data, and use block encryption technology for unstructured data. Combine symmetric encryption algorithms and asymmetric encryption technologies to generate a key through the user's voiceprint feature; Store the encrypted data in database tables according to types. The database tables include a VR tour guide table and an AR customization table. The VR tour guide table stores the hash value of the encrypted virtual hotel three-dimensional model, the ciphertext of the path nodes, and the ciphertext of the cost values. The AR customization table stores the encrypted RGB values, the ciphertext of the furniture layout plan, and the ciphertext of the user's historical operation logs. Verify the data integrity through a hash function. Calculate the ciphertext hash value and save it independently when storing data, and verify the hash consistency when accessing. Bind the access permission to voiceprint identity authentication, and unauthorized users cannot decrypt the data; After the user verifies their identity through voiceprint recognition, the virtual hotel tour guide customization system calls the pre-stored voiceprint feature to regenerate the encryption key. If the key verification fails, access to the encrypted data is prohibited. When the VR tour guide module and the AR customization module request data, the virtual hotel tour guide customization system decrypts the corresponding fields and data blocks. The decrypted data is temporarily stored in memory and cleared immediately after use.

[0015] Due to adopting the above technical solution, the technical progress achieved by the present invention compared with the prior art is: The present invention provides a virtual hotel tour guide customization system based on intelligent optimization and multi-modal interaction. By using the multi-source data acquisition module to obtain the detailed 3D space information of the hotel and user preferences, the VR tour guide module can accurately construct a virtual environment and provide intelligent path planning based on collision detection, greatly improving the user's navigation experience and immersion in the virtual hotel.

[0016] The present invention provides a virtual hotel tour customization system based on intelligent optimization and multimodal interaction. The AR customization module can intelligently adjust and optimize the room layout according to the user's customization behavior data, automatically select the best layout plan by calculating the scores of different layout plans, meet the personalized needs of users, and improve the customization and interactivity of virtual tours.

[0017] The present invention provides a virtual hotel tour customization system based on intelligent optimization and multimodal interaction. The user interaction interface integrates multiple interaction methods such as touch and voice, enabling users to interact with the virtual environment more naturally. At the same time, the data storage and management module uses encryption technology to ensure the security of hotel VR tour data and AR customization data, protecting user privacy and hotel business information. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required in the embodiments. Obviously, the drawings described below are only some embodiments recorded in the present invention. For those of ordinary skill in the art, other drawings can also be obtained based on these drawings.

[0019] Figure 1 It is a block diagram of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0020] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0021] Embodiment, as Figure 1 shown, the present invention provides a virtual hotel tour customization system based on intelligent optimization and multimodal interaction, including a multi-source data acquisition module, a VR tour module, an AR customization module, a user interaction interface, and a data storage and management module; Multi-source data acquisition module, which acquires the 3D spatial point cloud data and user-customized behavior data of the hotel, and performs preprocessing and feature extraction on them. LiDAR sensors are deployed at the central positions of the hotel lobby, corridor, and the top of the guest rooms, covering a 360-degree horizontal field of view downward at a fixed inclination angle, with a vertical field of view angle of 60 degrees. Depth cameras are deployed at the four corners of the guest rooms and on the columns in public areas, 2.5 meters above the ground, with a vertical viewing angle covering the height of the room and a horizontal field of view angle of 90 degrees. The LiDAR emits pulsed laser beams, measures the time difference between the emission and the reflection and reception of the beams, calculates the distance to the object surface, and generates 3D spatial point cloud data with three-dimensional coordinates and reflection intensity information for each point in space. The depth camera projects structured light patterns, calculates the depth value by capturing the pattern offset after the deformation of the object surface and combining the principle of triangulation, and generates dense 3D spatial point cloud data with a resolution of 0.5 cm to supplement the details of the furniture edges. The touch terminal is embedded in the hotel official website and mobile application interfaces, supporting touch screen operations, and acquiring the RGB values of the bedding colors selected by the user and the furniture translation instructions. The voice acquisition module is integrated into the mobile device microphone and VR headset, receiving voice commands in real-time and distinguishing user identities through voiceprint recognition. When the user touches the screen to select color blocks, the virtual hotel tour customization system extracts the RGB values corresponding to the screen pixel coordinates to acquire the user's color selection data. When the user drags the virtual furniture model, the virtual hotel tour customization system records the coordinate difference between the starting point and the ending point to obtain the layout adjustment instruction. Based on the statistical filtering principle, calculate the average distance and standard deviation between each point in the point cloud and its neighboring points. If there is a point whose average distance from its neighboring points exceeds the average distance plus 3 times the standard deviation, it is determined as an outlier noise point and removed. Use the point cloud data of adjacent complete wall surfaces to complement the missing wall area through plane equation fitting. In the preprocessed three-dimensional point cloud, mark the path nodes in the passable area and assign a cost value to each node. Combine the denoised point cloud data with the complemented wall data to generate a three-dimensional space model of the virtual hotel, and input it into the collision detection and path planning algorithms of the VR tour module. The cost value table of the path nodes is transmitted to the path planning algorithm as the weight basis for calculating the moving path. Convert the RGB values selected by the user to numerical values in the [0,1] interval through normalization processing, and convert the coordinate difference between the starting point and the ending point to absolute coordinates in combination with the room coordinate system. Associate the normalized color data and the absolute coordinate layout instruction with the user account and encrypt and store them. Input them into the AR customization module. Project the preprocessed three-dimensional point cloud data onto the horizontal plane, divide it into grids with a side length of 0.1 meter, count the number of grids covering the point cloud of the virtual hotel wall, multiply by the area of a single grid, and accumulate to obtain the total area of the room. Calculate the planar distance between the center points of each furniture based on the vertex coordinates of the furniture bounding box. Calculate the length, width, and height based on the vertex coordinates of the obstacle bounding box. In the point cloud of the passable area, extract the center point of the ground as the path node, and the distance between adjacent nodes is 0.5 meters, assign a corridor node cost value of 1 and a staircase node cost value of 2 according to the regional attributes of the path nodes, and use them as the weight parameters of the path planning algorithm. Statistically analyze the translation vectors in the user's historical adjustment data, calculate the high-frequency adjustment directions and amplitudes. Through the K-means clustering algorithm, classify the user's adjustment data, extract the translation vectors corresponding to the cluster centers, input the total room area, the distances between the centers of the furniture, and the obstacle sizes into the collision detection and path planning algorithm of the VR navigation module, and input the layout adjustment features into the AR customization module;. The VR navigation module, based on the 3D spatial point cloud data, uses a collision detection algorithm based on bounding boxes to calculate the positional relationship between the position of the user's virtual character and each bounding box, and plan the optimal movement path. From the preprocessed 3D spatial point cloud data, extract the vertex coordinates of the walls and furniture, calculate the minimum and maximum values of their three-dimensional coordinates respectively to form an axial bounding box. The user moves through the VR device, and the virtual hotel navigation customization system obtains the three-dimensional coordinates of the virtual character in real time based on the positioning system of the VR device, traverses the bounding boxes in the scene, and successively determines whether the current position of the virtual character overlaps with the spatial range of the bounding box. If the x-axis, y-axis, and z-axis of the virtual character's three-dimensional coordinates are all within the minimum and maximum values of the three-dimensional coordinates of the bounding box, it is determined that the virtual character collides with the bounding box. If a collision is detected, the virtual hotel navigation customization system enforces constraints according to the type of bounding box, prohibits the virtual character from entering the wall bounding box area, corrects the virtual character's coordinates to the nearest edge outside the bounding box, allows the virtual character to approach the furniture bounding box area but restricts penetration, and triggers a red light flashing prompt. The types of bounding boxes include walls and furniture. Receive the path node coordinates, cost value table, user start and end point coordinate data through the VR navigation module, accumulate the cumulative cost value from the start point to the current node according to the path node cost value to obtain the actual cost, calculate the straight-line distance from the current node to the end point using the Euclidean distance to obtain the estimated cost, and use the A* algorithm to define the total cost function as the sum of the actual cost and the estimated cost. Add the start point to the open list, record its total cost function value, and perform iterative search. The iterative search process includes selecting the node with the smallest total cost function value from the open list as the current node. If the current node is the end point, backtrack the path and output it. If the current node is not the end point, traverse its adjacent nodes, calculate the actual cost and the estimated cost for each adjacent node, and update the total cost function value. If the adjacent node is not in the open list and there is a better path, update the current node information and add it to the open list until there is no feasible path to terminate the iteration, output the optimal path node sequence, transmit the optimal path node sequence to the VR scene rendering engine to display it in real time with green lines, and dynamically adjust the node cost value according to the hotel traffic flow; The AR customization module selects and adjusts room layout elements based on user customization behavior data, calculates the layout scores of each plan, selects the best layout plan, matches the normalized user-selected RGB values with the bedding color library of the virtual hotel, calculates the Euclidean distance between the user's color and the colors in the bedding color library, selects the color with the smallest distance as the matching result, records its RGB value as the user preference feature, associates the RGB value of the optimally matched color with the user account, encrypts and stores it in the data storage and management module, and superimposes the matched bedding color onto the virtual guest room model in the AR scene in real time. Based on the preprocessed three-dimensional room model, furniture size data, and the user's historical layout preference features, an initial furniture layout plan is generated, and the furniture positions are defined using three-dimensional coordinates and orientation angles. The furniture spacing and the distance from the wall are comprehensively analyzed, and the furniture layout plan score is calculated. A position perturbation is applied to a random piece of furniture in the current layout, and the score of the new furniture layout plan is calculated. If the score of the new furniture layout plan is lower than the score of the current furniture layout plan then the new furniture layout plan is accepted; otherwise, the new plan is accepted with probability P. where T is the current temperature parameter, and the initial temperature is set to 100. When the furniture layout plan score has not improved after 10 consecutive iterations, the search is stopped, and the three-dimensional coordinates and orientation parameters of the current optimal layout plan are output and transmitted to the AR scene rendering engine to be superimposed onto the virtual guest room model for the user to preview. User interaction interface, which generates a touch and voice interaction interface for displaying VR tour results and AR customization effects, dynamically adjusts the display priority of interface modules according to user historical operation data, generates a touch and voice interaction interface through a template engine, supports gesture sliding and voice commands to switch tour and customization functions. The touch and voice interaction interface includes a tour control area, a customization operation area, and an information display area. The tour control area displays VR scene switching buttons and path planning status. The customization operation area embeds a touch color palette to display RGB values in real time and embeds a furniture layout adjustment panel to display translation vectors. The information display area dynamically displays hotel basic information, the RGB values of the matched bedding colors, and the scores of each furniture layout plan. When the user clicks on the color palette to select a color, the virtual hotel tour customization system captures the screen pixel coordinates, extracts the corresponding RGB value, and normalizes it to a value in the [0,1] interval. When the user drags a furniture model, the coordinate difference between the starting point and the ending point is recorded and converted into an absolute displacement amount. Voice signals are collected through a mobile device microphone and a VR headset, and voiceprint recognition technology is used to distinguish user identities. Voice command keywords are set to trigger corresponding operations. The optimal path node sequence output by the VR tour module is transmitted to the interface tour control area, and the movement path is dynamically displayed with green lines. The RGB values of the bedding color matching and the furniture layout plan of the AR customization module are superimposed on the information display area in real time, supporting perspective rotation and zoom preview. When the furniture layout score is lower than the set score threshold, it is prompted by red light flashing and pushed to the information display area; Data storage management module, which encrypts hotel VR tour data and AR customization data, and stores the encrypted data using big data storage and management technology. The VR tour data includes virtual hotel three-dimensional models, path node coordinates, cost value tables, and user movement trajectories. The AR customization data includes RGB values of bedding colors selected by users, three-dimensional coordinates and orientation angles of furniture layout plans, and translation vector clustering centers of historical adjustment records. Field-level encryption is performed on structured data in VR tour and AR customization data, and block encryption technology is used for unstructured data. A combination of symmetric encryption algorithms and asymmetric encryption technologies is employed to generate keys through user voiceprint features. The encrypted data is stored in database tables by type. The database tables include a VR tour table and an AR customization table. The VR tour table stores encrypted virtual hotel three-dimensional model hash values, path node ciphertexts, and cost value ciphertexts. The AR customization table stores encrypted RGB values, furniture layout plan ciphertexts, and user historical operation log ciphertexts. Data integrity is verified through a hash function. When storing data, the ciphertext hash value is calculated and saved independently. When accessing, the hash consistency is verified. Access permissions are bound to voiceprint identity authentication, and unauthorized users cannot decrypt the data. After the user verifies their identity through voiceprint recognition, the virtual hotel tour customization system calls the pre-stored voiceprint features to regenerate the encryption key. If the key verification fails, access to the encrypted data is prohibited. When the VR tour module and the AR customization module request data, the virtual hotel tour customization system decrypts the corresponding fields and data blocks. The decrypted data is temporarily stored in memory and cleared immediately after use.

[0022] First, the multi-source data acquisition module is responsible for collecting the 3D spatial point cloud data of the hotel and the customized behavior data of users, and preprocessing and feature extraction of these data. Guests download and open the software, register and log in to the account. Then, the user selects the VR hotel tour function on the software home page, connects the VR device, enters the VR virtual scene of the hotel, and starts a full-round tour. The system will construct a realistic virtual hotel environment based on the collected 3D data and use a collision detection algorithm based on bounding boxes to plan the best movement path for the user's virtual character. During the tour, guests can pause and switch areas at any time to view the details of the facilities and areas of interest. When the guest enters the guest room tour session, click the room customization button, turn on the camera of the mobile phone or tablet, scan the guest room space, enter the AR scene, and switch to the AR customization mode. In the AR customization mode, the guest selects different bedding colors and furniture layouts, etc. for virtual customization by touching the screen or voice commands. By calculating the layout scores of each plan, the best layout plan is selected for the user. After the customization is completed, the guest confirms the customization plan, and the software sends the guest's customization information to the hotel background. The user intuitively views the VR tour results and AR customization effects through the touch and voice interaction interface and performs interactive operations. Finally, the hotel VR tour data and AR customization data are securely stored by the data storage and management module using encryption technology. After receiving the customization information, the hotel arranges the room in advance according to the guest's requirements.

[0023] As described above, it is only the specific implementation manner of this application, but the protection scope of this application is not limited thereto. Any person skilled in the art within the technical scope disclosed by this application can easily think of changes or substitutions, which should all be covered within the protection scope of this application. Therefore, the protection scope of this application should be subject to the protection scope of the claims.

Claims

1. A virtual hotel tour customization system based on intelligent optimization and multi-modal interaction, characterized in that: It includes a multi-source data acquisition module, a VR tour guide module, an AR customization module, a user interaction interface, and a data storage and management module; The multi-source data acquisition module acquires the 3D spatial point cloud data and user customization behavior data of the hotel; The VR tour guide module, based on the 3D spatial point cloud data, uses a collision detection algorithm based on bounding boxes to calculate the positional relationship between the user's virtual character position and each bounding box, and plans the optimal movement path; The AR customization module, based on the user customization behavior data, selects and adjusts the room layout elements, calculates the layout scores of each plan, and selects the optimal layout plan; The user interaction interface generates a touch and voice interaction interface for displaying the VR tour guide results and AR customization effects; The data storage and management module encrypts the hotel VR tour guide data and AR customization data, and stores the encrypted data using big data storage and management technology.

2. The virtual hotel tour customization system based on intelligent optimization and multi-modal interaction according to claim 1, characterized in that: In the multi-source data acquisition module, the acquisition process of the 3D spatial point cloud data and user customization behavior data of the hotel includes: Deploy lidar sensors at the central positions of the tops of the hotel lobby, corridors, and guest rooms, with a fixed inclination angle to cover a 360-degree horizontal field of view downward, and a vertical field of view angle of 60 degrees. Deploy depth cameras at the four corners of the guest rooms and the columns in the public areas, 2.5 meters above the ground, with a vertical viewing angle covering the height of the room and a horizontal field of view angle of 90 degrees. The lidar generates 3D spatial point cloud data of a space where each point contains three-dimensional coordinate and reflection intensity information by emitting pulsed laser beams. The depth camera projects structured light patterns to generate dense 3D spatial point cloud data with a resolution of 0.5 cm; Embed a touch terminal into the hotel official website and mobile application interface to support touch screen operation, acquire the RGB values of the beddings selected by the user and the furniture translation instructions. Integrate a voice acquisition module into the mobile device microphone and VR headset to receive voice commands in real time, and distinguish user identities through voiceprint recognition. When the user touches the screen to select a color block, the virtual hotel tour guide customization system extracts the RGB value corresponding to the screen pixel coordinates to acquire the user's color selection data. When the user drags the virtual furniture model, the virtual hotel tour guide customization system records the coordinate difference between the starting point and the ending point to obtain a layout adjustment instruction.

3. The virtual hotel tour customization system based on intelligent optimization and multi-modal interaction according to claim 2, characterized in that: In the multi-source data acquisition module, the preprocessing process of the acquired 3D spatial point cloud data and user customization behavior data includes: Based on the statistical filtering principle, calculate the average distance and standard deviation between each point in the point cloud and its neighboring points. If there is a point whose average distance from its neighboring points exceeds the average distance plus 3 times the standard deviation, it is determined as an outlier noise point and removed. Use the adjacent complete wall point cloud data to complement the missing wall area through plane equation fitting. In the preprocessed three-dimensional point cloud, mark the path nodes in the passable area and assign a cost value to each node. Merge the denoised point cloud data with the complemented wall data to generate a three-dimensional space model of the virtual hotel, and input it into the collision detection and path planning algorithm of the VR tour guide module. The cost value table of the path nodes is transmitted to the path planning algorithm as the weight basis for calculating the movement path; The RGB values selected by the user are converted into values in the range of [0, 1] through normalization processing. The coordinate difference between the starting point and the ending point is converted into absolute coordinates in combination with the room coordinate system. The normalized color data and the absolute coordinate layout instructions are associated with the user account and encrypted for storage, and then input into the AR customization module.

4. The virtual hotel tour customization system based on intelligent optimization and multi-modal interaction according to claim 3, characterized in that: In the multi-source data acquisition module, the process of feature extraction from the preprocessed 3D spatial point cloud data and user customization behavior data includes: The preprocessed three-dimensional point cloud data is projected onto the horizontal plane, divided into grids with a side length of 0.1 meters, the number of grids covering the virtual hotel wall point cloud is counted, multiplied by the area of a single grid, and accumulated to obtain the total area of the room. For the vertex coordinates of the furniture bounding box, the planar distance between the centers of each piece of furniture is calculated. Based on the vertex coordinates of the obstacle bounding box, the length, width, and height are calculated. In the point cloud of the passable area, the center point of the ground is extracted as a path node, and the distance between adjacent nodes is 0.5 meters. The corridor node cost value of 1 and the stair node cost value of 2 are assigned according to the area attributes where the path nodes are located, and they are used as the weight parameters of the path planning algorithm; The translation vectors in the user's historical adjustment data are counted, the high-frequency adjustment directions and amplitudes are calculated, and the user adjustment data is classified through the K-means clustering algorithm, and the translation vectors corresponding to the cluster centers are extracted; The total area of the room, the distance between the centers of the furniture, and the obstacle size are input into the collision detection and path planning algorithm of the VR navigation module, and the layout adjustment features are input into the AR customization module.

5. The virtual hotel tour customization system based on intelligent optimization and multi-modal interaction according to claim 4, characterized in that: In the VR navigation module, the process of using the bounding box-based collision detection algorithm to calculate the positional relationship between the user's virtual character position and each bounding box includes: From the preprocessed 3D spatial point cloud data, the vertex coordinates of the wall and furniture are extracted, and the minimum and maximum values of their three-dimensional coordinates are calculated respectively to form an axial bounding box; The user moves through the VR device, and the virtual hotel navigation customization system obtains the three-dimensional coordinates of the virtual character in real time based on the positioning system of the VR device, traverses the bounding boxes in the scene, and judges one by one whether the current position of the virtual character overlaps with the spatial range of the bounding box. If the x-axis, y-axis, and z-axis of the virtual character's three-dimensional coordinates are all within the minimum and maximum values of the bounding box's three-dimensional coordinates, it is determined that the virtual character collides with the bounding box; If a collision is detected, the virtual hotel navigation customization system executes constraints according to the bounding box type, prohibits the virtual character from entering the wall bounding box area, corrects the virtual character coordinates to the nearest edge outside the bounding box, allows the virtual character to approach the furniture bounding box area but restricts penetration, and triggers a red light flashing prompt. The bounding box types include walls and furniture.

6. The virtual hotel tour customization system based on intelligent optimization and multi-modal interaction according to claim 5, characterized in that: In the VR navigation module, the process of planning the best movement path from the user's starting point to the search ending point includes: Receive the path node coordinates, cost value table, user starting point, and ending point coordinate data through the VR navigation module; Accumulate the cost value from the starting point to the current node according to the path node cost value to obtain the actual cost, calculate the straight-line distance from the current node to the ending point using the Euclidean distance to obtain the estimated cost, and use the A* algorithm to define the total cost function as the sum of the actual cost and the estimated cost; Add the starting point to the open list, record its total cost function value, and perform iterative search. The iterative search process includes selecting the node with the smallest total cost function value from the open list as the current node. If the current node is the end point, backtrack the path and output it. If the current node is not the end point, traverse its adjacent nodes. For each adjacent node, calculate the actual cost and the estimated cost, and update the total cost function value. If the adjacent node is not in the open list and there is a better path, update the current node information and add it to the open list. Terminate the iteration when there is no feasible path, output the optimal path node sequence, transmit the optimal path node sequence to the VR scene rendering engine to display it in real time with green lines, and dynamically adjust the node cost value according to the hotel footfall.

7. The virtual hotel tour customization system based on intelligent optimization and multi-modal interaction according to claim 6, characterized in that: In the AR customization module, the process of using a color recognition algorithm based on the RGB color space to match the bedding color selected by the user includes: Match the normalized RGB values selected by the user with the bedding color library of the virtual hotel, calculate the Euclidean distance between the user's color and the colors in the bedding color library, select the color with the smallest distance as the matching result, record its RGB value as the user preference feature, associate the RGB value of the optimal matching color with the user account, encrypt and store it in the data storage management module, and overlay the matched bedding color on the virtual guest room model in the AR scene in real time.

8. An intelligent optimization and multi-modal interaction-based virtual hotel tour customization system according to claim 7, characterized in that: In the AR customization module, the process of using the simulated annealing algorithm to calculate the scores of each furniture layout plan and select the best furniture layout plan includes: Generate an initial furniture layout plan based on the preprocessed three-dimensional room model, furniture size data, and the user's historical layout preference features. Define the furniture positions using three-dimensional coordinates and orientation angles, comprehensively analyze the furniture spacing and the distance from the wall, and calculate the score of the furniture layout plan. Apply position perturbations to random furniture in the current layout and calculate the score of the new furniture layout plan , if the score of the new furniture layout plan is lower than the score of the current furniture layout plan , then accept the new furniture layout plan; otherwise, accept the new plan with probability P , where T is the current temperature parameter, and the initial temperature is set to 100. When the score of the furniture layout plan has not improved for 10 consecutive iterations, stop the search, output the three-dimensional coordinates and orientation parameters of the current optimal layout plan, transmit them to the AR scene rendering engine, and superimpose them on the virtual guest room model for the user to preview.

9. An intelligent optimization and multi-modal interaction-based virtual hotel tour customization system according to claim 8, characterized in that: In the user interaction interface, the process of generating a touch and voice interaction interface for displaying the VR tour results and AR customization effects includes: Dynamically adjust the display priority of the interface module according to the user's historical operation data, generate a touch and voice interaction interface through a template engine, support gesture sliding and voice commands to switch the tour and customization functions. The touch and voice interaction interface includes a tour control area, a customization operation area, and an information display area. The tour control area displays the VR scene switching button and the path planning status. The customization operation area embeds a touch color palette to display the RGB value in real time, and embeds a furniture layout adjustment panel to display the translation vector. The information display area dynamically displays the basic hotel information, the RGB value of the matched bedding color, and the scores of each furniture layout plan. When the user clicks on the color palette to select a color, the virtual hotel tour customization system captures the screen pixel coordinates, extracts the corresponding RGB value, and normalizes it to a value in the range of [0,1]. When the user drags the furniture model, record the coordinate difference between the starting point and the end point, and convert it to an absolute displacement. Collect voice signals through the mobile device microphone and the VR headset, use voiceprint recognition technology to distinguish the user's identity, set voice command keywords, and trigger corresponding operations. Transmit the optimal path node sequence output by the VR tour guide module to the interface tour guide control area, dynamically display the movement path with green connection lines, and superimpose the RGB values of the bed color and the furniture layout plan of the AR customization module on the information display area in real time, support perspective rotation and zoom preview. When the furniture layout score is lower than the set score threshold, prompt with red light flashing and push it to the information display area.

10. An intelligent optimization and multi-modal interaction-based virtual hotel tour customization system according to claim 9, characterized in that: In the data storage and management module, the process of storing hotel VR tour guide data and AR customization data using encryption technology includes: The data storage and management module receives VR tour guide and AR customization data. The VR tour guide data includes the virtual hotel three-dimensional model, path node coordinates, cost value table, and user movement trajectory. The AR customization data includes the RGB values of the bed color selected by the user, the three-dimensional coordinates and orientation angles of the furniture layout plan, and the translation vector clustering center of the historical adjustment record. Field-level encryption is performed on the structured data in the VR tour guide and AR customization data, and block encryption technology is used for unstructured data. A combination of symmetric encryption algorithm and asymmetric encryption technology is used to generate a key through the user's voiceprint feature; Store the encrypted data in database tables by type. The database tables include a VR tour guide table and an AR customization table. The VR tour guide table stores the encrypted virtual hotel three-dimensional model hash value, path node ciphertext, and cost value ciphertext. The AR customization table stores the encrypted RGB values, furniture layout plan ciphertext, and user historical operation log ciphertext. Verify the data integrity through a hash function. Calculate the ciphertext hash value when storing the data and save it independently. Verify the hash consistency when accessing. Bind the access permission to voiceprint identity authentication, and unauthorized users cannot decrypt the data; After the user verifies their identity through voiceprint recognition, the virtual hotel tour guide customization system calls the pre-stored voiceprint feature to regenerate the encryption key. If the key verification fails, access to the encrypted data is prohibited. When the VR tour guide module and the AR customization module request data, the virtual hotel tour guide customization system decrypts the corresponding fields and data blocks. The decrypted data is temporarily stored in memory and cleared immediately after use.

Images