A scenic spot interactive simulation robot

Abstract

The present invention discloses a scenic spot interactive simulation robot. The robot includes a central control unit, a wireless communication unit, a detection unit, an execution unit, an AI agent, a human-computer interaction unit, a storage unit and a simulated human form display unit. The detection unit includes a binocular camera and a radar for real-time perception of the environment and tourists in the scenic spot. The execution unit adjusts the operating status of the equipment in the scenic spot according to the instructions of the central control unit. The AI ​​agent is based on large model technology, and can analyze the data of the detection unit and generate control instructions to achieve real-time interaction and personalized services with tourists. The simulated human form display unit simulates facial expressions and body movements through a mechanical structure to provide tourists with a realistic interactive experience. The invention combines advanced AI technology and simulated robots to provide tourists with a series of services such as inquiry, guided tours, ticket purchase, and photo taking, thereby improving the overall experience of tourists and the management efficiency of the scenic spot.

Description

Technical Field

[0001] The invention relates to an intelligent service robot, in particular to a scenic spot simulation service robot and an interactive control system thereof. Background Art

[0002] With the rapid development of science and technology and the improvement of people's living standards, the tourism industry has achieved remarkable development. As one of the important destinations for tourists, scenic spots are faced with the dual challenges of improving service quality and management efficiency. Traditional scenic spot service methods usually rely on manual operation. Service personnel need to face a large number of tourists, making it difficult to provide real-time interaction and personalized services, resulting in poor tourist experience. In addition, manual services are costly and inefficient, making it difficult to meet the needs of modern scenic spots. Therefore, it is particularly important to develop an intelligent service robot that can achieve real-time interaction and personalized services with tourists through AI technology.

[0003] In recent years, the rapid development of AI technology, especially deep learning and natural language processing technology, has provided new possibilities for intelligent service robots. AI technology based on large models can understand and process complex voice commands and generate natural language responses, thereby achieving efficient interaction with tourists. At the same time, the development of simulation robot technology enables robots to have human-like facial expressions and body movements, providing tourists with a more realistic and friendly interactive experience.

[0004] Despite this, most service robots on the market still have problems such as insufficient interactivity, lack of personalized services, limited environmental perception capabilities, and single service functions. These problems seriously restrict the application effect of service robots and make it difficult to meet the diverse needs of modern scenic spots. Summary of the invention

[0005] In order to solve the above problems, the present invention proposes an interactive simulation robot for scenic spots, which integrates advanced AI technology and simulation robot technology to provide efficient, intelligent and personalized services for scenic spot tourists, significantly improving tourist experience and scenic spot management efficiency.

[0006] The invention discloses a scenic spot interactive simulation robot, comprising two main components: a robot body and a control system.

[0007] The robot body comprises a humanoid display unit, a detection and perception unit, an execution unit and a human-machine interaction unit.

[0008] The simulated human form display unit comprises a mechanical structure facial expression control module and a mechanical structure limb movement control module.

[0009] Furthermore, the mechanical structure facial expression control module includes multiple micro servo motors for driving different facial areas to achieve expression changes; the facial expression control module generates specific expression control signals according to the instructions of the AI ​​agent through a facial expression control algorithm.

[0010] Furthermore, the mechanical structure limb motion control module includes multiple joint servo motors for driving the movement of each joint of the limb; the limb motion control module generates specific limb motion control signals according to the instructions of the AI ​​intelligent body through a limb motion control algorithm.

[0011] The detection and perception unit includes a binocular camera, a radar, a touch screen and a voice recognition module, which is used to perceive the environment and tourist information in the scenic area in real time.

[0012] Furthermore, the binocular camera is used to capture image information within the scenic area, the radar is used to detect the surrounding environment and the location and movement of tourists, the touch screen is used to receive touch input from tourists, and the voice recognition module is used to receive and process voice commands from tourists.

[0013] The execution unit includes a navigation module, a query module, a ticket purchasing module and a service module.

[0014] Furthermore, the navigation module provides navigation and tour services within the scenic area. Visitors can select scenic spots through the touch screen or voice input, and the robot will display detailed tour information on the touch screen, including scenic spot introductions, route maps and recommended tour routes. At the same time, the robot can also explain the history and cultural background of the scenic spot to tourists through voice.

[0015] Furthermore, the query module realizes the query of scenic spot information. Visitors can query through the touch screen or voice input, and the robot will display the query results on the touch screen.

[0016] Furthermore, the ticket purchase module provides the purchase of scenic spot tickets and service tickets. Visitors can select the required ticket type and quantity through the touch screen, and then display the payment code through the mobile payment interface to complete the ticket purchase. The robot will generate an electronic ticket in real time and display it on the touch screen. Visitors can directly scan and download it or receive the ticket via SMS.

[0017] Furthermore, the service module provides other personalized services, including a photo taking function. Visitors can select a photo taking mode through voice commands or touch screen, and the robot will adjust the angle and expression according to the command and take a photo with the visitor. After taking the photo, the robot will generate a QR code and display it on the touch screen, and visitors can download the photo or video by scanning the QR code.

[0018] The human-computer interaction unit includes a touch screen, a voice recognition module and a mobile payment interface.

[0019] Furthermore, the touch screen is used to display guide information and other interactive content; the voice recognition module is used to receive and process voice commands from tourists; and the mobile payment interface is used for tourists to purchase tickets by showing payment codes on their mobile phones.

[0020] The control system includes a central control unit, a wireless communication unit, an AI agent and a storage unit.

[0021] Furthermore, the central control unit is used to process and manage all data and instructions in the system; it is connected to each unit through an advanced microprocessor control method to achieve data collection, processing and issuance of instructions.

[0022] The wireless communication unit includes a WiFi module and a GPRS module.

[0023] Furthermore, the WiFi module is used to form a local area network for communication between devices in the system; the GPRS module is used for communication between the system and the scenic area management center.

[0024] The AI ​​agent includes an AI voice interaction processing module and an AI large model control center.

[0025] Furthermore, the AI ​​voice interaction processing module is used to process tourists' voice commands and generate corresponding system responses; the AI ​​large model control center is used to generate personalized service instructions based on tourists' needs and environmental parameters.

[0026] The storage unit includes a large-capacity memory.

[0027] Furthermore, the large-capacity memory is used to store system operation logs, visitor feedback data and environmental monitoring data; the large-capacity memory can save a large amount of data for a long time, providing data support for the operation and maintenance of the system.

[0028] The control method of the present invention comprises the following steps:

[0029] The detection and perception unit perceives the environment and visitor information in the scenic area in real time through binocular cameras, radars, touch screens and voice recognition modules, and transmits the data to the central control unit.

[0030] The central control unit transmits the data transmitted by the detection and perception unit to the AI ​​agent for analysis and decision-making; based on the large model technology, the AI ​​agent processes complex voice commands and environmental information to generate personalized service instructions.

[0031] The central control unit transmits the control instructions generated by the AI ​​agent to the execution unit and the humanoid display unit; the execution unit adjusts the operating state of the equipment according to the instructions, and the humanoid display unit realizes simulated facial expressions and body movements through a mechanical structure, providing tourists with a realistic interactive experience.

[0032] The storage unit records system operation logs, visitor feedback data and environmental monitoring data, providing data support for system analysis and optimization.

[0033] By combining advanced AI technology and simulation robot technology, the present invention provides efficient, intelligent and personalized services for tourists in scenic spots, significantly improving the overall experience of tourists and the management efficiency of scenic spots. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 A system structure block diagram provided for an embodiment of the present invention;

[0035] Figure 2 A schematic diagram of a workflow provided by an embodiment of the present invention;

[0036] Figure 3 A schematic diagram of an actual robot structure provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0037] The present invention is described in detail below in conjunction with specific embodiments. The following embodiments will help those skilled in the art to further understand the present invention, but are not intended to limit the present invention in any form. It should be noted that, for those of ordinary skill in the art, several variations and improvements may be made without departing from the concept of the present invention. These all belong to the protection scope of the present invention.

[0038] like Figure 1 As shown in the structural framework diagram, a scenic spot interactive simulation robot includes two main components: a robot body and a control system.

[0039] The robot body comprises a humanoid display unit, a detection and perception unit, an execution unit and a human-machine interaction unit.

[0040] The simulated human form display unit comprises a mechanical structure facial expression control module and a mechanical structure limb movement control module.

[0041] Furthermore, the mechanical structure facial expression control module includes a plurality of micro servo motors for driving different facial regions to achieve expression changes; the facial expression control module generates basic facial expression data using audio2face technology through a facial expression control algorithm. These data are represented in the form of blendshape, describing the deformation of each specific facial region. Blendshape is a set of numerical vectors, which define the weights of the movement amplitudes of various facial parts (such as opening the mouth and raising the eyebrows). Once the blendshape data is generated, the system needs to further convert this data into a control signal for the servo, and the value of the blendshape data (usually a floating point number between 0 and 1) will be mapped to the angle adjustment of the control servo. The rotation angle range of the servo is usually 0-180 degrees, so a blendshape value will correspond to a certain amount of angles, which reflect the micro-movements of the facial region, such as the raising of the corners of the mouth or the closing of the eyelids. The conversion mapping process uses a regression-type deep learning model, the model input is the expression weight (blendshape vector), and the output is the corresponding servo angle. In the training stage, a large amount of collected expression weights and corresponding servo angle data are input into the deep learning model for supervised learning. The training process uses the mean square error (MSE) as the loss function to minimize the difference between the model's predicted servo angle and the actual servo angle. The optimization method can use the Adam gradient descent optimizer to ensure that the model converges quickly, so that it can well derive the accurate servo angle from the input expression weight. In the execution stage, the model has been fully trained, and when given a new expression requirement (new blendshape weight vector), the model can very quickly reverse calculate the corresponding servo control parameters (angle position). The specific operation is: input the expression weight vector to be presented into the deep learning model, and the model will output a set of corresponding servo angles. These angle data will be sent to the servo through PWM signals, thereby driving various parts of the robot's face to achieve the target expression. The model can be continuously optimized based on the actual feedback effect. If a large error occurs during the execution process (the expression is not natural enough or does not meet expectations), the model can make the future inferred servo data more accurate through incremental learning or continued parameter adjustment.

[0042] Furthermore, the mechanical structure limb motion control module includes a plurality of servo motors for driving the joints of the robot. These servo motors are responsible for driving the movement of the mechanical limbs, such as the key joints of the arms, legs, shoulders, etc., to achieve realistic limb movements. The module processes the input audio or other instructions through a limb motion control algorithm based on deep learning, thereby generating limb motion skeleton data that meets the requirements. Then, these skeleton data are converted into corresponding servo motor control signals to drive the mechanical limbs to perform natural movements and posture changes. This process is partially achieved by using audio2gesture technology to map audio or motion instructions to skeleton motion data of joints. The input data can be audio signals (such as voice, background sound) or data from external input devices (such as motion controllers, sensors). Audio2gesture uses deep learning methods to learn to infer appropriate limb movements from these signals. Audio2gesture can generate accurate motion trajectories by learning the correlation between a large amount of voice / audio data and its corresponding limb movements (such as waving, nodding, and raising arms). These association information will generate a set of 3D skeleton motion data, which will describe the target motion position and target angle of each joint. In order to ensure the naturalness and coherence of limb movements, this module uses a deep learning model to generate high-level skeletal motion data from the input. A large amount of human limb motion data is collected through motion capture equipment, and the motion trajectory of each joint and the corresponding action command or audio input are recorded. At the same time, the rotation angle data of the corresponding joint servo motor is recorded, and a high-accuracy limb motion data set is generated based on the feedback mechanism. A regression-type deep learning model is used to capture and learn the complex association between limb movements and audio / command inputs, and the time dependency between actions is taken into account when designing the model. Through training, the model learns how to generate complex 3D motion data for each joint according to the target action command (while completing time coordination and transition). Similar to the mapping process in facial expression control, skeletal motion data needs to be converted into angle control signals for the actual operation of the servo motor. After the skeletal motion data generated by the front end, each joint has a corresponding target position or rotation angle. Through geometric transformation or coordinate transformation, these data are mapped to the pulse width modulation (PWM) control signal of the servo motor to determine the rotation angle of the servo motor. Considering the mechanical characteristics of the servo motor, each joint has a certain rotation constraint. For example, the shoulder joint may only be able to rotate within a limited angle range (such as 90 degrees), so the target position constraint should be set in the control algorithm to ensure that the output motor signal meets the motion limitations of the mechanical structure. In order to facilitate the coherent changes in limb movements, the system will interpolate the angle changes output by multiple servos. For example, when the arm moves from the 1 o'clock position to the 2 o'clock position, the servo motor will use a series of linear or nonlinear interpolations to make the joint transition in the form of a smooth curve to avoid the sense of disconnection of mechanical movements.The Bezier curve smoothing algorithm is added to optimize the acceleration and deceleration curves of the servo motor according to the nature of the limb movement to ensure that there will be no abrupt jumps or jerky movements during the operation of the servo motor. In order to further improve the accuracy of limb movements, a simple feedback control mechanism is added to fine-tune the movement by combining the potentiometer data and the real-time feedback of the servo motor: the servo motor will fine-tune to the target angle according to the feedback signal. If each joint cannot accurately reach the target position due to external forces or mechanical loads, the system will readjust the input of the execution signal according to the actual feedback signal until the joint reaches the desired posture. When the servo motor approaches its hardware limit, the feedback mechanism will also capture the information of the joint to ensure that it does not exceed the safety range. Since the robot may encounter various operations and external feedback in the actual environment, the system can make real-time adjustments to the actual scene to ensure the precise operation of the servo motor. As a closed-loop control, real-time joint position feedback can be used to correct the precision error when the servo motor performs the action. For example, when the servo motor movement is changed by load or external force, the system can readjust the posture of the limb through feedback to ensure the fidelity and accuracy of the action. If the robot has a series of unexpected action deviations, the system can be adaptively updated through incremental learning. In actual application scenarios, the system records new motion data and fine-tunes the deep learning model to improve subsequent athletic performance in similar situations.

[0043] The detection and perception unit includes a binocular camera, a radar, a touch screen (7-inch capacitive touch screen) and a voice recognition module, which are used to perceive the environment and tourist information in the scenic area in real time.

[0044] Furthermore, the binocular camera is used to capture image information in the scenic area, the radar is used to detect the surrounding environment and the location and movement of tourists, the touch screen is used to receive touch input from tourists, and the voice recognition module is used to receive and process voice commands from tourists. For example, tourists can ask for information about the scenic area through voice commands, and the robot will receive the command through the voice recognition module and transmit it to the central control unit.

[0045] The execution unit includes a navigation module, a query module, a ticket purchasing module and a service module.

[0046] Furthermore, the navigation module provides navigation and tour guide services within the scenic area. Visitors can select scenic spots through the touch screen or voice input, and the robot will display detailed tour guide information on the touch screen, including scenic spot introductions, route maps and recommended tour routes. At the same time, the robot can also explain the history and cultural background of the scenic spot to tourists through voice. For example, visitors can select a certain scenic spot through the touch screen, and the robot will display detailed information of the scenic spot and provide voice guide.

[0047] Furthermore, the query module enables querying of scenic spot information. Visitors can query through the touch screen or voice input, and the robot will display the query results on the touch screen. For example, visitors can query catering services through voice commands, and the robot will display nearby catering options and provide relevant information.

[0048] Furthermore, the ticket purchase module provides the purchase of scenic spot tickets and service tickets. Visitors can select the required ticket type and quantity through the touch screen, and then display the payment code through the mobile payment interface to complete the ticket purchase. The robot will generate an electronic ticket in real time and display it on the touch screen. Visitors can directly scan and download or receive the ticket via SMS. For example, visitors can select the ticket purchase option through the touch screen, and after completing the payment, the robot will generate and display the electronic ticket.

[0049] Furthermore, the service module provides other personalized services, including a photo taking function. Visitors can select a photo taking mode through voice commands or touch screen, and the robot will adjust the angle and expression according to the command and take a photo with the visitor. After the photo is taken, the robot will generate a QR code and display it on the touch screen, and visitors can download photos or videos by scanning the QR code. For example, visitors can select a photo taking mode through voice commands, and the robot will adjust the angle and take a photo with the visitor according to the command, and generate a QR code for the visitor to download the photo after the photo is taken.

[0050] The human-computer interaction unit includes a touch screen (7-inch capacitive touch screen), a voice recognition module and a mobile payment interface (supporting, for example, a QR code scanning module).

[0051] Furthermore, the touch screen is used to display guide information and other interactive content; the voice recognition module is used to receive and process voice commands from tourists; and the mobile payment interface is used for tourists to purchase tickets by displaying payment codes on their mobile phones. For example, tourists can select ticket purchase options on the touch screen and complete payment through the mobile payment interface.

[0052] The control system includes a central control unit, a wireless communication unit, an AI agent and a storage unit.

[0053] Furthermore, the central control unit is used to process and manage all data and instructions in the system; it is connected to each unit through an advanced microprocessor control method to achieve data collection, processing and issuance of instructions.

[0054] The wireless communication unit includes a WiFi module and a GPRS module.

[0055] Furthermore, the WiFi module is used to form a local area network for communication between devices in the system; the GPRS module is used for communication between the system and the scenic area management center.

[0056] The AI ​​agent includes an AI voice interaction processing module and an AI large model control center.

[0057] Furthermore, the AI ​​voice interaction processing module is used to process the voice instructions of tourists and generate corresponding system responses. Specifically, the module first converts the voice input of tourists into text format through the built-in voice recognition engine. This process involves signal processing and analysis of the recorded audio signal to ensure accurate recognition of the voice content and convert it into machine-readable text data.

[0058] After conversion to text, the text information will be passed to the AI ​​big model control center, for example, based on a large-scale pre-trained language model. After receiving this text input, the model uses its deep learning capabilities and understanding of the context, combined with the needs of tourists and environmental parameters, which are collected by sensors and converted into data descriptions and sent to the big model (such as traffic data), and then generates personalized service instructions (such as crowd congestion reminders). These instructions will be returned in text form, accurately reflecting the specific needs of tourists.

[0059] Next, the generated text instructions are converted into natural and fluent speech output through a text-to-speech (TTS) synthesis module to facilitate tourists' interaction.

[0060] Finally, the voice output is converted into control signals using audio2face technology and input into the humanoid display unit to achieve control of the robot.

[0061] The storage unit includes a large-capacity memory, and uses an SSD for large-capacity storage.

[0062] Furthermore, the large-capacity memory is used to store system operation logs, visitor feedback data and environmental monitoring data; the large-capacity memory can save a large amount of data for a long time, providing data support for the operation and maintenance of the system.

[0063] The following describes in detail a scenic spot interactive simulation robot provided by an embodiment of the present invention in conjunction with the accompanying drawings. Figure 2 , Figure 2 It is a schematic diagram of the working process of a scenic spot interactive simulation robot provided by an embodiment of the present invention.

[0064] Step S1: The detection and perception unit perceives the environment and tourist information in the scenic area in real time through binocular cameras, radars, touch screens and voice recognition modules, and transmits the data to the central control unit.

[0065] Step S2: The central control unit transmits the data transmitted by the detection and perception unit to the AI ​​agent for analysis and decision-making; the AI ​​agent processes complex voice commands and environmental information based on large model technology to generate personalized service instructions.

[0066] Step S3: The central control unit transmits the control instructions generated by the AI ​​agent to the execution unit and the humanoid display unit; the execution unit adjusts the operating state of the equipment according to the instructions, and the humanoid display unit realizes simulated facial expressions and body movements through a mechanical structure, providing tourists with a realistic interactive experience.

[0067] Step S4: The storage unit records the system operation log, visitor feedback data and environmental monitoring data to provide data support for system analysis and optimization.

[0068] It should be noted that the above-described embodiments are merely illustrative, wherein each module can be adjusted or combined according to specific needs, and the control unit, detection and perception unit, execution unit, AI agent, and human-computer interaction unit, etc. can select some or all of the modules according to actual needs to achieve the purpose of the present embodiment. Parts irrelevant to the content are omitted.

[0069] By combining advanced AI technology and simulation robot technology, the present invention provides efficient, intelligent and personalized services for tourists in scenic spots, significantly improving the overall experience of tourists and the management efficiency of scenic spots.

[0070] The above is only a preferred specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed by the present invention should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention shall be based on the protection scope of the claims.

Claims

1. A scenic spot interactive simulation robot, characterized in that: The robot comprises: A central control unit, which processes and manages all data and instructions in the robot's system; A wireless communication unit, used to realize data transmission and communication between various devices in the system; A detection unit, used to detect the environment and tourists in the scenic area; An execution unit, used to adjust the operating status of the equipment in the scenic area according to the instructions of the central control unit; AI agent, used to analyze and make decisions based on the data from the detection unit and generate control instructions; Human-computer interaction unit, used for information exchange and interaction between visitors and the system; A storage unit, used to store system operation data, visitor demand data and environmental parameter data; A humanoid display unit is used to display facial expressions and body movements of a humanoid image; the control algorithm of the facial expression includes: using audio2face technology to generate basic facial expression data, the facial expression data is expressed in the form of blendshape, describing the deformation of each specific area of ​​the face, converting the facial expression data into a control signal for a servo, and the value of the blendshape data is mapped to the angle adjustment of the servo; the mapping process uses a regression-type deep learning model, the model input is the expression weight, and the output is the corresponding servo angle; in the training stage, the collected expression weight and the corresponding servo angle data are input into the deep learning model for supervised learning; in the execution stage, the expression weight vector to be presented is input into the deep learning model, and a set of corresponding servo angles are output, and the servo angles are sent to the servo via a PWM signal, thereby driving various parts of the robot's face to achieve the target expression; The control algorithm of the limb movement first generates limb movement skeleton data according to the instructions of the AI ​​agent, converts the limb movement skeleton data into corresponding servo motor control signals, and drives the mechanical limb to perform movements and posture changes; wherein, When generating limb movement skeletal data, an audio2gesture technology based on deep learning is used to map audio or action instructions into skeleton movement data of joints; the audio2gesture technology based on deep learning generates a movement trajectory by learning the correlation between voice / audio data and its corresponding limb movements, including: firstly collecting a large amount of human limb movement data through a motion capture device, recording the movement trajectory of each joint and the corresponding action instruction or audio input; at the same time, recording the rotation angle data of the corresponding joint servo motor to generate a limb movement data set; using a regression-type deep learning model to capture and learn the correlation between limb movements and audio / command inputs, then training the deep learning model, and using the trained model to generate 3D movement data of each joint according to the target action instruction.

2. The scenic area interactive simulation robot according to claim 1, characterized in that: The simulated human form display unit comprises: A mechanical structure facial expression control module, used to control the facial expression changes of the simulated human image; The mechanical structure limb motion control module is used to control the limb motion of the simulated human image.

3. The scenic area interactive simulation robot according to claim 2, characterized in that: The mechanical structure facial expression control module comprises: Multiple micro servo motors to drive different areas of the face to achieve expression changes; Facial expression control algorithm is used to generate specific expression control signals according to the instructions of the AI ​​agent.

4. The scenic area interactive simulation robot according to claim 2, characterized in that: The mechanical structure limb motion control module comprises: Multiple joint servo motors are used to drive the movement of each joint of the limbs; The limb motion control algorithm is used to generate specific limb motion control signals according to the instructions of the AI ​​agent.

5. The scenic area interactive simulation robot according to claim 1, characterized in that: The control algorithm of the limb movement further includes: The feedback control mechanism combines the potentiometer data and the real-time feedback of the servo motor to fine-tune the movement. If each joint cannot reach the target position accurately, the input of the execution signal is readjusted according to the actual feedback signal until the joint reaches the desired posture. If the robot has movement deviations, adaptive updates are performed through incremental learning.

6. The scenic area interactive simulation robot according to claim 1, characterized in that: The AI ​​agent includes: AI voice interaction processing module, used to process visitors’ voice commands and generate corresponding system responses; The AI ​​large model control center is used to generate personalized service instructions based on tourist needs and environmental parameters.

7. The scenic area interactive simulation robot according to claim 6, characterized in that: The AI ​​voice interaction processing module first converts the visitor's voice input into text format through the built-in voice recognition engine to obtain text information; The text information is transmitted to the AI ​​large model control center, which generates personalized service instructions based on the needs of tourists and environmental parameters through deep learning capabilities and understanding of context, and returns the service instructions in text form to the speech synthesis module for conversion into speech output; The voice output is converted into a control signal and input into a humanoid display unit to realize control of the robot.

8. The scenic area interactive simulation robot according to any one of claims 1 to 7, characterized in that: The collaborative working modes between the robot units include: The detection unit senses the environment and visitor information through binocular cameras and radars, and transmits the data to the central control unit; The central control unit processes the data transmission from the detection unit into the input data required by the large model and provides it to the AI ​​agent for analysis and decision-making; The AI ​​agent generates control instructions based on the analysis results and transmits them back to the central control unit; The central control unit processes the control instructions into execution data, transmits them to the execution unit and the humanoid display unit, and adjusts the operating status of the equipment and the facial expressions and body movements of the humanoid in real time; The human-computer interaction unit receives the input from tourists and transmits the information to the central control unit, which then gives feedback to tourists based on the decision of the AI ​​agent; The storage unit records system operation logs, visitor feedback data and environmental monitoring data for system analysis and optimization.

Images