Merculation edge-equipped intelligent service vehicle system and control method

Abstract

The invention discloses a mercuration edge self-holding intelligent service vehicle system and a control method. The system comprises a man-machine interaction and voice instruction analysis module, a visual identification and distance measurement module, an autonomous navigation and obstacle avoidance module, a mechanical arm and manipulator collaborative grabbing module and a dual-core master control communication module. The man-machine interaction and voice instruction analysis module is used for receiving a voice instruction of a user, analyzing the voice instruction into structured data and sending the structured data to the main control board. The visual identification and distance measurement module is used for obtaining an environment image, carrying out object identification, and calculating a three-dimensional space coordinate of a target object. And the autonomous navigation and obstacle avoidance module is used for planning a motion path and controlling the trolley to move and avoid obstacles. The mechanical arm and manipulator cooperative grabbing module is used for grabbing a target object. And the dual-core master control communication module is used for coordinating communication among the modules and executing a motion control instruction.

Description

Technical Field

[0001] This invention relates to the field of intelligent robot control, specifically to the Ascend Edge Embossed Intelligent Service Vehicle System and its control method. Background Technology

[0002] With the increasing demand in the fields of smart elderly care and smart homes, various robotic devices have emerged to meet market needs. Currently, while methods exist for object position recognition in robot vision systems, they lack the ability to make autonomous decisions and classify objects. Simultaneously, there are intelligent catering robot systems and motion methods based on AI visual recognition; however, they have shortcomings in human-computer interaction, making it difficult for the robot to "communicate" with users and effectively understand their speech. Furthermore, there are obstacle avoidance path planning methods for multi-degree-of-freedom robotic arms based on 3D reconstruction, but these methods lack object grasping capabilities, making it impossible to use the robotic arm to complete object grasping and delivery tasks.

[0003] Currently, most mainstream intelligent service robots operate using manual remote control, fixed-position grasping, and simple command recognition modes, lacking key performance characteristics such as autonomy, adaptability, and interactivity. From a technological perspective, they are relatively mature in basic movement, obstacle avoidance, and simple command execution, with costs within a controllable range. However, their disadvantages are also quite obvious, primarily manifested in a lack of intelligent decision-making capabilities. When user language is unclear, the machine's automatic understanding may falter, leading to an inaccurate grasp of the user's intentions and the grasping of incorrect items. Furthermore, multiple functions are fragmented, with each module operating independently, failing to establish a complete system organically integrating human-computer interaction, visual recognition, motion navigation, and automatic grasping. Moreover, the interaction threshold is high; many robots require users to spend time learning complex remote control operations, which poses significant difficulties for the elderly, disabled, and other people with mobility impairments, failing to adequately meet their actual needs.

[0004] In conclusion, existing technologies are insufficient to effectively address the critical issue of "enabling people with mobility impairments to conveniently operate robots for accurate item positioning and delivery through simple and easy-to-use human-computer interaction."

[0005] Chinese invention patent application number 202111654350.6 discloses "An Autonomous Navigation Robot for Live-Line Operation in Power Distribution Networks and Its Working Method," comprising: a robot body, a target detection module, a motion planning module, and a grasping point detection module. The robot body is equipped with a grasping component and a vision platform. The target detection module is configured to identify target information and reconstruct a 3D scene based on visual data acquired by the vision platform to locate the work target. The motion planning module is configured to plan the obstacle avoidance path of the grasping component in the 3D scene using a node-controlled optimal fast search random tree algorithm based on the work target identification result and in a node-controlled manner. The grasping point detection module is configured to identify the position coordinates of the grasping point and convert them into joint angle values ​​of the grasping component, enabling the grasping component to complete the grasping task according to the obstacle avoidance path. Its technical solution is a dedicated robot that performs predefined tasks in a closed, structured industrial environment, and its intelligence is reflected in the optimization algorithm for specific work targets. Summary of the Invention

[0006] To address the technical problems existing in the background art, this invention provides an Ascend Edge Embossed Intelligent Service Vehicle System and Control Method. The technical solution adopted by this invention is as follows: The first aspect of the present invention provides an Ascend Edge Embossed Intelligent Service Vehicle System, the system comprising a human-computer interaction and voice command parsing module, a visual recognition and ranging module, an autonomous navigation and obstacle avoidance module, a robotic arm and robotic hand collaborative grasping module, and a dual-core main control communication module; The human-computer interaction and voice command parsing module is used to receive users' voice commands through mobile terminal applications and use artificial intelligence big data models to parse the voice commands into understandable structured data and send them to the main control board. The visual recognition and ranging module is used to acquire environmental images and use a target detection model to identify objects. At the same time, it integrates physical ranging and depth ranging schemes to calculate the three-dimensional spatial coordinates of the target object. The autonomous navigation and obstacle avoidance module includes a motion controller, a gyroscope, and an obstacle avoidance sensor, which are used to plan the motion path based on the coordinates of the target object and control the vehicle to avoid obstacles during movement. The robotic arm and robotic hand collaborative grasping module includes a robotic arm, a robotic hand, and a robotic arm camera, which is used to re-identify the target after the vehicle arrives near the target object and control the grasping action through inverse kinematics calculation. The dual-core main control communication module includes a main control board and an expansion board. The main control board is used to run visual recognition and path planning algorithms and coordinate communication between modules. The expansion board is used to execute motion control commands and exchange data with the main control board through a serial communication interface.

[0007] As a preferred embodiment, in the human-computer interaction and voice command parsing module, the mobile terminal application embeds an AI large model API, which is used to convert the user's voice command into text, construct specific prompt words and send them to the large model API, and parse the key command information from the JSON format data returned by the API and send it to the main control board.

[0008] As a preferred embodiment, in the visual recognition and ranging module, the physical ranging uses an ultrasonic sensor, and the depth ranging uses the depth flow of a depth camera for geometric ranging; the system prioritizes physical ranging at long distances and depth ranging at short distances based on the confidence level of the depth data.

[0009] As a preferred embodiment, in the autonomous navigation and obstacle avoidance module, the obstacle avoidance sensor is an ultrasonic sensor, and this sensor is only activated when the vehicle is moving to search for the target object, so as to avoid measurement errors caused by motor and power interference when the vehicle is stationary.

[0010] As a preferred embodiment, in the robotic arm and robotic hand collaborative grasping module, the robotic arm camera acquires images, performs target recognition and coordinate transformation through a preset target detection model, and then calls the inverse kinematics algorithm to calculate the target angles of each joint of the robotic arm and drives the joints to perform grasping.

[0011] As a preferred embodiment, the main control board is an Atlas 200I DK A2 board, and the expansion board is an i.MXRT1064 expansion board. The two communicate with each other via the UART protocol for commands and status data.

[0012] As a preferred embodiment, the system also includes a mapping and navigation module, which is used to directly retrieve all obstacles and targets in the current site by creating a map of the environment, and to plan the complete path from the current position of the vehicle to the target object in one go.

[0013] As a preferred embodiment, in the robotic arm and robotic hand collaborative grasping module, a deep learning-based end-to-end control method directly inputs and outputs servo control signals based on images from the robotic arm's camera through a pre-trained neural network model.

[0014] As a preferred option, the system also supports two wireless connection modes: In the first mode, the mobile terminal connects to the main control board through a router and a network card, and the mobile terminal directly accesses the Internet. In the second mode, the mobile terminal connects directly to the main control board via the AP, and the mobile terminal accesses the Internet by forwarding traffic through the main control board.

[0015] A second aspect of the present invention provides a control method for an Ascend Edge Embossed Intelligent Service Vehicle, the control method comprising: The system receives user voice commands via a mobile terminal application and uses a large artificial intelligence model to parse the voice commands into structured data, which is then sent to the main control board. The main control board starts the visual recognition and ranging module, acquires environmental color images and depth images through the depth camera, identifies target objects using the target detection model, and calculates the three-dimensional spatial coordinates of the target objects by combining physical ranging and depth ranging. The main control board plans the motion path based on the three-dimensional coordinates of the target object and the current position of the vehicle, and controls the movement of the vehicle through the expansion board; during the movement, obstacle avoidance sensors are used to monitor obstacles in real time to achieve obstacle avoidance and path replanning. When the car reaches a preset distance in front of the target object, it switches to the robotic arm camera to identify the target object again and calculate its coordinates relative to the robotic arm. The angles of each joint of the robotic arm are obtained through inverse kinematics calculation, and the robotic arm and robotic hand are driven to complete the grasping. After the capture is completed, the vehicle identifies the human body through the visual recognition module and autonomously returns to the user's location.

[0016] Compared with the prior art, the beneficial effects of this invention are: This invention introduces natural language interaction based on a large AI model as the system entry point and constructs prompt words to parse ambiguous user commands, achieving a paradigm shift from "executing predefined programs" to "understanding human intentions." This expands the system's application scenarios from closed industrial fields to open and dynamic home environments, significantly lowering the barrier to entry. By integrating physical ranging and depth-vision ranging schemes and dynamically selecting the appropriate one based on depth data confidence, it effectively overcomes the inherent limitations of single ranging technologies in terms of insufficient long-distance accuracy and short-distance data failure, achieving more stable and accurate 3D positioning of target objects in complex lighting and scenarios within a home environment. Through real-time obstacle avoidance, the complex global path planning problem is decomposed into simple sequential control steps, achieving efficient and reliable target approach, reducing the system's dependence on computing resources, and preventing collisions with target objects. By constructing a dual-core main control architecture consisting of a high-performance AI computing board and a real-time motion control board, and by implementing task division and collaboration, complex AI inference, visual processing, and highly reliable real-time motion control are decoupled, resulting in a significant improvement in overall system performance and stability. After the vehicle navigates and positions itself, the system switches to the robotic arm's camera for secondary visual recognition and positioning, and performs inverse kinematics calculations, thus enabling precise end-effector manipulation of the target object. By integrating and connecting multiple functional modules such as voice interaction, visual recognition, autonomous navigation, and robotic arm grasping into a complete automated process, the system overcomes the shortcomings of existing technologies where functions are fragmented. It constructs an end-to-end intelligent service system that can understand user commands, autonomously find targets, accurately grasp them, and safely return, achieving full automation of the delivery process for everyday items. Attached Figure Description

[0017] Figure 1 This is a structural diagram of the Ascend Edge Embossed Intelligent Service Vehicle system provided in this embodiment; Figure 2 This is a flowchart of the Ascend Edge Embossed Intelligent Service Vehicle control method provided in this embodiment; Explanation of reference numerals in the attached figures: 1. Plate area; 2. Camera carried on the robotic arm; 3. Robotic arm; 4. Astra Pro camera; 5. 3D printed shell. Detailed Implementation

[0018] The accompanying drawings are for illustrative purposes only and should not be construed as limiting the invention. It should be understood that the described embodiments are merely some, not all, of the embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of the embodiments of this application.

[0019] The terminology used in the embodiments of this application is for the purpose of describing particular embodiments only and is not intended to limit the embodiments of this application. The singular forms “a,” “the,” and “the” used in the embodiments of this application and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used herein refers to and includes any or all possible combinations of one or more of the associated listed items.

[0020] In the following description, when referring to the accompanying drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application as detailed in the appended claims. In the description of this application, it should be understood that the terms "first," "second," "third," etc., are used only to distinguish similar objects and are not necessarily used to describe a specific order or sequence, nor should they be construed as indicating or implying relative importance. Those skilled in the art can understand the specific meaning of the above terms in this application according to the specific circumstances.

[0021] Furthermore, in the description of this application, unless otherwise stated, "multiple" means two or more. "And / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, or B alone. The character " / " generally indicates that the preceding and following related objects are in an "or" relationship. The invention will be further described below with reference to the accompanying drawings and embodiments.

[0022] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0023] Example 1 Please refer to Figure 1 This embodiment provides the Ascend Edge Embossed Intelligent Service Vehicle System, which includes a human-computer interaction and voice command parsing module, a visual recognition and ranging module, an autonomous navigation and obstacle avoidance module, a robotic arm and robotic hand collaborative grasping module, and a dual-core main control communication module; The human-computer interaction and voice command parsing module is used to receive users' voice commands through mobile terminal applications and use artificial intelligence big data models to parse the voice commands into understandable structured data and send them to the main control board. The visual recognition and ranging module is used to acquire environmental images and use a target detection model to identify objects. At the same time, it integrates physical ranging and depth ranging schemes to calculate the three-dimensional spatial coordinates of the target object. The autonomous navigation and obstacle avoidance module includes a motion controller, a gyroscope, and an obstacle avoidance sensor, which are used to plan the motion path based on the coordinates of the target object and control the vehicle to avoid obstacles during movement. The robotic arm and robotic hand collaborative grasping module includes a robotic arm, a robotic hand, and a robotic arm camera, which is used to re-identify the target after the vehicle arrives near the target object and control the grasping action through inverse kinematics calculation. The dual-core main control communication module includes a main control board and an expansion board. The main control board is used to run visual recognition and path planning algorithms and coordinate communication between modules. The expansion board is used to execute motion control commands and exchange data with the main control board through a serial communication interface.

[0024] In one specific embodiment, in the human-computer interaction and voice command parsing module, the mobile terminal application has an embedded AI large model API, which is used to convert the user's voice command into text, construct specific prompt words and send them to the large model API, and parse the key command information from the JSON format data returned by the API and send it to the main control board.

[0025] In one specific embodiment, in the visual recognition and ranging module, the physical ranging uses an ultrasonic sensor, and the depth ranging uses the depth flow of a depth camera for geometric ranging; the system prioritizes physical ranging at long distances and depth ranging at short distances based on the confidence level of the depth data.

[0026] Specifically, the depth camera uses an Astra Pro camera, which can return a depth stream of up to 640*480*30fps and a color stream of up to 1920*1080*30fps. Target recognition is achieved using a YOLOv8 model for inference.

[0027] Specifically, by using both physical ranging and depth ranging, this effectively solves the shortcomings of insufficient accuracy at long distances in physical ranging and the vanishing depth flow at close distances in depth ranging. Furthermore, OpenImage V7 was chosen for the pre-training dataset, which offers advantages such as a wide variety of object types, rich annotations in the training images, small bounding boxes, and the ability to detect distant objects.

[0028] In one specific embodiment, in the autonomous navigation and obstacle avoidance module, the obstacle avoidance sensor is an ultrasonic sensor, and this sensor is only activated when the vehicle is moving to search for the target object, so as to avoid measurement errors caused by motor and power interference when the vehicle is stationary.

[0029] In one specific embodiment, in the robotic arm and robotic hand collaborative grasping module, the robotic arm camera acquires images, performs target recognition and coordinate transformation through a preset target detection model, and then calls the inverse kinematics algorithm to calculate the target angles of each joint of the robotic arm and drives the joints to perform grasping.

[0030] In one specific embodiment, the main control board is an Atlas 200I DK A2 board, and the expansion board is an i.MX RT1064 expansion board. The two communicate with each other via the UART protocol for commands and status data.

[0031] In one specific embodiment, the system further includes a mapping and navigation module, which is used to directly retrieve all obstacles and targets in the current site by creating a map of the environment, and to plan a complete path from the current position of the vehicle to the target object in one go.

[0032] In one specific embodiment, in the robotic arm and robotic hand collaborative grasping module, the deep learning-based end-to-end control method directly inputs and outputs servo control signals based on the images from the robotic arm's camera through a pre-trained neural network model.

[0033] In one specific embodiment, the system also supports two wireless connection modes: In the first mode, the mobile terminal connects to the main control board through a router and a network card, and the mobile terminal directly accesses the Internet. In the second mode, the mobile terminal connects directly to the main control board via the AP, and the mobile terminal accesses the Internet by forwarding traffic through the main control board.

[0034] Example 2 Please refer to Figure 1 as well as Figure 2 This embodiment provides a control method for the Ascend Edge Embossed Intelligent Service Vehicle, the control method including: S1: Receives user voice commands via mobile terminal application, and uses an artificial intelligence big data model to parse the voice commands into structured data and send them to the main control board; In one specific embodiment, the user launches the application via the HarmonyOS app on their mobile phone, and the app establishes a connection with the Atlas200I DK A2 main control board. The user speaks a voice command into the phone's microphone (e.g., "It's time to take your medicine").

[0035] In one specific embodiment, the speech recognition engine on the app side converts speech into text. The app then constructs a prompt (e.g., when the user gives the instruction "It's time to take your medicine," the prompt will return "medicine box") and sends it to the Alibaba Cloud Tongyi Qianwen Big Data Model API via an HTTP request. The API returns JSON data. The app parses this JSON, packages the processed information into an instruction, and sends it to the Atlas 200I DK A2 (e.g., after the app receives "It's time to take your medicine," the internal Tongyi Qianwen Big Data Model processes this information, understanding that "the user needs to take their medicine, and the car should look for medicine," thus sending the key information "medicine" as an instruction to the Atlas 200I DK A2).

[0036] S2: The main control board starts the visual recognition and ranging module, acquires environmental color images and depth images through the depth camera, identifies target objects using the target detection model, and calculates the three-dimensional spatial coordinates of the target objects by combining physical ranging and depth ranging. In one specific embodiment, upon receiving an instruction, the Atlas 200I DK A2 activates the Astra Pro camera to acquire color and depth images. The Atlas 200I DK A2 loads a YOLOv8 model converted to .om format and performs inference on the images. The model output includes bounding boxes, class labels, and confidence scores for multiple targets. The system selects the target with the highest confidence score that matches the target label, and calculates the three-dimensional spatial coordinates (X, Y, Z) of the medicine bottle in the cart's coordinate system based on its pixel coordinates in the image and the depth value in the depth image.

[0037] S3: The main control board plans the motion path based on the three-dimensional coordinates of the target object and the current position of the vehicle, and controls the movement of the vehicle through the expansion board; during the movement, it uses obstacle avoidance sensors to monitor obstacles in real time, and realizes obstacle avoidance and path replanning. In one specific embodiment, the Atlas 200I DK A2 plans a path from the current position to the target object based on the target coordinates (X, Y, Z) and the vehicle's current position (calculated by the i.MX RT1064 through encoder and gyroscope data fusion). Using coordinates for comprehension, this means first moving to the same Y coordinate as the target, then moving in a straight line to a position close to the target's X coordinate (precise distance control is achieved through an obstacle avoidance module to prevent collisions). The Atlas 200I DK A2 sends commands to the i.MX RT1064 via UART. Upon receiving the commands, the i.MX RT1064, combined with local sensor feedback, executes corresponding motion control or obstacle avoidance operations. Specifically, it controls the vehicle's steering based on gyroscope data, ensuring the front of the vehicle faces the target, and drives the motors via PWM signals to control the left and right wheel speeds to achieve straight-line movement. During movement, the i.MX RT1064 periodically calls the ultrasonic sensor to monitor the distance to obstacles ahead in real time. Simultaneously, during execution, the i.MX RT1064 also transmits status data and execution results back to the Atlas 200I DK A2 via UART, achieving closed-loop control of "perception-decision-execution". For example, when the detected distance is less than 20cm, the i.MX RT1064 immediately sends an "obstacle avoidance in progress" signal to the Atlas 200I DK A2, which then replans its path to avoid the obstacle.

[0038] S4: When the car reaches a preset distance in front of the target object, switch to the robotic arm camera, identify the target object again and calculate its coordinates relative to the robotic arm. Obtain the angles of each joint of the robotic arm through inverse kinematics calculation, and drive the robotic arm and robotic hand to complete the grasping. In one specific embodiment, when the vehicle successfully reaches approximately 30cm in front of the target object, the i.MX RT1064 stops moving and sends a signal to the Atlas 200I DK A2. The Atlas 200I DK A2 then switches to the robotic arm's camera view and repositions the target object. The system calculates the precise coordinates of the target object relative to the robotic arm and uses the Dofbot official library to perform inverse kinematics calculations, obtaining the target values ​​for each joint angle and driving the joints to execute, thus achieving closed-loop grasping.

[0039] S5: After the capture is completed, the car identifies the human body through the visual recognition module and autonomously returns to the user's location.

[0040] In one specific embodiment, after the grasping action is completed, the robotic arm drives the robotic hand to rotate and place the item into the box of the cart, initiating the return process. The cart uses the Astra Pro camera to identify the human body and accurately returns to the user who needs the item, stopping to deliver the item and completing the entire process.

[0041] Obviously, the above embodiments of the present invention are merely examples for clearly illustrating the present invention, and are not intended to limit the implementation of the present invention. Those skilled in the art can make other variations or modifications based on the above description. It is neither necessary nor possible to exhaustively describe all embodiments here. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the claims of the present invention.

Claims

1. The Ascend Edge Embedded Intelligent Service Vehicle System, characterized in that: The system includes a human-computer interaction and voice command parsing module, a visual recognition and ranging module, an autonomous navigation and obstacle avoidance module, a robotic arm and robotic hand collaborative grasping module, and a dual-core main control communication module. The human-computer interaction and voice command parsing module is used to receive users' voice commands through mobile terminal applications and use artificial intelligence big data models to parse the voice commands into understandable structured data and send them to the main control board. The visual recognition and ranging module is used to acquire environmental images and use a target detection model to identify objects. At the same time, it integrates physical ranging and depth ranging schemes to calculate the three-dimensional spatial coordinates of the target object. The autonomous navigation and obstacle avoidance module includes a motion controller, a gyroscope, and an obstacle avoidance sensor, which are used to plan the motion path based on the coordinates of the target object and control the vehicle to avoid obstacles during movement. The robotic arm and robotic hand collaborative grasping module includes a robotic arm, a robotic hand, and a robotic arm camera, which is used to re-identify the target after the vehicle arrives near the target object and control the grasping action through inverse kinematics calculation. The dual-core main control communication module includes a main control board and an expansion board. The main control board is used to run visual recognition and path planning algorithms and coordinate communication between modules. The expansion board is used to execute motion control commands and exchange data with the main control board through a serial communication interface.

2. The Ascend Edge Embossed Intelligent Service Vehicle System according to claim 1, characterized in that, In the human-computer interaction and voice command parsing module, the mobile terminal application has an embedded AI large model API, which is used to convert the user's voice command into text, construct specific prompt words and send them to the large model API, and parse the key command information from the JSON format data returned by the API and send it to the main control board.

3. The Ascend Edge Embossed Intelligent Service Vehicle System according to claim 1, characterized in that, In the visual recognition and ranging module, the physical ranging uses an ultrasonic sensor, and the depth ranging uses the depth flow of a depth camera for geometric ranging. Based on the confidence level of the depth data, the system prioritizes physical ranging at long distances and depth ranging at short distances.

4. The Ascend Edge Embossed Intelligent Service Vehicle System according to claim 1, characterized in that, In the autonomous navigation and obstacle avoidance module, the obstacle avoidance sensor is an ultrasonic sensor, and this sensor is only activated when the vehicle is moving to search for the target object, so as to avoid measurement errors caused by motor and power interference when the vehicle is stationary.

5. The Ascend Edge Embossed Intelligent Service Vehicle System according to claim 1, characterized in that, In the robotic arm and robotic hand collaborative grasping module, the robotic arm camera acquires images, performs target recognition and coordinate transformation through a preset target detection model, and then calls the inverse kinematics algorithm to calculate the target angles of each joint of the robotic arm and drives the joints to perform grasping.

6. The Ascend Edge Embossed Intelligent Service Vehicle System according to claim 1, characterized in that, The main control board is an Atlas 200I DK A2 board, and the expansion board is an i.MX RT1064 expansion board. The two communicate with each other via the UART protocol for commands and status data.

7. The Ascend Edge Embossed Intelligent Service Vehicle System according to claim 1, characterized in that, The system also includes a mapping and navigation module, which is used to create a map of the environment, directly retrieve all obstacles and targets in the current site, and plan the complete path from the vehicle's current position to the target object in one go.

8. The Ascend Edge Embossed Intelligent Service Vehicle System according to claim 1, characterized in that, In the robotic arm and robotic hand collaborative grasping module, the deep learning-based end-to-end control method directly inputs and outputs servo control signals based on the images from the robotic arm's camera through a pre-trained neural network model.

9. The Ascend Edge Embossed Intelligent Service Vehicle System according to claim 1, characterized in that, The system also supports two wireless connection modes: In the first mode, the mobile terminal connects to the main control board through a router and a network card, and the mobile terminal directly accesses the Internet. In the second mode, the mobile terminal connects directly to the main control board via the AP, and the mobile terminal accesses the Internet by forwarding traffic through the main control board.

10. The Ascend Edge Embossed Intelligent Service Vehicle Control Method, characterized in that, The control method includes: The system receives user voice commands via a mobile terminal application and uses a large artificial intelligence model to parse the voice commands into structured data, which is then sent to the main control board. The main control board starts the visual recognition and ranging module, acquires environmental color images and depth images through the depth camera, identifies target objects using the target detection model, and calculates the three-dimensional spatial coordinates of the target objects by combining physical ranging and depth ranging. The main control board plans the motion path based on the three-dimensional coordinates of the target object and the current position of the vehicle, and controls the movement of the vehicle through the expansion board; during the movement, obstacle avoidance sensors are used to monitor obstacles in real time to achieve obstacle avoidance and path replanning. When the car reaches a preset distance in front of the target object, it switches to the robotic arm camera to identify the target object again and calculate its coordinates relative to the robotic arm. The angles of each joint of the robotic arm are obtained through inverse kinematics calculation, and the robotic arm and robotic hand are driven to complete the grasping. After the capture is completed, the vehicle identifies the human body through the visual recognition module and autonomously returns to the user's location.

Images