Middleware adaptation development method for 4S store sales-assisted intelligent robots

Abstract

This invention relates to the field of robot middleware adaptation technology, providing a middleware adaptation development method for an embodied intelligent robot assisting sales in 4S stores. The method includes: Step 1, receiving a business scenario orchestration request, wherein the business scenario orchestration request carries a scenario identifier corresponding to the target sales business scenario; Step 2, determining the embodied intelligent robot hardware adaptation unit and the 4S store business system data interaction unit corresponding to the target sales business scenario through a middleware visual configuration interface; and Step 3, obtaining the robot multimodal capability description data corresponding to the hardware adaptation unit and the business interface description data corresponding to the data interaction unit. By introducing a large-scale middleware adaptation model and a visual configuration interface, developers do not need to deeply understand the robot's underlying motion control protocol and the complex database table structure of the 4S store backend. They only need to orchestrate and connect capability components on a canvas to automatically generate the middleware adaptation protocol stack and execution code package.

Description

Technical Field

[0001] This invention relates to the field of robot middleware adaptation technology, specifically to a middleware adaptation development method for a 4S store auxiliary sales intelligent robot. Background Technology

[0002] With the rapid development of artificial intelligence and robotics, embodied intelligent robots are gradually moving from the laboratory to commercial service scenarios. In the automotive 4S dealership sales and service sector, introducing embodied intelligent robots to handle tasks such as greeting customers, explaining vehicle parameters, recommending personalized options, and guiding customers through procedures has become an important trend in improving the technological feel and service efficiency of dealerships. However, currently available embodied intelligent robot hardware (such as bipedal, wheeled, and humanoid robots) and existing 4S dealership business systems (such as Dealer Management System (DMS), Customer Relationship Management System (CRM), and Enterprise Resource Planning (ERP)) are often heterogeneous and independent systems. A significant "gap" exists between the robot's perception, decision-making, and execution modules and the 4S dealership's backend data flow.

[0003] In existing development methodologies, developers typically employ a point-to-point customized interface development model, which involves writing dedicated bridging code for specific robot models and 4S store system versions. While this approach enables basic data interoperability, it faces two limitations in the complex and ever-changing physical environment and business interaction requirements of actual 4S stores: First, it suffers from low adaptation efficiency. When changing robot hardware manufacturers or upgrading store business systems, it often requires a complete overhaul or large-scale code refactoring, resulting in long deployment cycles. Second, the interaction logic is rigid, making it difficult to handle the dynamic coupling of natural language interaction, multimodal perception data, and business processes during the sales process. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides a middleware adaptation and development method for a 4S store assisted sales intelligent robot, which solves the problems of low efficiency in heterogeneous adaptation between the 4S store embodied intelligent robot and business systems and rigid interaction logic.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a middleware adaptation development method for a 4S store auxiliary sales intelligent robot, comprising the following steps: Step 1: Receive a business scenario orchestration request, wherein the business scenario orchestration request carries a scenario identifier corresponding to the target sales business scenario; Step 2: Through the middleware visual configuration interface, determine the embodied intelligent robot hardware adaptation unit and the 4S store business system data interaction unit corresponding to the target sales business scenario; Step 3: Obtain the robot multimodal capability description data corresponding to the hardware adaptation unit, and the business interface description data corresponding to the data interaction unit; Step 4: Input the robot multimodal capability description data and the business interface description data into the preset middleware adaptation model, and use the middleware adaptation model to execute the capability-data mapping fusion task to generate the middle layer adaptation protocol stack. The middle layer adaptation protocol stack defines the linkage logic between the robot's perception behavior and the 4S store's business data operation. Step 5: Based on the middleware layer adaptation protocol stack, dynamically generate an executable code package for deployment in the middleware gateway. The executable code package is configured to drive the embodied intelligent robot to perform auxiliary sales actions in conjunction with the 4S store business system at runtime.

[0006] Preferably, the business scenarios include: new car arrival reception and explanation scenario, competitor comparison analysis scenario, financial insurance plan trial calculation scenario, and after-sales maintenance appointment guidance scenario.

[0007] Preferably, in step three, the robot multimodal capability description data includes: robot motion control instruction set, speech output synthesis parameters, visual recognition interaction area coordinates, and touch screen interaction event callback function; The business interface description data includes: DMS system vehicle inventory query interface, CRM system customer profile acquisition interface, and ERP system work order creation interface.

[0008] Preferably, in step four, the middleware is used to adapt the large model execution capability-data mapping fusion task to generate an intermediate layer adaptation protocol stack, specifically including: S1. Based on the robot's multimodal capability description data, extract the robot's action primitive sequence and interaction triggering conditions in the physical space; S2. Based on the business interface description data, extract the input parameter constraints and return data structure of the business data request; S3. Through the attention mechanism network of the middleware adapted to the large model, calculate the spatiotemporal correlation weight matrix between the action primitive sequence and the returned data structure, and generate the capability-data mapping relationship based on the spatiotemporal correlation weight matrix, so that the robot's specific gestures, voice broadcasts or screen display content maintain semantic consistency and linkage with the real-time acquired 4S store business data.

[0009] Preferably, in step five, dynamically generating an executable code package for deployment in the middleware gateway based on the intermediate layer adaptation protocol stack includes: S1. Parse the state machine transition logic and data pipeline transmission rules defined in the intermediate layer adaptation protocol stack; S2. Compile the state machine transition logic into a JSON configuration file based on Node-RED streams, and encapsulate the data pipeline transmission rules into a topic subscription and publish configuration unit based on the MQTT protocol; S3. Package the JSON configuration file and the topic subscription and publishing configuration unit to generate the executable code package.

[0010] Preferably, the middleware visual configuration interface includes a robot capability library component panel, a 4S store business interface library component panel, and a canvas editing area; In particular, in response to the user dragging the first component in the robot capability library component panel to the canvas editing area, a corresponding hardware adaptation unit node is generated. In response to the user dragging the second component from the 4S store business interface library component panel to the canvas editing area, a corresponding data interaction unit node is generated. In response to the user's connection operation between the hardware adaptation unit node and the data interaction unit node, an initial capability-data mapping relationship is generated.

[0011] Preferably, it also includes: real-time monitoring of the interaction log stream generated during the runtime of the middleware gateway; When the response latency of the assisted sales action is detected to be greater than a preset threshold, the current CPU load rate, memory usage rate and network signal strength of the robot are obtained, and the CPU load rate, memory usage rate and network signal strength, as well as the currently executed middleware adaptation protocol stack fragment, are input into the middleware adaptation big model to obtain a simplified and optimized adaptation protocol stack strategy for the current business scenario.

[0012] Preferably, the executable code package includes formula-based... The calculated interaction time window control parameters; Where T represents the waiting time for the robot to actively ask follow-up questions after broadcasting vehicle data, D represents the text complexity index of the vehicle data, V represents the robot's preset speech rate constant, and C represents the environmental noise compensation correction value.

[0013] A middleware adapter development device for a 4S store auxiliary sales embodied intelligent robot, applied to edge computing devices, includes: The request receiving module is used to receive orchestration requests for business scenarios; The interface interaction module is used to determine the hardware adaptation unit and data interaction unit through the middleware visual configuration interface; The data acquisition module is used to acquire robot capability description data and business interface description data; The model processing module is used to input the description data into the middleware to adapt the large model to perform mapping and fusion tasks and generate the middle layer adaptation protocol stack. The code generation module is used to dynamically generate executable code packages based on the intermediate layer adaptation protocol stack and send them to the cloud-side gateway device for deployment.

[0014] A computing device includes a memory and a processor, the memory being used to store computer-executable instructions and the processor being used to execute the computer-executable instructions.

[0015] This invention provides a middleware adaptation development method for an intelligent robot that assists in sales at 4S stores. It has the following beneficial effects: 1. This invention introduces a middleware adaptation model and a visual configuration interface. Developers do not need to deeply understand the robot's underlying motion control protocol and the complex database table structure of the 4S store's backend. They only need to arrange and connect capability components on the canvas to automatically generate the middleware adaptation protocol stack and execution code package. When a store needs to replace with a new model of robot or upgrade the CRM system, it only needs to replace the corresponding component nodes on the visual interface and redeploy the protocol stack. There is no need to rewrite the underlying bridging code, which effectively solves the problems of low adaptation efficiency and large refactoring workload in the background technology.

[0016] 2. This invention utilizes a middleware-adapted large-model attention mechanism to calculate and establish dynamic spatiotemporal relationships between robot spatial actions (such as gestures pointing to vehicle wheel hubs or financial solutions displayed on the screen) and real-time business data (such as inventory status and personalized preferential policies). This allows the robot's voice content, gestures, and screen display information to be synchronized and consistent in real time when explaining vehicle models, rather than mechanically playing preset recordings.

[0017] 3. This invention generates adaptation logic during the development phase and also monitors the interaction logs of the middleware gateway in real time during runtime. When interaction stuttering or excessive latency is detected, it can combine the robot's real-time computing load and network conditions to infer a simplified adaptation protocol stack strategy for the current scenario using a large model. This ensures that even with limited network bandwidth in 4S stores or strained robot computing resources, the system can automatically downgrade non-core display effects to guarantee the smooth operation of the core sales process, reducing the difficulty of on-site maintenance and debugging. Attached Figure Description

[0018] Figure 1 This is a flowchart of the present invention. Detailed Implementation

[0019] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] Example: Please refer to the appendix Figure 1 This invention provides a middleware adaptation development method for a 4S store auxiliary sales embodied intelligent robot, including the following steps:

[0021] Step 1: Receive a business scenario orchestration request. The business scenario orchestration request carries the scenario identifier corresponding to the target sales business scenario. Step 2: Through the middleware visual configuration interface, determine the embodied intelligent robot hardware adaptation unit and the 4S store business system data interaction unit corresponding to the target sales business scenario; Step 3: Obtain the robot multimodal capability description data corresponding to the hardware adaptation unit, and the business interface description data corresponding to the data interaction unit; the robot multimodal capability description data includes: robot motion control instruction set, voice output synthesis parameters, visual recognition interaction area coordinates, and touch screen interaction event callback functions; the business interface description data includes: DMS system vehicle inventory query interface, CRM system customer profile acquisition interface, and ERP system work order creation interface. Step 4: Input the robot's multimodal capability description data and business interface description data into the preset middleware adaptation model. Utilize the middleware adaptation model to execute the capability-data mapping fusion task, generating the middleware adaptation protocol stack. The middleware adaptation protocol stack defines the linkage logic between the robot's perception behavior and the 4S store's business data operations. Specifically, the middleware adaptation protocol stack includes: S1. Based on the robot's multimodal capability description data, extract the robot's action primitive sequence and interaction triggering conditions in the physical space; S2. Based on the business interface description data, extract the input parameter constraints and return data structure of the business data request; S3. Through the middleware-adapted attention mechanism network of the large model, calculate the spatiotemporal correlation weight matrix between the action primitive sequence and the returned data structure, and generate the capability-data mapping relationship based on the spatiotemporal correlation weight matrix so that the robot's specific gestures, voice broadcasts or screen display content can maintain semantic consistency and linkage with the real-time acquired 4S store business data. Step 5: Based on the middleware layer's adapted protocol stack, dynamically generate the executable code package for deployment in the middleware gateway. This executable code package is configured to drive the embodied intelligent robot to perform auxiliary sales actions in conjunction with the 4S store's business system at runtime. The dynamically generated executable code package, based on the middleware layer's adapted protocol stack, includes: S1. Parse the state machine transition logic and data pipeline transmission rules defined in the intermediate layer adaptation protocol stack; S2. Compile the state machine transition logic into a JSON configuration file based on Node-RED streams, and encapsulate the data pipeline transmission rules into a topic subscription and publish configuration unit based on the MQTT protocol; S3. Package the JSON configuration file and topic subscription / publishing configuration unit to generate an executable code package; Specifically, the business scenario orchestration request in step one is typically initiated by the 4S store's IT administrator or system integration developer through the management backend on the terminal device. The "scenario identifier" carried in this request can be a predefined enumeration value, such as using the numerical code "1001" to represent new car reception, "1002" to represent competitor analysis, etc. Upon receiving this request, the system backend initializes a corresponding scenario adaptation task queue. The middleware visualization configuration interface in step two is a web-based graphical operating platform. In this platform, the "hardware adaptation unit" does not refer to a specific robot hardware entity, but rather to an abstract description of the robot's capabilities at the software level, such as the "UBTECH WalkerX motion control component" or the "OrionStar voice board output component." Similarly, the "data interaction unit" corresponds to the encapsulated metadata description of standard API interfaces, such as the "DMS vehicle information query interface." Users complete the logical association between the physical robot and the digital business flow by dragging and dropping and configuring attributes. The process of obtaining descriptive data in step three is automated. After the user drags a component into the canvas, the system backend retrieves the corresponding structured description file from the pre-built "Robot Capability Model Library" and "Business System Interface Registry Center". This description file not only contains the interface address and parameter types, but also the call frequency limit, authentication method, and data sensitivity level of the interface. This metadata provides a foundation for the subsequent generation of a secure and efficient adaptation protocol stack for large models. Steps four and five constitute an automated closed loop from logical design to physical deployment. The generated intermediate layer adaptation protocol stack is a lightweight, programming language-independent intermediate representation file. This file clearly records the action combination that the robot should perform when the CRM system returns "Customer preference is SUV"—namely, "nodding in confirmation and gesturing towards the SUV display area on the right side of the showroom." The final generated execution code package can be directly distributed via wireless network to the middleware gateway device deployed in the 4S store's server room without manual copying or service restart.

[0022] Business scenarios include: new car arrival reception and explanation, competitor comparison analysis, financial insurance plan calculation, and after-sales maintenance appointment guidance. Specifically, in the new car arrival reception and explanation scenario: when a customer enters the showroom, the robot confirms the customer's identity through visual recognition, triggering this scenario. The middleware adaptation protocol stack will call the CRM interface to obtain the customer's previous browsing history, and simultaneously call the DMS interface to obtain the inventory status of the corresponding car model. While the robot verbally announces, "Hello Mr. Wang, the new sedan you were interested in last time is currently available in the store," its screen will simultaneously display the car's 3D exterior and inventory information. Competitive product comparison analysis scenario: When a customer asks, "How does this car compare to a certain model from another brand?", the robot calls the knowledge base interface to obtain competitor parameters and invokes the sales pricing strategy from the ERP interface. The linkage logic in the protocol stack instructs the robot to adopt a "left-right split-screen comparison" screen layout. The left gesture points to the vehicle's parameters on the left screen, and the right gesture points to the competitor's parameters on the right screen. Based on the price difference returned by the data interface, the robot dynamically generates a voice prompt emphasizing "cost-effectiveness advantage." Financial insurance plan calculation scenario: The robot collects the customer's down payment ratio and installment term selection via a touchscreen. The middleware captures the data from the screen interaction event callback function and sends it in real time to the financial calculation service of the ERP system through a data pipeline. The monthly payment amount and total interest data returned by the calculation service are mapped by the large model to a scrolling animation of numbers on the robot's screen and a voice prompt saying "Your estimated monthly payment is XXXX yuan". After-sales maintenance appointment guidance scenario: The robot calls the ERP work order creation interface to query the available workstation time slots in the workshop. The protocol stack maps the returned time slot data to time selection buttons on the screen, and through the robot's motion control instruction set, the robot makes a slight head tilting motion when the customer hesitates, thus enhancing the interactivity.

[0023] The middleware visual configuration interface includes a robot capability library component panel, a 4S store business interface library component panel, and a canvas editing area; Among them, in response to the user dragging the first component in the robot capability library component panel to the canvas editing area, the corresponding hardware adaptation unit node is generated. In response to the user dragging the second component from the 4S store business interface library component panel to the canvas editing area, the corresponding data interaction unit node is generated. In response to the user's connection operation to the hardware adaptation unit node and the data interaction unit node, an initial capability-data mapping relationship is generated; Specifically, when a user drags the first component, "Gesture Actions," from the robot capability library to the canvas editing area, the system backend instantiates a hardware adaptation unit node object. This object not only contains enumeration values ​​for actions such as "waving" and "pointing," but also the range of inverse kinematics calculation parameters corresponding to that action. When the user drags the second component, "Vehicle Information Query," from the business interface library, a data interaction unit node object is generated. This object encapsulates the request's Uniform Resource Locator template and the returned JSON Schema definition. In response to a user's connection action, the system doesn't simply establish a visual connection; instead, it triggers a mapping configuration pop-up. In this pop-up, developers can fine-tune the trigger conditions. For example, the connection rule can be set as follows: when the "Vehicle Information Query" node returns a data field of "In Stock," the "Gesture Action" node will perform a thumbs-up gesture; if the value is "Out of Stock," a shrug gesture indicating regret will be executed. This visual condition configuration based on connection triggers significantly lowers the barrier to writing complex logical judgment code.

[0024] It also includes: real-time monitoring of the interaction log stream generated during the middleware gateway's runtime; when the response latency of the auxiliary sales action is detected to be greater than a preset threshold, obtaining the current robot's CPU load rate, memory usage rate, and network signal strength; and inputting the CPU load rate, memory usage rate, network signal strength, and the currently executed middleware adaptation protocol stack fragment into the middleware adaptation big model to obtain a simplified and optimized adaptation protocol stack strategy for the current business scenario. Specifically, when the monitoring agent detects that the end-to-end latency of the robot performing the action of "synchronously displaying high-definition vehicle images on the screen" exceeds a preset threshold (e.g., 800 milliseconds), it immediately triggers a self-optimization process. The monitoring agent obtains the current robot's CPU load rate (high load if it exceeds 85%), memory usage rate (memory shortage if the remaining memory is less than 200MB), and Wi-Fi signal strength (in dBm) through the system interface. The system inputs this set of vector data [85%, 180MB, -75dBm] along with the currently executing protocol stack segment (e.g., the rendering instruction set containing "displaying high-definition wheel hub detail images") into the middleware to adapt the large model.

[0025] The large model, based on a pre-trained compression and degradation policy library, outputs a simplified optimization policy, such as: "pause the high-resolution texture rendering of the wheel hub detail image and replace it with a preset vector contour image; temporarily reduce the sampling rate of the voice broadcast from 44.1 kHz to 22.05 kHz." This policy is injected into the protocol stack in real time, completing the degradation without interrupting the current session, thus ensuring the continuity of the dialogue.

[0026] The executable code package contains formula-based... The calculated interaction time window control parameters are as follows: T represents the waiting time for the robot to actively ask follow-up questions after broadcasting vehicle data, D represents the text complexity index of vehicle data, V represents the robot's preset speech rate constant, and C represents the environmental noise compensation correction value. Specifically, this formula calculates that after the robot reads out complex technical parameters, it will allow a longer waiting time to observe whether the customer asks any questions, while after reading out a simple quote, it will quickly move on to the next recommendation. This significantly enhances the robot's interactive rhythm, making it closer to the conversational habits of a real sales consultant.

[0027] A middleware adapter development device for a 4S store auxiliary sales embodied intelligent robot, applied to edge computing devices, includes: The request receiving module receives business scenario orchestration requests. This module listens to specific message queue topics to receive requests from the front-end interface or the upper-layer business orchestration system. Internally, the module includes a request parser to verify the validity of the scenario identifier and extract relevant metadata. The user interface module is used to determine the hardware adaptation units and data interaction units through a middleware visual configuration interface. This module is responsible for rendering the web front-end canvas and responding to interactive events. Its core is a directed graph data structure maintained in memory, where nodes represent hardware adaptation units and data interaction units, and edges represent mapping relationships. When the user drags on the canvas, the module calls the graphics engine to redraw the node positions and update the node coordinate attributes; when the user connects lines, the module inserts an edge object with a conditional function into the graph. The data acquisition module is used to acquire robot capability description data and business interface description data. This module maintains a long-lived connection to the "Robot Capability Model Library" and the "Business System Interface Registry Center". It adopts a lazy loading mechanism, only fetching the corresponding JSON Schema description file through the remote procedure call protocol when the node is instantiated, and caching it locally to reduce duplicate requests; The model processing module is used to input description data into the middleware to adapt the large model, perform mapping and fusion tasks, and generate an intermediate layer adaptation protocol stack. This module encapsulates the interaction logic with the middleware adapting the large model. It first vectorizes the heterogeneous description data passed in from the data acquisition module, and then calls the inference application programming interface provided by the large model. After obtaining the original attention weight matrix generated by the large model, this module also performs post-processing, mapping the floating-point values ​​in the matrix into discrete, machine-understandable action-data instruction pairs. The code generation module dynamically generates executable code packages based on the middleware adaptation protocol stack and sends them to the cloud-side gateway device for deployment. This module embeds a code generator based on a template engine. It reads the logic description from the middleware adaptation protocol stack and populates it into predefined Node-RED process templates and MQTT client configuration templates. The generated JSON file and configuration file are automatically compressed and packaged, and uploaded to a specified directory on the cloud-side gateway device via the Secure Shell protocol, triggering the gateway device's hot reloading process.

[0028] A computing device includes a memory and a processor, the memory being used to store computer-executable instructions and the processor being used to execute the computer-executable instructions.

[0029] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

A middleware adaptation development method for a 1.4S store auxiliary sales embodied intelligent robot, characterized in that, Includes the following steps: Step 1: Receive a business scenario orchestration request, wherein the business scenario orchestration request carries a scenario identifier corresponding to the target sales business scenario; Step 2: Through the middleware visual configuration interface, determine the embodied intelligent robot hardware adaptation unit and the 4S store business system data interaction unit corresponding to the target sales business scenario; Step 3: Obtain the robot multimodal capability description data corresponding to the hardware adaptation unit, and the business interface description data corresponding to the data interaction unit; Step 4: Input the robot multimodal capability description data and the business interface description data into the preset middleware adaptation model, and use the middleware adaptation model to execute the capability-data mapping fusion task to generate the middle layer adaptation protocol stack. The middle layer adaptation protocol stack defines the linkage logic between the robot's perception behavior and the 4S store's business data operation. Step 5: Based on the middleware layer adaptation protocol stack, dynamically generate an executable code package for deployment in the middleware gateway. The executable code package is configured to drive the embodied intelligent robot to perform auxiliary sales actions in conjunction with the 4S store business system at runtime.

2. The middleware adaptation development method for the 4S store auxiliary sales embodied intelligent robot according to claim 1, characterized in that, The business scenarios include: new car arrival reception and explanation, competitor comparison analysis, financial insurance plan calculation, and after-sales maintenance appointment guidance.

3. The middleware adaptation development method for the 4S store auxiliary sales embodied intelligent robot according to claim 1, characterized in that, In step three, the robot's multimodal capability description data includes: robot motion control instruction set, speech output synthesis parameters, visual recognition interaction area coordinates, and touch screen interaction event callback functions. The business interface description data includes: DMS system vehicle inventory query interface, CRM system customer profile acquisition interface, and ERP system work order creation interface.

4. The middleware adaptation development method for the 4S store auxiliary sales embodied intelligent robot according to claim 1, characterized in that, In step four, the middleware adaptation capability for large model execution—data mapping and fusion task—is used to generate an intermediate layer adaptation protocol stack, specifically including: S1. Based on the robot's multimodal capability description data, extract the robot's action primitive sequence and interaction triggering conditions in the physical space; S2. Based on the business interface description data, extract the input parameter constraints and return data structure of the business data request; S3. Through the attention mechanism network of the middleware adapted to the large model, calculate the spatiotemporal correlation weight matrix between the action primitive sequence and the returned data structure, and generate the capability-data mapping relationship based on the spatiotemporal correlation weight matrix, so that the robot's specific gestures, voice broadcasts or screen display content maintain semantic consistency and linkage with the real-time acquired 4S store business data.

5. The middleware adaptation development method for the 4S store auxiliary sales embodied intelligent robot according to claim 1, characterized in that, In step five, based on the intermediate layer adaptation protocol stack, an executable code package for deployment in the middleware gateway is dynamically generated, including: S1. Parse the state machine transition logic and data pipeline transmission rules defined in the intermediate layer adaptation protocol stack; S2. Compile the state machine transition logic into a JSON configuration file based on Node-RED streams, and encapsulate the data pipeline transmission rules into a topic subscription and publish configuration unit based on the MQTT protocol; S3. Package the JSON configuration file and the topic subscription and publishing configuration unit to generate the executable code package.

6. The middleware adaptation development method for the 4S store auxiliary sales embodied intelligent robot according to claim 1, characterized in that, The middleware visual configuration interface includes a robot capability library component panel, a 4S store business interface library component panel, and a canvas editing area. In particular, in response to the user dragging the first component in the robot capability library component panel to the canvas editing area, a corresponding hardware adaptation unit node is generated. In response to the user dragging the second component from the 4S store business interface library component panel to the canvas editing area, a corresponding data interaction unit node is generated. In response to the user's connection operation between the hardware adaptation unit node and the data interaction unit node, an initial capability-data mapping relationship is generated.

7. The middleware adaptation development method for the 4S store auxiliary sales embodied intelligent robot according to claim 1, characterized in that, Also includes: Real-time monitoring of the interaction log stream generated during the runtime of the middleware gateway; When the response latency of the assisted sales action is detected to be greater than a preset threshold, the current CPU load rate, memory usage rate and network signal strength of the robot are obtained, and the CPU load rate, memory usage rate and network signal strength, as well as the currently executed middleware adaptation protocol stack fragment, are input into the middleware adaptation big model to obtain a simplified and optimized adaptation protocol stack strategy for the current business scenario.

8. The middleware adaptation development method for the 4S store auxiliary sales embodied intelligent robot according to claim 1, characterized in that, The executable code package contains formula-based... The calculated interaction time window control parameters; Where T represents the waiting time for the robot to actively ask follow-up questions after broadcasting vehicle data, D represents the text complexity index of the vehicle data, V represents the robot's preset speech rate constant, and C represents the environmental noise compensation correction value.

9. A middleware adaptation development device for a 4S store assisted sales embodied intelligent robot, using the middleware adaptation development method for a 4S store assisted sales embodied intelligent robot as described in any one of claims 1-8, characterized in that, Applications in edge computing devices, including: The request receiving module is used to receive orchestration requests for business scenarios; The interface interaction module is used to determine the hardware adaptation unit and data interaction unit through the middleware visual configuration interface; The data acquisition module is used to acquire robot capability description data and business interface description data; The model processing module is used to input the description data into the middleware to adapt the large model to perform mapping and fusion tasks and generate the middle layer adaptation protocol stack. The code generation module is used to dynamically generate executable code packages based on the intermediate layer adaptation protocol stack and send them to the cloud-side gateway device for deployment.

10. A computing device, using the middleware adaptation development method for a 4S store assisted sales embodied intelligent robot as described in any one of claims 1-8, characterized in that, It includes a memory and a processor, wherein the memory is used to store computer-executable instructions and the processor is used to execute the computer-executable instructions.

Images