Data processing method, system and device of intelligent display stand system and storage medium

Abstract

The application provides a data processing method, system and device of an intelligent display stand system and a storage medium. The method comprises: monitoring personnel in a preset area of the intelligent display stand by a radar device of the intelligent display stand, automatically generating a temporary tracking ID for the personnel when it is monitored that the personnel enter the preset area of the intelligent display stand, determining whether the personnel is a consumer based on the temporary tracking ID, obtaining multi-dimensional behavior data generated by the consumer in the preset area of the intelligent display stand when the personnel is determined to be a consumer, performing consumer intention analysis on the multi-dimensional behavior data by using a preset consumer intention analysis model, and generating consumer intention data; and generating promotion content according to the consumer intention data and feeding back the promotion content to the intelligent display stand for display. The method can realize accurate statistics of passenger flow based on the role separation of consumers and salespersons, and can realize accurate prediction of consumer intention by performing intention analysis by fusing multi-dimensional data features.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of commodity sales equipment, and particularly relates to a data processing method and system of an intelligent display stand system, equipment and a storage medium. BACKGROUND

[0002] In recent years, more and more sales institutions try to optimize store management with the help of various intelligent tools. The intelligent display stand is a common intelligent sales tool. The data processing method of the existing intelligent display stand is generally to analyze the flow of people by deploying cameras, Wi-Fi probes and the like, and to recommend hot-selling commodities to consumers. However, the data processing method based on the existing intelligent display stand is insufficient in capturing intent when collecting data, and usually relies on manual annotation and the like to separate salespersons and consumers when counting the flow of people, which has poor separation effect and cannot accurately filter out data pollution caused by repeated inspection and service behavior of employees or salespersons. When analyzing behavior, the recognition of implicit demand of long natural language is not complete, which cannot accurately predict the intent of consumers and accurately recommend commodity content that meets the intent to consumers. SUMMARY

[0003] Therefore, the embodiments of the present application provide a data processing method and system of an intelligent display stand system, equipment and a storage medium, which can realize accurate statistics of the flow of people and accurate prediction of the intent of consumers.

[0004] The first aspect of the embodiments of the present application provides a data processing method of an intelligent display stand system, comprising: monitoring personnel in a preset area of an intelligent display stand by a radar device of the intelligent display stand, and automatically generating a temporary tracking ID for the personnel when it is monitored that the personnel enter the preset area of the intelligent display stand; synchronously scanning the preset area of the intelligent display stand by a Bluetooth positioning base station device to determine whether there is a Bluetooth positioning tag ID at the position where the temporary tracking ID is located; if there is a Bluetooth positioning tag ID, it is determined that the personnel is a salesperson, otherwise it is determined that the personnel is a consumer; in the case of determining that the personnel is a consumer, obtaining multi-dimensional behavior data generated by the consumer in the preset area of the intelligent display stand, and using a preset consumer intent analysis model to analyze the multi-dimensional behavior data for consumer intent, to generate consumer intent data; and generating promotion content according to the consumer intent data and feeding back the promotion content to the intelligent display stand for display.

[0005] The second aspect of the embodiment of the present application provides a smart display stand system, which is used to realize the data processing method of the smart display stand system as described in the first aspect, and comprises a server end and a smart display stand, wherein: the server end is used to receive multi-dimensional behavior data transmitted by the smart display stand, perform consumption intention analysis processing based on the multi-dimensional behavior data, and generate promotion content and feed back to the smart display stand; the smart display stand comprises a display device, a radar device, a microphone device and a communication device, the display device comprises a display for displaying product information and promotion content and a display stand for placing products, and is used to collect human-computer interaction data generated when a consumer interacts with the smart display stand; the radar device is used to collect radar signal data generated when a consumer is in a preset area of the smart display stand; the microphone device comprises a Bluetooth microphone and a Bluetooth positioning base station, and is used to collect shopping guide conversation data and positioning information of shopping guides generated when a consumer is in a preset area of the smart display stand by connecting the Bluetooth microphone and the Bluetooth positioning base station; and the communication device is used to package data collected by the display device, the radar device and the microphone device in the smart display stand into multi-dimensional behavior data and transmit the multi-dimensional behavior data to the server end.

[0006] The third aspect of the embodiment of the present application provides a device comprising a memory, a processor and a computer program stored in the memory and executable on the device, wherein the processor implements each step of the data processing method of the smart display stand system provided in the first aspect when executing the computer program.

[0007] The fourth aspect of the embodiment of the present application provides a computer readable storage medium, which stores a computer program, wherein the computer program is executed by a processor to implement each step of the data processing method of the smart display stand system provided in the first aspect.

[0008] The fifth aspect of the embodiment of the present application provides a computer program product, which, when executed on a device, enables the device to implement each step of the data processing method of the smart display stand system provided in the first aspect.

[0009] The data processing method, system, device and storage medium of the smart display stand system provided by the embodiment of the present application have the following beneficial effects:

[0010] By identifying the identity of the person entering the preset area of the intelligent display stand, it is judged whether the person is a consumer, and in the case where the person is determined to be a consumer, multi-dimensional behavior data generated by the consumer in the preset area of the intelligent display stand is acquired; a preset consumer intention analysis model is used to analyze the multi-dimensional behavior data for consumer intention, and consumer intention data is generated; and promotion content is generated according to the consumer intention data and fed back to the intelligent display stand for display. Based on the method, the roles of consumers and salespersons can be accurately separated, data pollution caused by salesperson's goods arrangement and service behavior can be effectively filtered out, and accurate customer flow statistics can be achieved. Moreover, the consumer intention analysis model introduces a large model semantic and structured label fusion technology, fuses multi-dimensional data features, and can accurately predict the intention of the consumer. BRIEF DESCRIPTION OF DRAWINGS

[0011] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed to be used in the embodiments or prior art description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0012] Figure 1 An implementation flowchart of a data processing method of an intelligent display stand system provided by an embodiment of the present application is provided.

[0013] Figure 2 An implementation flowchart of a method for collecting radar signal data in a data processing method of an intelligent display stand system provided by an embodiment of the present application is provided.

[0014] Figure 3 An implementation flowchart of a method for multiplexing a temporary tracking ID in a data processing method of an intelligent display stand system provided by an embodiment of the present application is provided.

[0015] Figure 4 An implementation flowchart of a method for using a consumer intention analysis model to analyze consumer intention in a data processing method of an intelligent display stand system provided by an embodiment of the present application is provided.

[0016] Figure 5 An implementation flowchart of a method for triggering intelligent display stand and consumer deep human-computer interaction in a data processing method of an intelligent display stand system provided by an embodiment of the present application is provided.

[0017] Figure 6 An implementation flowchart of a method for updating a dialogue knowledge vector library in a data processing method of an intelligent display stand system provided by an embodiment of the present application is provided.

[0018] Figure 7 A structural schematic diagram of an intelligent display stand system provided by an embodiment of the present application is provided.

[0019] Figure 8 This application provides an example of the external structural diagram of an intelligent display stand in an intelligent display stand system.

[0020] Figure 9 A basic structural block diagram of a data processing device for an intelligent display stand system provided in an embodiment of this application;

[0021] Figure 10 This is a basic structural block diagram of a device provided in an embodiment of this application. Detailed Implementation

[0022] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of this application. However, those skilled in the art will understand that this application may also be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, devices, circuits, and methods are omitted so as not to obscure the description of this application with unnecessary detail.

[0023] It should be understood that, when used in this application specification and the appended claims, the term "comprising" indicates the presence of the described features, integrals, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or a collection thereof.

[0024] It should also be understood that the term “and / or” as used in this application specification and the appended claims means any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0025] As used in this application specification and the appended claims, the term "if" may be interpreted, depending on the context, as "when," "once," "in response to determination," or "in response to detection." Similarly, the phrases "if determined" or "if detected [the described condition or event]" may be interpreted, depending on the context, as meaning "once determined," "in response to determination," "once detected [the described condition or event]," or "in response to detection [the described condition or event]."

[0026] Furthermore, in the description of this application and the appended claims, the terms "first," "second," "third," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0027] References to "one embodiment" or "some embodiments" in this specification mean that one or more embodiments of this application include a specific feature, structure, or characteristic described in connection with that embodiment. Therefore, the phrases "in one embodiment," "in some embodiments," "in other embodiments," "in still other embodiments," etc., appearing in different parts of this specification do not necessarily refer to the same embodiment, but rather mean "one or more, but not all, embodiments," unless otherwise specifically emphasized. The terms "comprising," "including," "having," and variations thereof mean "including but not limited to," unless otherwise specifically emphasized. "A plurality" means "two or more."

[0028] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0029] Intelligent display systems for retail use specifically analyze foot traffic and heat maps by deploying customer flow statistics and behavior analysis devices (such as cameras and Wi-Fi probes). Utilizing big data and simple machine learning methods, they recommend popular products or customized information to consumers. Existing technical solutions typically collect consumer preferences and intentions through structured questionnaires, CRM tag scoring, and human customer service scripts. The dialogue text is then transformed using template matching, keyword detection, and manual post-annotation. However, this approach suffers from high noise levels, low data utilization, and insufficient capture of complex contexts or implicit intentions. Introducing NLP algorithms to assist in understanding customer intentions also faces limitations such as limited coverage, limited contextual understanding, and poor data processing for multi-turn dialogues. Rule engines suffer from limited expressive power, incomplete recognition of long passages of natural language, ambiguous expressions, and implicit needs, and difficulty in large-scale application of manual intervention. Furthermore, deep learning-based dialogue understanding has failed to deeply integrate with actual store scenarios and marketing processes, hindering the achievement of a commercial closed-loop system. Furthermore, current consumer behavior analysis products primarily focus on overall customer flow distribution analysis; however, their support for separating consumer behavior from the influence of sales staff is limited. Relying on manual labeling and device tagging to separate sales staff and consumers fails to filter out data contamination resulting from repeated patrols, inventory management, and service activities by employees or sales staff. The user experience is poor, and the systems are not scalable; camera-based recognition also raises privacy concerns. This application provides an intelligent display stand system that aims to offer a data processing method to address the aforementioned problems of existing intelligent display stand systems.

[0030] In some embodiments of this application, please refer to Figure 1 , Figure 1This is a flowchart illustrating the implementation of a data processing method for an intelligent display stand system provided in an embodiment of this application. Figure 1 As shown, it may specifically include steps S11 to S13.

[0031] S11: The radar device of the smart display stand monitors the personnel in the preset area of ​​the smart display stand. When a person is detected entering the preset area of ​​the smart display stand, a temporary tracking ID is automatically generated for the person.

[0032] S12: The preset area of ​​the smart display stand is scanned synchronously by the Bluetooth positioning base station device to determine whether a Bluetooth positioning tag ID exists at the location of the temporary tracking ID.

[0033] S13: If a Bluetooth location tag ID exists, the person is identified as a sales guide; otherwise, the person is identified as a consumer.

[0034] In this embodiment, a radar device is installed in the smart display stand, specifically a millimeter-wave radar. When a person enters the preset area of ​​the smart display stand, the radar device can identify the person's location through human detection and automatically generate a temporary tracking ID for target tracking. It can be understood that the preset area of ​​the smart display stand is a location area centered on the location of the smart display stand, such as within 3 meters of the smart display stand. The smart display stand is also equipped with multiple Bluetooth positioning base station devices. Each salesperson responsible for selling products on the smart display stand wears a Bluetooth microphone with a unique Bluetooth positioning tag ID. The Bluetooth positioning tag ID is registered and pre-stored in the registration ID list in the system backend. The Bluetooth microphone communicates with the Bluetooth positioning base station devices to form a Bluetooth network.

[0035] In this embodiment, the intelligent display stand system can use a radar device in conjunction with a Bluetooth positioning base station to identify whether the person corresponding to the temporary tracking ID is a salesperson wearing a Bluetooth microphone. Specifically, the radar device on the intelligent display stand monitors personnel in a preset area. When a person enters the preset area, the radar device identifies the person's location through human detection and automatically generates a temporary tracking ID for target tracking. Simultaneously, the Bluetooth positioning base station scans the preset area of ​​the intelligent display stand. The Bluetooth microphone is equipped with a Bluetooth positioning tag, which contains a unique Bluetooth positioning tag ID. The Bluetooth positioning base station can scan the preset area of ​​the intelligent display stand for Bluetooth positioning tags based on the Bluetooth positioning tag IDs recorded in the registered ID list stored in the system backend. This determines whether a Bluetooth positioning tag ID recorded in the registered ID list exists in the preset area of ​​the intelligent display stand. When a Bluetooth positioning tag ID recorded in the registered ID list is detected in the area, the spatial relationship between the Bluetooth positioning tag ID and the temporary tracking ID is used to determine whether a Bluetooth positioning tag ID exists at the location of the temporary tracking ID. When the temporary tracking ID and the Bluetooth positioning base station ID are sufficiently close, the person corresponding to both can be considered the same person. In this case, the person is identified as a sales associate, and their data generated within the preset area of ​​the smart display stand is not included in the customer flow statistics or used for purchase intention analysis. It's understandable that determining whether the temporary tracking ID and the Bluetooth positioning tag ID are sufficiently close can be measured by whether the distance between them is less than a preset distance value. For example, a distance of 0.2 meters between the Bluetooth positioning tag ID and the temporary tracking ID can be considered sufficiently close. The preset distance value can be adjusted based on actual conditions and is not limited to 0.2 meters. When no Bluetooth positioning tag ID recorded in the registered ID list is detected within the detection area, the person corresponding to the temporary tracking ID is identified as a consumer, and this person is considered a valid customer for statistical purposes. The data generated by this person within the preset area of ​​the smart display stand is then input into the customer intention analysis model for behavioral analysis, providing personalized product promotion content and promoting sales orders.

[0036] In one specific implementation, when the radar device and Bluetooth positioning base station equipment detect the presence of multiple people, such as multiple salespersons walking together or salespersons mixed with consumers, the radar algorithm can distinguish the number of people, and the relative position of the salespersons can be determined based on the Bluetooth positioning tag ID, thereby distinguishing consumers from salespersons. Furthermore, by processing clustering and error in real time, the system can also simultaneously identify multiple people, and then determine the identity of each person based on their spatial location.

[0037] S14: If the person is identified as a consumer, obtain multi-dimensional behavioral data generated by the consumer when the consumer is in the preset area of ​​the smart display stand, and use a preset consumer intention analysis model to perform consumer intention analysis on the multi-dimensional behavioral data to generate consumer intention data.

[0038] In this embodiment, if the person is identified as a consumer, multi-dimensional behavioral data generated by that person within a preset area of ​​the smart display stand can be acquired. This multi-dimensional behavioral data can be used for consumer intention analysis to provide personalized product promotion content and facilitate sales orders. It is understood that the multi-dimensional behavioral data is real-time collected time-series data.

[0039] In this embodiment, the preset consumer intention analysis model is a large language model capable of consumer intention analysis, pre-trained using a temporal neural network combined with an attention mechanism. In one specific implementation, the consumer intention analysis model can be trained in stages based on the open-source Deepseek-R1 architecture. Deepseek-R1 is an open-source large language model released by Deepseek. The training process of the consumer intention analysis model is as follows: In the data preparation stage, product knowledge graph data, high-quality sales guide text data, radar trajectory feature data, customer interaction behavior sequence data, and inventory sales order data are collected as basic data for model training. These basic data undergo preprocessing such as deduplication and cleaning, entity annotation, intent classification, and multimodal alignment to obtain model training data. It is understood that the product knowledge graph data contains product category information and complementary relationships between products. In the model training stage, model training is divided into three stages: pre-training, domain adjustment, and strategy optimization. First, in the pre-training phase, the basic language understanding capabilities of the model are pre-trained using a general corpus. The training data used in pre-training includes 50% general corpus data such as books and web pages, 30% retail-related text data such as product manuals and industry reports, and 20% multimodal aligned data such as product images and descriptions. Then, in the domain adjustment phase, the model is trained using data specific to the shopping guide scenario to achieve domain adjustment. Specifically, knowledge graph data can be converted into TransE vector embeddings for the model's input layer, and the RAG interface can be used to obtain vectors of one or more of the most relevant knowledge fragments in the domain for further training, thus achieving the goal of domain adjustment. Retrieval Augmented Generation (RAG) is a cutting-edge technology that integrates the advantages of retrieval-based AI models and generative AI models, providing more accurate, relevant, and human-language-featured responses. The core idea of ​​RAG is to supplement the massive amount of existing knowledge in large language models with targeted domain-specific information, making the model's generated responses not only coherent and natural but also based on real and timely data. In this embodiment, vector retrieval enhancement is achieved by calling the RAG interface, and efficient information retrieval is realized using a vector database. When information is acquired by calling the RAG interface, the external knowledge base is converted into a high-dimensional vector space, and each piece of information is represented as a dense vector. This process is called vectorization or embedding, which can quickly identify and retrieve the most relevant information based on the similarity between the user's query and the stored vector.Finally, in the strategy optimization stage, experienced sales consultants can score the dialogues in the high-quality sales consultant scripts based on three dimensions: guidance effectiveness, professionalism, and friendliness. These dialogues are then labeled using a tiered reward system based on the scores, and the labeled data is used for further model training to optimize the model's strategies. The trained consumer intention analysis model involves three steps in consumer intention analysis: data fusion, feature inference, and result output. The data fusion step extracts data features from multi-dimensional behavioral data, obtaining text features representing sales consultant scripts, behavioral features representing consumer trajectories, and product features representing product information such as inventory and price. A cross-attention mechanism is used to weight and fuse these text features, behavioral features, and product features to obtain fused features. The feature inference step then uses the fused features to perform intent recognition, knowledge retrieval, strategy generation, and effect prediction, resulting in consumer intention analysis results including, but not limited to, script suggestions, intended product recommendation rankings, consumer purchase probability distributions, and the weights of product attributes that consumers focus on. The output step uses a JSON structure to encapsulate the consumer intention analysis results, including suggested sales pitches, ranking of recommended products, consumer purchase probability distribution, and weights of product attributes that consumers are interested in, generating consumer intention data output. It's understood that JSON (JavaScript Object Notation) is a lightweight data exchange format commonly used for transmitting data between front-end and back-end interactions. This embodiment, based on the aforementioned consumer intention analysis model, packages the collected multi-dimensional behavioral data into a contextual input model, enabling the model to understand the products currently being viewed by the consumer and their attributes, as well as the consumer's current core questions and preferences.

[0040] S15: Generate promotional content based on the consumer intention data and feed it back to the smart display stand for display.

[0041] In this embodiment, after obtaining consumer intention data through consumer intention analysis, customized promotional content such as product comparisons, usage suggestions, and recommendations of related or complementary products can be generated based on this data. This promotional content is then fed back to the smart display stand for display. For example, the smart display stand is equipped with a display device, which includes a monitor. The smart display stand displays the promotional content to consumers through the monitor.

[0042] In one specific implementation, when generating promotional content based on consumer intention data, the recommendation priority of each product can be determined according to the clarity of consumer intention, the potential for real-time conversion rate improvement, and the product inventory warning coefficient. Promotional content is then generated according to this recommendation priority. For example, the product recommendation priority = 0.5 * clarity of consumer intention + 0.3 * potential for real-time conversion rate improvement + 0.2 * product inventory warning coefficient. Furthermore, different promotional content can be generated for consumers with different intention types. For example, for high-intent consumers, precise sales pitches to facilitate transactions can be generated and displayed as promotional content; for hesitant consumers, comparative analysis and decision-making basis of the products they are interested in and similar products can be generated and displayed as promotional content; for low-intent consumers, information on relevant alternative products corresponding to the products they are interested in can be generated and displayed as promotional content.

[0043] In one specific implementation, the intelligent display stand system can also record consumer reactions and final sales results after each feedback of promotional content to consumers. Based on consumer reactions and final sales results, it performs reinforcement learning to continuously optimize recommendation strategies and regularly integrates and updates the knowledge graph and sales script vector library. The intelligent display stand system's large language model algorithm logic, through multi-dimensional data fusion and real-time learning mechanisms, achieves continuous accumulation and precise application of sales guide knowledge. This ensures efficient knowledge reuse while dynamically adjusting according to actual scenarios, ultimately achieving rapid improvement in sales guide capabilities and effective reduction in training costs.

[0044] As can be seen from the above, the data processing method of the intelligent display rack system provided in this application can separate the roles of consumers and sales staff through personnel identification. Only consumer data will be entered into the consumer intention analysis model for analysis, filtering out data pollution caused by repeated patrols, stocking, and service activities of sales staff, thus achieving data security and compliance and ensuring accurate customer flow statistics. Without collecting human image information, it can accurately determine whether there are people in the monitoring area, statistically output the number of target personnel monitored in each area, and distinguish between sales staff and consumers, thus better protecting consumer privacy and security. Moreover, the training of the consumer intention analysis model introduces large-scale model semantics and structured label fusion technology, integrating multi-dimensional data features to achieve accurate prediction of consumer intentions.

[0045] In some embodiments of this application, multi-dimensional behavioral data refers to behavioral data obtained through real-time analysis of multimodal sensor signals using various forms or sensing channels. This behavioral data is specifically generated through various modes of expression, communication, and understanding, including but not limited to visual, auditory, textual, and tactile means. In this embodiment, multi-dimensional behavioral data can be obtained in three forms: The first form is the collection of human-computer interaction data generated when consumers interact with the smart display stand via its display device. Specifically, the interaction between consumers and the smart display stand includes, but is not limited to, consumers picking up goods on the display stand, the smart display stand quickly locating the product ID via NFC reader identification, and the display instantly displaying detailed information about the picked-up product, such as name, price, promotional status, and related reviews; consumers touching the display, the display showing different dimensions of product information based on the consumer's touch location; and the smart display stand recording the product dwell time when consumers pick up goods or when the display shows product information. Based on the interaction with consumers, the smart display stand's display device can collect product information data displayed during the interaction process and product dwell time data recorded by the display device when consumers pay attention to products. The second method involves collecting sales dialogue data generated when consumers are within a preset area of ​​the smart display stand using its microphone device. Specifically, sales staff wear Bluetooth microphones, which capture the sales dialogue data generated during their interactions with consumers. It's important to note that the Bluetooth microphone-collected dialogue data exists in voice format, while the dialogue data included in the multi-dimensional behavioral data exists in text format. Therefore, in this embodiment, after collecting the dialogue data via Bluetooth, a series of processing steps, including voiceprint feature extraction, voiceprint feature clustering analysis, and sound source classification, are performed on the voice-based dialogue data to achieve voice separation, obtaining the consumer's voice data and the sales staff's voice data. These two voice data are then converted into text data. It is understood that voiceprint features include MFCC (Mel-frequency cepstral coefficients) features, spectral features, and prosodic features. Furthermore, after converting the salesperson's voice data into text data, natural language understanding can be used to filter out irrelevant audio noise, verbal embellishments, environmental noise, and other invalid dialogue fragments, retaining only valid and complete behavioral and semantic information. The voice data is only stored locally on the smart display stand; it is only uploaded to the model for analysis after being converted into text. Moreover, after the model's analysis, the voice data stored locally on the smart display stand is deleted, effectively protecting consumer privacy. The third method involves collecting radar signal data generated when consumers are within a preset area of ​​the smart display stand using its radar device.Specifically, the radar device installed on the smart display stand monitors the preset area of ​​the smart display stand in real time. When a consumer is detected entering the preset area, the radar tracks the consumer, collecting radar signal data that characterizes the consumer's movement trajectory and body contour within the preset area. It is understood that radar signal feature extraction is the process of obtaining parameters such as the target's position, speed, and attitude, as well as extracting the target's shape, material, and other physical characteristics by analyzing radar echo signals. Therefore, in this embodiment, radar signal extraction can be achieved by analyzing radar echo signals in the radar signal data, obtaining consumer movement trajectory data and consumer body contour data. Based on the above three methods of acquiring multi-dimensional behavioral data, a non-intrusive behavioral analysis system is proposed for the smart display stand system. This system can collect data in a way that collects only anonymous trajectories and IDs throughout the entire process, without collecting facial or sensitive biometric information. Temporary IDs also expire periodically, effectively protecting personnel privacy. Furthermore, without collecting human image information, it can accurately determine whether there are people in the monitored area, statistically output the number of target personnel detected in each area, and distinguish between sales staff and consumers, achieving seamless separation of roles such as consumers and sales staff. Automatically identify consumers and filter out data pollution caused by repeated inspections, inventory management, and service activities by employees or sales guides.

[0046] In this embodiment, the gesture data corresponding to the hand gestures generated when consumers interact with the smart display stand can also be collected by the radar device of the smart display stand. Specifically, the radar device collects radar signal data representing the consumer's hand gestures by recognizing the consumer's hand movements. For the radar signal data representing the consumer's gestures, the radar echo signal can first be filtered and converted into a grayscale image, and then features can be extracted based on the grayscale image and a convolutional neural network model can be used for gesture recognition to obtain the consumer's gesture data. For example, the network architecture of the convolutional neural network model specifically includes an input layer, a feature learning layer, and an output layer. The input layer uses a 3D radar echo data matrix. The feature learning layer includes a 3×3 convolutional layer, a 3×3 pooling layer, a 7×7 convolutional layer, and an Inception module, which includes 1×1 convolution, 3×3 convolution, 5×5 convolution, and max pooling operations. The output layer outputs the gesture category through feature concatenation and a classifier.

[0047] In some embodiments of this application, the intelligent display stand system can pre-establish a mapping relationship between gestures, consumer intentions, and system response operations. This mapping relationship includes, but is not limited to: a "point + hold for 1 second" gesture corresponding to the consumer intention of "following the product," and the corresponding system response operation being "displaying the product details page"; a "swipe up and down" gesture corresponding to the consumer intention of "switching options," and the corresponding system response operation being "showing similar products on a flip page"; a "clenched fist + released" gesture corresponding to the consumer intention of "confirming the selection," and the corresponding system response operation being "adding to favorites / comparison list," etc. Based on this mapping relationship, the intelligent display stand system can trigger system response operations by recognizing gestures, thereby realizing the interaction function between consumers and the intelligent display stand. Specifically, the interaction logic between consumers and the intelligent display stand is as follows: the interaction confidence level between consumers and the intelligent display stand is calculated based on the gesture recognition probability, intention matching degree, and environmental stability. When the interaction confidence level reaches a preset confidence threshold (e.g., 0.85), the system response operation is triggered. It is understood that the intention matching degree can be dynamically adjusted based on user historical behavior data, and environmental stability can be evaluated using the radar signal-to-noise ratio.

[0048] In some embodiments of this application, please refer to Figure 2 , Figure 2 This is a flowchart illustrating one method for acquiring radar signal data in the data processing method of the intelligent display stand system provided in this application embodiment. Figure 2 As shown, it may specifically include steps S21 to S23.

[0049] S21: Monitor the preset area of ​​the smart display stand using the radar device of the smart display stand;

[0050] S22: If the radar device detects that a consumer has entered the preset area of ​​the smart display stand, the radar device of the smart display stand generates radar signal data to characterize the consumer's movement trajectory and body outline, and associates the radar signal data with a temporary tracking ID automatically generated for the consumer.

[0051] S23: If the radar device detects that a consumer in the preset area of ​​the smart display stand meets the condition of leaving the smart display stand, the ID retention mechanism is activated to retain the temporary tracking ID associated with the radar signal data. The radar device monitors whether the consumer returns to the preset area of ​​the smart display stand within a first preset time period. If so, the temporary tracking ID is reused and the radar signal data generated when the consumer returns to the preset area of ​​the smart display stand is merged with the radar signal data associated with the temporary tracking ID.

[0052] In this embodiment, when collecting radar signal data generated by consumers within a preset area of ​​the smart display stand using its radar device, the radar device can monitor the preset area and identify the entry and exit of people within that area for data collection. Specifically, if the radar device detects a consumer entering the preset area of ​​the smart display stand, it generates radar signal data characterizing the consumer's movement trajectory and body shape, and associates this radar signal data with a temporary tracking ID automatically generated for the consumer. Further, in this embodiment, when collecting radar signal data, if the same consumer briefly leaves and then returns to the smart display stand, the temporary tracking ID can be automatically reused based on radar path and spatiotemporal characteristics to prevent duplicate counting and achieve accurate customer flow statistics. In one specific implementation, if the radar device detects that a consumer within the preset area of ​​the smart display stand meets the conditions for leaving the smart display stand, an ID retention mechanism can be activated to retain the temporary tracking ID associated with the radar signal data. After activating the ID retention mechanism, radar devices can monitor whether consumers return to the preset area of ​​the smart display stand within a first preset time period. If so, the temporary tracking ID is reused, and the radar signal data generated when the consumer returns to the preset area of ​​the smart display stand is merged with the radar signal data associated with the temporary tracking ID. Specifically, the ID retention mechanism sets an expiration window, retaining the temporary tracking ID for the first preset time period. After the first preset time period expires, the temporary tracking ID is invalidated to protect consumer privacy. The conditions for determining whether a consumer has left the smart display stand include the distance between the consumer and the smart display stand detected by the radar device being greater than a preset distance threshold, the duration for which the distance between the consumer and the smart display stand is greater than the preset distance threshold being greater than a second preset time period, and the direction of the consumer's movement trajectory and body outline being away from the smart display stand area. For consumers who return, by merging the radar signal data generated after the consumer returns to the preset area of ​​the smart display stand with the radar signal data previously generated and associated with the temporary tracking ID, duplicate counting of customer flow can be prevented, achieving accurate customer flow statistics.

[0053] In some embodiments of this application, please refer to Figure 3 , Figure 3 A flowchart illustrating a method for reusing temporary tracking IDs in the data processing method of the intelligent display stand system provided in this application embodiment. (See flowchart for example.) Figure 3 As shown, it may specifically include steps S31 to S33.

[0054] S31: If the radar device detects that a consumer enters the preset area of ​​the smart display stand within a first preset time period, then calculate the trajectory similarity between the consumer's movement trajectory and the movement trajectory of the consumer corresponding to the temporary tracking ID, the body shape similarity between the consumer's body outline and the body shape outline of the consumer corresponding to the temporary tracking ID, the feature similarity between the consumer's movement pattern and the movement pattern of the consumer corresponding to the temporary tracking ID, and the time interval coefficient generated based on the consumer's departure time.

[0055] S32: Based on the trajectory similarity, body shape similarity, motion feature similarity, and time interval coefficient, calculate the identity matching degree between consumers who enter the preset area of ​​the smart display stand within the first preset time period and the person corresponding to the temporary tracking ID;

[0056] S33: If the identity matching degree reaches a preset matching degree threshold, it is determined that the consumer returns to the preset area of ​​the smart display stand within a first preset time period, the temporary tracking ID is reused, and the radar signal data generated when the consumer returns to the preset area of ​​the smart display stand is merged with the radar signal data associated with the temporary tracking ID.

[0057] In this embodiment, when the radar device detects a person entering a preset area of ​​the smart display stand within a first preset time period, it can calculate the identity matching degree between the person entering the preset area of ​​the smart display stand within the first preset time period and the consumer corresponding to the temporary tracking ID. Based on the identity matching degree, it can determine whether the consumer has returned to the preset area of ​​the smart display stand within the first preset time period. Specifically, the formula for calculating the identity matching degree is: Identity matching degree = 0.4 * trajectory similarity + 0.3 * body shape similarity + 0.2 * time interval coefficient + 0.1 * motion feature similarity. In a specific implementation, by using the Dynamic Time Warping (DTW) algorithm to compare the path sequences of the consumer entering the preset area of ​​the smart display stand within the first preset time period with the path sequences of the consumer corresponding to the temporary tracking ID entering the preset area of ​​the smart display stand, and calculating the Euclidean distance deviation rate between the two path sequences, the trajectory similarity between the movement trajectory of the consumer entering the preset area of ​​the smart display stand within the first preset time period and the movement trajectory of the consumer corresponding to the temporary tracking ID can be obtained. By analyzing radar echo intensity distribution, the body contour features of consumers entering the preset area of ​​the smart display stand within a first preset time period and the body contour features of consumers corresponding to temporary tracking IDs are generated. The cosine similarity between the two body contour features yields the body contour similarity between the consumers entering the preset area of ​​the smart display stand within the first preset time period and the consumers corresponding to temporary tracking IDs. Similarly, by analyzing radar echo intensity distribution, the gait frequency, turning angle, and other characteristics of consumers entering the preset area of ​​the smart display stand within the first preset time period and the consumers corresponding to temporary tracking IDs are analyzed. The cosine similarity between the gait frequency, turning angle, and other characteristics of the two consumers yields the motion feature similarity between the motion patterns of consumers entering the preset area of ​​the smart display stand within the first preset time period and the motion patterns of consumers corresponding to temporary tracking IDs. Based on the time the consumer corresponding to the temporary tracking ID leaves the smart display stand and the time the consumer enters the preset area of ​​the smart display stand within the first preset time period, the consumer's departure time is calculated. Then, based on the consumer's departure time and a preset attenuation function, a time interval coefficient is calculated. It can be understood that the longer the departure time, the smaller the time interval coefficient. After calculating the identity matching degree, the identity matching degree is compared with the preset matching degree threshold. If the identity matching degree reaches the preset matching degree threshold, it is determined that the consumer returns to the preset area of ​​the smart display stand within the first preset time period. At this time, it can be determined that the consumer who enters the preset area of ​​the smart display stand within the first preset time period is the same consumer as the consumer corresponding to the temporary tracking ID. Therefore, the temporary tracking ID is reused and the radar signal data generated when the consumer returns to the preset area of ​​the smart display stand is merged with the radar signal data associated with the temporary tracking ID.

[0058] In one specific implementation, the following decision logic can be set based on the identity matching degree to achieve data merging: When the identity matching degree is ≥0.8, the temporary tracking ID is directly reused, and the radar signal data generated when the consumer returns to the preset area of ​​the smart display stand is merged with the radar signal data associated with the temporary tracking ID; when 0.6≤identity matching degree<0.8, auxiliary verification is triggered, such as sending a verification prompt message to the salesperson to request the salesperson to confirm whether the consumer is the consumer who returned after leaving according to the temporary tracking ID. After the salesperson confirms, the temporary tracking ID is reused, and the radar signal data generated when the consumer returns to the preset area of ​​the smart display stand is merged with the radar signal data associated with the temporary tracking ID; when the identity matching degree is <0.6, a new temporary tracking ID is automatically generated for the consumer who entered the preset area of ​​the smart display stand within the first preset time period, and the radar signal data generated when the consumer entered the preset area of ​​the smart display stand is associated with the new temporary tracking ID.

[0059] In one specific implementation, for radar tracking anomalies where multiple people simultaneously enter the preset area of ​​the smart display stand within a first preset time period, K-means clustering can be used to separate the movement trajectories of each person entering the preset area simultaneously, and the identity matching degree can be calculated and compared one by one. For anomalies where the movement trajectories of consumers entering the preset area of ​​the smart display stand within the first preset time period cannot be identified due to scene occlusion, a trajectory prediction model trained based on LSTM can be used to fill in the missing movement trajectory data. For radar tracking anomalies where the duration of continuous tracking failure exceeds a third preset time period (e.g., 5 seconds), the radar recognition process can be restarted to obtain the consumer's movement trajectory.

[0060] In some embodiments of this application, please refer to Figure 4 , Figure 4 The flowchart illustrates a method for analyzing consumer intentions using a consumer intention analysis model in the data processing method of the intelligent display stand system provided in this application embodiment. (For example...) Figure 4 As shown, it may specifically include steps S41 to S43.

[0061] S41: Extract information from the multi-dimensional behavioral data according to the dimension categories to obtain behavioral data of multiple categories;

[0062] S42: Weighted fusion of the behavioral data from the multiple categories is performed to obtain a multi-dimensional feature vector;

[0063] S43: The consumer intention data is generated by performing consumer intention analysis on the multi-dimensional feature vector using a pre-trained temporal neural network and attention mechanism in the preset consumer intention analysis model.

[0064] In this embodiment, the multi-dimensional behavioral data includes data collected in various forms based on multimodal approaches. Specifically, multi-dimensional behavioral data includes, but is not limited to, sales guide dialogue data, human-computer interaction data, and radar signal data. In this embodiment, when using a consumer intention analysis model to analyze consumer intentions from multi-dimensional behavioral data, it is first necessary to extract key information representing human behavior from the multi-dimensional behavioral data. Specifically, information can be extracted from the multi-dimensional behavioral data according to dimension categories to obtain multiple categories of behavioral data. For example, according to dimension categories, the key information representing human behavior can be divided into recent sales data of similar products, consumer browsing dwell time data, touch interaction data, sales guide service time data, price range attention data, sales guide guidance effect data, consumer movement trajectory data, consumer body contour data, and consumer gesture data, etc. Based on the above categorization, the weighting of recent sales data for similar products can be set by calculating recent sales trends using a time decay factor; the weighting of consumer browsing dwell time data can be set using nonlinear function transformation; the weighting of touch interaction data can be set based on consumer touch frequency and duration; the weighting of sales guide service time data can be set by establishing a correlation model with the probability of conversion; the weighting of price range attention data can be set by training price sensitivity curves using historical conversion data; the weighting of sales guide effect data can be dynamically set based on model testing results of consumer intention analysis models; and the weighting of consumer movement trajectory data and consumer body contour data can be set using fixed preset values. After obtaining behavioral data for multiple categories and their corresponding weighting values, the behavioral data for these multiple categories is weighted according to their respective weighting values ​​to achieve data fusion and obtain corresponding multi-dimensional feature vectors. It can be understood that the consumer browsing dwell time data is used to determine the intensity of consumer interest; the longer the dwell time, the greater the weighting value. If a consumer picks up a product and then puts it back in its original position, it can be considered a misoperation or low interest, in which case the weighting value of the consumer's browsing dwell time data is smaller. After obtaining the multi-dimensional feature vector, the pre-trained temporal neural network and attention mechanism in the consumer intention analysis model are used to analyze the consumer intention of the multi-dimensional feature vector, thereby generating consumer intention data.

[0065] In some embodiments of this application, please refer to Figure 5 , Figure 5 This document presents a flowchart illustrating a method for triggering deeper human-computer interaction between the smart display stand and consumers within the data processing method of the smart display stand system provided in this embodiment. Figure 5 As shown, it may specifically include steps S51 to S52.

[0066] S51: Determine whether the consumer meets the conditions for initiating deep human-computer interaction based on the multi-dimensional behavioral data;

[0067] S52: If the conditions are met, the smart display stand will be triggered to engage in deeper human-computer interaction with the consumer.

[0068] In this embodiment, since the multi-dimensional behavioral data is real-time collected time-series data, data collection begins when a consumer enters the preset area of ​​the smart display stand and continues until the consumer leaves. In this embodiment, after generating promotional content based on consumer intention data and displaying it on the smart display stand, the consumer's behavior can be further analyzed based on the real-time collected multi-dimensional behavioral data to determine whether the consumer meets the initiation conditions for deeper human-computer interaction. This allows for the initiation of deeper human-computer interaction as needed to guide the consumer to ask further questions, recommend bundled promotions, and provide decision-making support services, thus providing more in-depth services, improving the consumer's service experience, and promoting sales orders. Specifically, an interaction initiation threshold can be predefined to determine whether a consumer meets the initiation conditions for deeper human-computer interaction. When the interaction initiation threshold reaches a preset value, the consumer is considered to meet the initiation conditions for deeper human-computer interaction. The interaction initiation threshold can be obtained by weighting the dwell time target rate, product interaction frequency, and silence time. For example, the formula for calculating the interaction initiation threshold is: Interaction Initiation Threshold = 0.6 * Dwell Time Target Rate + 0.3 * Product Interaction Frequency + 0.1 * Silence Duration. It should be noted that the Dwell Time Target Rate = Actual Dwell Time / Average Consumer Decision Time. Based on the above interaction initiation threshold, when a consumer's dwell time on a product reaches a preset value, or when a consumer touches two or more similar products consecutively without making a clear behavioral response such as asking questions or using gestures, it can be considered that the consumer meets the initiation conditions for deep human-computer interaction. Deep human-computer interaction can then be initiated to guide and assist the consumer in making relevant decisions.

[0069] When intelligent display stands engage in deeper human-computer interaction with consumers, personalized guiding scripts can be generated based on multi-dimensional behavioral data. These scripts guide consumers to engage in deeper human-computer interaction with the intelligent display stands. Specifically, when generating personalized guiding scripts, open-ended questions can be designed based on the products that consumers are interested in and the core selling points of the products. For example, for skincare products, the guiding script could be designed as, "This moisturizing serum contains natural hyaluronic acid. Are you more concerned about daytime protection or nighttime repair?" For makeup products, the guiding script could be designed as, "This lipstick's matte texture is perfect for autumn and winter. Would you like me to test the staying power of similar shades?" Furthermore, guiding scripts can also be designed based on the relevant attributes of the product in the product knowledge graph.

[0070] When the smart display stand engages in deeper human-computer interaction with consumers, it can also set up multi-round guidance strategies. The first round of guidance can be generated by focusing on the product's functional attributes, such as "Would you like to know which skin type it is suitable for?" If no response is received from the consumer, the second round of guidance can be generated by using scenario-based applications, such as "This is suitable for touch-ups during commutes. Would you like a demonstration of how to use it?" If no response is received from the second round of guidance, a prompt for the sales associate to intervene and provide manual guidance.

[0071] When intelligent display stands engage in deeper human-computer interaction with consumers, they can also be configured with bundled display strategies. The display can be divided into multiple areas to showcase different products, including the main product and complementary products. Furthermore, it can dynamically display promotional information for the main product and complementary products, such as "Save 30 yuan on the bundle" or "Free sample of the same product." It can also provide relevant contextual descriptions of the main product and complementary products when combined, such as "A three-piece morning skincare set, from cleansing to protection in one step." A real-time adjustment mechanism can be set up for the bundled display strategy. When a consumer touches a complementary product, the bundle's weight is increased in real time, and a detailed comparison of the products is displayed; if a consumer explicitly refuses a bundled product, similar bundled products are no longer recommended. When intelligent display stands engage in deeper human-computer interaction with consumers, they can also identify consumer decision-making obstacles. Semantic analysis is performed on text data converted from consumer voice data in sales guide dialogues to extract hesitant keywords such as "too oily," "high price," and "uncertain effects," which prevent consumers from making a decision, thus identifying consumer decision-making obstacles. Behavioral analysis is also performed on consumer gesture data from radar signal data, identifying obstacles based on findings such as "repeatedly switching between two products (≥3 times) with similar dwell times." Furthermore, corresponding obstacle tags are generated based on these obstacles, including but not limited to "texture concerns," "cost-effectiveness considerations," and "doubtful effects." Then, solutions tailored to the consumer are provided based on these obstacle tags. For example, for texture concerns, an ingredient comparison table and skin type compatibility can be displayed to alleviate concerns; for price concerns, unit prices can be broken down and a promotion countdown timer can be displayed to address price concerns; for effects concerns, short videos comparing before and after use and access to nearby stores for trial appointments can be shown to alleviate concerns.

[0072] When intelligent display stands engage in deeper human-computer interaction with consumers, they can also be configured with decision-making path optimization strategies. For two-choice scenarios, a comparison matrix of the two products can be generated and displayed, highlighting their core differences. Focusing on only 2 to 3 key differences avoids information overload. For multi-choice scenarios, products can be sorted based on consumer intention scores, retaining only the top N products for recommendation. For scenarios requiring consumer verification, sample trial guidance can be provided.

[0073] In some embodiments of this application, please refer to Figure 6 , Figure 6 A flowchart illustrating one method for updating the dialogue knowledge vector library in the data processing method of the intelligent display stand system provided in this application embodiment. Figure 6 As shown, it may specifically include steps S61 to S63.

[0074] S61: Based on the human-computer interaction data generated when consumers interact with the smart display stand and the sales guide dialogue data generated when consumers communicate with sales guides in the multi-dimensional behavioral data, new sales guide dialogue is generated in combination with the pre-stored dialogue templates.

[0075] S62: Score the new sales script according to the preset script quality scoring rules to obtain the script quality score corresponding to the new sales script;

[0076] S63: If the quality score of the sales script reaches the preset score threshold, the new sales script will be updated to the sales script knowledge vector library.

[0077] In this embodiment, the intelligent display stand system can also utilize the intelligent display stand and the big data model algorithm to collect and accumulate knowledge from the data generated during the shopping guide service process, thereby empowering new shopping guides. Based on the products sold by the intelligent display stand, the big data model stores knowledge graph data related to the sold products. This knowledge graph data is used to display to consumers when they interact with the intelligent display stand. The knowledge graph data includes structured semantic relationship data of products such as "Product A belongs to category B" and "Product A and Product D are complementary products," as well as structured shopping guide script data related to products such as the corresponding scripts for the "whitening" effect of Product A and the corresponding scripts for the "moisturizing" effect of Product B. In this embodiment, after the service to a consumer is completed, new shopping guide scripts can be generated based on the multi-dimensional behavioral data obtained from the consumer, including the human-computer interaction data generated when the consumer interacts with the intelligent display stand and the shopping guide dialogue data generated when the consumer talks with the shopping guide, combined with pre-stored script templates. For example, the pre-stored script templates can be extracted from high-rated historical script data. When generating new sales scripts, real-time data such as consumer-focused product tags, current product inventory, and current product promotions can be extracted from the multi-dimensional behavioral data of consumers who have completed their service. This real-time data is then populated into a pre-stored script template to generate the new sales script. The new sales script is scored according to a preset script quality scoring rule, specifically using the following formula: Quality Score = 0.3 * Conversion Rate + 0.2 * Consumer Dwell Time + 0.2 * Interaction Frequency + 0.15 * Average Order Value Increase + 0.15 * Customer Satisfaction. When the script quality score reaches a preset threshold, the new sales script is updated to the script knowledge vector library. Simultaneously, the knowledge graph data is updated, for example, by adding mappings between product attributes and scripts, and strengthening the association paths for high-conversion products. After the new sales script is updated to the script knowledge vector library, it can be distributed to sales associates' mobile devices, allowing them to recommend the new script to customers. New sales associates can quickly learn and transfer skills through mobile prompts, thus filling gaps in their abilities, significantly improving their performance and reducing training costs.

[0078] It is understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0079] In some embodiments of this application, please refer to Figure 7 and Figure 8 , Figure 7 This is a schematic diagram of a smart display stand system provided in an embodiment of this application. Figure 8This is a structural diagram of an intelligent display stand in an intelligent display stand system provided in an embodiment of this application. (See diagram below.) Figure 7 and Figure 8 As shown, the intelligent display stand system includes a server and an intelligent display stand. The intelligent display stand system is used to implement the data processing method described above. In this embodiment, the server receives multi-dimensional behavioral data transmitted from the intelligent display stand, performs consumer intention analysis based on the multi-dimensional behavioral data, and generates promotional content to be fed back to the intelligent display stand. The intelligent display stand is controlled by a main control module and includes a display device, a radar device, a microphone device, and a communication device. The display device includes a monitor displaying product information and promotional content, and a display stand for placing products, and is used to collect human-computer interaction data generated when consumers interact with the intelligent display stand. The radar device is used to collect radar signal data generated when consumers are within a preset area of ​​the intelligent display stand. The microphone device includes a Bluetooth microphone and a Bluetooth positioning base station device, used to connect with the Bluetooth positioning base station device to collect sales guide dialogue data and sales guide location information generated when consumers are within the preset area of ​​the intelligent display stand. It is understood that the Bluetooth microphone is equipped with a Bluetooth positioning tag for positioning, and the Bluetooth positioning base station device can collect the sales guide's location information within the preset area of ​​the intelligent display stand by scanning the Bluetooth positioning tag. The communication device is used to package the data collected by the display devices, radar devices, and microphone devices in the smart display rack into multi-dimensional behavioral data and transmit it to the server. Furthermore, in this embodiment, the smart display rack system may also include a mobile terminal for sales associates. The server sends updated sales scripts to the sales associate's mobile terminal, recommending new sales scripts to the sales associate to serve consumers. New sales associates can quickly learn and transfer skills through mobile terminal prompts, thus completing their skill gaps, significantly improving their performance capabilities, and reducing training costs.

[0080] In one specific implementation, please refer to Figure 9 , Figure 9 This is a basic structural block diagram of a data processing device for an intelligent display stand system provided in an embodiment of this application. Deploying this device in an intelligent display stand system allows the various units included in the device to execute the steps in the above-described method embodiments. Please refer to the relevant descriptions in the above-described method embodiments for details. For ease of explanation, only the parts relevant to this embodiment are shown. Figure 9As shown, the data processing device of the intelligent display stand system includes: a personnel monitoring module 91, a Bluetooth scanning module 92, an identity determination module 93, an intention analysis module 94, and a content display module 95. Specifically: the personnel monitoring module 91 monitors personnel in a preset area of ​​the intelligent display stand using its radar device, automatically generating a temporary tracking ID for each person entering the preset area. The Bluetooth scanning module 92 scans the preset area of ​​the intelligent display stand synchronously using a Bluetooth positioning base station device to determine if a Bluetooth positioning tag ID exists at the location of the temporary tracking ID. The identity determination module 93 determines the person as a sales guide if a Bluetooth positioning tag ID exists, otherwise, it determines them as a consumer. The intention analysis module 94, when determining the person as a consumer, acquires multi-dimensional behavioral data generated by the consumer within the preset area of ​​the intelligent display stand, and uses a preset consumer intention analysis model to analyze the multi-dimensional behavioral data to generate consumer intention data. The content display module 95 generates promotional content based on the consumer intention data and displays it on the intelligent display stand.

[0081] It should be understood that the data processing device of the aforementioned intelligent display stand system corresponds one-to-one with the data processing method of the aforementioned intelligent display stand system, and will not be described in detail here.

[0082] In some embodiments of this application, please refer to Figure 10 , Figure 10 This is a basic structural block diagram of a device provided in an embodiment of this application. For example... Figure 10 As shown, the device 10 in this embodiment includes: a processor 101, a memory 102, and a computer program 103 stored in the memory 102 and executable on the processor 101, such as a program for a data processing method for an intelligent display stand system. When the processor 101 executes the computer program 103, it implements the steps in each embodiment of the data processing method for the intelligent display stand system described above. Alternatively, when the processor 101 executes the computer program 103, it implements the functions of each module in the embodiment corresponding to the data processing device for the intelligent display stand system described above. Please refer to the relevant descriptions in the embodiments for details, which will not be repeated here.

[0083] For example, the computer program 103 can be divided into one or more modules (units) for performing the various steps in the above method embodiments. The one or more modules are stored in the memory 102 and executed by the processor 101 to complete this application. The one or more modules can be a series of computer program instruction segments capable of performing specific functions, which describe the execution process of the computer program 103 in the device 10.

[0084] The device may include, but is not limited to, a processor 101 and a memory 102. Those skilled in the art will understand that... Figure 10 This is merely an example of device 10 and does not constitute a limitation on device 10. It may include more or fewer components than shown, or combine certain components, or different components. For example, the device may also include input / output devices, network access devices, buses, etc.

[0085] The processor 101 can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor.

[0086] The memory 102 can be an internal storage unit of the device 10, such as a hard disk or RAM of the device 10. The memory 102 can also be an external storage device of the device 10, such as a plug-in hard disk, Smart Media Card (SMC), Secure Digital (SD) card, or FlashCard equipped on the device 10. Furthermore, the memory 102 can include both internal and external storage units of the device 10. The memory 102 is used to store the computer program and other programs and data required by the device. The memory 102 can also be used to temporarily store data that has been output or will be output.

[0087] It should be noted that the information interaction and execution process between the above-mentioned devices / units are based on the same concept as the method embodiments of this application. For details on their specific functions and technical effects, please refer to the method embodiments section, and they will not be repeated here.

[0088] This application also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, can implement the steps in the various method embodiments described above. In this embodiment, the computer-readable storage medium can be either non-volatile or volatile.

[0089] This application provides a computer program product that, when run on a mobile terminal, enables the mobile terminal to implement the steps described in the various method embodiments.

[0090] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is merely an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiments can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit. Furthermore, the specific names of the functional units and modules are only for easy differentiation and are not intended to limit the scope of protection of this application. The specific working process of the units and modules in the above device can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0091] If the integrated module / unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include: any entity or device capable of carrying the computer program code, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc. It should be noted that the content included in the computer-readable medium can be appropriately added or removed according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media do not include electrical carrier signals and telecommunication signals.

[0092] In the above embodiments, the descriptions of each embodiment have different focuses. For parts that are not described in detail or recorded in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0093] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.

Claims

1. A data processing method for an intelligent display stand system, characterized in that, include: The radar device on the smart display stand monitors personnel in a preset area of ​​the smart display stand. When a person is detected entering the preset area of ​​the smart display stand, a temporary tracking ID is automatically generated for the person. The Bluetooth positioning base station device synchronously scans the preset area of ​​the smart display stand to determine whether a Bluetooth positioning tag ID exists at the location of the temporary tracking ID; If a Bluetooth location tag ID exists, the person is identified as a sales guide; otherwise, they are identified as a consumer. Once the person is identified as a consumer, multi-dimensional behavioral data generated by the consumer within a preset area of ​​the smart display stand is acquired. This multi-dimensional behavioral data includes human-computer interaction data collected by the display device of the smart display stand when the consumer interacts with the smart display stand, sales guide dialogue data collected by the microphone device of the smart display stand when the consumer is within the preset area of ​​the smart display stand, and radar signal data collected by the radar device of the smart display stand when the consumer is within the preset area of ​​the smart display stand. The acquisition of radar signal data includes monitoring the preset area of ​​the smart display stand through the radar device of the smart display stand. If the radar device detects that a consumer has entered the preset area of ​​the smart display stand, radar signal data representing the consumer's movement trajectory and body shape is generated through the radar device of the smart display stand, and the radar signal data is associated with a temporary tracking ID automatically generated for the consumer. Information is extracted from the multi-dimensional behavioral data according to the dimension categories to obtain behavioral data of multiple categories. The behavioral data from the multiple categories are weighted and fused to obtain a multi-dimensional feature vector. The consumer intention data is generated by analyzing the multi-dimensional feature vectors using a pre-trained temporal neural network and attention mechanism in a preset consumer intention analysis model. Promotional content is generated based on the consumer intention data and fed back to the smart display stand for display.

2. The data processing method of the intelligent display stand system according to claim 1, characterized in that, When collecting radar signal data, it also includes: If the radar device detects that a consumer within the preset area of ​​the smart display stand meets the conditions for leaving the smart display stand, an ID retention mechanism is activated to retain the temporary tracking ID associated with the radar signal data. The radar device then monitors whether the consumer returns to the preset area of ​​the smart display stand within a first preset time period. If so, the temporary tracking ID is reused, and the radar signal data generated when the consumer returns to the preset area of ​​the smart display stand is merged with the radar signal data associated with the temporary tracking ID.

3. The data processing method of the intelligent display stand system according to claim 2, characterized in that, The step of monitoring whether a consumer returns to the preset area of ​​the smart display stand within a first preset time period using the radar device, and if so, reusing the temporary tracking ID and merging the radar signal data generated when the consumer returns to the preset area of ​​the smart display stand with the radar signal data associated with the temporary tracking ID, includes: If the radar device detects that a consumer enters the preset area of ​​the smart display stand within a first preset time period, then the similarity between the consumer's movement trajectory and the movement trajectory of the consumer corresponding to the temporary tracking ID, the similarity between the consumer's body outline and the body outline of the consumer corresponding to the temporary tracking ID, the similarity between the consumer's movement pattern and the movement pattern of the consumer corresponding to the temporary tracking ID, and the time interval coefficient generated based on the consumer's departure time are calculated. Based on the trajectory similarity, body shape similarity, motion feature similarity, and time interval coefficient, the identity matching degree between consumers who enter the preset area of ​​the smart display stand within the first preset time period and the person corresponding to the temporary tracking ID is calculated. If the identity matching degree reaches a preset matching degree threshold, it is determined that the consumer returns to the preset area of ​​the smart display stand within a first preset time period. The temporary tracking ID is reused, and the radar signal data generated when the consumer returns to the preset area of ​​the smart display stand is merged with the radar signal data associated with the temporary tracking ID.

4. The data processing method of the intelligent display stand system according to claim 1, characterized in that, After the step of generating promotional content based on the consumer intention data and feeding it back to the smart display stand for display, the method further includes: Based on the multi-dimensional behavioral data, determine whether the consumer meets the conditions for initiating deep human-computer interaction; If the conditions are met, the smart display stand will be triggered to engage in deeper human-computer interaction with the consumer.

5. The data processing method for the intelligent display stand system according to claim 1, characterized in that, After the step of generating promotional content based on the consumer intention data and feeding it back to the smart display stand for display, the method further includes: Based on the human-computer interaction data generated when consumers interact with the smart display stand and the sales guide dialogue data generated when consumers communicate with sales guides, new sales guide dialogues are generated by combining the pre-stored dialogue templates. The new sales script is scored according to the preset script quality scoring rules to obtain the corresponding script quality score; If the quality score of the sales script reaches the preset scoring threshold, the new sales script will be updated to the sales script knowledge vector library.

6. An intelligent display stand system, characterized in that, The intelligent display stand system is used to implement the method as described in any one of claims 1-5, and the intelligent display stand system includes a server and an intelligent display stand, wherein: The server is used to receive multi-dimensional behavioral data transmitted by the smart display stand, perform consumer intention analysis and processing based on the multi-dimensional behavioral data, and generate promotional content to be fed back to the smart display stand. The intelligent display stand includes a display device, a radar device, a microphone device, and a communication device. The display device includes a monitor for displaying product information and promotional content, as well as a display stand for placing products, and is used to collect human-computer interaction data generated when consumers interact with the intelligent display stand. The radar device is used to collect radar signal data generated when consumers are within a preset area of ​​the intelligent display stand. The microphone device includes a Bluetooth microphone and a Bluetooth positioning base station device, used to connect the Bluetooth microphone to the Bluetooth positioning base station device to collect sales guide dialogue data and sales guide location information generated when consumers are within the preset area of ​​the intelligent display stand. The communication device is used to package the data collected by the display device, radar device, and microphone device in the intelligent display stand into multi-dimensional behavioral data and transmit it to the server.

7. An apparatus comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the method as described in any one of claims 1-5.

8. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method as described in any one of claims 1-5.

Images