Control system and method for jewelry smart terminal modalities

Abstract

This invention belongs to the field of intelligent control technology for jewelry terminals. It discloses a control system and method for the modalities of jewelry intelligent terminals, including: collecting a multi-dimensional control signal dataset, parsing the control intent, and obtaining a control state feature map; predicting modal switching, identifying potential modal transition nodes, and constructing a modal transition probability matrix; generating cross-modal control signal mapping rules to obtain inter-modal control signal transmission channels; performing real-time state synchronization, evaluating the continuity of control response, identifying control interruption nodes, and performing intelligent adjustment processing to form a seamless control path; optimizing multi-modal collaborative adjustment strategies to generate modal linkage control schemes; intelligently orchestrating functional control units to obtain personalized control flows; and executing personalized control flows to achieve collaborative operation and perform real-time monitoring and dynamic adjustment; significantly improving the user experience of jewelry intelligent terminals.

Description

Technical Field

[0001] This invention relates to the field of intelligent control technology for jewelry terminals, and more specifically, to a control system and method for intelligent jewelry terminal modes. Background Technology

[0002] With the rapid development of intelligent terminal control technology, the intelligent terminal equipment in the jewelry retail industry is undergoing a critical period of transformation from single-function control to multimodal collaborative control. Modern jewelry intelligent terminals integrate multiple functional control modules such as AR display control unit, AI computing control unit, and customized processing control unit to build a complex multimodal control system architecture, which puts forward higher technical requirements for the system's collaborative control capability and status monitoring and adjustment capability.

[0003] Existing smart jewelry terminals suffer from significant shortcomings in the design of multimodal collaborative control systems, severely hindering the overall operational efficiency and continuity of control response. Specifically, when the system needs to switch between different control modes such as AR display control, AI computation control, and customized processing control, the lack of effective control signal transmission channels and state synchronization mechanisms between control units leads to interruptions and fragmentation of the control process. For example, after the AR display control unit completes the rendering of the try-on effect and outputs user preference parameters, switching to the AI ​​computation control unit fails to automatically inherit the control state and transfer control parameters, requiring re-initialization of parameter acquisition and computation. Similarly, when switching to the customized processing control unit, the output signals and computation results of the preceding control unit cannot be automatically imported as input parameters, necessitating repeated signal acquisition and parameter parsing processes. This signal isolation between control units not only increases the response delay of the control system but also prevents the effective reuse of a large amount of valuable control state data and intermediate computation results. Existing solutions mostly employ simple control command jumps and manual parameter passing methods, lacking a unified multimodal collaborative control architecture and closed-loop feedback adjustment mechanism, thus failing to achieve intelligent linkage control and seamless state connection between various functional control units. At the same time, existing systems lack real-time monitoring devices and dynamic adjustment units for multimodal collaborative operation status, making it impossible to identify control interruption nodes in a timely manner and implement compensation adjustments, ultimately affecting the overall stability, control efficiency, and response consistency of the jewelry smart terminal control system.

[0004] In view of this, the present invention proposes a control system and method for a smart jewelry terminal mode to solve the above problems. Summary of the Invention

[0005] To overcome the aforementioned deficiencies of the prior art and to achieve the above objectives, the present invention provides the following technical solution: a control method for a jewelry smart terminal mode, comprising:

[0006] Step S1: Collect multi-dimensional control signal datasets from each control unit in the smart jewelry terminal, analyze the control intent of the multi-dimensional control signal datasets, and obtain a control state feature map;

[0007] Step S2: Based on the control state feature map, predict mode switching, identify potential mode transition nodes, and construct a mode transition probability matrix; generate cross-mode control signal mapping rules based on the mode transition probability matrix to obtain the inter-mode control signal transmission channel;

[0008] Step S3: Real-time state synchronization is performed through the inter-modal control signal transmission channel to obtain a multimodal control state consistency view; based on the multimodal control state consistency view, the continuity of control response is evaluated, control interruption nodes are identified, and intelligent adjustment processing is performed on the control interruption nodes to form a seamless control path;

[0009] Step S4: Optimize the multimodal collaborative adjustment strategy based on the seamless control path to generate a modal linkage control scheme; intelligently orchestrate the functional control units based on the modal linkage control scheme to obtain a personalized control flow;

[0010] Step S5: Execute the personalized control process to achieve the coordinated operation of AR display control, AI calculation control and customized processing control functions; and monitor and dynamically adjust the coordinated operation effect in real time.

[0011] The control system for the smart terminal mode of jewelry includes:

[0012] Data parsing module: Collects multi-dimensional control signal datasets from various control units in the jewelry smart terminal, analyzes the control intent of the multi-dimensional control signal datasets, and obtains a control state feature map;

[0013] Modality mapping module: Based on the control state feature map, it predicts mode switching, identifies potential mode transition nodes, and constructs a mode transition probability matrix; based on the mode transition probability matrix, it generates cross-modal control signal mapping rules to obtain the inter-modal control signal transmission channel;

[0014] State assessment module: Real-time state synchronization is performed through the inter-modal control signal transmission channel to obtain a consistent view of the multimodal control state; based on the consistent view of the multimodal control state, the continuity of control response is assessed, control interruption nodes are identified, and intelligent adjustment is performed on the control interruption nodes to form a seamless control path;

[0015] Process formulation module: Optimizes multimodal collaborative adjustment strategies based on seamless control paths to generate modal linkage control schemes; intelligently orchestrates functional control units based on modal linkage control schemes to obtain personalized control processes;

[0016] Process Implementation Module: Executes personalized control processes to achieve coordinated operation of AR display control, AI computation control, and customized processing control functions; monitors and dynamically adjusts the coordinated operation effect in real time.

[0017] The technical effects and advantages of the control system and method for the smart terminal mode of jewelry in this invention are as follows:

[0018] This invention extracts explicit control demand features and implicit control preference features from a multi-dimensional control signal dataset through time-series correlation analysis, providing a unified control representation foundation for multi-modal collaborative control. By identifying high-frequency conversion paths, it establishes a cross-modal control signal dictionary and a control signal mapping rule base, forming an inter-modal control signal transmission channel. This enables automatic conversion and transmission of source modal output signals to target modal input signals, reducing the probability of control interruption during modal switching. Conflict resolution at state conflict points ensures consistent control states. Differentiated intelligent adjustment processing, including preloading, state inheritance, and intelligent prompts, is employed for different interruption types, forming a seamless control path. This solves the problems of control flow interruption and fragmentation in existing technologies, allowing the output signals and calculation results of preceding control units to be automatically imported as input parameters into subsequent control units. A real-time inter-unit communication mechanism is established to monitor response delay, signal synchronization rate, and functional smoothness during collaborative operation. Based on performance indicators, bottlenecks are identified, and optimization measures such as caching strategy adjustment, parallel processing optimization, and interface transition smoothing are implemented, forming a closed-loop control mechanism. This improves the overall stability, control efficiency, and response consistency of the multi-modal control system for jewelry smart terminals. Attached Figure Description

[0019] Figure 1 This is a schematic diagram of the control method for the smart jewelry terminal mode of the present invention;

[0020] Figure 2 This is a schematic diagram of the control system for the smart jewelry terminal mode of the present invention. Detailed Implementation

[0021] 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. Example 1

[0022] Please see Figure 1 As shown, the control method for the jewelry smart terminal mode in this embodiment includes:

[0023] Step S1: Collect multi-dimensional control signal datasets from each control unit in the jewelry smart terminal, analyze the control intent of the multi-dimensional control signal datasets, and obtain a control state feature map.

[0024] Modern smart jewelry terminals integrate multiple functional control modules, including AR display control units, AI computing control units, and customized processing control units. Users generate various types of control signals during operation, reflecting their control intentions and operational preferences. To achieve multimodal collaborative control, it is first necessary to comprehensively collect control signal data generated by each control unit and perform in-depth analysis.

[0025] In this embodiment, the jewelry smart terminal is equipped with a touch screen, an eye tracker, a voice recognition module, and a gesture recognition camera to collect user control signals. Specifically, it collects the touch signal trajectory, gaze focus signal, voice command signal, and gesture action signal of each control unit in the jewelry smart terminal to form a multi-dimensional control signal dataset.

[0026] The touch signal trajectory includes the coordinates of the user's click position on the touchscreen, the swipe trajectory, and the duration of the pause; the gaze focus signal includes the coordinate sequence of the user's gaze point and the duration of the gaze; the voice command signal includes the command text after voice recognition and semantic tags; and the gesture action signal includes the gesture type encoding and the action amplitude parameters. This embodiment uses a 50-millisecond sampling interval for signal acquisition. This sampling interval meets the response requirements of real-time control while avoiding data redundancy caused by excessively high sampling frequencies.

[0027] A time-series correlation analysis is performed on the multidimensional control signal dataset to obtain control sequence patterns. The specific method for time-series correlation analysis is as follows: The collected control signals are aligned according to timestamps to form a multi-channel time-series signal matrix; the signal matrix is ​​segmented using a sliding time window. In this embodiment, the time window length is set to 3 seconds and the sliding step size is set to 1 second. This parameter setting is based on the fact that the average time for a user to complete a single control operation is approximately 2.5 seconds, and a 3-second window length can completely capture the entire process of a control operation; within each time window, the sequence and time interval relationship of the signals from each channel are analyzed to extract signal combination patterns, such as a typical control sequence pattern like "looking at a product image → asking for a price via voice → clicking the details button".

[0028] This study identifies control intentions based on control sequence patterns, extracting explicit control demand features and implicit control preference features. The method for extracting explicit control demand features involves extracting key signals from the control sequence patterns, identifying user-initiated control operation signals, including explicit clicks, voice commands, and confirmation gestures, and constructing these signals into an explicit control signal set. The frequency of occurrence and duration of each type of signal in the explicit control signal set are analyzed to determine the user's control focus, generating an explicit control demand feature vector. For example, if a user clicks the AR try-on function 3.2 times per minute and the customization function 0.8 times per minute, the component value corresponding to the AR try-on function will be higher in the explicit control demand feature vector.

[0029] The method for extracting implicit control preference features is as follows: Micro-signal capture is performed on control sequence patterns to identify unconscious control tendency signals, including involuntary operation signals such as gaze duration, swipe deceleration, and repeated re-viewing. These signals constitute an implicit control signal set. By analyzing the dwell time distribution and re-viewing frequency of each signal in the implicit control signal set, the user's potential interests are inferred, and an implicit control preference feature vector is extracted. For example, if a user's average gaze duration on a ring image is 4.5 seconds, and they re-view the ring 3 times after browsing other products, then the implicit control preference component value corresponding to that ring is relatively high.

[0030] The explicit control demand characteristics and implicit control preference characteristics are weighted to construct a comprehensive control intent representation. In this embodiment, the weight of the explicit control demand characteristics is set to 0.6, and the weight of the implicit control preference characteristics is set to 0.4. This weight allocation is based on the fact that explicit operations directly reflect user intent and should be given higher weight, while implicit preferences, as auxiliary judgment criteria, have a lower weight.

[0031] Semantic fusion is performed on explicit control demand features and implicit control preference features to generate a control intent vector. The semantic fusion method involves weighted concatenation of the explicit control demand feature vector and the implicit control preference feature vector to form a fused feature vector; a pre-trained semantic encoder is then used to encode the fused feature vector, mapping it to a unified semantic space to obtain the control intent vector. The control intent vector is set to 128 dimensions, which fully represents the semantic information of the user's control intent while maintaining computational efficiency.

[0032] A control state topology is constructed based on the control intent vectors to obtain a control association network. The construction method for the control state topology is as follows: control intent vectors extracted from different time windows are used as nodes, and directed edges are established according to the chronological order of the time windows; the cosine similarity of the control intent vectors between adjacent nodes is calculated, and association edges are added between nodes with a similarity greater than a preset threshold of 0.7 to form the control association network. The preset threshold of 0.7 is set based on the following: when the cosine similarity of two control intent vectors exceeds 0.7, it indicates that the two control states have a strong semantic correlation and a high probability of a control evolution relationship.

[0033] Control evolution paths are deduced through a control association network to form a control state feature map. The method for deducing control evolution paths is as follows: in the control association network, the evolution patterns of control states are analyzed based on the in-degree and out-degree of nodes to identify frequently occurring control evolution paths; these control evolution paths are then integrated and summarized to form the control state feature map. The control state feature map is stored in a graph structure, containing control state nodes, transition edges between states, and transition probability weights, providing a data foundation for subsequent mode switching prediction.

[0034] Step S2: Based on the control state feature map, predict mode switching, identify potential mode transition nodes, and construct a mode transition probability matrix; generate cross-mode control signal mapping rules based on the mode transition probability matrix to obtain the inter-mode control signal transmission channel.

[0035] Existing smart jewelry terminals lack effective control signal transmission channels and state synchronization mechanisms between control units when switching between control modes, leading to interruptions in the control flow. To solve this problem, it is necessary to predict mode switching and establish a control signal transmission mechanism between modes.

[0036] Control trajectory analysis is performed based on the control state feature map to identify control transfer patterns and obtain control migration paths. The method for control trajectory analysis is as follows: extract all complete control trajectories from the initial state to the final state from the control state feature map; statistically analyze the control mode sequences traversed in each trajectory to identify the transfer patterns between modes, such as typical control migration paths like "AR display control mode → AI computation control mode → customized processing control mode".

[0037] Modal correlation is calculated for the control migration path to determine the conversion tendency between modes. The method for calculating modal correlation is as follows: the frequency of occurrence of each modal pair in the control migration path is statistically analyzed, and the proportion of the frequency of conversion from mode A to mode B to the total number of conversions out of mode A is calculated. This proportion is used as the conversion tendency index from mode A to mode B. In this embodiment, the jewelry smart terminal includes three control modes: AR display control mode, AI computing control mode, and customized processing control mode, constituting a total of 9 modal conversion pairs (including self-to-self maintenance).

[0038] A mode switching predictor is constructed based on switching propensity to assess the likelihood of mode switching in real time and obtain the mode switching probability distribution. The modal switching predictor is constructed as follows: using the current control state feature vector and historical control sequences as input, a recurrent neural network is used to predict the modal state at the next time step; the predicted output is the switching probability value of each control mode, forming the mode switching probability distribution vector. The predictor's training data comes from user historical control behavior records, with a training sample size of no less than 10,000 control sequences to ensure the generalization ability of the prediction model.

[0039] High-probability transition nodes are identified based on the modality switching probability distribution, forming a set of potential modality switching nodes. The identification method for high-probability transition nodes is as follows: a modality switching probability threshold of 0.3 is set. When the switching probability of a non-current mode exceeds this threshold, the current moment is determined to be a potential modality switching node, and this node is added to the potential modality switching node set. The threshold of 0.3 is based on the analysis of historical data, which shows that when the modality switching probability exceeds 0.3, the actual probability of modality switching occurring reaches over 85%, demonstrating high predictive accuracy.

[0040] The potential modality transition node set is processed to extract transition conditions, generating modality transition trigger rules. The method for extracting transition conditions involves analyzing the control signal characteristics at the occurrence of each potential modality transition node, extracting key signal combinations that trigger the modality transition, and encoding these key signal combinations into modality transition trigger rules, which are then stored in a rule base. For example, continuously gazing at a product for more than 3 seconds and issuing the voice command "I want to try it on" serves as a trigger rule for switching from AI computation control mode to AR display control mode.

[0041] A mode transition probability matrix is ​​constructed based on the mode transition triggering rules and the mode switching probability distribution. The mode transition probability matrix is ​​an N×N square matrix, where N is the number of control modes; in this embodiment, N=3. The element in the i-th row and j-th column of the matrix represents the probability value of transitioning from mode i to mode j. The matrix is ​​constructed by weighting the historical transition frequencies and real-time prediction probabilities of each mode pair, with a weight of 0.3 for historical frequencies and a weight of 0.7 for real-time prediction probabilities. This weighting setting balances the stability of historical patterns with the timeliness of real-time predictions.

[0042] Control signal dependency analysis is performed based on the mode transition probability matrix to identify cross-modal control signal requirements, generate a control signal mapping rule base, and establish inter-modal control signal transmission channels. The specific method is as follows:

[0043] The high-frequency transition paths in the modality transition probability matrix are analyzed to determine the core modality transition scenarios. The criterion for determining high-frequency transition paths is that transition pairs with a modality transition probability greater than 0.4 are considered high-frequency transition paths. In this embodiment, typical core modality transition scenarios include: the transition from AR display control mode to AI computation control mode (after the user completes a virtual try-on, AI recommends similar styles), and the transition from AI computation control mode to customization processing control mode (after the user accepts the AI ​​recommendation, they enter the customization process).

[0044] For core mode transition scenarios, control signal flow tracing is performed to identify the source mode output signal type and the target mode input signal requirements. The signal flow tracing method involves: analyzing the output signals generated by the source mode when completing the control task, including state variables, intermediate calculation results, and user preference parameters; simultaneously analyzing the input signal type and format requirements required when the target mode starts; and establishing the correspondence between the source mode output signals and the target mode input signals.

[0045] Semantic alignment is performed on signal types to establish a cross-modal control signal dictionary. The semantic alignment method involves classifying control signals from different modalities according to semantic categories to establish a unified signal classification system; for signals with the same semantics but different formats, signal equivalence mapping relationships are defined to form a cross-modal control signal dictionary. For example, the satisfaction score of the virtual try-on effect output by the AR display control modality and the user preference weight coefficient of the AI ​​computation control modality input have a semantic mapping relationship.

[0046] Signal format conversion rules are defined based on a cross-modal control signal dictionary to ensure signal compatibility. These rules include: data type conversion rules (e.g., converting floating-point satisfaction ratings to integer preference levels), data range normalization rules (e.g., mapping ratings from 0-100 to a weight range of 0-1), and data structure conversion rules (e.g., converting list-type signals to vector-type signals).

[0047] A signal priority evaluation mechanism is constructed to classify the importance of transmitted signals. The signal priority evaluation method is as follows: based on the degree of influence of the signal on the target modality control effect, the signals are divided into three priority levels: high priority signals (must be transmitted, the absence of which will cause the target modality to fail to start normally), medium priority signals (recommended to be transmitted, the absence of which will reduce control efficiency), and low priority signals (optional to be transmitted, used to optimize the control experience).

[0048] A control signal mapping rule base is generated based on signal priority and transition rules; an inter-modal control signal transmission channel is established through this rule base. The control signal transmission channel is implemented in the form of a message queue, supporting asynchronous signal transmission and signal buffering to ensure reliable transmission of control signals during mode switching.

[0049] Step S3: Real-time state synchronization is performed through the inter-modal control signal transmission channel to obtain a multimodal control state consistency view; based on the multimodal control state consistency view, the continuity of control response is evaluated, control interruption nodes are identified, and intelligent adjustment processing is performed on the control interruption nodes to form a seamless control path.

[0050] Because each control unit maintains its own control state independently during operation, inconsistencies may occur when mode switching happens, leading to interruptions in the control flow. To ensure the continuity of control response, a real-time synchronization mechanism for multi-modal control states needs to be established.

[0051] Real-time signal synchronization is achieved through an inter-modal control signal transmission channel to obtain the current control status information of each mode. The real-time signal synchronization method is as follows: each control mode publishes its current status information to the signal transmission channel at fixed time intervals (100 milliseconds in this embodiment); the status information includes the control mode identifier, the currently executing task, the task completion progress, the output buffer data, and the resource usage; the signal transmission channel aggregates and organizes the status information published by each mode.

[0052] The control state information is standardized to generate a unified control state descriptor. The standardization method is as follows: a unified control state descriptor format is defined, including a timestamp field, a modality identifier field, a state category field, a state value field, and metadata fields; the original state information of each modality is converted and filled according to the descriptor format to generate a standardized control state descriptor sequence.

[0053] State consistency verification is performed based on a unified control state descriptor to identify state conflict points. The method for state consistency verification is as follows: compare the control state descriptors of each modality at the same time to check for logical conflicts; the types of state conflicts include: resource contention conflicts (multiple modalities simultaneously request exclusive resources), parameter contradiction conflicts (different modalities have contradictory settings for the same parameter), and timing disorder conflicts (subsequent modalities start before the preceding modalities complete). Detected state conflicts are recorded as state conflict points.

[0054] Conflict resolution is performed on state conflict points to obtain coordinated control state data. Different strategies are used for conflict resolution depending on the conflict type: for resource contention conflicts, a priority scheduling strategy is used, with higher-priority modes acquiring resources first; for parameter inconsistencies, a latest-value overwrite strategy is used, with the most recently updated parameter value as the standard; for timing disorder conflicts, a waiting synchronization strategy is used, with subsequent modes waiting for the preceding mode to complete before starting.

[0055] By integrating and coordinating the control status data, a multimodal control status consistency view is constructed. This view provides a global description of the system's current control status, including the operating status of each mode, the signal flow relationships between modes, the queue of signals to be transmitted, and resource allocation. This view is presented in the form of a visual dashboard, facilitating system monitoring and status diagnosis.

[0056] Control operation flow analysis is performed based on a multimodal control state consistency view to identify operation interruption nodes and waiting times. The method of control operation flow analysis is as follows: extract the time series of user control operations from the consistency view, calculate the time interval between adjacent operations; mark the positions where the time interval exceeds a preset threshold (set to 2 seconds in this embodiment) as operation interruption nodes; calculate the actual waiting time at each operation interruption node to evaluate the continuity of control response.

[0057] The operation interruption nodes are classified by interruption type to determine the control interruption node set. The interruption type classification method is as follows: analyze the system state and user behavior at the operation interruption node, and divide the interruption into three types: data missing interruption (caused by the incomplete input data required by the target modality), function switching interruption (caused by the delay in function loading during modality switching), and cognitive jump interruption (caused by the user needing to readjust to the new modality interface after switching modality). All operation interruption nodes are classified by type to form the control interruption node set.

[0058] A smart compensation and adjustment strategy is designed for the set of control interruption nodes, employing preloading, state inheritance, and intelligent prompts for adjustment. The specific method is as follows:

[0059] The causes of control interruptions are analyzed to identify the distribution of data-missing interruptions, function-switching interruptions, and cognitive-jumping interruptions. Based on historical data, in this embodiment, data-missing interruptions account for approximately 35%, function-switching interruptions for approximately 45%, and cognitive-jumping interruptions for approximately 20%.

[0060] A preloading adjustment strategy is implemented for data-missing interruptions, preparing the necessary control data in advance before mode switching. The preloading adjustment strategy is implemented as follows: based on the mode transition probability matrix, when the switching probability of a certain mode exceeds the preloading trigger threshold (set to 0.25 in this embodiment), the system preloads the input data required for the target mode in the background. The preloaded data is stored in a cache and can be directly accessed when the mode switch actually occurs, avoiding data waiting delays. The preloading trigger threshold of 0.25 is set because it is lower than the mode transition probability threshold of 0.3, allowing preloading to be completed before mode switching confirmation, but not so low as to cause excessive invalid preloading operations.

[0061] For function-switching interrupts, a state inheritance adjustment strategy is implemented to automatically transfer the control results and parameter settings of the previous mode. The implementation method of the state inheritance adjustment strategy is as follows: at the moment of mode switching, the system automatically saves a snapshot of the control state of the source mode to the state inheritance cache; when the target mode starts, the relevant state data is read from the state inheritance cache, and parameter initialization and context recovery are automatically completed; the scope of state inheritance is determined by the control signal mapping rule base to ensure that only the valid state data required by the target mode is transferred.

[0062] A smart prompt adjustment strategy is implemented to address cognitive jump-type interruptions, generating transitional control guidance signals. The smart prompt adjustment strategy is implemented as follows: after a mode switch, the system generates personalized guidance prompts based on the user's actions in the original mode; these prompts are presented as an interface overlay or voice broadcast, helping the user quickly understand the operation method of the new mode and the current task progress; the display duration of the prompts adaptively adjusts according to the user's response speed, and they are automatically hidden once the user resumes normal operation.

[0063] The integrated adjustment results form a seamless control path. A seamless control path describes the control flow after intelligent adjustment, including the control node sequence, the adjustment strategy configuration for mode switching nodes, and the expected response time. After intelligent compensation adjustment, the expected control interruption duration is reduced by more than 60%, and the continuity of control response is significantly improved.

[0064] Step S4: Optimize the multimodal collaborative adjustment strategy based on the seamless control path to generate a modal linkage control scheme; intelligently orchestrate the functional control units based on the modal linkage control scheme to obtain a personalized control process.

[0065] To further improve the efficiency of multimodal collaborative control, it is necessary to optimize the collaborative methods of each mode and realize the intelligent orchestration of functional control units.

[0066] Key control nodes are extracted based on the seamless control path, and the dependencies between nodes are analyzed. The method for extracting key control nodes is as follows: control nodes with longer user dwell time (more than 5 seconds) and higher operation frequency (more than 3 times) are identified from the seamless control path and marked as key control nodes; the execution order and data dependencies between key control nodes are analyzed to construct a control node dependency graph.

[0067] Modal cooperative timing planning is performed based on dependencies to determine the optimal execution order. The method of modal cooperative timing planning is as follows: topological sorting is performed on the control node dependency graph to obtain the execution sequence of nodes that satisfy the dependency constraints; for nodes without dependencies, parallel execution optimization can be performed; considering the node execution time and resource consumption, the optimal execution order is determined to minimize the overall control completion time.

[0068] Capability assessments were performed on each modal functional control unit to identify complementary and overlapping functional areas. The assessment method involved: establishing a capability matrix for each functional control unit, with the matrix representing the functional control units and listing them as functional capability dimensions (e.g., image processing capability, data computation capability, user interaction capability); evaluating the performance of each functional control unit across these dimensions and filling the capability matrix; comparing the capability vectors of different functional control units to identify complementary unit pairs (where one unit performs weakly in a certain dimension while the other performs strongly) and overlapping unit pairs (where both units perform strongly in a certain dimension).

[0069] The collaborative enhancement adjustment strategy is designed based on complementary functional points, and the resource sharing adjustment strategy is designed based on overlapping functional areas. The design method for the collaborative enhancement adjustment strategy is as follows: For functionally complementary unit pairs, a task allocation mechanism is designed to assign each sub-stage of the task to the functional control unit with the best matching capabilities, achieving capability complementarity. For example, the AR display control unit excels at image rendering, while the AI ​​computing control unit excels at data analysis. For the task of "recommending suitable sizes based on user body shape and demonstrating the trial effect," the data analysis stage is assigned to the AI ​​computing control unit, and the effect rendering stage is assigned to the AR display control unit. The design method for the resource sharing adjustment strategy is as follows: For functionally overlapping unit pairs, a resource reuse mechanism is designed to avoid redundant calculations and storage. For example, when both the AR display control unit and the AI ​​computing control unit need to acquire 3D model data of the product, the unit executing first loads the data and stores it in a shared cache, while the unit executing later reads directly from the shared cache.

[0070] By integrating collaborative enhancement regulation strategies and resource sharing regulation strategies, a modal linkage control scheme is generated. The modal linkage control scheme includes: the task division definition of each modality, data sharing rules between modalities, the timing arrangement of modal collaborative execution, and contingency plans for handling anomalies.

[0071] Based on the modal linkage control scheme, functional control units are reorganized to create dynamic control combinations. The method for creating dynamic control combinations is as follows: according to the requirements of the current control task, the required functional control units are selected from each mode and combined to form a set of dedicated control units for the current task; the members of the dynamic control combination can be dynamically adjusted according to the progress of the task, realizing on-demand loading and releasing.

[0072] Personalized adjustments are made to dynamic control combinations based on control state feature maps to generate personalized control flows. The personalized adjustment method is as follows: extract the current user's control preference information from the control state feature map, including preferred functions, operating habits, and interface style; adjust the parameters of the dynamic control combination according to these preferences, including default function configurations, shortcut operation settings, and interface layout adjustments; and integrate the adjusted dynamic control combination and its configuration parameters to form a personalized control flow.

[0073] Step S5: Execute the personalized control process to achieve the coordinated operation of AR display control, AI calculation control and customized processing control functions; and monitor and dynamically adjust the coordinated operation effect in real time.

[0074] The personalized control process is executed to activate the AR display control unit, AI computing control unit, and customized processing control unit. The activation method is as follows: each functional control unit is initialized sequentially according to the startup order defined in the personalized control process; personalized configuration parameters of each unit are loaded; and communication connections between each unit and the system control center are established, putting each unit into a ready state.

[0075] A real-time communication mechanism between units is established to ensure the immediate sharing of control signals and status. The method for establishing this mechanism is as follows: a message bus based on a publish-subscribe model is created, with each functional control unit acting as a publisher and subscriber to the message bus; a unified message format specification is defined, including message type, message content, sender identifier, and timestamp; and a priority queue for message transmission is set up to ensure that high-priority messages are delivered first. In this embodiment, the message transmission delay is controlled within 20 milliseconds to meet the requirements of real-time communication.

[0076] Monitor the switching behavior between units and dynamically adjust the unit response priority. The method for monitoring switching behavior is as follows: record the time, direction, and frequency of user switching between various functional control units; analyze the switching behavior data to identify the user's current main focus of operation; dynamically adjust the response priority of each unit according to the focus of operation, setting the unit currently of user attention as the highest priority to ensure that the unit receives the most system resources and the fastest response speed.

[0077] Performance metrics during collaborative operation are collected, including response latency, signal synchronization rate, and functional seamlessness. The response latency is collected by recording the time the user initiates a control operation and the time the system completes the response, calculating the difference as the response latency; in this embodiment, the target value for response latency is within 500 milliseconds. The signal synchronization rate is collected by counting the expected number of control signals transmitted and the actual number of successfully transmitted control signals, calculating the ratio as the signal synchronization rate; in this embodiment, the target value for the signal synchronization rate is above 99%. The functional seamlessness is collected by evaluating the user experience at mode switching points, calculating a functional seamlessness score based on the continuity of user operations and waiting time; in this embodiment, the target value for functional seamlessness is above 0.9 (out of 1).

[0078] Bottleneck identification is based on performance metrics to pinpoint key factors affecting control consistency. The bottleneck identification method involves: statistically analyzing collected performance metrics to identify time periods and functional modules exhibiting abnormal metrics; deeply analyzing the causes of these abnormalities, including hardware resource bottlenecks, software processing bottlenecks, and communication transmission bottlenecks; and prioritizing the identified bottlenecks by their degree of impact to determine the key factors requiring optimization.

[0079] Optimization measures are implemented to address the identified key factors, including caching strategy adjustment, parallel processing optimization, and interface transition smoothing. The caching strategy adjustment method involves dynamically adjusting the scope and timing of pre-loaded data based on user behavior prediction results to improve cache hit rate; in this embodiment, the target cache hit rate is above 85%. The parallel processing optimization method involves enabling multi-threaded parallel processing for control tasks with no dependencies to fully utilize system computing resources; in this embodiment, the parallelism is set to 4, meaning a maximum of 4 independent control tasks can be executed simultaneously. The interface transition smoothing method involves adding transition animation effects during modal switching, with the animation duration set to 300 milliseconds. This effectively masks the brief delay during switching without significantly affecting control efficiency.

[0080] Continuous monitoring of optimization effects forms a closed-loop control mechanism. The implementation method of the closed-loop control mechanism is as follows: the effects of optimization measures are fed back to the performance monitoring module to evaluate the effectiveness of the optimization measures; the parameter configuration of the optimization strategy is dynamically adjusted based on the evaluation results; for optimization measures with insignificant effects, a secondary optimization process is triggered to try other optimization schemes; a continuous improvement cycle of "monitoring-analysis-optimization-evaluation" is formed to continuously improve the overall performance of the multimodal collaborative control system.

[0081] It should be noted that the parameter settings in this embodiment (such as time window length, sampling interval, probability threshold, etc.) are all empirical values ​​determined based on actual application scenarios and performance tests. Implementers can make adaptive adjustments according to the specific hardware configuration of the jewelry smart terminal and the characteristics of the user group.

[0082] This embodiment establishes a multimodal collaborative control architecture and a closed-loop feedback adjustment mechanism, realizing intelligent linkage control and seamless state connection between functional modules such as AR display control, AI computing control, and customized processing control. It effectively solves the problems of control flow interruption and control process fragmentation in the multimodal collaborative control of existing jewelry smart terminals, and significantly improves the overall stability, control efficiency, and response consistency of the jewelry smart terminal control system. Example 2

[0083] Please see Figure 2 As shown, parts not described in detail in this embodiment are described in Embodiment 1. A control system for a jewelry smart terminal mode is provided, including:

[0084] Data parsing module: Collects multi-dimensional control signal datasets from various control units in the jewelry smart terminal, analyzes the control intent of the multi-dimensional control signal datasets, and obtains a control state feature map;

[0085] Modality mapping module: Based on the control state feature map, it predicts mode switching, identifies potential mode transition nodes, and constructs a mode transition probability matrix; based on the mode transition probability matrix, it generates cross-modal control signal mapping rules to obtain the inter-modal control signal transmission channel;

[0086] State assessment module: Real-time state synchronization is performed through the inter-modal control signal transmission channel to obtain a consistent view of the multimodal control state; based on the consistent view of the multimodal control state, the continuity of control response is assessed, control interruption nodes are identified, and intelligent adjustment is performed on the control interruption nodes to form a seamless control path;

[0087] Process formulation module: Optimizes multimodal collaborative adjustment strategies based on seamless control paths to generate modal linkage control schemes; intelligently orchestrates functional control units based on modal linkage control schemes to obtain personalized control processes;

[0088] Process Implementation Module: Executes personalized control processes to achieve coordinated operation of AR display control, AI computation control, and customized processing control functions; monitors and dynamically adjusts the coordinated operation effect in real time.

[0089] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

[0090] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0091] In the description of this invention, it should be understood that the terms "first," "second," etc., are used only for distinguishing descriptions and should not be construed as indicating or implying relative importance.

[0092] In the description of this invention, unless otherwise stated, "a plurality of" means two or more.

[0093] In the description of this invention, "several" means one or more, and "a large number" means two or more.

[0094] The terms "example," "specific example," or "some examples," etc., refer to specific features, structures, materials, or characteristics described in connection with that embodiment or example, which are included in at least one embodiment or example of the present invention. In this specification, illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0095] All formulas in this manual are dimensionless and calculated numerically. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters and thresholds in the formulas are set by those skilled in the art according to the actual situation.

[0096] Although embodiments of the invention have been shown and described, those skilled in the art will understand 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 claims and their equivalents.

Claims

1. A control method of a jewelry smart terminal mode, characterized in that, include: Step S1: Collect multi-dimensional control signal datasets from each control unit in the smart jewelry terminal, analyze the control intent of the multi-dimensional control signal datasets, and obtain a control state feature map; Step S2: Based on the control state feature map, predict mode switching, identify potential mode transition nodes, and construct a mode transition probability matrix; generate cross-mode control signal mapping rules based on the mode transition probability matrix to obtain the inter-mode control signal transmission channel; Step S3: Real-time state synchronization is performed through the inter-modal control signal transmission channel to obtain a multimodal control state consistency view; based on the multimodal control state consistency view, the continuity of control response is evaluated, control interruption nodes are identified, and intelligent adjustment processing is performed on the control interruption nodes to form a seamless control path; Step S4: Optimize the multimodal collaborative adjustment strategy based on the seamless control path to generate a modal linkage control scheme; Intelligent orchestration of functional control units based on modal linkage control scheme to obtain personalized control flow; Step S5: Execute the personalized control process to achieve the coordinated operation of AR display control, AI calculation control, and customized processing control functions; and monitor and dynamically adjust the coordinated operation effect in real time. Step S2 includes: performing control trajectory analysis based on the control state feature map to identify control transfer patterns and obtain control migration paths; calculating modal correlation of the control migration paths to determine the conversion tendency between modes; constructing a mode switching predictor based on the conversion tendency to evaluate the possibility of mode switching in real time and obtain a mode switching probability distribution; identifying high-probability conversion nodes based on the mode switching probability distribution to form a potential mode switching node set; extracting conversion conditions from the potential mode switching node set to generate mode switching trigger rules; constructing a mode switching probability matrix based on the mode switching trigger rules and the mode switching probability distribution; and performing control signal dependency analysis based on the mode switching probability matrix to identify cross-modal control signal requirements, generate a control signal mapping rule base, and establish an inter-modal control signal transmission channel. The process of performing control signal dependency analysis based on the mode transition probability matrix, identifying cross-modal control signal requirements, generating a control signal mapping rule base, and establishing inter-modal control signal transmission channels includes: This process involves analyzing high-frequency transition paths in the mode transition probability matrix to identify core mode transition scenarios; tracing control signal flow for these scenarios to identify source mode output signal types and target mode input signal requirements; semantically aligning signal types to establish a cross-modal control signal dictionary; defining signal format conversion rules based on the dictionary to ensure signal compatibility; constructing a signal priority evaluation mechanism to classify the importance of transmitted signals; generating a control signal mapping rule base based on signal priority and transition rules; and establishing inter-modal control signal transmission channels through the control signal mapping rule base.

2. The control method for the smart terminal mode of jewelry according to claim 1, characterized in that, Step S1 includes: The system collects touch signal trajectories, gaze focus signals, voice command signals, and gesture signals from each control unit in a smart jewelry terminal to form a multidimensional control signal dataset. Temporal correlation analysis is performed on this dataset to obtain control sequence patterns. Control intent is identified based on these patterns, extracting explicit control demand features and implicit control preference features. Semantic fusion is then performed on these features to generate control intent vectors. A control state topology is constructed based on these vectors to obtain a control association network. Finally, the control evolution path is deduced through this network to form a control state feature map.

3. The control method for the smart terminal mode of jewelry according to claim 2, characterized in that, The control intent recognition based on control sequence patterns extracts explicit control demand features and implicit control preference features, including: Key signals are extracted from control sequence patterns to identify active control operation signals, resulting in a set of explicit control signals. The frequency and duration of signals in the explicit control signal set are analyzed to determine control focus and generate explicit control demand features. Micro-signal capture is performed on the control sequence patterns to identify unconscious control tendency signals, resulting in a set of implicit control signals. The distribution of signal dwell time and review frequency are analyzed using the implicit control signal set to infer potential control points and extract implicit control preference features. The explicit control demand features and implicit control preference features are weighted and assigned to construct a comprehensive control intent representation.

4. The control method for the smart jewelry terminal mode according to claim 1, characterized in that, Step S3 includes: Real-time signal synchronization is achieved through inter-modal control signal transmission channels to acquire current control state information for each mode. This control state information is then standardized to generate a unified control state descriptor. State consistency verification is performed based on the unified control state descriptor to identify state conflict points. Conflict resolution is then implemented to obtain coordinated control state data. The coordinated control state data is integrated to construct a multimodal control state consistency view. Control operation flow analysis is performed based on this view to identify operation interruption nodes and their waiting durations. Interruption nodes are classified by interruption type to determine the control interruption node set. An intelligent compensation adjustment strategy is designed for the control interruption node set, employing preloading, state inheritance, and intelligent prompts for adjustment processing. The adjustment results are then integrated to form a seamless control path.

5. The control method for the smart terminal mode of jewelry according to claim 4, characterized in that, The intelligent compensation and adjustment strategy design for the control interruption node set includes adjustment processing through preloading, state inheritance, and intelligent prompts, including: The causes of control interruptions are analyzed to identify data-missing interruptions, function-switching interruptions, and cognitive-jumping interruptions. For data-missing interruptions, a pre-loading adjustment strategy is implemented to prepare the required control data in advance before mode switching. For function-switching interruptions, a state inheritance adjustment strategy is implemented to automatically transfer the control results and parameter settings of the previous mode. For cognitive-jumping interruptions, an intelligent prompt adjustment strategy is implemented to generate transitional control guidance signals.

6. The control method for the smart terminal mode of jewelry according to claim 1, characterized in that, Step S4 includes: Key control nodes are extracted based on the seamless control path, and the dependencies between nodes are analyzed. Modal collaborative timing planning is performed based on the dependencies to determine the optimal execution order. Capacity assessment is conducted on each modal functional control unit to identify functional complementarity points and functional overlap areas. Collaborative enhancement regulation strategies are designed based on functional complementarity points, and resource sharing regulation strategies are designed based on functional overlap areas. The collaborative enhancement regulation strategies and resource sharing regulation strategies are integrated to generate a modal linkage control scheme. Functional control units are reorganized according to the modal linkage control scheme to create a dynamic control combination. The dynamic control combination is personalized based on the control state feature map to generate a personalized control process.

7. The control method for the smart terminal mode of jewelry according to claim 6, characterized in that, Step S5 includes: Execute personalized control processes, activating the AR display control unit, AI computing control unit, and customized processing control unit; establish a real-time communication mechanism between units to ensure instant sharing of control signals and status; monitor the switching behavior between units and dynamically adjust unit response priorities; collect performance indicators during collaborative operation, including response latency, signal synchronization rate, and smoothness of function transitions; identify bottlenecks based on performance indicators to pinpoint key factors affecting control consistency; implement optimization and adjustment measures for identified key factors, including caching strategy adjustment, parallel processing optimization, and interface transition smoothing; continuously monitor the optimization effect to form a closed-loop control mechanism.

8. A control system for a jewelry smart terminal mode, used to implement the control method for the jewelry smart terminal mode according to any one of claims 1 to 7, characterized in that, include: Data parsing module: Collects control signals and status data from each control unit in the smart jewelry terminal, analyzes the control signals and status data to obtain a control status feature map; Modality mapping module: Based on the control state feature map, it predicts mode switching, identifies potential mode transition nodes, and constructs a mode transition probability matrix; based on the mode transition probability matrix, it generates cross-modal control signal mapping rules to obtain the inter-modal control signal transmission channel; State assessment module: Real-time state synchronization is performed through the inter-modal control signal transmission channel to obtain a consistent view of the multimodal control state; based on the consistent view of the multimodal control state, the continuity of control response is assessed, control interruption nodes are identified, and intelligent adjustment is performed on the control interruption nodes to form a seamless control path; Process formulation module: Optimizes multimodal collaborative adjustment strategies based on seamless control paths to generate modal linkage control schemes; Intelligent orchestration of functional control units based on modal linkage control scheme to obtain personalized control flow; Process Implementation Module: Executes personalized control processes to achieve coordinated operation of AR display control, AI computation control, and customized processing control functions; monitors and dynamically adjusts the coordinated operation effect in real time.

Images