Star flash linkage control method and system based on three-mode remote controller and intelligent gateway

Abstract

The invention relates to the technical field of remote controllers, in particular to a star flash linkage control method and system based on a three-mode remote controller and an intelligent gateway, and the method comprises the steps: generating a scene triggering instruction based on the interaction operation on the three-mode remote controller; sending the scene triggering instruction to an intelligent gateway through star flash wireless communication, and analyzing the scene triggering instruction based on the intelligent gateway to obtain a logic operation sequence; according to the logic operation sequence and the device network connection state, after a target communication protocol corresponding to the target device is determined, the logic operation sequence is converted into a device control instruction corresponding to the target device according to the target communication protocol; the equipment control instruction is sent to the target equipment through the intelligent gateway to drive the target equipment to execute the linkage control operation corresponding to the scene triggering instruction, and the technical problem that an existing linkage method of the three-mode remote controller and the intelligent gateway cannot effectively guarantee the real-time performance and stability of equipment linkage is effectively solved.

Description

Technical Field

[0001] This invention relates to the field of remote control technology, specifically to a star-flash linkage control method and system based on a three-mode remote control and a smart gateway. Background Technology

[0002] Currently, with the popularization of smart home devices, there are often smart devices from different brands and using different communication protocols in the home environment, such as TVs, lights, curtains, and air conditioners. In order to achieve unified control and scenario-based linkage of these heterogeneous devices, the solution based on centralized scheduling of smart gateways and equipped with three-mode remote controls as the interaction entry point is gradually becoming a trend.

[0003] Existing methods for linking tri-mode remote controls with smart gateways typically rely on preset protocol mapping rules. Upon receiving a remote control command, the smart gateway directly sends control commands to the target device using a fixed communication protocol based on pre-configured device binding relationships. However, because home wireless network environments and device connection states are constantly changing, fixed protocol mapping mechanisms cannot adaptively select the optimal communication path. This leads to decreased reliability and increased response latency when interference or delays occur in specific protocol links, making it difficult to guarantee the real-time performance and stability of command execution in cross-device linkage scenarios, thus impacting user experience. Summary of the Invention

[0004] To address the technical problem that existing methods for linking three-mode remote controllers and smart gateways cannot effectively guarantee the real-time performance and stability of device linkage, this application provides a star-flash linkage control method and system based on a three-mode remote controller and a smart gateway.

[0005] The star-flash linkage control method and system based on a three-mode remote controller and a smart gateway provided in this application adopts the following technical solution: A star-flash linkage control method based on a three-mode remote controller and a smart gateway includes: Generating scene trigger commands based on interactive operations on a three-mode remote control; The scene trigger command is sent to the smart gateway via StarFlash wireless communication. The smart gateway then parses the scene trigger command to obtain the logical operation sequence. Based on the logical operation sequence and the device network connection status, after determining the target communication protocol corresponding to the target device, the logical operation sequence is converted into device control commands corresponding to the target device according to the target communication protocol. The smart gateway sends device control commands to the target device to drive the target device to perform linkage control operations corresponding to the scene trigger commands.

[0006] Furthermore, the steps for generating scene trigger commands based on the interactive operations on the three-mode remote control include: The system captures user-triggered interactive operations using a three-mode remote control and generates raw operation data. Feature extraction is performed on the original operational data to obtain the interaction feature vector; The user's intent is identified by matching the interaction feature vector with a preset intent recognition model. Based on user intent, scene trigger commands are mapped from a preset scene command library.

[0007] Furthermore, the steps for extracting features from the original operational data to obtain the interaction feature vector include: Multimodal data fusion is performed on the original operational data to obtain a multimodal data stream; Spatiotemporal feature analysis is performed on multimodal data streams to extract temporal and spatial features, resulting in multidimensional feature vectors. The multidimensional feature vectors are reduced in dimensionality and standardized to generate interactive feature vectors for intent recognition.

[0008] Furthermore, the steps for parsing scene trigger commands based on the smart gateway to obtain the logical operation sequence include: Semantic parsing of scene trigger commands yields the initial operation intent; Based on the current environmental context information, the initial operation intent is modified to generate an optimized operation intent; Based on the optimization operation intent, the system queries the preset equipment capability knowledge base and generates a logical operation sequence.

[0009] Furthermore, the steps for determining the target communication protocol corresponding to the target device based on the logical operation sequence and the device network connection status include: Extract the device identifier of the target device based on the logical operation sequence; Based on the device identifier, several candidate communication protocols supported by the target device are obtained by querying the preset dynamic protocol mapping table; Obtain the current network status parameters of each candidate communication protocol and their connection status with the target device's network. Based on the connection status of each candidate communication protocol's network, calculate the connection quality score for each candidate communication protocol. Based on the preset protocol decision-making strategy and the latency sensitivity of the scenario triggering command, the quality scores of each connection are weighted and calculated to obtain the selection degree of each candidate communication protocol. Based on the selection criteria, the target communication protocol is selected from the candidate communication protocols.

[0010] Furthermore, the steps of converting the logical operation sequence into device control instructions corresponding to the target device, based on the target communication protocol, include: Invoke the protocol abstraction layer corresponding to the target communication protocol, and convert the logical operation sequence into device operation instructions based on the protocol abstraction layer; Based on the device identifier of the target device, the device instruction template that matches the device operation instruction is retrieved from the preset instruction template library; Fill the device operation instructions with the instruction parameters of each instruction into the device instruction template to generate device control instructions.

[0011] This application also provides a star-flash linkage control system based on a three-mode remote controller and a smart gateway, including: The data generation module is used to generate scene trigger commands based on the interactive operations on the three-mode remote controller; The data parsing module is used to send scene triggering commands to the smart gateway via StarFlash wireless communication, and to parse the scene triggering commands based on the smart gateway to obtain the logical operation sequence; The data conversion module is used to determine the target communication protocol corresponding to the target device based on the logical operation sequence and the device network connection status, and then convert the logical operation sequence into device control commands corresponding to the target device according to the target communication protocol. The device driver module is used to send device control commands to the target device through the smart gateway, so as to drive the target device to perform linkage control operations corresponding to the scene trigger commands.

[0012] Beneficial effects achieved: This application provides a StarFlash linkage control method based on a three-mode remote controller and a smart gateway, comprising: generating a scene trigger command based on interactive operations on the three-mode remote controller; sending the scene trigger command to the smart gateway via StarFlash wireless communication; parsing the scene trigger command on the smart gateway to obtain a logical operation sequence; determining the target communication protocol corresponding to the target device based on the logical operation sequence and the device network connection status; converting the logical operation sequence into a device control command corresponding to the target device according to the target communication protocol; and sending the device control command to the target device via the smart gateway to drive the target device to execute the linkage control operation corresponding to the scene trigger command.

[0013] In this application, firstly, a scene trigger command is generated based on the interactive operation on the three-mode remote controller, ensuring accurate capture of the user's intent and providing clear input for subsequent control. Next, the scene trigger command is sent to the smart gateway via StarSpark wireless communication, leveraging its low latency and high reliability to guarantee the real-time performance and stability of the scene trigger command transmission at the transmission level. Then, the smart gateway parses the scene trigger command to obtain a logical operation sequence, providing a structured basis for communication protocol decision-making. Finally, based on the logical operation sequence and the device network connection status, the target communication protocol corresponding to the target device is determined by dynamically evaluating network conditions in real time. By selecting the optimal communication protocol, delays or interruptions caused by static protocol mapping when the link changes are avoided, thus directly improving the real-time adaptability of the protocol and the stability of the link. Then, the logical operation sequence is converted into device control commands corresponding to the target device according to the target communication protocol, ensuring that the command format accurately matches the dynamically selected protocol and reducing the risk of transmission errors and retransmissions. Finally, the device control commands are sent to the target device through the intelligent gateway, and the optimized protocol path drives the device to perform linkage control operations. Overall, the adaptive protocol selection mechanism significantly reduces the command transmission latency and packet loss probability, thereby systematically enhancing the real-time performance and stability of device linkage. Attached Figure Description

[0014] Figure 1 This is a flowchart illustrating the steps of a star-flash linkage control method based on a three-mode remote controller and a smart gateway according to this application; Figure 2 A flowchart illustrating the steps involved in generating scenario triggering instructions for this application; Figure 3 A flowchart illustrating the steps involved in generating the logical operation sequence for this application; Figure 4 A flowchart illustrating the steps involved in generating device control instructions for this application; Figure 5 This is a schematic diagram of the Star Flash Linkage Control System based on a three-mode remote controller and a smart gateway, as described in this application.

[0015] Explanation of reference numerals in the attached figures: 10. Data generation module; 20. Data parsing module; 30. Data conversion module; 40. Device driver module. Detailed Implementation

[0016] The following combination Figures 1-5 This application will be described in further detail.

[0017] 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 a part of the embodiments of the present invention, and not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0018] It should be noted that if the embodiments of the present invention involve directional indications (such as up, down, left, right, front, back, etc.), the directional indications are only used to explain the relative positional relationship and movement of the components in a specific posture. If the specific posture changes, the directional indications will also change accordingly.

[0019] Furthermore, if the embodiments of this invention involve descriptions such as "first" or "second," these descriptions are for descriptive purposes only and should not be construed as indicating or implying their relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined with "first" or "second" may explicitly or implicitly include at least one of those features. Additionally, the use of "and / or" or "and / or" throughout the text includes three parallel solutions. For example, "A and / or B" includes solution A, solution B, or a solution where both A and B are satisfied simultaneously. Furthermore, the technical solutions of the various embodiments can be combined with each other, but this must be based on the ability of those skilled in the art to implement them. When the combination of technical solutions is contradictory or impossible to implement, it should be considered that such a combination of technical solutions does not exist and is not within the scope of protection claimed by this invention.

[0020] This application discloses a star-flash linkage control method based on a three-mode remote controller and a smart gateway.

[0021] Please refer to Figure 1 The proposed star-flash linkage control method based on a three-mode remote controller and a smart gateway in this embodiment includes steps S10 to S40: Step S10: Generate scene trigger commands based on the interactive operations on the three-mode remote control.

[0022] This step captures user input via button presses, voice commands, or gestures using a three-mode remote control, and generates them as scene trigger commands. This provides an initial standard trigger signal for the entire linkage control process, ensuring accurate transmission of user intent and reliability of subsequent processing. It enables users to efficiently trigger scenes through simple and intuitive operations, improving the system's usability and responsiveness, and laying the foundation for the smooth execution of subsequent steps.

[0023] In this embodiment, the three-mode remote control is a remote control device that integrates three wireless communication modes: infrared communication, Bluetooth communication, and Star Flash wireless communication. Its core design purpose is to achieve universal compatibility and intelligent control through diversified communication capabilities. The infrared communication mode ensures backward compatibility with traditional non-smart home appliances, the Bluetooth communication mode provides universal connectivity with mainstream smart devices, and the Star Flash wireless communication mode provides optimal transmission guarantee for high-level linkage scenarios with its ultra-low latency and high reliability. This combination enables the three-mode remote control to serve as a unified control entry point, seamlessly operating various devices with different protocols, brands, and eras. It fundamentally solves the pain points of multiple remote controls coexisting and fragmented cross-device control, laying the physical foundation for centralized scheduling and scenario-based linkage of smart homes.

[0024] Step S20: The scene trigger command is sent to the smart gateway via StarFlash wireless communication. The smart gateway parses the scene trigger command to obtain the logical operation sequence.

[0025] First, the scene trigger command is sent to the smart gateway via StarSpark wireless communication. Leveraging the low latency and high reliability of StarSpark wireless communication, the real-time performance and stability of the scene trigger command transmission process are ensured. Then, the smart gateway parses the received scene trigger command, thereby breaking down the user's scene intent and converting it into operational steps that the device can understand and execute, i.e., a logical operation sequence. This realizes the transformation of user commands from abstract intent to specific operational logic, laying the foundation for subsequent cross-communication protocol and cross-device scheduling, thus ensuring the accuracy of control logic and the integrity of system response in linked scenarios.

[0026] In this embodiment, the smart gateway is the control center and protocol conversion core of the system. It is an embedded hardware device with multi-mode communication capabilities such as Star Flash wireless communication, Bluetooth communication, and infrared communication. It acts as a relay station between the three-mode remote control and various smart home appliances. By receiving scene trigger commands sent by the three-mode remote control via Star Flash wireless communication, and performing parsing, decision-making, and protocol conversion on the smart gateway, the unified scene trigger commands are converted into logical operation sequences that can be recognized by different smart home appliances. This enables centralized and unified scheduling and scene-based linkage control of heterogeneous devices across brands and protocols, solving the problem of isolated and uncoordinated devices in the smart home ecosystem.

[0027] Step S30: After determining the target communication protocol corresponding to the target device based on the logical operation sequence and the device network connection status, the logical operation sequence is converted into device control commands corresponding to the target device according to the target communication protocol.

[0028] In this step, the target communication protocol corresponding to the target device is first determined based on the logical operation sequence and the device network connection status, breaking through the limitations of static protocol mapping. By evaluating network conditions in real time, the most stable and efficient communication protocol is dynamically selected for each logical operation sequence. Then, the logical operation sequence is converted into device control commands according to the determined target communication protocol. This conversion process ensures that the logical operation sequence is accurately converted into specific commands that conform to the target device communication protocol, realizing real-time optimization of the transmission path of the device control command sequence and adaptation of the command content. This significantly improves the reliability of device control command transmission, reduces response latency, and ultimately ensures the overall stability and real-time performance of multi-device linkage in complex home environments.

[0029] Step S40: Send device control commands to the target device through the smart gateway to drive the target device to perform linkage control operations corresponding to the scene trigger commands.

[0030] The intelligent gateway sends the generated device control commands to the target device. As the execution terminal of the entire control process, the intelligent gateway reliably distributes the control commands, which have been optimized by the previous steps, to the corresponding target device through the communication path corresponding to the target communication protocol.

[0031] Specifically, the smart gateway calls the driver interface corresponding to the target communication protocol, encapsulates the device control commands into data frames that conform to the target device's communication protocol, and sends them out through the corresponding wireless or wired physical link. When the target device successfully receives and parses these device control commands, it will execute the specific operations defined by the device control commands, so as to accurately and reliably transform the logical commands that have been optimized and processed in the previous steps into the actual linkage actions of cross-brand and cross-protocol smart devices, and ultimately ensure that all target devices can complete precise collaborative control according to the requirements of the scenario trigger commands.

[0032] Additionally, it should be noted that this application can also construct an end-to-end system of a StarSpark tri-mode remote control + a StarSpark module on the target device. Taking a TV terminal as an example, the tri-mode remote control is equipped with a StarSpark SLE / SLP dual-protocol chip and a dedicated live channel switching button. The TV terminal has a built-in StarSpark Bluetooth Combo module, which improves signal reception sensitivity to -97dBm through a self-developed butterfly antenna array, ensuring seamless communication with the tri-mode remote control within a 10-meter range. It should be noted that the tri-mode remote control has an independent "live one-click switch" physical button, and the tri-mode remote control has a unique instruction code corresponding to this physical button, thereby avoiding conflicts with other function buttons.

[0033] The specific process is as follows: After the user presses the "One-Click Switch to Live Broadcast" physical button, the tri-mode remote control will directly generate a device operation command to "Switch to Live Broadcast Channel" based on this interaction, and then use its set StarSpark SLE protocol at 20 The device operation commands are sent to the TV terminal and set-top box with a latency of 99.9%. Polar channel coding and adaptive frequency modulation technology are used to resist interference, ensuring that the device operation commands are successfully delivered at a rate of 99.9%, effectively avoiding abnormal display phenomena such as black screens and stuttering.

[0034] After receiving the device operation command, the StarSpark Bluetooth Combo module on the TV terminal will call its built-in "APP-Live Signal Mapping Table" to automatically match the ASI bitstream or digital signal source of the corresponding CCTV channel. This enables the automatic matching of the channel interface type without the need for manual selection by the user. The set-top box will then send an "Activate / Switch" command to output the signal of the corresponding channel, realizing the linkage switching between the TV terminal and the set-top box.

[0035] In one feasible implementation, refer to Figure 2 As shown, step S10 may specifically include steps S11 to S14: Step S11: Capture user-triggered interactive operations using the three-mode remote control to generate raw operation data.

[0036] The tri-mode remote control captures user interactions by utilizing multiple built-in sensors and interfaces. Specifically, when a user presses a button, inputs voice, or performs a specific gesture, the remote control simultaneously acquires the corresponding button electrical signals, audio sampling data, and motion sensor readings through the button circuit, microphone array, and inertial measurement unit. Using the built-in system clock as a time reference, a hardware interrupt mechanism is immediately triggered when any sensor in the button circuit, microphone array, or inertial measurement unit generates data. The microcontroller's interrupt service routine captures the current system clock count as a global timestamp and encapsulates this global timestamp along with the corresponding button electrical signals, audio sampling data, and / or motion sensor readings into a data packet. Subsequently, the data packet is stored in a pre-divided multimodal circular buffer using direct memory access technology, thereby generating original operation data that maintains time alignment. This process completely transforms various forms of user interaction intentions into machine-readable data, providing an input source for subsequent feature extraction and intention recognition, and ensuring the data foundation quality of the system's interaction entry point.

[0037] Step S12: Extract features from the original operation data to obtain the interaction feature vector.

[0038] By extracting features from the raw operational data, an interaction feature vector is obtained. This extracts the interaction feature vector that can characterize the essence of user interaction from the raw operational data, which may contain noise and redundant information. This provides input for the subsequent preset intent recognition model, thereby improving the accuracy and robustness of the system in recognizing user intent. In turn, it ensures that the triggering point of the entire linkage control process is more accurate and reliable.

[0039] Furthermore, step S12 may include steps S121 to S123: Step S121: Perform multimodal data fusion on the original operation data to obtain a multimodal data stream.

[0040] By processing the raw operation data generated by the tri-mode remote controller, the raw operation data collected by the tri-mode sensors is first aligned on the time axis and divided into uniform processing frames based on the global timestamp in the raw operation data. Then, the raw operation data in each frame is standardized in format. For example, ① for audio sampling data, the signal after pre-emphasis, framing, and windowing is subjected to Fast Fourier Transform to obtain the spectrum, then filtered by a Mel filter bank and the logarithmic energy is taken. Finally, the dimension-reduced Mel spectrum coefficient vector is extracted by Discrete Cosine Transform; ② for motion sensor readings, the raw values ​​of the triaxial gyroscope and accelerometer are subjected to coordinate transformation and a complementary filter is used to remove noise and drift. Then, the corresponding values ​​are calculated by quaternion operations. Euler angle vectors such as roll angle, pitch angle, and yaw angle; ③ For button electrical signals, the button electrical signals captured within the current frame time are mapped to a fixed-length binary vector, where only the index position of the button corresponding to the button electrical signal is set to 1 and the rest are set to 0, forming a one-hot encoded vector; Finally, the Mel spectrum coefficient vector, Euler angle vector, and one-hot encoded vector obtained from the above processing are concatenated in the feature dimension to generate a multimodal feature vector corresponding to each time frame, thereby forming a continuous multimodal data stream, realizing the transformation of raw operational data with different physical meanings and dimensions into a multimodal data stream with a unified time base and comparable feature representation, providing a structurally consistent and informationally complete input for subsequent spatiotemporal feature analysis.

[0041] Step S122: Perform spatiotemporal feature analysis on the multimodal data stream, extract temporal and spatial features, and obtain multidimensional feature vectors.

[0042] By employing a sliding time window to traverse the multimodal data stream, ① for the Mel spectrum coefficient vectors of consecutive frames within the sliding time window, the Mel spectrum coefficient vectors corresponding to each frame are first arranged into an input sequence according to their chronological order. Then, this sequence is input into a pre-trained Long Short-Term Memory (LSTM) network model. The LSM network model processes each Mel spectrum coefficient vector in the sequence step by step of the sliding time window through its internal gating mechanisms such as input gates, forget gates, and output gates. After receiving the Mel spectrum coefficient vector of a frame corresponding to that time step at each time step, the LSM network model calculates the Mel spectrum coefficient vector of this frame together with the hidden state from the previous time step through its internal gating structure: deciding which old information to forget from the cell state of the previous time step → deciding which new information from the current Mel spectrum coefficients to store in the cell state → outputting the hidden state of the current time step based on the updated cell state. By updating the cell state and generating a new hidden state at each time step based on the current Mel spectral coefficient vector and the hidden state of the previous time step, the long-term dependence and short-term dynamic change trend of audio features in the time dimension are effectively captured. Finally, the hidden state is extracted from the last time step of the Long Short-Term Memory Network model. This hidden state encodes the dynamic evolution information of the Mel spectral coefficient vector within the entire sliding time window and serves as a temporal feature representing the dynamic change pattern of audio features.

[0043] ② For the combination of Euler angle vector and one-hot encoded vector within the same sliding time window, each frame of data within the sliding time window is first regarded as a graph node. The feature vector of the graph node is formed by concatenating the Euler angle vector and one-hot encoded vector of the frame data. The edges between graph nodes are predefined according to temporal proximity and modal correlation to construct an initial graph structure. The constructed graph data is input into a graph attention network. The graph attention network calculates the attention coefficient between each pair of connected graph nodes through a learnable attention mechanism, thereby dynamically measuring the correlation strength between different graph nodes, i.e., gesture postures and key events at different times. Then, the graph attention network performs weighted aggregation based on these attention coefficients and the features of adjacent graph nodes, updating the node features of each graph node so that the node features of each graph node incorporate information from other graph nodes related to it. Finally, all the updated node features are aggregated into a unified graph-level feature vector through a global pooling operation. This graph-level feature vector encodes the spatial correlation between gesture postures and key events within the entire sliding time window, and this spatial correlation is used as a spatial feature.

[0044] Finally, the extracted temporal and spatial features are fused and spliced ​​to form a multi-dimensional feature vector that can comprehensively represent the user's interaction intent. This enables the extraction of discriminative dynamic behavior patterns and static correlation information from multimodal data streams, providing a key information foundation for the subsequent generation of interaction feature vectors.

[0045] In this context, the cell state is like a conveyor belt in a long short-term memory network model. It runs through the entire time series and is responsible for transmitting selected long-term information between time steps. For example, in audio sampling data, the 261.6 Hz sound component of middle C exists stably for a period of time, and its corresponding Mel spectrum coefficient vector will show a stable high value in multiple consecutive time steps. The update of this cell state is controlled by the forget gate and the input gate.

[0046] The hidden state is the output calculated by the output gate based on the current cell state and the Mel spectrum coefficient vector input at the current time step. It contains a sequence information summary that the long short-term memory network model considers useful for subsequent calculations up to the current time step, and serves as part of the input for the next time step.

[0047] Step S123: Perform dimensionality reduction and standardization on the multidimensional feature vector to generate an interactive feature vector for intent recognition.

[0048] ① When performing dimensionality reduction on a multidimensional eigenvector, it is multiplied by the dimensionality reduction matrix of a pre-trained principal component analysis model. The mean vector of the principal component analysis model is subtracted from the multidimensional eigenvector to achieve centering. Then, the centered multidimensional eigenvector is multiplied by the dimensionality reduction matrix composed of the first k principal component eigenvectors retained by the principal component analysis model. This linearly projects the multidimensional eigenvector into the low-dimensional subspace spanned by these k principal component eigenvectors, ultimately obtaining a k-dimensional eigenvector that retains the most important variance information.

[0049] ② When standardizing the k-dimensional reduced feature vector, the k-dimensional reduced feature vector is input into a standardization processor based on the Z-score method. By subtracting a k-dimensional mean vector that is pre-calculated and stored from a large amount of data during the principal component analysis model training phase, and then dividing element by element by the same pre-stored k-dimensional standard deviation vector, each dimension of the k-dimensional reduced feature vector is centered and scaled, so that the numerical distribution of each dimension is transformed into a standard normal distribution with a mean of 0 and a standard deviation of 1. Finally, an interactive feature vector with uniform scale, elimination of dimensional differences and normal distribution is generated.

[0050] This step enables the compression and normalization of high-dimensional multidimensional feature vectors into a compact representation that is more suitable for processing and matching by the preset intent recognition model, significantly improving the system's ability to accurately and in real time determine user intent.

[0051] Step S13: Match the interaction feature vector with the preset intent recognition model to identify the user intent.

[0052] It should be noted that the preset intent recognition model is a multi-class neural network pre-trained with a large number of labeled interaction feature vectors. The specific matching process is as follows: when the interaction feature vector is fed into the preset intent recognition model as an input vector, the input vector is first linearly transformed with the weight matrix of the input layer and a bias vector is added. Then, the result is passed to the subsequent fully connected layers and mapped by non-linear activation functions. Finally, at the output layer, the Softmax function is used to normalize the linear output of the last fully connected layer to calculate the exponent, thereby obtaining the probability distribution of the input vector corresponding to each preset intent category. The category with the highest probability is determined as the currently identified user intent, realizing semantic understanding of user interaction operations and mapping low-level features to high-level intents. This provides a key basis for subsequently mapping the corresponding scene trigger commands from the preset scene command library.

[0053] Step S14: Based on the user's intent, map the scene trigger command from the preset scene command library.

[0054] It should be noted that the preset scene instruction library is a lookup table that stores the mapping relationship between multiple user intentions and corresponding scene trigger instructions. By using the user intention as the index key, the mapping relationship defined in the preset scene instruction library can be directly queried to retrieve the corresponding executable scene trigger instruction. This realizes the transformation of semantic user intentions into scene trigger instructions that the system can parse and execute, providing clear and reliable instruction input for subsequent intelligent linkage control processes.

[0055] In one feasible implementation, refer to Figure 3 As shown, step S20 may specifically include steps S21 to S23: Step S21: Semantic parsing of the scene trigger command to obtain the initial operation intent.

[0056] The scenario-triggered command is taken as text input and encoded and understood using a pre-defined natural language processing model. Key semantic components such as core operation verbs, target devices, and modifier parameters are extracted from the scenario-triggered command. These semantic components are then structurally combined into a machine-readable initial operation intent, such as a data structure containing fields for "operation type," "target object," and "operation parameters." This process transforms text commands from users or a pre-defined scenario command library, which may have diverse natural language characteristics, into a standardized and unambiguous expression of operation intent. This lays the foundation for subsequent context-based intent correction and logical operation sequence matching.

[0057] Among them, the preset natural language processing model refers to a machine learning model that has been trained on a large-scale corpus and is used to parse scene trigger commands for smart homes. For example, the BERT model based on the Transformer architecture is used to automatically understand the semantics of scene trigger commands.

[0058] Step S22: Based on the current environment context information, the initial operation intention is modified to generate an optimized operation intention.

[0059] By acquiring and integrating data from environmental sensors in real time, such as current time, indoor temperature and humidity, light intensity, and the presence of people, the initial operation intent is modified based on predefined contextual logic rules. Specifically, the operation parameters (such as target temperature and light brightness) contained in the initial operation intent are compared and calculated with the current environmental context information. If the current environmental context information already meets or exceeds the initial operation intent, the operation of that step is canceled, and the initial operation intent is directly used as the optimized operation intent to execute step S23. If there is a potential conflict between the current environmental context information and the initial operation intent, such as turning on the lights during the day but with sufficient light, the operation parameters in the initial operation intent are adjusted or operation information representing secondary confirmation is output to the user. Finally, an optimized operation intent that takes into account real-time environmental factors and is more executable and reasonable is generated, thereby ensuring that the subsequent generated logical operation sequence not only meets the user's fundamental expectations but also fits the actual situation, effectively improving the accuracy of control and user experience.

[0060] Among them, the predefined context logic rules refer to a set of "condition-action" judgment rules that are pre-set in the design phase based on domain knowledge or user habits. These rules are used to automatically correct and optimize the initial operation intention based on the current environmental context information.

[0061] Step S23: Query the preset equipment capability knowledge base according to the optimization operation intention to generate a logical operation sequence.

[0062] It should be noted that the preset equipment capability knowledge base is a structured database that stores the functional descriptions, status interfaces, performance parameters and constraints of all controllable devices.

[0063] By using the optimized operation intent as the query input, a logical operation sequence is generated through retrieval and matching in a preset device capability knowledge base. Specifically, based on the operation type specified in the optimized operation intent (i.e., the core action to be performed, such as turning on or adjusting) and the target object (i.e., the device affected by the core action, such as the main living room light or air conditioner), the preset device capability knowledge base is matched to find target devices with corresponding capabilities and currently reachable states. Then, based on the device operation protocols, parameter ranges, and mutual exclusion or dependency relationships between target devices defined in the preset device capability knowledge base, the optimized operation intent is decomposed and arranged into one or more device control instructions. Finally, the executability of the device control instructions under physical conditions is verified and necessary status check instructions are added, thus forming a complete logical operation sequence. This transforms the context-optimized optimized operation intent into a logical operation sequence that the underlying physical devices can understand and execute, serving as a key bridge connecting intelligent decision-making with specific device linkage.

[0064] The logical operation sequence includes one or more operation entries arranged in a specific order, such as setting the air conditioner temperature to 25 degrees Celsius, as well as necessary equipment status check instructions inserted to ensure the reliability of equipment control commands, such as querying the current status of the air conditioner.

[0065] In one feasible implementation, refer to Figure 4 As shown, step S30 may specifically include steps S31 to S38: Step S31: Extract the device identifier of the target device according to the logical operation sequence.

[0066] By parsing the logical operation sequence, the target device field defined in each operation entry of the logical operation sequence is directly accessed, thereby reading and extracting the string or code of the device identifier used to uniquely identify the target device. This enables accurate binding of business logic with the target device entity, providing a precise device addressing basis for querying the device communication protocol and generating device operation instructions in subsequent steps, and ensuring that the entire control flow is deployed for the correct target device.

[0067] Step S32: Based on the device identifier, query the preset dynamic protocol mapping table to obtain several candidate communication protocols supported by the target device.

[0068] It should be noted that the default dynamic protocol mapping table is a lookup table that stores the mapping relationship between the unique identifier of each registered device and one or more communication protocols it supports.

[0069] By using the extracted device identifier as a query key to match and search in a preset dynamic protocol mapping table, all currently available communication protocols of the target device can be directly retrieved and identified as candidate communication protocols. This enables accurate and rapid determination of all available communication paths for the target device in the face of an IoT environment composed of heterogeneous devices. This provides the necessary selection range for subsequent adaptive selection of the optimal target communication protocol and is the foundation for ensuring that device control commands can be effectively delivered.

[0070] Step S33: Obtain the current network status parameters of each candidate communication protocol and the device network connection status of the target device respectively, and calculate the connection quality score of each candidate communication protocol based on the device network connection status corresponding to each candidate communication protocol.

[0071] By actively probing the network monitoring service built into the StarFlash Linkage Control System or subscribing in real time from the device management platform, the system obtains current network status parameters corresponding to each candidate communication protocol, such as signal strength, transmission delay, and data packet loss rate, as well as the device network connection status of the target device based on each candidate communication protocol and the current network status parameters, such as connection activity and historical stability indicators. Qualitative indicators (such as connection activity) and quantitative indicators (such as historical connection success rate) reflecting the device network connection status are input into a scoring function. This scoring function first subtracts the pre-calculated mean from each connection status parameter and divides it by its standard deviation to eliminate differences in units and magnitudes among different connection status parameters. The different dimensions of measurement ensure that all connection status parameters are on the same order of magnitude. Then, based on the weighting coefficient table determined through domain knowledge or data analysis, a specific weighting coefficient is assigned to each standardized connection status parameter. This weighting coefficient directly reflects the importance of the corresponding connection status parameter to the assessment of connection stability and reliability. For example, signal strength and historical connection success rate may be given higher weighting coefficients, while some auxiliary parameters are given lower weighting coefficients. Finally, a connection quality score is calculated by weighted summation, realizing the comprehensive calculation of the connection status information between the target device and the network into a comparable numerical score, thereby providing a clear and objective quantitative basis for subsequent protocol selection.

[0072] Step S34: Based on the preset protocol decision strategy and the latency sensitivity of the scenario triggering command, the quality scores of each connection are weighted and calculated to obtain the selection degree of each candidate communication protocol.

[0073] It should be noted that the preset protocol decision strategy is a set of predefined calculation rules used to comprehensively evaluate the priority of communication protocol selection.

[0074] Based on the latency sensitivity level of the scenario-triggered command, the latency weight adjustment coefficients corresponding to each candidate communication protocol are mapped from the preset protocol decision strategy. Then, the connection quality score of each candidate communication protocol and its corresponding latency weight adjustment coefficient are linearly weighted and combined to calculate the selection degree of each candidate communication protocol. This enables dynamic consideration of network real-time performance and specific business needs when selecting protocols, ensuring that high latency-sensitive commands are prioritized for low latency protocols, thereby optimizing the transmission timeliness and overall reliability of control commands in complex IoT environments.

[0075] The determination of latency sensitivity level involves pre-defining and storing a latency sensitivity level label for each scenario trigger instruction in the design of the preset scenario instruction library. This latency sensitivity level label is classified according to the urgency of the user intent corresponding to the scenario trigger instruction and the business logic requirements. For example, scenario trigger instructions such as "emergency alarm" or "real-time adjustment" are defined as high latency sensitivity, scenario trigger instructions such as "scenario switching" or "scheduled task" are defined as medium latency sensitivity, and scenario trigger instructions such as "data reporting" or "log recording" are defined as low latency sensitivity. When a specific scenario trigger instruction is obtained based on the user intent mapping, the pre-stored latency sensitivity level of that scenario trigger instruction is retrieved from the preset scenario instruction library.

[0076] Step S35: Select the target communication protocol from the candidate communication protocols according to the selection degree.

[0077] By directly comparing the selection degree of each candidate communication protocol, the candidate communication protocol with the highest selection degree value is selected as the target communication protocol used in this control. This enables the automatic and deterministic execution of an optimal protocol decision in each round of device control, thereby dynamically adapting the transmission of device control commands to the communication channel with the best current network conditions and the highest service matching degree, ensuring the efficiency and success rate of device control command issuance.

[0078] Step S36: Invoke the protocol abstraction layer corresponding to the target communication protocol, and convert the logical operation sequence into device operation instructions based on the protocol abstraction layer.

[0079] The protocol management layer within the StarFlash linkage control system loads and instantiates a protocol abstraction layer that strictly corresponds to the target communication protocol, based on the type of the target communication protocol. This protocol abstraction layer encapsulates the unique communication specifications, message formats, and encoding rules of the target communication protocol. According to the standardized conversion interface provided by this protocol abstraction layer, each operation item in the logical operation sequence is filled in and assembled into a device operation instruction in binary or text format, containing a specific protocol header, command word, device address, parameter list, and checksum, according to the specifications of the target communication protocol. This transforms the abstract logical operation sequence into device control instructions that can be transmitted through the communication link of the target communication protocol and directly parsed by the target device. This effectively shields the complexity of the underlying heterogeneous communication technology and ensures the reliable transmission and execution of user intentions.

[0080] Step S37: Based on the device identifier of the target device, retrieve the device instruction template that matches the device operation instruction from the preset instruction template library.

[0081] It should be noted that the preset instruction template library is a pre-defined, precise data format of instructions that the target device can understand and execute, specifically for each target device model or type. This includes a structured data model of instruction headers, opcodes, parameter field positions and lengths, checksum fields, etc.

[0082] By using the target device's device identifier as a unique query key, a matching search is performed in a preset instruction template library to obtain the corresponding device instruction template. Specifically, the device identifier is used to search the index of the preset instruction template library to find one or more device instruction templates that completely match the target device. This enables the device operation instructions, which have been converted by the protocol abstraction layer, to be further adapted to the manufacturer-specific instruction format framework required by the target device. This provides a format specification for the subsequent generation of device control instructions that the target device can directly parse, and is a key step in ensuring that device control instructions can be correctly understood and executed by heterogeneous devices.

[0083] Step S38: Fill the instruction parameters in the equipment operation instructions into the equipment instruction template to generate equipment control instructions.

[0084] By parsing the specific instruction parameters contained in the device operation instructions, such as the target brightness value, temperature setpoint, or on / off status, and filling these instruction parameter values ​​one by one into the corresponding instruction parameter placeholder fields in the device instruction template according to the format specifications defined in the device instruction template, a device control instruction is assembled into a format that fully conforms to the target device's communication protocol and can be directly received and parsed by its network interface. This instantiates the parameterized user control intent into a device control instruction that can be transmitted over the physical link and correctly recognized and executed by the target device's hardware or firmware. This is the final key step in ensuring that intelligent decisions can accurately reach the target device and enable it to complete the expected actions.

[0085] This application also provides a star-flash linkage control system based on a three-mode remote controller and a smart gateway, referring to... Figure 5 As shown, the StarFlash linkage control system based on a three-mode remote controller and a smart gateway includes: Data generation module 10 is used to generate scene trigger commands based on interactive operations on the three-mode remote controller; The data parsing module 20 is used to send scene triggering instructions to the smart gateway via StarFlash wireless communication, and to parse the scene triggering instructions based on the smart gateway to obtain a logical operation sequence; The data conversion module is used to determine the target communication protocol corresponding to the target device based on the logical operation sequence and the device network connection status, and then convert the logical operation sequence into device control commands corresponding to the target device according to the target communication protocol. The device driver module 40 is used to send device control commands to the target device through the smart gateway, so as to drive the target device to perform linkage control operations corresponding to the scene trigger commands.

[0086] Optionally, the data generation module 10 is also used for: The system captures user-triggered interactive operations using a three-mode remote control and generates raw operation data. Feature extraction is performed on the original operational data to obtain the interaction feature vector; The user's intent is identified by matching the interaction feature vector with a preset intent recognition model. Based on user intent, scene trigger commands are mapped from a preset scene command library.

[0087] Optionally, the data generation module 10 is also used for: Multimodal data fusion is performed on the original operational data to obtain a multimodal data stream; Spatiotemporal feature analysis is performed on multimodal data streams to extract temporal and spatial features, resulting in multidimensional feature vectors. The multidimensional feature vectors are reduced in dimensionality and standardized to generate interactive feature vectors for intent recognition.

[0088] Optionally, the data parsing module 20 is also used for: Semantic parsing of scene trigger commands yields the initial operation intent; Based on the current environmental context information, the initial operation intent is modified to generate an optimized operation intent; Based on the optimization operation intent, the system queries the preset equipment capability knowledge base and generates a logical operation sequence.

[0089] Optionally, the data conversion module is also used for: Extract the device identifier of the target device based on the logical operation sequence; Based on the device identifier, several candidate communication protocols supported by the target device are obtained by querying the preset dynamic protocol mapping table; Obtain the current network status parameters of each candidate communication protocol and their connection status with the target device's network. Based on the connection status of each candidate communication protocol's network, calculate the connection quality score for each candidate communication protocol. Based on the preset protocol decision-making strategy and the latency sensitivity of the scenario triggering command, the quality scores of each connection are weighted and calculated to obtain the selection degree of each candidate communication protocol. Based on the selection criteria, the target communication protocol is selected from the candidate communication protocols.

[0090] Optionally, the data conversion module is also used for: Invoke the protocol abstraction layer corresponding to the target communication protocol, and convert the logical operation sequence into device operation instructions based on the protocol abstraction layer; Based on the device identifier of the target device, the device instruction template that matches the device operation instruction is retrieved from the preset instruction template library; Fill the device operation instructions with the instruction parameters of each instruction into the device instruction template to generate device control instructions.

[0091] The above are all preferred embodiments of this application, and are not intended to limit the scope of protection of this application. Therefore, all equivalent changes made in accordance with the structure, shape and principle of this application should be covered within the scope of protection of this application.

Claims

1. A star-flash linkage control method based on a three-mode remote controller and a smart gateway, characterized in that, include: Generating scene trigger commands based on the interactive operations on the three-mode remote controller; The scene trigger command is sent to the smart gateway via StarFlash wireless communication. The smart gateway then parses the scene trigger command to obtain a logical operation sequence. Based on the logical operation sequence and the device network connection status, after determining the target communication protocol corresponding to the target device, the logical operation sequence is converted into a device control command corresponding to the target device according to the target communication protocol. The smart gateway sends the device control command to the target device to drive the target device to perform a linkage control operation corresponding to the scene trigger command.

2. The star-flash linkage control method based on a three-mode remote controller and a smart gateway according to claim 1, characterized in that, The step of generating scene trigger commands based on the interactive operation on the three-mode remote controller includes: The interactive operation triggered by the user is captured by the three-mode remote control, and raw operation data is generated. Feature extraction is performed on the original operational data to obtain an interaction feature vector; The interaction feature vector is matched with a preset intent recognition model to identify the user intent; Based on the user's intent, the scene trigger command is mapped from the preset scene command library.

3. The star-flash linkage control method based on a three-mode remote controller and a smart gateway according to claim 2, characterized in that, The step of extracting features from the original operational data to obtain the interaction feature vector includes: The original operational data is fused using multimodal data fusion to obtain a multimodal data stream; Spatiotemporal feature analysis is performed on the multimodal data stream to extract temporal and spatial features, resulting in a multidimensional feature vector. The multidimensional feature vector is subjected to dimensionality reduction and standardization to generate the interaction feature vector for intent recognition.

4. The star-flash linkage control method based on a three-mode remote controller and a smart gateway according to claim 1, characterized in that, The step of parsing the scene trigger command based on the smart gateway to obtain the logical operation sequence includes: The initial operation intent is obtained by semantically parsing the scene trigger command; Based on the current environmental context information, the initial operation intention is modified to generate an optimized operation intention; Based on the optimized operation intent, a preset device capability knowledge base is queried to generate the logical operation sequence.

5. The star-flash linkage control method based on a three-mode remote controller and a smart gateway according to claim 1, characterized in that, The step of determining the target communication protocol corresponding to the target device based on the logical operation sequence and the device network connection status includes: Based on the logical operation sequence, extract the device identifier of the target device; Based on the device identifier, several candidate communication protocols supported by the target device are obtained by querying a preset dynamic protocol mapping table; Obtain the current network status parameters of each candidate communication protocol and the device network connection status of the target device, and calculate the connection quality score of each candidate communication protocol based on the device network connection status corresponding to each candidate communication protocol. Based on the preset protocol decision-making strategy and the latency sensitivity of the scenario triggering command, the connection quality scores of each candidate communication protocol are weighted and calculated to obtain the degree of selection of each candidate communication protocol. Based on the selection criteria, the target communication protocol is selected from the candidate communication protocols.

6. The star-flash linkage control method based on a three-mode remote controller and a smart gateway according to claim 1, characterized in that, The step of converting the logical operation sequence into device control instructions corresponding to the target device according to the target communication protocol includes: Invoke the protocol abstraction layer corresponding to the target communication protocol, and convert the logical operation sequence into device operation instructions based on the protocol abstraction layer; Based on the device identifier of the target device, a device instruction template matching the device operation instruction is retrieved from the preset instruction template library; The device control instructions are generated by filling the device operation instructions with the instruction parameters of the device instruction template.

7. A star-flash linkage control system based on a three-mode remote controller and a smart gateway, characterized in that, include: The data generation module is used to generate scene trigger commands based on the interactive operations on the three-mode remote controller; The data parsing module is used to send the scene triggering command to the smart gateway via StarFlash wireless communication, and parse the scene triggering command based on the smart gateway to obtain a logical operation sequence; The data conversion module is used to determine the target communication protocol corresponding to the target device based on the logical operation sequence and the device network connection status, and then convert the logical operation sequence into device control instructions corresponding to the target device according to the target communication protocol. The device driver module is used to send the device control command to the target device through the smart gateway, so as to drive the target device to perform the linkage control operation corresponding to the scene trigger command.

Images