Method, device and medium for expressing constraint of ultra-low bandwidth interaction interface for whole-house intelligence

Abstract

The application discloses a kind of expression constraint methods, equipment and medium of ultra-low bandwidth interactive interface for whole-house intelligence, by collecting home environment state parameters, available expression budget value is generated as the ruling admission threshold by weighted mapping;Receive the expression intention package initiated by each intelligent agent, arrange into the queue to be ruled in descending order of priority;Intention package estimated expression cost is compared with residual budget in turn, and three-state ruling of output permission, degradation or delay, when degradation, expression cost is compressed in equal proportion;For permission and degraded intention package, it is mapped to pixel dot array frame sequence or atmosphere light driving vector by semantic symbol compiler, and ultra-low bandwidth output frame is generated;After rendering, according to actual area, brightness, time length, true consumption is calculated back, and residual budget is updated, forming scheduling closed loop.The application realizes the ordered scheduling of multi-agent expression on ultra-low bandwidth physical channel, guarantees the priority of emergency information transmission, avoids expression congestion and resource waste.

Description

Technical Field

[0001] This invention relates to the field of smart home technology, specifically to a method, device, and medium for expressing constraints of an ultra-low bandwidth interactive interface for whole-house intelligence. Background Technology

[0002] With the popularization of whole-house smart systems and the introduction of Large Language Model (LLM) technology, smart homes are evolving from traditional rule-driven systems to multi-agent collaborative systems. In traditional architectures, user interface output is mainly through app push notifications, central control screen cards, or voice broadcasts. The display channel has sufficient bandwidth and resolution, and information presentation is not constrained by the size of the display. However, to achieve "environmental interaction" under the Calm Technology concept, the new generation of whole-house smart systems has begun to deploy ultra-low bandwidth physical display channels on a large scale, such as pixel-matrix panels embedded in furniture surfaces, monochrome strip e-ink screens, and wall ambient lighting strips. These channels have an extremely low number of pixels that can be expressed per frame, limited color dimensions, and a much lower amount of information that can be carried per message than conventional displays, making them typical narrow-bandwidth perception channels.

[0003] The existing technology has two technical shortcomings in the above scenarios: First, the output of large-scale multi-agent models has a natural "language flood" characteristic. Home intelligent agents driven by large language models tend to generate rich text, structured JSON, or multiple suggestion statements during tool calls and inference. Even with constraints on the brevity of prompts, the length and form of the output remain uncontrollable. When such unconstrained output is directly mapped to notification bubbles, voice broadcasts, or low-bandwidth display channels, the home environment degenerates from a "quiet background" into an "information noise source," violating the non-intrusive principle of environmental interaction.

[0004] Secondly, existing low-bandwidth interface degradation mechanisms rely on static template switching, lacking dynamic expression limit auditing capabilities. For example, some publicly available low-bandwidth communication UI solutions only switch between two preset appearances (network-enabled / network-inactive), failing to dynamically prioritize and control the access of concurrent intent flows from multiple agents. Furthermore, the "bandwidth" concept in some smart home resource scheduling solutions refers only to network link capacity quotas, not covering perceptual domain constraints such as "the proportion of illuminated area on the interface," "brightness limits in nighttime scenes," and "expression frequency during sleep periods." In other words, existing technologies either manage network bandwidth, device computing resources, or perform static appearance degradation, but none quantify "home perception intrusion" into a calculable expression budget, using this as a unified access control basis for multi-agent expression channels. When multiple agents simultaneously initiate expression requests within the same time window, the lack of a calculable limit auditing mechanism to determine who can express, at what intensity, and for how long ultimately leads to either information overload disrupting home tranquility or abrupt global silencing causing the omission of urgent matters. Therefore, there is an urgent need for an expression constraint method for ultra-low bandwidth interfaces that can dynamically quantify the budget for home quietness and perform hierarchical adjudication and constraint compilation of concurrent expression requests from multiple agents, so as to maintain low intrusion and quietness of whole-house interaction while ensuring that key information is not lost. Summary of the Invention

[0005] This invention provides a method, device, and medium for expression constraints of an ultra-low bandwidth interactive interface for whole-house intelligence. By realizing the orderly scheduling of multi-agent expression on an ultra-low bandwidth physical channel, it ensures the priority transmission of emergency information, avoids expression congestion and resource waste, and solves the problems in the background art.

[0006] To achieve the above objectives, the technical solution of the present invention is as follows: A method for expressing constraints in an ultra-low bandwidth interactive interface for whole-house smart home systems involves executing the following steps via a computer device: S1. Collect home environment status parameters in real time, calculate the status parameters using a preset weighted mapping function to generate the available expression budget value for the current scheduling cycle, and use the available expression budget value as the adjudication admission threshold. S2. Receive expression intent packets initiated by intelligent agents in all areas of the house. Each expression intent packet carries a priority field, a semantic content field, and an estimated expression cost. Arrange all expression intent packets in descending order of priority according to the priority field to generate a queue to be decided. S3. Read the expression intent packets in the queue to be adjudicated in sequence, compare the estimated expression cost of each expression intent packet with the available expression budget value, and output the three-state adjudication result: If the estimated expression cost is less than or equal to the remaining available expression budget, it is marked as permitted and the corresponding estimated cost is deducted from the remaining budget; if the estimated expression cost is greater than the remaining available expression budget but the priority reaches the preset emergency threshold, it is marked as degraded and its expression cost is compressed proportionally to the remaining budget; otherwise, it is marked as delayed and moved to the pending queue of the next scheduling cycle. S4. For the expression intent packets that were decided as permitted or downgraded in S3, call the semantic symbol compiler to map the natural language semantics in the semantic content field into pixel matrix frame sequences or ambient light driving vectors, and generate output frame data that conforms to the specifications of ultra-low bandwidth physical expression channels. S5. Send the output frame data to the corresponding ultra-low bandwidth physical expression channel to perform rendering; at the same time, calculate the actual expression consumption based on the actual rendered area, brightness, and duration parameters, and update the remaining available expression budget value in S1 based on the actual expression consumption to complete the single expression scheduling closed loop.

[0007] Preferably, S1 specifically includes: S11. Through human presence sensors, ambient light sensors, distance sensors and system event log interfaces distributed throughout the house, collect data on personnel occupancy status, ambient light intensity level, physical distance between the user and the display interface, and the number of disturbance events within the current sliding time window. S12. Read the user's preset sleep period configuration table and historical quietness credit score. The quietness credit score is the cumulative reward value for no expression rendering in the previous N consecutive scheduling cycles. At the same time, collect the user's interaction feedback data on past expressions, including active viewing behavior, manual extinguishing behavior and expression interface dwell time. S13. Input the parameters collected in S11 and S12 into a preset weighted mapping function to calculate the available expression budget value with a range of 0 to 1, which is used as the adjudication admission threshold for the current scheduling cycle.

[0008] Preferably, S13 specifically includes: S131. Perform exponential smoothing on the current periodic disturbance event count collected in S11 to calculate the smoothed disturbance event count; the calculation formula is:

[0009] In the formula, Let be the number of smoothed disturbance events in the t-th scheduling period. This is the smoothed value from the previous period. Count the new disturbance events added in the current period. The preset smoothing coefficient; S132. Read the previous period's quietness credit score from S12, combine it with the current period's expression status and user interaction feedback, perform credit decay and feedback update, and obtain the current period's quietness credit score. The calculation formula is:

[0010] In the formula, The quietness credit score at the end of period t. This is the integral at the end of the previous period. To preset the credit decay coefficient, This is the increase in the base reward for this period. For feedback direction coefficient, The confidence coefficient is... To preset the single feedback adjustment step size, This is the maximum score. S133. After normalizing the number of smoothed disturbance events obtained in S131, the personnel occupancy status and user physical distance obtained in S11, and the quietness credit score obtained in S132, the data are weighted and superimposed according to time period to generate the available expression budget value for the current scheduling cycle and stored in the shared storage unit. The calculation formula is:

[0011] In the formula, Let $t$ be the available representation budget value for the $t$-th scheduling period. To preset the perturbation saturation threshold, For personnel occupancy status coefficient, The physical distance between the user and the interface. To preset reference distance, The weighted coefficients are and satisfy the following conditions: , The preset time period weighting coefficient.

[0012] Preferably, S2 specifically includes: S21. Monitor the output bus of the multi-agent orchestrator, perform protocol normalization conversion on the heterogeneous expression requests initiated by agents in various domains, and encapsulate them into expression intent packets. Each expression intent packet contains at least a priority field, a semantic content field, and an estimated expression cost field. S22. Sort all encapsulated expression intent packets in descending order according to the priority field. Packets with the same priority are arranged according to their arrival timestamps. Generate a queue to be decided and cache it in the message queue to wait for scheduling.

[0013] Preferably, S3 specifically includes: S31. Take an expression intent packet from the head of the queue to be decided and read the current remaining available expression budget value; S32. Compare the estimated expression cost of the expression intent packet with the remaining available expression budget value: if the estimated cost is less than or equal to the remaining budget, mark it as an allowed state and deduct the corresponding estimated cost value from the remaining budget; if the estimated cost is greater than the remaining budget and the priority reaches the preset emergency threshold, mark it as a degraded state and compress its expression cost proportionally according to the remaining budget; otherwise, mark it as a delayed state and move it to the queue of the next scheduling cycle. S33. Repeat S31 to S32 until the pending queue is empty or the remaining available expression budget value is exhausted.

[0014] Preferably, the proportional compression of expression cost in the degraded state in S32 specifically includes: S321. Read the available expression budget value and the current remaining available expression budget value in the shared memory unit, and calculate the budget saturation for this period; the calculation formula is:

[0015] In the formula, Let t be the current budget saturation level in period t. To implement the remaining available expression budget prior to this ruling, The available expression budget value generated for S133; S322. Based on the remaining available expression budget and the estimated expression cost of the expression intent package to be decided, combined with the budget saturation obtained in S321, calculate the base compression ratio; the calculation formula is:

[0016] In the formula, Let i be the base compression ratio of the i-th expressive packet. Let be the original estimated representation cost of the i-th representation intent packet. This is the preset saturation adjustment coefficient; S323. Introduce a priority gain factor to correct the base compression ratio, obtain the final compression ratio, and convert it into the actual expression cost, simultaneously deducting the remaining budget; the calculation formula is:

[0017]

[0018]

[0019] In the formula, This is the priority gain factor. Let i be the priority value of the i-th packet expressing intent. To preset the emergency priority threshold, This is the highest priority value in the system. Preset priority gain coefficient, For the final compression ratio, This represents the actual cost of the compressed expression.

[0020] Preferably, S4 specifically includes: S41. Read the semantic content field in the expression intent package where the ruling is approved or downgraded, call the pre-built semantic-symbol mapping dictionary to extract keywords, and match to obtain the corresponding pixel icon template index or ambient light effect template index. S42. Based on the actual expression cost after the ruling, the three constraint parameters of maximum allowable rendering area ratio, peak brightness level, and longest duration are calculated. S43. Based on the template index matched in S41 and the constraint parameters converted in S42, the template is scaled and cropped to generate a pixel matrix frame sequence or ambient light driving vector that conforms to the ultra-low bandwidth physical expression channel specification, and written into the frame buffer.

[0021] Preferably, S5 specifically includes: S51. Read the output frame data from the frame buffer, distribute it to the corresponding ultra-low bandwidth physical expression channel driver interface according to the channel type identifier, and start rendering and display. S52. After rendering is completed, the actual number of lit pixels, the actual brightness level and the actual duration are counted, the actual expression consumption is calculated, and the actual expression consumption is written back to update the remaining available expression budget value, thus completing a single scheduling loop.

[0022] In another aspect, the present invention also discloses a computer-readable storage medium storing a computer program, which, when executed by a processor, causes the processor to perform the steps of the method described above.

[0023] In another aspect, the present invention also discloses a computer device, including a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor performs the steps of the method described above.

[0024] As can be seen from the above technical solution compared with the prior art, the present invention has the following beneficial effects: 1. This invention integrates multiple environmental parameters, including smoothing of disturbance events, dynamic accumulation of quietness credit scores, personnel occupancy status, and physical distance, to generate a real-time available expression budget value using a weighted mapping method. This ensures that the release of expression resources is closely linked to the actual state of the home environment. When the home is in a quiet resting period or the user is far from the expression interface, the budget automatically converges to reduce unnecessary disturbances; when user activity is detected and the distance is appropriate, the budget increases accordingly to ensure timely expression response. This effectively solves the technical problem of expression strategies being out of sync with the environment and easily interfering with users under traditional timed or single-threshold control methods.

[0025] 2. This invention, under the constraint of ultra-low bandwidth physical expression channels, ensures that high-priority urgent intentions can still be expressed through degradation when budget is tight by using three-state adjudication and priority-urgency joint gain compression, while preserving the identifiability of key semantics. Based on this, after rendering, the actual consumption is recalculated based on the actual area, brightness, and duration, and the remaining budget is updated, forming a closed-loop correction of "budget generation—adjudication compression—rendering recalculation." This effectively suppresses resource overdraft or multi-round scheduling conflicts caused by the accumulation of prediction deviations, improving scheduling reliability and resource utilization efficiency in multi-agent concurrent expression scenarios. Attached Figure Description

[0026] Figure 1 This is a schematic diagram of the method steps in an embodiment of the present invention; Figure 2 This is a schematic diagram of the data change process in S13 of the present invention; Figure 3 This is a schematic diagram of the data change process in S32 of the present invention. Detailed Implementation

[0027] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, 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 some embodiments of the present invention, but not all embodiments.

[0028] The embodiments of the present invention will be described in further detail below with reference to the accompanying drawings and examples. The following examples are used to illustrate the present invention, but should not be used to limit the scope of the present invention.

[0029] This invention provides a method for expressing constraints in an ultra-low bandwidth interactive interface for whole-house smart home systems, such as... Figure 1 As shown, perform the following steps using a computer device: S1. Collect home environment status parameters in real time, calculate the status parameters using a preset weighted mapping function to generate the available expression budget value for the current scheduling cycle, and use the available expression budget value as the adjudication admission threshold. S2. Receive expression intent packets initiated by intelligent agents in all areas of the house. Each expression intent packet carries a priority field, a semantic content field, and an estimated expression cost. Arrange all expression intent packets in descending order of priority according to the priority field to generate a queue to be decided. S3. Read the expression intent packets in the queue to be adjudicated in sequence, compare the estimated expression cost of each expression intent packet with the available expression budget value, and output the three-state adjudication result: If the estimated expression cost is less than or equal to the remaining available expression budget, it is marked as permitted and the corresponding estimated cost is deducted from the remaining budget; if the estimated expression cost is greater than the remaining available expression budget but the priority reaches the preset emergency threshold, it is marked as degraded and its expression cost is compressed proportionally to the remaining budget; otherwise, it is marked as delayed and moved to the pending queue of the next scheduling cycle. S4. For the expression intent packets that were decided as permitted or downgraded in S3, call the semantic symbol compiler to map the natural language semantics in the semantic content field into pixel matrix frame sequences or ambient light driving vectors, and generate output frame data that conforms to the specifications of ultra-low bandwidth physical expression channels. S5. Send the output frame data to the corresponding ultra-low bandwidth physical expression channel to perform rendering; at the same time, calculate the actual expression consumption based on the actual rendered area, brightness, and duration parameters, and update the remaining available expression budget value in S1 based on the actual expression consumption to complete the single expression scheduling closed loop.

[0030] Example: In this embodiment, taking a whole-house smart system as an example, a 16×16 resolution monochrome LED dot matrix screen is installed on the living room wall as a pixel-level physical expression channel, and independently addressable RGB LED light strips are arranged at the bottom of the sofa and the edge of the ceiling as ambient light expression channels. The aforementioned expression channels are all connected to the central scheduling controller through an RS-485 low-speed bus. The upper limit of the single frame data payload is about 200 bytes, and the effective communication bandwidth is about 9.6kbps, constituting a typical ultra-low bandwidth physical expression channel environment.

[0031] Millimeter-wave human presence sensors, BH1750 ambient light sensors, UWB positioning base stations, and capacitive proximity sensors are deployed in each functional area of ​​the house, and interact with a multi-agent orchestrator via an event bus. The multi-agent orchestrator runs multiple domain-specific agents, including security agents, lighting agents, health monitoring agents, and voice assistant agents, each capable of generating expressive requests based on its own logic.

[0032] The expression constraint method described in this embodiment is executed cyclically by the central scheduling controller at a frequency of one scheduling cycle every 5 seconds. The single scheduling process is as follows: S11. Through human presence sensors, ambient light sensors, distance sensors and system event log interfaces distributed throughout the house, the system collects data on personnel occupancy status, ambient light intensity level, physical distance between the user and the display interface, and the number of disturbance events within the current sliding time window.

[0033] For example, a millimeter-wave human presence sensor can be deployed in the living room to detect the occupancy status of people by reporting once every 2 seconds; a BH1750 ambient light sensor can be integrated into the display panel to read the current illuminance value in real time and quantify it into 0 to 3 levels (corresponding to dark, weak, medium, and bright). A UWB ranging base station is set up in the sofa area to measure the physical distance between the user and the living room wall using a positioning tag carried by the user, with an accuracy of 0.1 meters. The system event log interface is connected to the whole-house event bus to count the number of abnormal alarms triggered, the number of device status changes, and the number of user manual interventions triggered within the past 5 minutes of the sliding time window, and summarize them into a disturbance event count.

[0034] S12. Read the user's preset sleep period configuration table and historical quietness credit score. The quietness credit score is the cumulative reward value for no expression rendering in the previous N consecutive scheduling cycles. At the same time, collect the user's interaction feedback data on past expressions, including active viewing behavior, manual extinguishing behavior, and the duration of stay on the expression interface.

[0035] For example, the sleep time period setting of "23:00-07:00" is read from the user profile; the quietness credit score saved at the end of the previous scheduling cycle is read from the shared storage unit. This score has accumulated to 65 points (maximum 100 points) under the condition that there is no expression rendering in the previous 3 consecutive cycles; in terms of interactive feedback, the user's approach viewing behavior is detected by the built-in capacitive proximity sensor on the panel, the user's operation of actively clicking the turn-off button is recorded by the touch module, and the duration of the user's stay in front of the expression interface is recorded by the infrared pyroelectric sensor. The three types of feedback data are encoded into feedback direction coefficient and confidence coefficient.

[0036] S13. Input the parameters collected in S11 and S12 into a preset weighted mapping function to calculate the available expression budget value ranging from 0 to 1, which is used as the adjudication threshold for the current scheduling cycle. Figure 2 As shown.

[0037] S131. Perform exponential smoothing on the current periodic disturbance event count collected in S11 to calculate the smoothed disturbance event count; the calculation formula is:

[0038] In the formula, Let be the number of smoothed disturbance events in the t-th scheduling period. This is the smoothed value from the previous period. Count the new disturbance events added in the current period. The preset smoothing coefficient; S132. Read the previous period's quietness credit score from S12, combine it with the current period's expression status and user interaction feedback, perform credit decay and feedback update, and obtain the current period's quietness credit score. The calculation formula is:

[0039] In the formula, The quietness credit score at the end of period t. This is the integral at the end of the previous period. To preset the credit decay coefficient, This is the increase in the base reward for this period. For feedback direction coefficient, The confidence coefficient is... To preset the single feedback adjustment step size, This is the maximum score. S133. After normalizing the number of smoothed disturbance events obtained in S131, the personnel occupancy status and user physical distance obtained in S11, and the quietness credit score obtained in S132, the data are weighted and superimposed according to time period to generate the available expression budget value for the current scheduling cycle and stored in the shared storage unit. The calculation formula is:

[0040] In the formula, Let $t$ be the available representation budget value for the $t$-th scheduling period. To preset the perturbation saturation threshold, For personnel occupancy status coefficient, The physical distance between the user and the interface. To preset reference distance, The weighted coefficients are and satisfy the following conditions: , The preset time period weighting coefficient.

[0041] Compared to existing technologies that switch expression modes based solely on a preset schedule or a single light threshold, this step integrates perturbation smoothing, quietness credit accumulation, and multi-source environmental parameters to enable the expression budget value to dynamically and adaptively adjust according to the actual home environment. When the home is quiet and no one is sleeping, the budget automatically decreases to reduce unnecessary expression; when the user is active and at a suitable distance, the budget remains at a higher level to ensure the timeliness and richness of expression.

[0042] S21. Monitor the output bus of the multi-agent orchestrator, perform protocol normalization conversion on the heterogeneous expression requests initiated by agents in various domains, and encapsulate them into expression intent packets. Each expression intent packet contains at least a priority field, a semantic content field, and an estimated expression cost field.

[0043] For example, a multi-agent orchestrator can simultaneously host a lighting agent, a security agent, a health monitoring agent, and a voice assistant agent. The lighting agent initiates a request stating, "Notify the user that the living room light color temperature has automatically switched to reading mode." The system encapsulates this request into an intent package, with a priority field value of 40 (medium priority), a semantic content field in JSON format of {"Category":"Lighting Reminder","Decoration":"Reading Mode"}, and an estimated expression cost field calculated as 0.25 based on pixel usage and estimated display duration.

[0044] S22. Sort all encapsulated expression intent packets in descending order according to the priority field. Packets with the same priority are arranged according to their arrival timestamps. Generate a queue to be decided and cache it in the message queue to wait for scheduling.

[0045] For example, in the current cycle, a total of 4 intent packets were received: the security agent's "Door Lock Not Closed Alarm" with a priority of 90, the health agent's "Sedentary Reminder" with a priority of 60, the lighting agent's "Color Temperature Switching Prompt" with a priority of 40, and the voice assistant's "Weather Report" with a priority of 30. The sorted queue is: [Security Alarm, Health Reminder, Lighting Prompt, Weather Report], with packets of the same priority enqueued according to their timestamps.

[0046] Through a unified protocol conversion and priority sorting mechanism, the original independent and protocol-different domain-specific intelligent agent expression requests are converged into a single-dimensional ordered scheduling, avoiding interface contention and user cognitive confusion caused by concurrent expression by multiple intelligent agents.

[0047] S31. Retrieve an expression intent packet from the head of the pending queue and read the current remaining available expression budget value. For example, retrieve the security alarm intent packet from the head of the queue and read the current remaining budget in the shared storage unit.

[0048] S32. Compare the estimated expression cost of the expression intent package with the remaining available expression budget: if the estimated cost is less than or equal to the remaining budget, mark it as permitted and deduct the corresponding estimated cost from the remaining budget, such as... Figure 3 As shown; if the estimated cost is greater than the remaining budget and the priority reaches the preset emergency threshold, it is marked as a degraded state and its expressed cost is compressed proportionally according to the remaining budget; otherwise, it is marked as a delayed state and moved to the queue of the next scheduling cycle.

[0049] Compared to a simple first-come, first-served scheduling strategy, the three-state adjudication mechanism ensures that high-priority urgent messages can still be displayed through downgrading when the budget is tight, while low-priority non-urgent messages are automatically deferred, so that limited message resources always serve the most important information delivery needs at present.

[0050] S321. Read the available expression budget value and the current remaining available expression budget value in the shared memory unit, and calculate the budget saturation for this period; the calculation formula is:

[0051] In the formula, Let t be the current budget saturation level in period t. To implement the remaining available expression budget prior to this ruling, The available expression budget value generated for S133; S322. Based on the remaining available expression budget and the estimated expression cost of the expression intent package to be decided, combined with the budget saturation obtained in S321, calculate the base compression ratio; the calculation formula is:

[0052] In the formula, Let i be the base compression ratio of the i-th expressive packet. Let be the original estimated representation cost of the i-th representation intent packet. This is the preset saturation adjustment coefficient; S323. Introduce a priority gain factor to correct the base compression ratio, obtain the final compression ratio, and convert it into the actual expression cost, simultaneously deducting the remaining budget; the calculation formula is:

[0053]

[0054]

[0055] In the formula, This is the priority gain factor. Let i be the priority value of the i-th packet expressing intent. To preset the emergency priority threshold, This is the highest priority value in the system. Preset priority gain coefficient, For the final compression ratio, This represents the actual cost of the compressed expression.

[0056] For example, the estimated cost of a security alarm intent package is 0.25, and the current remaining budget is 0.5935 (≥ 0.25), so it is marked as permitted. After deduction, the remaining budget is updated to 0.3435. Next, retrieve the health reminder intent package. The remaining budget is 0.3435 < 0.35, but its priority of 60 has reached the emergency threshold of 50, so it is marked as a downgraded state and enters the downgrade compression process.

[0057] S33. Repeat S31 to S32 until the pending decision queue is empty or the remaining available expression budget is exhausted. For example, after processing security alarms and health alerts, the remaining budget has dropped to near 0, and the lighting cues and weather broadcast intent packets cannot meet the permission or degrade conditions, are marked as delayed, and moved to the pending decision queue for the next scheduling week.

[0058] Compared to a uniform compression scheme, the introduction of a differentiated degradation strategy based on budget saturation awareness and priority gain results in a smaller compression ratio and higher fidelity when the budget is sufficient, while automatically adjusting the compression intensity based on the urgency of the intent when the budget is tight, thus preventing high-priority expressions from losing their recognizability due to excessive compression.

[0059] S41. Read the semantic content field in the expression intent package where the ruling is approved or downgraded, call the pre-built semantic-symbol mapping dictionary to extract keywords, and match to obtain the corresponding pixel icon template index or ambient light effect template index.

[0060] For example, if the semantic content of the security alarm intent package is "door lock not closed alarm", the system queries the semantic-symbol mapping dictionary and matches the lock icon template with the number #012 corresponding to the keyword "door lock". The keyword "alarm" triggers the red flashing ambient light effect template with the number #F03.

[0061] S42. Based on the actual expression cost after the ruling, the three constraint parameters of maximum allowable rendering area ratio, peak brightness level, and longest duration are calculated.

[0062] For example, the actual cost of a security alarm is 0.25. According to the system's preset cost-resource mapping table, this cost corresponds to a maximum rendering area of ​​30% of the screen area, a peak brightness level of 3 (out of 4), and a maximum duration of 8 seconds. In contrast, the cost of a downgraded health alert corresponds to a rendering area of ​​15%, a peak brightness level of 2, and a duration of 5 seconds.

[0063] S43. Based on the template index matched in S41 and the constraint parameters converted in S42, the template is scaled and cropped to generate a pixel matrix frame sequence or ambient light driving vector that conforms to the ultra-low bandwidth physical expression channel specification, and written into the frame buffer.

[0064] For example, the #012 lock icon was scaled to 30% of its area and rendered on the LED dot matrix screen for 8 seconds at the highest brightness level 3; the #F03 red ambient light effect was cropped to the corresponding 15% strip area and rendered on the LED light strip for 5 seconds at brightness level 2. Ultimately, the data size per frame was controlled to within approximately 200 bytes, allowing for real-time transmission via a low-speed serial bus.

[0065] Through a two-stage transformation of semantic-symbolic compilation and cost-resource mapping, high-dimensional natural language semantics are compressed into a pixel instruction stream with extremely low data volume. This enables physical expression channels that only support a few kbps bandwidth (such as monochrome LED dot matrix and light strip) to carry semantically distinguishable information expression, breaking through the limitation of traditional screens relying on high-bandwidth transmission.

[0066] S51. Read the output frame data from the frame buffer, distribute it to the corresponding ultra-low bandwidth physical expression channel driver interface according to the channel type identifier, and start rendering and display.

[0067] For example, the output frame data of the security alarm is read from the frame buffer and distributed to the 16×16 monochrome LED dot matrix screen driver interface on the living room wall through the channel type identifier "LED_DOT_MATRIX_CH1"; at the same time, the ambient light driver vector of the health reminder is distributed to the light strip under the sofa through the identifier "LED_STRIP_CH2", and each microcontroller performs rendering independently.

[0068] S52. After rendering is completed, the actual number of lit pixels, the actual brightness level and the actual duration are counted, the actual expression consumption is calculated, and the actual expression consumption is written back to update the remaining available expression budget value, thus completing a single scheduling loop.

[0069] For example, a security alarm actually lit up 76 out of 256 pixels (29.7%), with an average brightness of 2.8 and a duration of 8.2 seconds (because the user manually turned it off in advance). According to the preset energy consumption model, the actual consumption was 0.27. This value is slightly higher than the estimated cost of 0.25. The system writes it back to the remaining budget and includes it in the overspending compensation during the next cycle scheduling to prevent budget overdraft in consecutive cycles.

[0070] By employing a closed-loop mechanism that recalculates the actual resource consumption after rendering and updates the budget, the problem of accumulated discrepancies between estimated costs and actual rendering results is resolved. When users actively shorten the rendering time or sensors detect occlusion leading to a reduction in the actual rendered area, the unused budget will naturally flow back, making the system's measurement of rendering resources more accurate. Over long-term operation, the budget allocation will better align with the user's actual perceived needs.

[0071] This embodiment utilizes multi-dimensional information from the home environment, including human presence, ambient light, physical distance, disturbance frequency, and user interaction feedback, to form a dynamic expression budget through disturbance smoothing and quietness credit accumulation. This budget then constrains the heterogeneous expression requests of multiple agents throughout the house, performing three-state adjudication and differentiated compression, ultimately compiling them into extremely low-bandwidth physical expression instructions. Compared to existing fixed-rule scheduling or single-threshold control schemes, this method has significant advantages in the following aspects: First, the decision-making process can adapt to the user's environment and behavior, automatically converging when the user is quiet and responding promptly when the user is active. Second, it ensures the priority delivery of urgent information under limited expressive resources and retains the recognizability of key semantics through a joint gain compression mechanism. Third, it forms a closed-loop correction by back-calculating the actual consumption after rendering, so that the budget allocation remains consistent with the user's actual perception in the long term.

[0072] In another aspect, the present invention also discloses a computer-readable storage medium storing a computer program, which, when executed by a processor, causes the processor to perform the steps of the method described above.

[0073] In another aspect, the present invention also discloses a computer device, including a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor performs the steps of the method described above.

[0074] In another embodiment provided in this application, a computer program product containing instructions is also provided, which, when run on a computer, causes the computer to execute any of the expression constraint methods for ultra-low bandwidth interactive interfaces for whole-house intelligence in the above embodiments.

[0075] It is understood that the systems, devices, and storage media provided in the embodiments of the present invention correspond to the methods provided in the embodiments of the present invention, and the explanations, examples, and beneficial effects of the relevant content can be referred to the corresponding parts of the above methods.

[0076] In the above embodiments, implementation can be achieved, in whole or in part, through software, hardware, firmware, or any combination thereof. When implemented in software, it can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transferred from one computer-readable storage medium to another.

[0077] For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that a computer can access, or a data storage device such as a server or data center that integrates one or more available media.

[0078] The available media may be magnetic media (e.g., floppy disks, hard disks, magnetic tapes), optical media (e.g., DVDs), or semiconductor media (e.g., solid state disks (SSDs)).

[0079] It should be noted that in this document, relational terms such as first and second are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations.

[0080] Furthermore, 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. Without further limitation, 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.

[0081] The various embodiments in this specification are described in a related manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the system embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions of the method embodiments.

[0082] The embodiments of the present invention are given for the purposes of illustration and description. Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of the present invention.

Claims

1. A method for expressing constraints in an ultra-low bandwidth interactive interface for whole-house smart home systems, characterized in that, Perform the following steps using a computer device: S1. Collect home environment status parameters in real time, calculate the status parameters using a preset weighted mapping function to generate the available expression budget value for the current scheduling cycle, and use the available expression budget value as the adjudication admission threshold. S2. Receive expression intent packets initiated by intelligent agents in all areas of the house. Each expression intent packet carries a priority field, a semantic content field, and an estimated expression cost. Arrange all expression intent packets in descending order of priority according to the priority field to generate a queue to be decided. S3. Read the expression intent packets in the queue to be adjudicated in sequence, compare the estimated expression cost of each expression intent packet with the available expression budget value, and output the three-state adjudication result: If the estimated expression cost is less than or equal to the remaining available expression budget, it is marked as permitted and the corresponding estimated cost is deducted from the remaining budget; if the estimated expression cost is greater than the remaining available expression budget but the priority reaches the preset emergency threshold, it is marked as degraded and its expression cost is compressed proportionally to the remaining budget; otherwise, it is marked as delayed and moved to the pending queue of the next scheduling cycle. S4. For the expression intent packets that were decided as permitted or downgraded in S3, call the semantic symbol compiler to map the natural language semantics in the semantic content field into pixel matrix frame sequences or ambient light driving vectors, and generate output frame data that conforms to the specifications of ultra-low bandwidth physical expression channels. S5. Send the output frame data to the corresponding ultra-low bandwidth physical expression channel to perform rendering; at the same time, calculate the actual expression consumption based on the actual rendered area, brightness, and duration parameters, and update the remaining available expression budget value in S1 based on the actual expression consumption to complete the single expression scheduling closed loop.

2. The expression constraint method for an ultra-low bandwidth interactive interface for whole-house intelligence as described in claim 1, characterized in that: S1 specifically includes: S11. Through human presence sensors, ambient light sensors, distance sensors and system event log interfaces distributed throughout the house, collect data on personnel occupancy status, ambient light intensity level, physical distance between the user and the display interface, and the number of disturbance events within the current sliding time window. S12. Read the user's preset sleep period configuration table and historical quietness credit score. The quietness credit score is the cumulative reward value for no expression rendering in the previous N consecutive scheduling cycles. At the same time, collect the user's interaction feedback data on past expressions, including active viewing behavior, manual extinguishing behavior and expression interface dwell time. S13. Input the parameters collected in S11 and S12 into a preset weighted mapping function to calculate the available expression budget value with a range of 0 to 1, which is used as the adjudication admission threshold for the current scheduling cycle.

3. The expression constraint method for an ultra-low bandwidth interactive interface for whole-house intelligence as described in claim 2, characterized in that: S13 specifically includes: S131. Perform exponential smoothing on the current periodic disturbance event count collected in S11 to calculate the smoothed disturbance event count; the calculation formula is: In the formula, Let be the number of smoothed disturbance events in the t-th scheduling period. This is the smoothed value from the previous period. Count the new disturbance events added in the current period. The preset smoothing coefficient; S132. Read the previous period's quietness credit score from S12, combine it with the current period's expression status and user interaction feedback, perform credit decay and feedback update to obtain the current period's quietness credit score; the calculation formula is: In the formula, The quietness credit score at the end of period t. This is the integral at the end of the previous period. To preset the credit decay coefficient, This is the increase in the base reward for this period. For feedback direction coefficient, The confidence coefficient is... To preset the single feedback adjustment step size, This is the maximum score. S133. After normalizing the number of smoothed disturbance events obtained in S131, the personnel occupancy status and user physical distance obtained in S11, and the quietness credit score obtained in S132, the data are weighted and superimposed according to time period to generate the available expression budget value for the current scheduling cycle and stored in the shared storage unit. The calculation formula is: In the formula, Let $t$ be the available representation budget value for the $t$-th scheduling period. To preset the perturbation saturation threshold, For personnel occupancy status coefficient, The physical distance between the user and the interface. As a preset reference distance, The weighted coefficients are and satisfy the following conditions: , The weighting coefficients are preset for different time periods.

4. The expression constraint method for an ultra-low bandwidth interactive interface for whole-house intelligence as described in claim 3, characterized in that: S2 specifically includes: S21. Monitor the output bus of the multi-agent orchestrator, perform protocol normalization conversion on the heterogeneous expression requests initiated by agents in various domains, and encapsulate them into expression intent packets. Each expression intent packet contains at least a priority field, a semantic content field, and an estimated expression cost field. S22. Sort all encapsulated expression intent packets in descending order according to the priority field. Packets with the same priority are arranged according to their arrival timestamps. Generate a queue to be decided and cache it in the message queue to wait for scheduling.

5. The expression constraint method for an ultra-low bandwidth interactive interface for whole-house intelligence as described in claim 4, characterized in that: S3 specifically includes: S31. Take an expression intent packet from the head of the queue to be decided and read the current remaining available expression budget value; S32. Compare the estimated expression cost of the expression intent packet with the remaining available expression budget value: if the estimated cost is less than or equal to the remaining budget, mark it as an allowed state and deduct the corresponding estimated cost value from the remaining budget; if the estimated cost is greater than the remaining budget and the priority reaches the preset emergency threshold, mark it as a degraded state and compress its expression cost proportionally according to the remaining budget; otherwise, mark it as a delayed state and move it to the queue of the next scheduling cycle. S33. Repeat S31 to S32 until the pending queue is empty or the remaining available expression budget value is exhausted.

6. The expression constraint method for an ultra-low bandwidth interactive interface for whole-house intelligence as described in claim 5, characterized in that: The proportional compression of expression cost in the degraded state in S32 specifically includes: S321. Read the available expression budget value and the current remaining available expression budget value in the shared memory unit, and calculate the budget saturation for this period; the calculation formula is: In the formula, Let t be the current budget saturation level in period t. To implement the remaining available expression budget prior to this ruling, The available expression budget value generated for S133; S322. Based on the remaining available expression budget and the estimated expression cost of the expression intent package to be decided, combined with the budget saturation obtained in S321, calculate the base compression ratio; the calculation formula is: In the formula, Let i be the base compression ratio of the i-th expressive packet. Let be the original estimated representation cost of the i-th representation intent packet. This is the preset saturation adjustment coefficient; S323. Introduce a priority gain factor to correct the base compression ratio, obtain the final compression ratio, and convert it into the actual expression cost, simultaneously deducting the remaining budget; the calculation formula is: In the formula, This is the priority gain factor. Let i be the priority value of the i-th packet expressing intent. To preset the emergency priority threshold, This is the highest priority value in the system. Preset priority gain coefficient, For the final compression ratio, This represents the actual cost of the compressed expression.

7. The expression constraint method for an ultra-low bandwidth interactive interface for whole-house intelligence as described in claim 1, characterized in that: S4 specifically includes: S41. Read the semantic content field in the expression intent package where the ruling is approved or downgraded, call the pre-built semantic-symbol mapping dictionary to extract keywords, and match to obtain the corresponding pixel icon template index or ambient light effect template index. S42. Based on the actual expression cost after the ruling, the three constraint parameters of maximum allowable rendering area ratio, peak brightness level, and longest duration are calculated. S43. Based on the template index matched in S41 and the constraint parameters converted in S42, the template is scaled and cropped to generate a pixel matrix frame sequence or ambient light driving vector that conforms to the ultra-low bandwidth physical expression channel specification, and written into the frame buffer.

8. The expression constraint method for an ultra-low bandwidth interactive interface for whole-house intelligence as described in claim 1, characterized in that: S5 specifically includes: S51. Read the output frame data from the frame buffer, distribute it to the corresponding ultra-low bandwidth physical expression channel driver interface according to the channel type identifier, and start rendering and display. S52. After rendering is completed, the actual number of lit pixels, the actual brightness level and the actual duration are counted, the actual expression consumption is calculated, and the actual expression consumption is written back to update the remaining available expression budget value, thus completing a single scheduling loop.

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

10. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the computer program is executed by the processor, it causes the processor to perform the steps of the method as described in any one of claims 1 to 8.

Images