A Smart Community Service Interaction System and Method Based on Information Collection

Abstract

This invention relates to the field of electronic digital data processing technology, and discloses a smart community service interaction system and method based on information acquisition. The system includes: after a sensor state change is triggered, immediately requesting synchronization status from other logically related members of the sensor group; performing a weighted logical consistency judgment based on the communication channel reliability of each state response; and outputting an abnormal event command only when the judgment result is logically inconsistent. The decision-making basis of this invention shifts from trusting the absolute value of a single sensor data point to a self-consistent judgment of the inherent logical relationship of a set of uncertain information. This avoids a large number of invalid alarms caused by false triggering of single-point sensors or unstable communication links in terms of information processing methods. Simultaneously, by utilizing the system's normal operation data, it gains the ability to predict the sub-health state of equipment and the trend changes in user behavior patterns.

Description

TECHNICAL FIELD

[0001] The present application relates to an information collection-based smart community service interaction system and method, and belongs to the technical field of electronic digital data processing. BACKGROUND

[0002] In the construction and application of the current smart community system, there is a technical problem between its core architecture and the actual deployment conditions. The contradiction is that the system's information processing architecture depends on the accuracy and effectiveness of the single sensor data received in the design, but in the cost-sensitive large-scale deployment, it must use a large number of economic sensors whose output state is uncertain.

[0003] The difference between this design premise and the operation reality directly leads to the continuous false alarm caused by sensor false triggering or performance drift, and the judgment failure in complex scenarios that require multi-dimensional information corroboration, thereby eroding the application reliability and practical value of the system. The path of trying to solve this problem by simply replacing high-precision sensors not only faces economic cost challenges, but also does not respond to the limitations of single-dimensional information judgment ability in complex situations.

[0004] The background of this technical contradiction is that the industry generally adopts a centralized data processing architecture of terminal-cloud. After receiving the raw data uploaded by the field sensor, the cloud platform conducts centralized analysis and decision. Specifically, the contradiction leads to the following problems in the application of the existing technology: the decision logic of the system depends on the effectiveness of the single data source in structure. Once the accuracy of the data source cannot be guaranteed, the credibility of the output result will decrease. The existing architecture lacks an information processing mechanism to form a state judgment with higher confidence by exploring the internal logical correlation among a group of uncertain or even conflicting heterogeneous data without significantly increasing hardware costs. Therefore, how to design a new information processing architecture that can get rid of the dependence on the effectiveness of single sensor data and instead generate reliable state perception and decision instructions by examining the logical consistency between a group of localized event-triggered and multi-source heterogeneous sensor data has become a technical problem to be solved by the present application. SUMMARY

[0005] The present application provides an information collection-based smart community service interaction system and method, which mainly aims to solve the technical problem that the existing information processing architecture lacks reliable state perception generated by examining the logical consistency among multi-source data when facing the data uncertainty brought by low-cost sensors due to its dependence on the effectiveness of single sensor data.

[0006] To achieve the above object, the application provides a smart community service interaction system based on information collection, comprising a plurality of sensors, which are divided into at least one sensor group in advance, each sensor group comprising one trigger sensor and at least one verifier sensor defined based on physical proximity or logical correlation, and at least one local gateway in communication with the sensor group, wherein the local gateway is configured to:

[0007] in response to a trigger signal received from the trigger sensor and indicating a change in its own state, sending a state query request to all verifier sensors in the sensor group to which the trigger sensor belongs;

[0008] based on the number of data communication retransmissions required for successful reception of the state response of each verifier sensor, determining a decision weight representing the real-time reliability of the communication channel for the state response;

[0009] According to a logical rule set in the system initialization phase, the state change of the trigger sensor and the state response of each verifier sensor are subjected to a weighted logical consistency judgment, wherein the contribution of each state response in the judgment is associated with its corresponding decision weight;

[0010] Only when the result of the weighted logical consistency judgment is logically inconsistent, an abnormal event interaction instruction representing logical inconsistency is generated and output.

[0011] Preferably, the sensor group is configured for the safety scene of the elderly living alone, wherein the door magnet sensor is defined as the trigger sensor, and at least one of the infrared detector, the bedside pressure sensor and the intelligent power equipment state sensor located in the same residential space as the door magnet sensor is defined as the verifier sensor; and the local gateway determines the decision weight of the state response The rule is defined by the following relationship: wherein, is the number of data communication retransmissions, is a non-negative integer; is a non-negative sensitivity coefficient set in the system initialization phase for adjusting the influence degree of the retransmission number on the decision weight; the value range of the decision weight is .

[0012] Preferably, the local gateway is further configured to: for each event determined to be logically consistent, extract an event timing feature associated with the specific monitoring object; based on the event timing feature, dynamically calculate and continuously roll update a statistical baseline channel locally for the monitoring object to represent its normal behavior pattern, the statistical baseline channel being jointly defined by the mean and standard deviation of the specific behavior of the monitoring object within a recent time period; and monitor the long-term unidirectional drift trend of the statistical baseline channel itself, and generate and output an early warning instruction indicating a trend change in the behavior pattern of the monitoring object when the speed of the drift trend continuously exceeds a preset trend threshold.

[0013] Preferably, the local gateway is further configured to: when the result of the weighted logical consistency determination is a new logical consistent pattern outside a preset rule, activate a secondary verification; the secondary verification includes: instructing an actuator physically adjacent to the sensor group and having a light-emitting or sound-emitting function to generate a standardized physical disturbance; and synchronously requesting a sensor capable of sensing the physical disturbance to report its state; if the sensor capable of sensing the physical disturbance fails to report its state of having sensed the physical disturbance within a predetermined time, determining that the sensor is in an unusable state; if the state of having sensed the physical disturbance is reported, determining that the new logical consistent pattern is a real abnormal event to be reported.

[0014] Preferably, the local gateway is further configured to: at each time of executing the state query request, measure and record the end-to-end response delay required for each verifier sensor to return its state response; for each verifier sensor, locally maintain a statistical model about its historical response delay, the model including the recent average delay and recent delay jitter of the sensor; when detecting that the recent average delay of any verifier sensor appears to be continuously increasing or its recent delay jitter appears to be continuously amplifying and exceeds the normal range determined by its own historical performance, generate and output a pre-diagnosis instruction indicating that the sensor is in a sub-healthy state.

[0015] Preferably, the local gateway is further configured to: for each logical rule in the logical rule library, associate a dynamically adjustable rule confidence score; when the abnormal event interaction instruction triggered by a certain logical rule is not confirmed by any manual intervention or higher-level alarm within a specified observation period, automatically lower the rule confidence score of the logical rule; when the rule confidence score is lower than a preset dormancy threshold, automatically disable the logical rule.

[0016] Preferably, the logic rules set in the local gateway include: when the door magnetic sensor responds to a continuously closed state for a period of more than twelve hours, and at least two of the infrared detector, bedside pressure sensor, and smart electrical device status sensor indicate no human activity or no change in device status, then it is determined to be a logical inconsistency.

[0017] Preferably, the local gateway is further configured to: after each determination of logical consistency, for a numerical sensor, calculate the predicted value of its theoretical state response based on the synchronous state response of other sensors in the group; the difference between the predicted value of the theoretical state response and the actual state response of the numerical sensor is defined as the response residual of the event; locally, continuously calculate the energy value of the response residual sequence generated by each numerical sensor over a long period; when the energy value of the response residual sequence of a sensor shows a continuous unidirectional increase and deviates from its own historical baseline to a predetermined extent, generate a pre-diagnostic instruction indicating that the sensor is in a sub-healthy state due to performance degradation.

[0018] Preferably, the local gateway is also configured to: process or record all events determined to be logically consistent and their associated raw sensor data on the local network; send only the interaction instructions for abnormal events that represent logical inconsistencies to a central service platform; thereby restricting the vast majority of raw sensor data to the local network in the system's information flow, and selectively reporting only cross-validated abnormal events.

[0019] A smart community service interaction method based on information collection is applied to a system containing multiple sensors, wherein the sensors are pre-divided into at least one sensor group, and each sensor group includes a trigger sensor and at least one verifier sensor. The method includes the following steps:

[0020] Step a: When the state of one of the trigger sensors changes, immediately send a state query request to all the verifier sensors in the sensor group to which the trigger sensor belongs, and receive the state response from the verifier sensors.

[0021] Step b: Based on the number of data communication retransmissions required to successfully receive the state response of each of the verifier sensors, determine a decision weight characterizing the real-time reliability of its communication channel for the state response.

[0022] Step c: Based on a logical rule set during the system initialization phase, a weighted logical consistency judgment is performed on the state changes of the trigger sensor and the state response of each verifier sensor, wherein the contribution of each state response in the judgment is associated with its corresponding decision weight.

[0023] Step d: Only when the result of the weighted logical consistency judgment is logically inconsistent, generate and output an abnormal event interaction instruction representing the logical inconsistency.

[0024] Compared with the prior art, the beneficial effects of the present invention are:

[0025] 1. This technical solution transforms the data retransmission count parameter in the communication protocol stack into a decision weight for application-layer logic judgment through a specific mathematical formula, thereby establishing dynamic coupling between the decision logic and the physical communication environment. This information processing method endows the system with a scenario-adaptive decision-making capability. Specifically, when the communication channel quality of a specific sensor deteriorates, the system does not rudely determine its failure or adopt potentially erroneous signals, but can automatically and smoothly reduce the contribution of that information source in the final decision, prioritizing the acceptance of the status of sensors with other unobstructed channels. As a result, the abnormal event interaction command output by the system is transformed from a simple Boolean logic result into a decision basis with higher confidence that inherently contains a real-time physical reliability assessment of each information source. Furthermore, the system retains all logically consistent normal event data locally for further long-term trend analysis of the behavior pattern baseline of the monitored object, and achieves pre-diagnosis of the sub-health state of the equipment by analyzing characteristics such as sensor response delay or response residual. Ultimately, a closed-loop information processing system integrating real-time judgment, trend prediction, and self-health assessment is formed.

[0026] 2. This invention provides an information processing method whose decision-making basis no longer depends on whether the state reading of any single sensor is absolutely valid. When the state of any sensor changes, the change only serves as the starting point for triggering a local information correlation verification. The system obtains the synchronization state of other members in its logically related sensor group in real time and performs a weighted logical decision based on the real-time reliability of each state response transmission channel. In this way, the system's perception credibility is established from the self-consistency of the inherent logical relationship between a set of low-cost, uncertain components. This allows the system to transform the nature of its alarm output from a suspected signal that requires prior manual verification into a responsive decision-making basis when facing common realities such as instantaneous false triggering of single-point sensors or unstable communication links.

[0027] 3. This invention utilizes data from each logically consistent normal event, discarding information that would otherwise be obsolete in traditional architectures, to build and continuously update a personalized behavioral pattern baseline for a specific monitored object locally. The system focuses not only on whether an isolated instantaneous state exceeds a certain fixed threshold, but also on whether the monitored object's own behavioral baseline channel exhibits a long-term, slow, but directional drift. This gives the system a longitudinal timescale insight capability, enabling it to identify those events that appear normal in any single time slice, but whose cumulative trends indicate deep risks in critical state transitions. Attached Figure Description

[0028] Fig. 1 This is a flowchart of the event processing and analysis based on weighted logic judgment in this invention;

[0029] Fig. 2 This is a diagram illustrating the interaction architecture of the system's functional modules and data model in this invention.

[0030] Fig. 3 This is a schematic diagram illustrating the dynamic evolution of the confidence score of the logical rules in this invention. Detailed Implementation

[0031] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be described in further detail below. Obviously, the described embodiments are only some embodiments of this invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0032] The smart community service interaction system and method based on information acquisition disclosed in this invention comprises an information processing architecture including multiple pre-configured sensors, at least one local gateway responsible for edge-side data processing, and a central service platform. Data processing tasks such as correlation analysis and logical consistency judgment of multi-source heterogeneous sensor data are configured to be executed within the local gateway, thereby restricting the processing of the vast majority of raw sensor data to the local network. Only when an abnormal event confirming logical inconsistency occurs does the local gateway generate and output corresponding interaction commands to the central service platform. This electronic digital data processing architecture operates within a distributed information processing flow. This approach addresses the technical limitations of existing centralized processing methods, which rely on the validity of data from a single sensor and struggle to make reliable status assessments in the face of data uncertainty and communication channel fluctuations. In a specific application scenario, such as a safety monitoring service for elderly people living alone, multiple sensors deployed within a residence are pre-divided into one or more sensor groups based on their physical proximity or logical correlation in safety monitoring. Each sensor group is defined as a logical set containing a trigger sensor and at least one verifier sensor. For example, a door magnetic sensor can be defined as a trigger sensor, and an infrared sensor located in the same residential space as the door magnetic sensor... At least one of the following: a pressure sensor, a bedside pressure sensor, and a status sensor for electrical equipment connected via a smart socket, can be defined as a validator sensor. The local gateway establishes this grouping configuration and association during system initialization, laying the foundation for subsequent event-triggered local information consistency verification. When the system is running, if the state of any sensor defined as a trigger changes—for example, a door magnetic sensor reports a change from closed to open—this state change signal, after being sent to the local gateway, is not directly treated as an alarm event. Instead, it is defined by the local gateway as a trigger signal for a verification. This is because the state report of a single sensor may be affected by its own... If inaccuracy occurs due to performance drift or environmental interference, the local gateway, upon responding to the trigger signal, immediately sends a status query request to all validator sensors within the sensor group to which the trigger sensor belongs. This allows it to obtain a multi-dimensional synchronous status description of the current physical environment within a short time window. For example, upon receiving a door opening signal, the local gateway queries whether the infrared detector in the same group has detected movement of a person, whether the bedside pressure sensor has sensed a person leaving the bed, and whether the smart socket of the table lamp has reported its power-on status. In this way, the system's decision-making basis shifts from judging the value of a single sensor to examining the inherent logical relationship between a set of multi-source heterogeneous data triggered by an event.

[0033] In environments such as community deployments, the quality of wireless communication channels between the local gateway and each validator sensor can be dynamically affected by factors such as co-channel interference or physical obstruction. To address this challenge, when the local gateway executes a status query request and waits for status responses from each validator sensor, its communication protocol stack is configured to record the number of data communication retransmissions required to successfully receive each status response. ,Should This is a non-negative integer whose value characterizes the real-time reliability of the corresponding communication channel; the local gateway determines a decision weight for each received status response based on the following formula. , ,in, The range of values ​​is , This is a non-negative sensitivity coefficient set during the system initialization phase to adjust the influence of the number of retransmissions on the decision weights. The calibration can be accomplished through a deterministic procedure, namely, during the system debugging phase, actively injecting a controllable packet loss rate into the test channel, observing the communication success rate under different retransmission counts, and setting... The value of makes the calculated weights such that, under a typical communication quality degradation boundary, such as 3 retransmissions, the weights are... If the value falls within a preset low confidence interval, such as less than 0.3, the weights are adjusted to reflect changes in channel quality. This procedure allows for the determination of the decision weights for a state response that requires multiple retransmissions to achieve successful reception. The weight of a successful one-time communication response will be significantly reduced; conversely, the weight of a successful one-time communication response will be significantly reduced. The value is 1, and the communication quality of the physical layer is thus converted into a quantitative parameter that can be used for application layer decision-making, enabling logical judgments to adapt to dynamic changes in the communication environment.

[0034] After acquiring the state changes of the trigger sensor, as well as the state responses of all validator sensors and their corresponding decision weights. Subsequently, the local gateway performs a weighted logical consistency judgment based on a logical rule set during the system initialization phase. During this judgment, the contribution of each validator sensor's state response to the final decision is determined by its corresponding decision weight. For example, a logical rule for the nighttime safety of elderly people living alone could be set as follows: if, for a continuous period of more than twelve hours, the door magnetic sensor remains continuously closed, and at least two of the infrared detector, bedside pressure sensor, and smart device status sensor all indicate no human activity or no change in device status, then a logical inconsistency is determined. When executing this rule, the system will comprehensively consider the weight of each validator sensor response. If the infrared detector indicating no human activity has a higher weight... The weight of the other sensor, which indicates weak activity, is very high. If the value is lower due to channel issues, the system will primarily accept the former state in the logical decision. The local gateway is configured to generate and output an abnormal event interaction command representing the logical inconsistency only when the result of the weighted logical consistency judgment is logically inconsistent. This decision mechanism reduces invalid alarms caused by false triggering of single-point sensors or unstable communication links.

[0035] For each logically consistent normal event, the associated raw sensor data is processed or recorded on the local network for continuous analysis of the monitored object's behavioral patterns and equipment health status. Given that some key risks do not originate from sudden events but rather from long-term, slow state drift, the local gateway is configured to extract time-series features from events associated with a specific monitored object. Based on this confirmed normal event data, a statistical baseline channel characterizing the monitored object's normal behavioral patterns is dynamically calculated and continuously updated locally. This baseline channel is composed of the average value of specific behaviors of the monitored object, such as daily wake-up time, over a recent time period. with standard deviation By jointly defining it, a dynamic normal range can be formed. The local gateway further monitors the long-term unidirectional drift trend of the baseline channel itself, for example, by calculating the baseline channel average value. The system measures the rate of change of the weekly moving average and generates and outputs an early warning instruction indicating a trend change in the behavior pattern of the monitored object when the rate of change continuously exceeds a preset trend threshold. By executing this procedure, the system can identify and warn of potential risks indicated by cumulative changes in behavior patterns; the statistical baseline channel width multiplier factor. It is not a preset fixed value; its value is determined by a parameter calibration procedure associated with the service level agreement. This procedure is executed during the system initialization phase and includes the following steps: First, the local gateway reads a value representing the target's false alarm tolerance, i.e., the target false alarm rate. This value is a system configuration parameter that defines the acceptable percentage of false alarm events within a specific monitoring period. Subsequently, in background data learning mode, the system collects and stores the quantitative feature values ​​of all logically consistent specific behavioral events within a period (e.g., 30 consecutive days), forming an initial behavioral sample set. Finally, the local gateway calculates the data distribution based on this sample set. For samples with an approximately Gaussian distribution, it queries the inverse cumulative distribution function of the standard normal distribution to find the probability of a two-sided test. The corresponding standard quantiles are used as The value of , or, for non-Gaussian distributed samples identified by built-in association rules or lightweight machine learning models, by calculating the th value of the empirical distribution of that sample. The difference between the quantile and the mean is then divided by the standard deviation to determine the value. This value allows the setting of this key parameter to be directly anchored to quantifiable service quality requirements and the distribution of real data collected on-site.

[0036] Simultaneously, the system also possesses the ability to pre-diagnose the health status of the sensors themselves. This capability is achieved through the analysis of implicit features in the system's normal operation data. To address the slow performance degradation of sensors due to aging or contamination, i.e., a sub-healthy state, the local gateway is configured to execute two parallel pre-diagnostic mechanisms. First, upon each execution of a status query request, it measures and records the end-to-end response delay required for each validator sensor to return its status response, and maintains a statistical model locally for each validator sensor regarding its historical response delay. This model includes the sensor's recent average delay. With recent delay jitter When the system detects a continuous increase in the recent average delay or a continuous amplification in the recent delay jitter of any validator sensor, exceeding the normal range determined by its own historical performance, it generates and outputs a pre-diagnostic instruction indicating that the sensor is in a sub-healthy state. Secondly, for numerical sensors, after each determination of logical consistency, the local gateway calculates the predicted value of its theoretical state response based on the synchronous state response of other sensors in the group. The difference between the predicted value and the actual state response of the sensor is defined as the response residual. The local gateway continuously calculates the energy value of the response residual sequence generated by each numerical sensor over a long period. When the energy value of the response residual sequence of a sensor shows a continuous unidirectional increase and deviates from its own historical baseline to a predetermined degree, another pre-diagnostic instruction indicating that the sensor is in a sub-healthy state due to performance degradation is generated. Through these two parallel pre-diagnostic mechanisms, equipment maintenance is transformed from post-fault repair to predictive management in the early stages of performance degradation.

[0037] To address new logical consistency patterns that may emerge in the system and are outside the preset rule base, the local gateway is also configured with a secondary verification capability. When the result of a weighted logical consistency judgment is a new logical consistency pattern outside the preset rules, the system activates a secondary verification. This verification involves instructing an actuator with light or sound emission capabilities, such as a corridor light, that is physically adjacent to the sensor group to generate a momentary standardized physical disturbance and simultaneously requesting a sensor capable of sensing this physical disturbance to report its status. If the sensor fails to report that it has sensed the physical disturbance within a predetermined time, it can be determined that the sensor has failed or its physical sensing channel has been blocked. Conversely, if it successfully reports, the new logical consistency pattern can be determined as a real abnormal event to be reported. By performing secondary verification, the system obtains a means of actively verifying the unobstructedness of the physical channel, which is used to assist in judging the authenticity of a state combination when encountering a state combination outside the preset rules. The standardized physical disturbances commanded by the mechanism and their corresponding predetermined response times are obtained through an enhanced actuator-sensor pair calibration procedure executed during the system pre-calibration phase. This procedure, while establishing a physical interaction mapping table, measures the following parameters for each valid association pair: the local gateway commands the actuator to output at multiple discrete intensity steps and simultaneously records the response readings of the corresponding sensors, thereby plotting the sensor's response curve and automatically identifying the linear operating region of the curve—the range of commands within which a change in actuator command can cause stable and predictable changes in sensor readings. The standardized physical disturbance is defined as a precise command capable of driving the actuator to produce a state transition of a specific amplitude within this linear operating region. Simultaneously, the system repeatedly executes the standardized physical disturbance command a specific number of times (e.g., 100 times) and records the end-to-end delay from command issuance to the sensor's return of a valid response for each instance. Then, the average value of this set of delay data is calculated. with standard deviation Ultimately, the scheduled response time It is determined deterministically by the following relation: , here This is a dimensionless statistical confidence coefficient, used as a global system configuration parameter, typically ranging from 4 to 6. It defines the fault tolerance level for response timeouts. The execution of this procedure ensures that the judgment basis for secondary verification originates from the measured physical characteristics of each device in its actual deployment environment. To ensure that the system's logical rule base can adapt to the long-term evolution of the environment and user behavior, the local gateway is also configured to perform self-optimization of logical rules. The system associates a dynamically adjustable rule confidence score locally with each logical rule in the logical rule base. When an abnormal event interaction command triggered by a logical rule is not confirmed by any manual intervention or higher-level alarm within a specified observation period, the system considers this as an implicit negative feedback and automatically lowers the confidence score of that logical rule. When the rule confidence score is lower than a preset dormancy threshold, the logical rule is automatically disabled, thereby achieving dynamic management and self-correction of the rule base to ensure the long-term effectiveness of system alarms.

[0038] Example 1: In a residential environment providing security monitoring services for elderly people living alone, the system is deployed as described above. Door magnetic sensors, bedroom infrared detectors, bedside pressure sensors, and bathroom floor water sensors are configured in the same sensor group and managed by a local gateway deployed in the living room. At night, the wireless communication channel in the bathroom experiences unstable signal quality degradation due to electromagnetic interference from other household appliances. Simultaneously, the bedroom infrared detector, due to prolonged operation, enters a sub-optimal state, intermittently failing to detect small movements. At this time, the elderly person slips and falls in the bathroom due to the wet floor and is unable to get up. Under these circumstances, the water sensor's status response data packet is only sent after multiple attempts because the elderly person's body blocks the signal path to the local gateway, resulting in numerous data communication retransmissions. The infrared detector in the bedroom, which detects sub-optimal health, failed to capture the elderly person's movement when they left the bedroom, maintaining a status of no movement. The bedside pressure sensor, however, accurately reported that the bed was unoccupied. After a period of time, the logic rule set in the local gateway regarding prolonged periods of vacancy and inactivity in the bedroom was triggered. The trigger was the bedside pressure sensor, whose status changed to "unoccupied." The local gateway immediately sent a status query request to all validator sensors in the sensor group, including the bedroom infrared detector and the bathroom floor water sensor. The status responses received by the local gateway and their channel conditions were as follows: the bedside pressure sensor's status was "unoccupied," and the bedroom infrared detector's status was "no movement." Both sensor status responses were received only once. All readings were 0, indicating no water accumulation on the bathroom floor water sensor, but its response was retransmitted four times. The value is 4.

[0039] Local gateway based on relational For each state response, a decision weight is determined, where the sensitivity coefficient is... The default value is 0.8. Based on this, the state response weights of the bedside pressure sensor and the bedroom infrared detector are calculated. All are 1, while the state response weight of the bathroom floor water sensor is 1. Then it is The weighted logical consistency judgment is approximately 0.238. When performing the weighted logical consistency judgment, the local gateway assigns the highest confidence level to the "no one in bed" state (weight 1), while correspondingly reducing the contribution of the "no water accumulation" state (weight only 0.238). The system correlates the two high-weighted states—"no one in bed for a long time" and "no movement in the bedroom"—and the result indicates logical inconsistency. The confidence level of this judgment remains high because it is not significantly affected by the unreliable channel and the low-weighted state. Ultimately, based on this logical inconsistency judgment, the local gateway generates and outputs an abnormal event interaction command representing a potential disability risk. Upon receiving this high-confidence command, the central service platform immediately notifies community property management personnel for on-site verification. The property management personnel respond promptly to the command, confirming the elderly person's disability status and providing assistance. This avoids a potential missed event caused by a combination of factors, including sensor sub-optimal health and communication channel uncertainty.

[0040] Example 2: To objectively verify the technical effectiveness of the weighted logic consistency judgment mechanism of the present invention in suppressing invalid alarms caused by single-point sensor failure or unstable communication channels, the following experimental platform was built. This platform simulates a safety monitoring scenario for an elderly person living alone in a residential environment. It includes a local gateway and a sensor group consisting of a trigger sensor S1 (door magnetic sensor), two verifier sensors S2 (infrared detector), and S3 (bedside pressure sensor). The platform injects discrete test events through an event sequence generator and uses a network simulator to programmatically control the communication channel quality between the local gateway and each verifier sensor to simulate data communication retransmission. In the experiment, System B of the present invention was compared with a control group system A that uses traditional Boolean logic judgment. Both systems received the same sensor data input. The logic rule of the control group system A is: if the trigger S1 is on and the state of any verifier S2 or S3 does not conform to the preset normal mode, then an abnormal event command is output. System B of the present invention adopts the weighted logic consistency judgment mechanism disclosed in the aforementioned specific embodiments, wherein the decision weight... From the relation Determine the sensitivity coefficient. The setting affects the system's response to communication channel quality degradation and instantaneous network fluctuations. The setting rule is that when the number of retransmissions on a channel reaches 3 or more, its corresponding decision weight should be suppressed to below 0.4. Based on this, the setting in this experiment is... The value is 0.8.

[0041] Experimental results show that, compared to the control group system A, the present invention system B outputs fewer abnormal event commands when dealing with spurious states caused by communication instability or momentary false alarms from the sensor itself. For example, when injecting an event simulating a user going out normally but the communication channel of the verifier sensor S3 being unstable, the state response of sensor S3 needs to be retransmitted four times. The control group system A directly judges this communication anomaly as a logical inconsistency and generates an invalid alarm, while the present invention system B calculates the decision weight of the state response based on the four retransmissions. With a value of only 0.238, the contribution of S3's state was effectively reduced in the weighted judgment, thus the system was not affected by communication instability and made a logically consistent and correct judgment. Similarly, in another test event, when the verifier sensor S2 gave a momentary false alarm, contradicting the real scene and the states of other sensors, the control group system A directly output an abnormal command, while the logical decision-making mechanism of the present invention system B was able to identify the contradiction based on other high-weight, consistent sensor states and maintain a logically consistent judgment, thereby avoiding another invalid alarm. After the entire test sequence containing 1000 event injections was completed, the two systems were compared. The outputs were statistically analyzed. The control group system A output a total of 215 abnormal event commands, of which 188 were invalid alarms after comparison with the real scenario, resulting in an invalid alarm rate of 87.4%. The present invention system B output a total of 31 abnormal event commands, of which 4 were invalid alarms, resulting in an invalid alarm rate of 12.9%. The experimental data confirms that by determining a decision weight that characterizes the real-time reliability of the communication channel for the state response, and performing a weighted logical consistency judgment based on this weight, the output of invalid event interaction commands caused by single-point failure of sensors or unstable communication links can be reduced without reducing the sensitivity to real abnormal events.

[0042] Example 3: This example combines Figs. 1 to 3 This describes a smart community service interaction system and method based on information collection, such as... Fig. 1As shown, the process begins with a sensor state change trigger event. This event does not directly generate an alarm, but instead initiates a correlation verification, i.e., querying the verifier within the sensor group to obtain the multi-dimensional synchronization status in real time. Next, the system evaluates the reliability of the communication channel, calculates decision weights for each state response by analyzing the stability of data transmission, and fuses these weighted state information. The core of the process lies in a decision node for weighted logical consistency judgment. This node makes a decision based on preset rules and decision weights. If the judgment result is logical inconsistency, it directly outputs an abnormal event interaction command to report to the central platform. Conversely, if the judgment result is logical consistency, the event is judged as a normal event, and its data is stored locally for two parallel analysis processes. The first is behavioral pattern baseline analysis, which updates the statistical baseline by extracting event time-series features and continuously monitors the long-term drift of the baseline channel. Once the trend exceeds a preset threshold, a trend change warning command is output. The second is sensor health pre-diagnosis, which analyzes the long-term statistical characteristics of response delay and response residual. When the delay or residual exceeds the historical range and the characterization performance continues to deteriorate, a sub-health state pre-diagnosis command is output.

[0043] like Fig. 2 As shown, after receiving a status query request, the sensor, acting as the data source, sends its raw status signal / response to the 1.0 data acquisition and weighted sensor data module for processing. The weighted status information generated by this module is input to the 2.0 weighted logical consistency judgment module. This module judges the data based on the logical rules retrieved from the data storage D1 logical rule base. If the judgment result is logically inconsistent, an abnormal event interaction command is generated and sent to the central service platform. If the data is logically consistent, the logically consistent event data is passed to the 3.0 behavior analysis and device status module. This module is responsible for analyzing historical data. On the one hand, it feeds back the historical behavior baseline and updated behavior characteristics obtained from the analysis to the data storage D2 behavior pattern baseline. On the other hand, it feeds back the historical health model and updated health data to the data storage D3 device health model. Based on this analysis, module 3.0 can generate trend change early warning commands and sub-health state pre-diagnosis commands, and also sends these two types of commands to the central service platform.

[0044] like Fig. 3As shown in the figure, the vertical axis represents the reliability score of the rule, and the horizontal axis represents a series of events occurring in chronological order. The figure shows the evolution path of the reliability scores of three rules: the reliability score of the active rule 2 remains at 70 points in all events, presented as a stable solid line, indicating that it is continuously effective. The reliability score of the disabled rule 1 is initially 70 points, but after experiencing events 3, 5 and 7, its score is successively reduced, eventually falling below the dormant threshold and being disabled, presented as a step-like decline and eventually stabilizing dashed line. The reliability score of the recovered rule 3 is also reduced after experiencing events 5 and 6, but its score rebounds after event 13, indicating that it was re-confirmed as effective by the system after a period of low confidence, presented as a dashed line that first declines and then recovers.

[0045] Example 4: In an unattended data center with strict requirements for temperature and humidity, the system of the present invention was deployed to monitor the operating status of critical server racks. Three numerical temperature sensors were arranged at different locations around the same rack. , , And a numerical humidity sensor Together, these sensors form a sensor array to achieve redundant sensing of the micro-environment surrounding the server rack. One technical challenge here is how to optimize the sensors... Before complete failure, the sensor's sub-health state due to performance degradation can be identified in advance through electrical digital data processing; this means that its measured values ​​begin to show small but continuously increasing random deviations or drifts. To achieve this, a pre-diagnostic command based on response residuals needs to be enabled. The prerequisite for executing this command is that the sensor... Establish a theoretical state response prediction model that reflects its intrinsic relationship with other members in the sensor group, and set a non-arbitrary anomaly judgment threshold for the energy value of its response residual sequence. To this end, after the system has been physically installed and running stably for 24 hours, the local gateway performs the following offline calibration procedure once.

[0046] After the procedure is initiated, the local gateway first synchronously collects and records the complete reading sequence of sensors T1, T2, T3, and H1 at 10-second intervals over a continuous period of one hour, thereby obtaining an initial dataset containing 360 synchronous data points. Subsequently, the local gateway uses this dataset to... The reading value is used as the target variable, with , , Using the numerical values ​​as input variables, the standard multiple linear regression algorithm is applied to solve a problem of the form . coefficients in the prediction model After the predictive model is established, the local gateway enters a 72-hour baseline establishment phase. During this period, the system continuously collects data at 1-minute intervals and calculates the baseline using the established predictive model. Theoretical state response prediction value Real-time response every minute The difference between the predicted value and the actual value is defined as the response residual for that event. ,Right now The local gateway further calculates the sum of squares of all response residuals within a given hour, using this sum as the energy value of the response residual sequence for that time period. After the 72-hour baseline period, the local gateway obtained 72 consecutive energy value samples and calculated the average of its historical baseline based on this sample sequence. and standard deviation Finally, the trigger threshold for the pre-diagnostic command used to determine if sensor T1 has entered a sub-healthy state. From the formula Calculations show that after this calibration procedure is completed, the local gateway used for The sensor's pre-diagnostic logic possesses specific model parameters and quantified judgment thresholds, and enters a continuous monitoring operation state. In subsequent operation, when the sensor... The energy value of the response residual sequence exceeds the threshold twice consecutively. At that time, the system generates and outputs an indicator for the sensor. Pre-diagnosis instructions for those in a sub-healthy state.

[0047] Example 5: When the system of the present invention is first deployed in an industrial environment with a specific physical layout and operating rhythm, in order for the local gateway to have the ability to perform secondary verification and apply initial logic rules, the following on-site deployment pre-calibration procedure needs to be executed. After the procedure is started, the local gateway first enters an actuator mapping mode. In this mode, the local gateway will sequentially and independently send instructions to each actuator with light-emitting or sound-emitting function within its jurisdiction, causing it to generate a momentary and standardized physical disturbance. At the same time as each actuator generates a physical disturbance, the local gateway broadcasts a status query request to all sensors within its coverage area and records the sensors that reported status changes during the disturbance. If the warning light... When flashing, only the photosensitive sensor is present. When a corresponding pulse appears in the reading, the local gateway internally establishes and stores a mapping record, which will then be used by the actuator. Its physical disturbance type and the unique sensor that responds The process involves associating all actuators until a sensor-actuator physical interaction mapping table is automatically generated for the specific physical environment, providing a basis for the accurate routing of subsequent secondary verification commands.

[0048] After completing the actuator mapping, the local gateway immediately enters a 48-hour background data learning mode. In this mode, the system only records the complete state response sequence of all events triggered by the trigger sensors and their associated verifier sensors, without performing any logical judgments or alarm outputs. After the learning period ends, the local gateway performs a statistical analysis on all recorded event sequence data and identifies the repetitive state combination patterns that appear in the top 95% of the time through a frequency analysis algorithm. These high-frequency state combination patterns are automatically extracted and solidified by the system into the initial logical rule base for this specific environment, representing the operating mode that has been empirically confirmed as normal in this scenario. Based on this, the local gateway completes initialization and enters normal monitoring operation.

[0049] Example 6: To ensure that the statistical baseline channel for monitoring the behavioral patterns of specific monitored objects and the rule reliability score for logical rule self-optimization in the system of this invention have operable, non-arbitrary operating parameters, after the system is deployed and the initialization and calibration of the aforementioned examples are completed, the following application layer parameter configuration procedure must be executed once more. This procedure transforms a qualitative service level agreement into a set of quantitative algorithm control parameters. In a scenario of monitoring elderly people living alone, if the service level agreement stipulates that when the average daily wake-up time is delayed by more than 30 minutes each week for three consecutive weeks, an early warning should be generated, then the trend monitoring algorithm in the local gateway is configured as follows, which calculates the average value of the daily wake-up time series on a weekly basis. The moving average is calculated, and the first time derivative of the moving average is defined as the drift velocity. The preset trend threshold is set to 30 minutes / week, and the condition for generating the warning instruction is determined to be that the drift velocity is greater than this threshold for three consecutive calculation periods.

[0050] Furthermore, the procedure configures the response characteristics of the logical rule self-optimization mechanism. If the service requirement stipulates that a rule should enter a suppressed state after generating three consecutive abnormal event interaction commands that are not verified by manual intervention, the system parameters are configured as follows: the initial reliability score of all logical rules solidified in the initial stage is set to 70, which is within the range of 0 to 100. The single reduction increment is set to 15 points. Correspondingly, the dormancy threshold for disabling the logical rule is set to 25. Subsequently, if the system encounters the aforementioned three consecutive unverified alarms during operation, the reliability score of the corresponding rule will decrease from 70 to 55, 40, and 25 respectively, and will be automatically disabled after falling below the dormancy threshold.

[0051] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0052] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. An information collection-based intelligent community service interaction system, characterized in that, The system comprises a plurality of sensors, which are pre-divided into at least one sensor group, each sensor group comprising one trigger sensor and at least one verifier sensor defined based on physical proximity or logical correlation, and at least one local gateway in communication with the sensor group, wherein the local gateway is configured to: in response to receiving a trigger signal from the trigger sensor indicating a change in its own state, immediately sending a state query request to all verifier sensors in the sensor group to which the trigger sensor belongs; based on the number of data communication retransmissions required for successfully receiving a state response of each verifier sensor, determining a decision weight representing real-time reliability of a communication channel of the state response; according to a logical rule set in the system initialization phase, performing a weighted logical consistency judgment on the state change of the trigger sensor and the state response of each verifier sensor, wherein the contribution of each state response in the judgment is associated with its corresponding decision weight; only when the result of the weighted logical consistency judgment is logical inconsistency, generating and outputting an abnormal event interaction instruction representing the logical inconsistency, otherwise, if the judgment result is logical consistency, it is determined as a normal event, and its data is stored locally for two parallel analysis processes; and the local gateway is further configured to: for each event determined as logical consistency, extract event timing features associated with a specific monitoring object; based on the event timing features, dynamically calculate and continuously update a statistical baseline channel representing the normal behavior pattern of the monitoring object locally for the monitoring object; the statistical baseline channel is jointly defined by the average value and standard deviation of the specific behavior of the monitoring object in the recent time period; and monitor the long-term one-way drift trend of the statistical baseline channel itself, and when the speed of the drift trend continuously exceeds a preset trend threshold, generate and output a warning instruction indicating that the behavior pattern of the monitoring object has changed in trend. 2.The intelligent community service interaction system based on information collection of claim 1, wherein, The sensor group is configured for the safety scene of the elderly living alone, wherein the door magnet sensor is defined as a trigger sensor, and at least one of the infrared detector, the bedside pressure sensor and the intelligent power equipment state sensor located in the same residential space as the door magnet sensor is defined as a verifier sensor; and the local gateway determines the decision weight of the state response according to the rules, which are defined by the following relationship: , wherein, is the number of data communication retransmissions, is a non-negative integer; is a non-negative sensitivity coefficient set in the system initialization stage for adjusting the influence degree of the retransmission number on the decision weight; the value range of the decision weight is .​ 3.The intelligent community service interaction system based on information collection of claim 1, wherein, The local gateway is further configured to: when the result of the weighted logical consistency judgment is a new logical consistency mode other than a preset rule, activate a secondary verification; the secondary verification comprises: instructing an actuator physically adjacent to the sensor group and having a light-emitting or sound-emitting function to generate a standardized physical disturbance; and synchronously requesting a sensor capable of sensing the physical disturbance to report its state; if the sensor capable of sensing the physical disturbance fails to report its state of having sensed the physical disturbance within a predetermined time, it is determined that the sensor is in an unusable state; if the state of having sensed the physical disturbance is reported, it is determined that the new logical consistency mode is a real abnormal event to be reported. 4.The intelligent community service interaction system based on information collection of claim 1, wherein, The local gateway is further configured to measure and record the end-to-end response delay required for each validator sensor to return its state response each time a state query request is executed; for each validator sensor, maintain a statistical model of its historical response delay locally, which contains the recent average delay and the recent delay jitter of the sensor; when detecting that the recent average delay of any validator sensor is continuously increasing or its recent delay jitter is continuously amplifying, and exceeds the normal range determined by its own historical performance, generate and output a pre-diagnosis instruction indicating that the sensor is in a sub-healthy state.

5. The intelligent community service interaction system based on information collection according to claim 1, characterized in that, The local gateway is further configured to associate a dynamically adjustable rule confidence score with each logical rule in the logical rule library; when the abnormal event interaction instruction triggered by a certain logical rule is not confirmed by any manual intervention or higher-level alarm within the specified observation period, automatically lower the rule confidence score of the logical rule; when the rule confidence score is lower than a preset dormancy threshold, automatically disable the logical rule. 6.The intelligent community service interaction system based on information collection of claim 2, wherein, The logical rules set in the local gateway include: when the state response of the door magnetic sensor is continuously closed for more than twelve hours, and the state responses of at least two of the infrared detector, the bedside pressure sensor, and the intelligent electrical equipment state sensor all indicate no personnel activity or no equipment state change, it is determined that the logic is inconsistent. 7.The intelligent community service interaction system based on information collection of claim 1, wherein, The local gateway is further configured to, for each event determined to be logically consistent, for numerical sensors, calculate the predicted value of the theoretical state response based on the synchronous state responses of other sensors in the group; the difference between the predicted value of the theoretical state response and the true state response of the numerical sensor is defined as the response residual of the current event; for each numerical sensor, continuously calculate the energy value of the response residual sequence generated in a long period; when the energy value of the response residual sequence of a sensor continuously rises in one direction and deviates from its own historical baseline by a predetermined degree, a pre-diagnosis instruction indicating that the sensor is in a sub-healthy state due to performance degradation is generated. 8.The intelligent community service interaction system based on information collection of claim 1, wherein, The local gateway is further configured to process or record all events determined to be logically consistent and their associated raw sensor data locally; only send abnormal event interaction instructions indicating logical inconsistency to a central service platform; thereby limiting raw sensor data to the local network and only selectively reporting abnormal events after cross-validation in the information flow of the system. 9.A smart community service interaction method based on information collection, applied to the smart community service interaction system based on information collection in claim 1, characterized in that, The method is applied to a system comprising a plurality of sensors, the sensors being pre-divided into at least one sensor group, each sensor group comprising a trigger sensor and at least one validator sensor, the method comprising the following steps: Step a, when the state of a trigger sensor changes, immediately send a state query request to all validator sensors in the sensor group to which the trigger sensor belongs, and receive state responses from the validator sensors; Step b, when the state of a validator sensor changes, immediately send a state query request to the trigger sensor in the sensor group to which the validator sensor belongs, and receive a state response from the trigger sensor; Step b, determining a decision weight for each status response of the verifier sensors, based on the number of data communication retransmissions required for successfully receiving the status response; Step c, performing a weighted logical consistency judgment on the status change of the trigger sensor and the status responses of the verifier sensors according to a logical rule set in the system initialization stage, wherein the contribution degree of each status response in the judgment is associated with the corresponding decision weight; Step d, only when the result of the weighted logical consistency judgment is logical inconsistency, generating and outputting an abnormal event interaction instruction representing the logical inconsistency, if the result is logical consistency, determining as a normal event, and its data is stored locally for two parallel analysis processes.

Images