Alarm system and method based on intelligent monitoring of chicken farm

Abstract

The invention discloses an alarm system and method based on intelligent monitoring of a chicken farm, and belongs to the technical field of anti-theft alarm. According to an anti-theft early warning scheme, a corresponding base line is generated and recorded into an anti-theft base line library based on daily data of a perimeter conducting ring, an access control lock spring sensing unit, a footstep pressure array, a chicken flock voiceprint array and feeding trough impedance bait; then carrying out differential operation, extracting an open circuit event, an access control lock spring unlocking sequence abnormity, a footstep pressure array track, voiceprint disturbance and an impedance touch trace, and aggregating into an evidence set; when the primary trigger condition is established, a feeding password is broadcasted, a chicken flock voiceprint response time sequence is recorded, and a feeding password confirmation result is obtained; and matching an intrusion criterion rule in combination with the evidence set and a verification result, generating alarm information containing a trigger type identifier, a trigger position, a monitoring section index and an evidence sequence, finally determining high, medium and low early warning levels according to a grading rule table and the alarm information, and binding the corresponding monitoring section index.

Description

Technical Field

[0001] This invention relates to the field of burglar alarm technology, specifically to an alarm system and method based on intelligent monitoring of chicken farms. Background Technology

[0002] Existing security technologies for livestock farms primarily rely on a combination of perimeter protection facilities and traditional monitoring systems. For example, some farms install infrared beam detectors, video surveillance cameras, electronic fences, and access control sensors along their boundaries to achieve real-time monitoring and recording of personnel and the environment. Simultaneously, the monitoring center typically receives data transmitted from sensors and cameras via wired or wireless communication networks and issues alarms when intrusion signals are detected. While this type of alarm system can improve the farm's security capabilities to some extent, its technical characteristics are relatively simplistic. Most rely solely on trigger signals or image changes from a single sensor to determine intrusion events, lacking multi-dimensional intelligent analysis and dynamic adaptation capabilities. Therefore, in the complex environment of livestock farms, existing technologies suffer from insufficient monitoring accuracy and rigid alarm strategies.

[0003] In actual chicken farm operations, besides natural factors, theft and vandalism are the main threats to farm security. However, existing alarm systems often have a high false alarm rate in such scenarios. These false alarms not only reduce farmers' trust in the alarm system but also increase the manpower costs of security response. Therefore, existing alarm systems based on single-signal triggers are insufficient for effective detection of theft and vandalism in the dynamic and complex environment of chicken farms. There is an urgent need to introduce more intelligent, multi-dimensional sensing and analysis alarm methods to address the shortcomings of existing technologies. Summary of the Invention

[0004] The purpose of this invention is to provide an alarm system and method for intelligent monitoring of chicken farms, thereby solving the problems in the background art: The objective of this invention can be achieved through the following technical solutions: An alarm method based on intelligent monitoring of chicken farms includes: S1: Generate corresponding baselines from the daily data of the perimeter conductive ring, access control spring sensing unit, footstep pressure array, chicken flock voiceprint array and feed trough impedance bait, and write them into the anti-theft baseline library. S2: Perform differential operations on real-time data and the anti-theft baseline database to obtain open circuit events, abnormal access control lock spring unlocking sequences, footstep pressure trajectories, voiceprint disturbances and impedance touch traces, and aggregate them into an evidence set; S3: When the primary triggering condition is met, the feeding command is broadcast, the response sequence of the chicken flock's voiceprint array is recorded, and the consistency of the voiceprint perturbation and footstep array trajectory in the evidence set is checked accordingly to obtain the feeding command verification result. S4: Based on the evidence set and the feeding password verification result, match the intrusion judgment rule and generate alarm information. The alarm information includes trigger type identifier, trigger location, monitoring segment index and evidence sequence. S5: Determine the high, medium or low warning level according to the hierarchical rule table and in combination with the trigger type identifier and evidence sequence in the alarm information, and bind the monitoring segment index; S6: Based on the warning level and monitoring section index, drive the fence electric lock, corridor lighting, buzzer and communication dialer to perform access control locking, lighting guidance, sound and light prompts and communication reporting.

[0005] As a further aspect of the present invention: In step S1, the process of generating corresponding baselines from the daily data of the perimeter conductive ring, the access control lock spring sensing unit, the footstep pressure array, the chicken flock voiceprint array, and the feed trough impedance bait, and writing them into the anti-theft baseline library is as follows: During daily feeding and nocturnal roosting periods when no intrusion is confirmed, data from the perimeter conductive ring, access control spring sensor unit, footstep array, chicken flock acoustic array, and feed trough impedance bait are continuously collected and archived to the raw data storage area. Data is extracted from the original data storage area to generate on / off state sequences, access control lock spring release timing sequences, footstep pressure trajectory segments, chicken flock voiceprint response segments, and feed trough impedance touch curves, and a daily feature table is generated according to a unified field format. Based on the quiet period of nighttime roosting, short-term noise segments are removed, and daily features are aggregated according to time period to obtain on / off templates, detangle sequence templates, gait templates, voiceprint templates and impedance templates. Based on the monitoring section index, channel number, and time period label, each template is solidified into a corresponding baseline, written into the anti-theft baseline library, and the baseline version label and effective identifier are recorded.

[0006] As a further aspect of the present invention: In step S2, the process of performing differential operations on real-time data and the anti-theft baseline database to obtain open circuit events, abnormal access control lock spring unlocking sequences, footstep tracking trajectories, voiceprint disturbances, and impedance touch traces, and aggregating them into an evidence set, is as follows: Real-time data is segmented according to the monitoring segment index and the unified time window, and templates with the same segment and the same time window are retrieved in the anti-theft baseline library to form one-to-one aligned data pairs. Perform differential operations on the aligned data pairs to output on / off deviation markers, de-tethering timing deviation markers, array position deviation markers, acoustic frequency band deviation markers, and impedance touch deviation markers. A circuit break event is generated based on the on / off deviation marker; an access control lock spring release sequence anomaly is generated based on the release timing deviation marker; and a foot pressure trajectory is generated based on the pressure position deviation marker. Based on the acoustic frequency band deviation mark, acoustic disturbance is generated; based on the impedance touch deviation mark, impedance touch trace is generated; and the monitoring segment index and time window number are bound together. Using the monitoring section index as the primary key, the evidence set is generated by aggregating circuit breakers, abnormal access control lock spring unlocking sequences, footstep tracking trajectories, voiceprint disturbances, and impedance touch traces in chronological order according to time windows.

[0007] As a further aspect of the present invention: in step S3, the process of obtaining the feeding command verification result is as follows: After determining that the primary triggering condition is met, the feeding command is broadcast through the loudspeaker according to the monitoring segment index, and the start time of the command is recorded. The chicken flock's voiceprint array records the start time and sequence of responses, while the footstep array synchronously captures the footstep trajectory within the same time window. Based on the start time of the password, the time window of the voiceprint disturbance and the footstep trajectory is aligned to verify the consistency of the monitoring section and generate a consistency judgment result. Based on the consistency determination results and the monitoring segment index, a feeding password verification result is generated and bound to the relevant evidence sequence number.

[0008] As a further aspect of the present invention: the specific content of determining that the primary triggering condition is met is as follows: When a circuit breaker event or an abnormal access control lock spring release sequence occurs, or when footsteps and voiceprint disturbances, footsteps and impedance touch traces, or voiceprint disturbances and impedance touch traces are present simultaneously in the same monitoring segment and time window, the primary triggering condition is determined to be met.

[0009] As a further aspect of the present invention: in step S4, the process of matching the intrusion judgment rules and generating alarm information based on the evidence set and the feeding password verification result is as follows: Align the evidence set with the feeding command inquiry results according to the monitoring segment index and time window, and merge them into an evidence mapping table; Based on the feeding password verification results, locate the corresponding rule branch and its matching field in the intrusion criterion rules; The rule fields are compared item by item in the evidence mapping table to generate a trigger type identifier, and the trigger location is determined based on the monitoring segment index. Organize the evidence entries for the monitored area in chronological order of the time window, generate an evidence sequence, and associate it with the trigger type identifier; Integrate trigger type identifier, trigger location, monitoring segment index, and evidence sequence to generate structured alarm information.

[0010] As a further aspect of the present invention: In step S5, the process of determining the high, medium, or low warning level according to the hierarchical rule table and combining the trigger type identifier and evidence sequence in the alarm information, and binding the monitoring segment index, is as follows: Analyze alarm information to extract trigger type identifiers, evidence sequences, and monitoring segment indexes, which serve as inputs for hierarchical judgment. Based on the trigger type identifier, locate the corresponding rule entry in the hierarchical rule table and obtain the evidence verification requirements and warning level mapping for that entry; The evidence sequence is verified item by item according to the rules to confirm the consistency between the evidence and the time window, and to determine the high, medium or low warning level. The warning level is bound to the monitoring segment index, and a graded record is generated together with the trigger type identifier and evidence sequence.

[0011] An alarm system based on intelligent monitoring of chicken farms includes: The baseline modeling module is used to generate corresponding baselines from the daily data of the perimeter conductive ring, access control lock spring sensing unit, footstep pressure array, chicken flock voiceprint array and feed trough impedance bait, and write them into the anti-theft baseline library. The anomaly extraction module is used to perform differential operations on real-time data and the anti-theft baseline library to obtain open circuit events, abnormal access control lock spring unlocking sequences, footstep pressure trajectories, voiceprint disturbances and impedance touch traces, which are aggregated into a set of evidence. The password verification module is used to broadcast the feeding password when the primary triggering condition is met, record the response sequence of the chicken flock's voiceprint array, and perform consistency verification on the voiceprint perturbation and footstep array trajectory in the evidence set based on the response sequence to obtain the feeding password verification result. The rule matching module is used to match intrusion criteria rules based on the evidence set and the feeding password query results, and generate alarm information. The alarm information includes trigger type identifier, trigger location, monitoring segment index and evidence sequence. The early warning classification module is used to determine the high, medium or low warning level according to the classification rule table and in combination with the trigger type identifier and evidence sequence in the alarm information, and bind the monitoring segment index. The linkage execution module is used to drive the fence electric lock, corridor lighting, buzzer and communication dialer according to the warning level and monitoring section index to perform access control locking, lighting guidance, sound and light prompts and communication reporting.

[0012] The beneficial effects of this invention are: This invention significantly reduces the false alarm rate of alarm systems in chicken farms by introducing a multi-dimensional sensing unit and baseline database comparison mechanism. In its implementation, multiple sensing methods, including perimeter conductive rings, access control spring sensing units, footstep patterns, chicken acoustic signature arrays, and feed trough impedance baits, are used to collect baseline data during daily operation and write it into an anti-theft baseline database, providing a standard reference for subsequent data analysis. Circuit breakers, abnormal access control spring unlocking sequences, footstep patterns, acoustic signature disturbances, and impedance touch traces obtained through differential calculations can be compared temporally with the anti-theft baseline. This allows the alarm system to no longer rely on the triggering results of a single sensor, but rather generate an evidence set through the aggregation of multi-source data. Compared with existing technologies, this invention effectively overcomes the false alarm problem caused by wind disturbances or the chickens' own activities in traditional alarm methods, enhancing the accuracy of intrusion detection. In particular, the response characteristics based on the chicken flock's voiceprint array can be used to verify the consistency of abnormal voiceprint disturbances, thereby identifying whether there is human interference. This mechanism improves the accuracy of theft and vandalism detection and ensures that alarm information has sufficient credibility.

[0013] Furthermore, this invention achieves tiered early warning and intelligent response control through joint matching of evidence sets and feeding command query results. The alarm system can not only generate structured alarm information based on trigger type identifiers, trigger locations, monitoring segment indexes, and evidence sequences, but also classify alarm events into high-level, medium-level, or low-level categories according to a tiered rule table and bind them to specific monitoring segments. In this way, farm managers can quickly understand the severity and location of the alarm upon receiving it, facilitating targeted responses. Attached Figure Description

[0014] The invention will now be further described with reference to the accompanying drawings.

[0015] Figure 1 This is a flowchart illustrating an alarm method for intelligent monitoring of chicken farms according to the present invention.

[0016] Figure 2 This is a schematic diagram of the structure of an alarm system based on intelligent monitoring of chicken farms according to the present invention. Detailed Implementation

[0017] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0018] Example 1 Please see Figure 1 As shown, this invention is an alarm method based on intelligent monitoring of chicken farms, comprising: S1: Generate corresponding baselines from the daily data of the perimeter conductive ring, access control spring sensing unit, footstep pressure array, chicken flock voiceprint array and feed trough impedance bait, and write them into the anti-theft baseline library. S2: Perform differential operations on real-time data and the anti-theft baseline database to obtain open circuit events, abnormal access control lock spring unlocking sequences, footstep pressure trajectories, voiceprint disturbances and impedance touch traces, and aggregate them into an evidence set; S3: When the primary triggering condition is met, the feeding command is broadcast, the response sequence of the chicken flock's voiceprint array is recorded, and the consistency of the voiceprint perturbation and footstep array trajectory in the evidence set is checked accordingly to obtain the feeding command verification result. S4: Based on the evidence set and the feeding password verification result, match the intrusion judgment rule and generate alarm information. The alarm information includes trigger type identifier, trigger location, monitoring segment index and evidence sequence. S5: Determine the high, medium or low warning level according to the hierarchical rule table and in combination with the trigger type identifier and evidence sequence in the alarm information, and bind the monitoring segment index; S6: Based on the warning level and monitoring section index, drive the fence electric lock, corridor lighting, buzzer and communication dialer to perform access control locking, lighting guidance, sound and light prompts and communication reporting.

[0019] In step S1, the process of generating corresponding baselines from the daily data of the perimeter conductive ring, access control lock spring sensing unit, footstep pressure array, chicken flock voiceprint array, and feed trough impedance bait, and writing them into the anti-theft baseline library is as follows: Data collection was conducted under the premise of no intrusion. This embodiment selected two relatively stable time periods: one was the daily feeding period (e.g., 6:30-7:30 and 16:30-17:30 daily), and the other was the nocturnal roosting period (e.g., 20:00-05:00 the following day). To ensure no intrusion, cross-verification was performed using video playback of the past seven days and historical alarm records to confirm that no one entered the chicken coop or enclosure during these periods. Subsequently, the on / off current values ​​of the perimeter conductive ring, the force and unlocking displacement of the access control spring sensor unit, the pressure distribution matrix of the footstep array, the audio signal of the chicken flock's vocalization array, and the combined resistance-capacitance change value of the feed trough impedance bait were continuously collected at a uniform sampling frequency (e.g., 10 times per second). The collected data was written to the "raw data storage area" in real time in the format of "timestamp-monitoring segment index-channel number-raw reading". If a short-term anomaly occurs during the data collection process (such as a sudden strong wind causing the fence to sway slightly, or a rodent passing by causing a momentary change in a certain pressure point), only an anomaly is marked without triggering an alarm, and the original record is retained for subsequent cleaning.

[0020] After extracting data from the original data storage area, feature processing is performed to form a unified daily feature table. Specifically: For the perimeter conductive ring, the continuous electrical signal is discretized into an "on-off" state sequence based on the current threshold, similar to writing the state of a switch throughout the day as a series of "1 / 0" time trajectories; for the access control spring sensing unit, the start time, peak displacement, and reset time of the unlocking action are extracted to generate two key fields: "unlocking sequence" and "intensity," used to distinguish between normal door opening by a feeder and abnormal prying; for footsteps, the two-dimensional pressure matrix is ​​spliced ​​on the time axis to identify the continuous landing points of the gait and the movement path of the force center, thus obtaining a "footstep trajectory segment," which can be understood as a "light-heavy-light" force trace curve left by a person walking on a carpet; for the chicken flock's voiceprint array, according to the chicken flock's response pattern to feeding commands or fixed environmental sounds (such as daily timed feeding sounds), "voiceprint response segments" are cut out, and parameters such as the dominant frequency, resonant peak spacing, and energy envelope are extracted; for the feed trough impedance bait, the impedance touch curve caused by hand or beak contact is recorded, including the rising slope, peak impedance, and recovery time. Finally, a structured daily characteristic table is generated using time tag, segment index, channel number, on / off status, detangle timing, gait segment, voiceprint segment, impedance curve, and anomaly marker as unified fields. For example, if a normal feeding door opening occurs at 7:05 AM in segment A1, channel 03, the record will show: the on / off sequence is continuously "on", the detangle timing has a single peak and resets quickly, the gait segment has a "short step, fast frequency" characteristic, the voiceprint segment has obvious group calls, and the impedance curve shows a rapid drop after a short touch.

[0021] Template aggregation was performed based on the nocturnal roosting quiet period to obtain a typical and stable quiet baseline. First, noise reduction was performed: short-duration noise segments with durations below a set threshold (e.g., 1-2 seconds) were removed, and each channel was smoothed using a sliding window to avoid occasional spikes affecting the overall judgment. Second, aggregation was performed by time period: the period from 20:00 to 05:00 each day was divided into several time blocks (e.g., every 30 minutes), and the stable patterns of each feature were statistically analyzed. For the perimeter conductive ring, the on / off template during the quiet period should maintain a flat shape of "continuous on"; for the access control spring, the unlocking sequence template during the quiet period should be empty or only have weak fluctuations generated by device self-testing; for footsteps, the gait template showed a low-frequency, scattered distribution, reflecting the small pressure changes caused by the occasional movement of the perch by the flock at night, without exhibiting "continuous, directional, and heavy" human gait trajectories; for the voiceprint template, the dominant frequency energy should be significantly lower than during the day, and there should be no group response peak to commands; for the impedance template, the curve should be mainly stable at the baseline, with occasional short touches. This "noise removal-blocking-statistics" method summarizes complex and messy data into "on / off templates, detuning sequence templates, gait templates, voiceprint templates, and impedance templates." For example, the voiceprint template for the 02:00-02:30 time period often shows slight fluctuations in background noise without group call peaks, which becomes the reference standard for judging abnormal voiceprint disturbances at night.

[0022] Finally, the templates are solidified according to the monitoring segment index, channel number, and time period label, and written into the "anti-theft baseline library," while recording the "baseline version label" and "effectiveness identifier." During implementation, a unique version label is assigned to each baseline (e.g., "V2025-09-A1-03-Nighttime 1"), and the "effectiveness identifier" is only set to "effective" after the new template passes quality verification. Quality verification includes: whether the data coverage meets the standard (e.g., complete sampling for more than three consecutive days), whether the template stability meets the threshold (e.g., the variance of each statistical indicator is lower than the set value), and consistency comparison with historical effective baselines (to prevent erroneous updates due to occasional environmental changes). If the verification fails, the previous version remains effective to avoid affecting subsequent real-time comparisons and alarms. When writing to the library, the "monitoring segment index" (e.g., A1, A2, etc.), "channel number" (e.g., 03 corresponds to a certain conductive ring or a certain pressure plate), and "time period label" (e.g., "daytime feeding one," "nighttime roosting two") are also associated to ensure that the corresponding reference baseline can be quickly found at any time and location. In this way, when performing real-time differential calculation in step S2, the corresponding five types of baselines, namely "on / off, detangle, gait, voiceprint, and impedance", can be accurately called according to the current time period and specific channel, thereby improving the reliability and interpretability of anomaly identification; once a deviation occurs, the version and parameters used can be traced back.

[0023] In step S2, the process of performing differential operations on real-time data and the anti-theft baseline database to obtain open circuit events, abnormal access control lock spring unlocking sequences, footstep tracking trajectories, voiceprint disturbances, and impedance touch traces, and aggregating them into an evidence set, is as follows: In this embodiment, the real-time data streams from each sensor channel are first categorized according to the "monitoring segment index" (e.g., A1, A2, etc.), and then sliced ​​using a unified time window, for example, each window is set to 60 seconds and updated on a rolling basis. To reduce clock errors, the same time source is used to correct the timestamps within each segment, and linear interpolation is performed on the sampling points when necessary, so that the sampling points of different sensors fall into the same window boundary. Subsequently, templates for "same segment, same time window label"—including on / off templates, unlocking sequence templates, gait templates, voiceprint templates, and impedance templates—are retrieved from the anti-theft baseline library to form a one-to-one data pair with the real-time data of that window. If a window appears for the first time and there is no template yet, it reverts to the "adjacent window template" or "historical template of the same time period" in the same segment and is marked as "low confidence alignment"; it is automatically replaced when a new template is generated in the next round of offline training. For example, the real-time slice of segment A1 from 20:00 to 20:01 will be aligned with the baseline template of segment A1 during the night roosting period from 20:00 to 20:01, thus providing a stable reference for subsequent differential comparisons.

[0024] After alignment, differential analysis is performed on each type of sensing feature. For the perimeter conductive loop, the real-time "on-off" sequence is compared with the continuous segment of the template. If an inconsistency with the template occurs and continues for more than a set threshold (e.g., 2 seconds), an "on-off deviation mark" is generated. For the door lock spring release timing, key parameters such as the start time of the action, peak amplitude, duration, and reset gradient are compared. If any indicator deviates from the template tolerance band (e.g., ±10%) and the combinational logic is abnormal, an "release timing deviation mark" is output. For footsteps, the spatial difference between the force center trajectory and the template trajectory is calculated, and the changes in stride length, rhythm, and force peak are considered. If the trajectory deviation is continuous, highly directional, and inconsistent with the "nighttime sporadic movement" template, an "array position deviation mark" is output. For the chicken flock voiceprint array, the energy distribution of the main frequency bands is compared with the spectral envelope of the template. If a sudden increase in frequency bands outside the template, continuous human voice characteristics, or misalignment of the group response timing occurs, it is recorded as a "voiceprint frequency band deviation mark". For trough impedance decoys, compare the rising slope, peak value, and recovery time of the touch curve. If a wide peak resembling a "long press" or multiple repeated touches are formed while the template is "silent," it is marked as an "impedance touch deviation mark." For example, if the pressure pattern within a window shows continuous "heavy-heavy-heavy" landing points and a stride similar to an adult gait, while the template shows "slight scattered points," it will be judged as a significant deviation.

[0025] After obtaining various deviation markers, they are converted into event or trajectory objects with clear semantics. For continuity deviations, if the conductive loop experiences continuous disconnections without self-test records, an "open circuit event" is generated, along with the start and end times and estimated length. For access control springs, if the unlocking parameter combination differs significantly from the "normal door opening" template (e.g., excessively long duration, slow reset, or abnormal unlocking count), an "access control spring unlocking sequence anomaly" is generated, and suspected method characteristics (such as multiple slight attempts) are recorded. For footsteps, discrete "footstep position deviation markers" are concatenated into a continuous "footstep trajectory," providing the trajectory segment, estimated stride length, rhythm, and direction. For example, if segment A2 continuously exhibits eastward force center migration, a stride length of approximately 0.7 meters, and a stable rhythm within the 23:15–23:16 window, a "suspicious walking trajectory" will be formed; however, normal nighttime chicken activity would not form such a regular trajectory.

[0026] For deviations in both voiceprint and impedance, "voiceprint disturbance" and "impedance touch trace" objects are generated respectively. Voiceprint disturbances include the frequency band range, peak energy, duration, and tags indicating whether it resembles a human voice or is related to feeding commands. Impedance touch traces include the number of touches, the peak value of each touch, and its duration, used to distinguish between "short pecking" and "human touch." All objects are bound to a "monitoring segment index" and "time window number" upon creation, facilitating subsequent cross-source association and tracing. For example, if a sustained and significant peak is detected near 1.8 kHz in the A1 segment's 22:30–22:31 window, accompanied by harmonic structures not unique to chickens, a "voiceprint disturbance" will be generated, along with a "suspected whistle or human voice" determination. If two high-impedance contacts, each lasting more than 3 seconds, are simultaneously recorded in the feed trough impedance channel, two "impedance touch traces" will be generated. This type of simultaneous "sound-touch" interaction is common in scenarios where people approach the feed trough to investigate.

[0027] Finally, using the "monitoring segment index" as the primary key, various objects generated in consecutive windows within the same segment and time period are aggregated into an "evidence set" in chronological order. The aggregation process includes: removing duplicates (e.g., only retaining overlapping segments of the same trajectory across two windows), associative integration (e.g., associating "voiceprint disturbance" and "resistance touch trace" within the same time period as the same chain), and weighted sorting (e.g., the weight of a circuit breaker event is higher than that of a single touch trace). The generated evidence set uses windows as sequence units, clearly showing the evolutionary relationship from "precursor-trigger-continuation-end," facilitating subsequent verification and graded early warning. For example, if "voiceprint disturbance," "resistance touch trace," and "footsteps" appear successively within three minutes from 22:30 to 22:33 in segment A1, without a normal feeding command, it should be considered a continuous suspicious link; the evidence set records this link completely as a time sequence of "sound, touch, movement," and presents it together with any possible "circuit breaker event" or "access control anomaly."

[0028] In step S3, when the primary triggering condition is met, a feeding command is broadcast, the response sequence of the chicken flock's voiceprint array is recorded, and a consistency check is performed on the voiceprint perturbation and footstep trajectory in the evidence set to obtain the feeding command verification result. Upon meeting the initial triggering conditions, the loudspeaker within the designated monitoring segment is immediately selected based on the "monitoring segment index" for directional broadcasting to avoid interference with other segments. Feeding commands use fixed and clear Chinese phrases, such as "assemble" or "feed," to reduce ambiguity. A unified timestamp is generated and recorded as the "command start time" at the same moment the command control instruction is issued. To ensure time consistency between different devices, time synchronization is performed before broadcasting, and the "command end time" is recorded at the end of the broadcast. If multiple loudspeakers exist within a segment, the one closest to the feeding trough is selected according to the priority table, and its number and volume level are written into the log. For example, if a "door lock spring unlocking sequence abnormality" occurs in segment A1, the command is broadcast on loudspeaker 01 in segment A1, and the command start time is recorded as 22:31:05, facilitating subsequent alignment and verification of the flock's responses and footsteps.

[0029] Specifically, the primary triggering conditions determine whether to initiate "password verification." The specific criteria are as follows: First, if an "open circuit event" is detected (such as a persistent disconnection of the perimeter conductive ring), the condition is triggered. Second, if an "abnormal access control spring release sequence" occurs (such as excessively long release duration, repeated releases without resetting), the condition is also triggered. Third, if "footstep patterns" and "voiceprint disturbances" (potentially indicating someone approaching while making a sound) are observed simultaneously within the same monitoring segment and time window, the condition is also considered triggered. Fourth, if "footstep patterns" and "impedance touch marks" occur simultaneously (potentially indicating approaching and touching the feed trough or fence), the condition is triggered. Fifth, if "voiceprint disturbances" and "impedance touch marks" occur concurrently (potentially indicating someone making a sound while touching the equipment), the primary triggering condition is activated; subsequently, a password is broadcast for verification. If the flock does not respond normally, and the strait continues to move towards the exit, it will likely be judged as "inconsistent" in subsequent steps, providing strong evidence for tiered early warning and follow-up handling. Through the above mechanism, the chain of "initial trigger - password verification - consistency judgment - result binding" is closed, which can quickly eliminate disturbances caused by normal feeding and promptly identify suspicious intrusion behavior.

[0030] After the command is given, the chicken flock's voiceprint array enters "response monitoring" mode, automatically marking the time of the first obvious response peak as the "response start time." Simultaneously, it continuously records the response timing and intensity changes over the next few seconds (e.g., 10-20 seconds), forming a "response timing curve." To eliminate environmental noise interference, the stability of the dominant frequency, harmonic structure, and energy envelope are considered in the acoustic characteristics. A response is only confirmed when all these indicators meet a threshold. Simultaneously, the footstep tracking module captures data from the same time window overlapping with the command, outputting the footstep tracking trajectory, including the movement path of the force center, stride length, and rhythm. Intuitively, this step is like "listening to the sound and observing the reaction" while simultaneously "observing the ground and observing the footsteps": if the flock gathers due to the command, the tracking diagram will show multiple slight convergences near the feeding trough; if someone approaches, the trajectory will show a regular, highly directional path with greater force. For example, at 22:31:07, the acoustic array detected a short, abrupt peak in the group response, while the pressure array showed scattered clusters in front of the feed trough, which is usually considered a normal feeding response.

[0031] Using the "command start time" as a benchmark, the time windows of voiceprint perturbation and footstep tracking are aligned. If there is a slight drift in the timestamps of the two data sources, correction is performed within the allowable tolerance range (e.g., ±2 seconds). Subsequently, it is verified whether the two types of data come from the same "monitoring segment index" to avoid overlapping situations such as "the command is broadcast in segment A1, but the response comes from segment A2". After alignment, a "consistency judgment result" is given based on temporal matching degree, spatial aggregation degree, and behavioral pattern matching degree, which can be divided into three types: "consistent", "inconsistent", and "insufficient evidence". For example, if a chicken flock-specific response peak appears within 2-5 seconds after the command, and the footstep tracking forms multiple slight aggregations around the feed trough, and both come from the same segment of the command broadcast, it is judged as "consistent"; conversely, if the voiceprint shows long-term human voice characteristics while the footstep tracking shows a single path quickly passing through the feed trough area, it is judged as "inconsistent".

[0032] After obtaining a consistency determination, a "feeding command verification result" is generated, which includes at least: monitoring segment index, command start and end times, response start time and response sequence summary, footstep trajectory summary (such as stride length, rhythm, and convergence position), consistency level, and associated evidence sequence number. The evidence sequence number is used to point to the "circuit breaker event, access control lock spring unlocking sequence anomaly, footstep trajectory, voiceprint disturbance, and impedance touch trace" formed in step S2, realizing bidirectional traceability between "command verification" and "original evidence." A brief explanation is also provided: if "consistent," it indicates that the current behavior is closer to a normal feeding pattern; if "inconsistent," it suggests abnormal intervention; if "insufficient evidence," it is recommended to extend the monitoring time or repeat the command for secondary verification. For example, the verification result of section A1 shows "consistent" and binds evidence numbers A1-20250915-223105-01 (voiceprint) and A1-20250915-223105-02 (backup), which can be directly referenced in the rule matching stage later.

[0033] In step S4, based on the evidence set and the feeding password verification result, the intrusion judgment rules are matched to generate alarm information. The alarm information includes a trigger type identifier, trigger location, monitoring segment index, and evidence sequence. The process is as follows: In this embodiment, the monitoring segment index and time window number are used as a unified primary key to strictly align the evidence set formed in step S2 (including circuit breakers, abnormal access control lock spring unlocking sequences, footstep tracking, voiceprint disturbances, and impedance touch traces) with the feeding password query results output in step S3. The alignment strategy is as follows: first match the segment index, then match the time window; if there is a slight deviation in the window boundary, it is merged within the allowable tolerance (e.g., ±2 seconds). After alignment, the two types of information are merged into an evidence mapping table, where each row corresponds to a specific segment and window. Fields include: window timestamp, evidence item list, password consistency level (e.g., consistent / inconsistent / insufficient evidence), and the numbers that the evidence items point to. For example, the mapping row of segment A1 in the window from 22:31:00 to 22:31:30 contains both a summary of "footstep tracking, voiceprint disturbances, and impedance touch traces" and the conclusion that "password consistency = inconsistency," facilitating subsequent one-stop comparison and judgment.

[0034] The system prioritizes reading the feeding password verification results from this window to select the corresponding branch in the intrusion criterion rule base. The rule base contains at least three categories: "password consistency branch," "password inconsistency branch," and "insufficient evidence branch." When the passwords match, the system favors normal feeding or flock gathering scenarios. When the passwords don't match, it applies strict criteria, focusing on combinations of fields such as "human gait," "abnormal touch," "access control anomaly," and "perimeter disconnection." When the evidence is insufficient, a conservative strategy is adopted, and the system prompts for continued observation. After locating a branch, the system automatically loads the set of matching fields to be examined for that branch, such as "whether there is a continuous human gait trajectory," "whether there is a long period of high-resistance touch," "whether the access control unlock exceeds the threshold," "whether the perimeter is continuously disconnected," and "whether the voiceprint contains human voice characteristic peaks," and assigns a threshold and weight to each field. In this way, locating the rule branch is like selecting a "special ruler," which will then be used to measure the evidence in the mapping table item by item.

[0035] After determining the rule branch, each field in the "Evidence Mapping Table" is compared. If the deviation of the perimeter conductive loop from the open / closed state reaches a sustained threshold, a "Trigger Type Identifier = Open Circuit" is generated. If the access control unlocking duration exceeds the limit or there are multiple abnormal attempts, it is identified as "Access Control Abnormality". If the tracking trajectory is regular, directional, and the stride is close to the adult range, accompanied by a human voice characteristic peak or impedance long-press curve in the voiceprint, it is classified as "Suspicious Intrusion". When multiple conditions occur concurrently, the primary trigger type is output according to the priority strategy, while secondary labels are retained to ensure clear meaning and no conflict. The trigger location is determined based on the "Monitoring Segment Index" combined with the geographical mapping relationship of the segment. For example, segment A1 corresponds to "Eastern Section of the North Fence". For example, if "Human Gait + Long-Press Impedance + Inconsistent Password" appears in the window at 22:31 of segment A1, the primary trigger type is identified as "Suspicious Intrusion", and the trigger location is located as "A1 - Eastern Section of the North Fence".

[0036] Subsequently, evidence items within the same segment, occurring consecutively over time, are sequentially organized to form an "evidence sequence." The organization rules are as follows: sorting by window sequence; splicing trajectories that extend across windows; chaining mutually pointing "voiceprint-touch-gait" items; and deduplicating and merging duplicate or overlapping items. The sequence head typically represents a precursor (e.g., weak voiceprint disturbance at a distance), the middle section represents the main behavior (e.g., a stable human gait approaching and triggering a long-press touch), and the tail may represent the result (e.g., access control anomaly or perimeter circuit failure). This evidence sequence is bound to a "trigger type identifier," forming a clear chain of "type-process-result." Taking A1 as an example, from 22:30 to 22:33, "voiceprint disturbance, footsteps tracing a path, impedance long press, and access control anomaly" appear sequentially. This sequence is organized into a complete sequence and saved along with the "suspicious intrusion" trigger type for later review and verification.

[0037] Finally, the key elements are integrated into a structured alarm message. This message contains at least four items: 1. Trigger type identifier (e.g., circuit breaker, access control anomaly, suspicious intrusion); 2. Trigger location (mapped from section to specific location, such as "A1—East section of north fence"); III. Monitoring segment index (e.g., A1); IV. Evidence sequence (evidence items arranged by time window and their numbers).

[0038] To facilitate understanding, alarm information is output in a structured field format, allowing for both automatic processing by the platform and quick manual interpretation. Example: Trigger type identifier: Suspicious intrusion; Trigger location: A1—East section of the north fence; Monitoring segment index: A1; Evidence sequence: [22:30—Voiceprint disturbance (No. A1-223000-01); 22:31—Footprints (No. A1-223100-02); 22:31—Impeded touch trace (No. A1-223100-03, press and hold for 3 seconds); 22:32—Abnormal access control lock spring unlocking sequence (No. A1-223200-04)]

[0039] In step S5, the process of determining the high, medium, or low warning level according to the hierarchical rule table and combining the trigger type identifier and evidence sequence in the alarm information, and binding the monitoring segment index, is as follows: In this embodiment, the classification module first receives the structured alarm information generated in step S4 and parses its fields. It extracts three key input items in a predetermined order: first, a "trigger type identifier," used to describe the main cause of the alarm (e.g., circuit breaker, access control anomaly, suspicious intrusion, etc.); second, an "evidence sequence," a list of evidence items arranged by time window (including evidence number, source channel, start and end times, intensity index, and confidence level); and third, a "monitoring segment index," used to pinpoint the physical location of the alarm. To avoid misjudgment across segments, the parsing phase verifies whether the segment labels of all items in the "evidence sequence" match the segment index in the alarm information. If cross-segment items exist, they are marked as "requires manual review" or "evidence removed" before proceeding to classification. For example, if the alarm information shows that the trigger type is identified as "suspicious intrusion", the monitoring segment index is A1, and the evidence sequence is "22:30 voiceprint disturbance, 22:31 footsteps, 22:31 impedance long press, 22:32 access control abnormality", then the hierarchical module will cache these four pieces of evidence along with segment A1 in chronological order as input for subsequent rule comparison.

[0040] After parsing, the corresponding rule entry is located in the "Hierarchical Rule Table" using the "Trigger Type Identifier" as the search key. The rule table uses rows to represent trigger types and columns to show the mapping relationship between the required evidence verification requirements and the warning level. Taking "Suspicious Intrusion" as an example, its rule entry typically requires: the existence of a continuous human gait trajectory or equivalent strong evidence; at least one piece of contact-related evidence (such as impedance long press reaching a threshold); and the formation of a causal chain with voiceprint anomalies or access control anomalies within a limited time window. This entry also provides mappings from "meeting all conditions, high-level warning," "meeting some key conditions, medium-level warning," and "meeting only weak conditions, low-level warning." For "Circuit Break" triggers, the rule entry will emphasize the duration and scope of impact; for "Access Control Anomalies" triggers, it will place more emphasis on parameters such as the duration of unlocking, the number of repeated attempts, and the failure to reset. After locating the entry, the corresponding numerical threshold (such as continuous stride range, lower limit of touch duration in seconds, and time window merging width) and weight coefficient will be loaded simultaneously as the benchmark for subsequent item-by-item verification.

[0041] In this step, the evidence sequence is verified item by item and its temporal consistency is checked according to the identified rule entries. First, a "temporal consistency" check is performed: based on the start and end times of the evidence, it is determined whether multi-source evidence appears consecutively within the specified merging window (e.g., the start and end interval between any two key pieces of evidence does not exceed 90 seconds, and the overall link is not interrupted by a window exceeding a threshold). Second, a spatial consistency check is performed: the monitoring segment index and channel number of the evidence should be consistent. If the trajectory crosses into an adjacent segment, it needs to extend in the same direction and meet the additional condition of "cross-segment intrusion" to be counted. Third, a "strength consistency" check is performed: such as stride length and rhythm falling within the adult gait range, impedance touch duration reaching the long press criterion, significant human voice characteristic peaks and obvious differences from chicken responses, and continuous exceeding of access control unlocking limits. The above verification results are weighted and summarized into a comprehensive score, and then the warning level is determined according to the mapping given by the rule entries. For example: If segment A1 continuously exhibits "human gait trajectory, impedance long press ≥3 seconds, access control anomaly" within three minutes, and the feeding command verification result is "inconsistent," then it satisfies the strong causal chain and strength criterion, and is directly judged as "high-level warning"; if only "human gait trajectory + voiceprint disturbance" occurs and the duration is short, it can be judged as "medium-level warning"; if there is only a single voiceprint disturbance without other supporting evidence, it is "low-level warning." In cases of insufficient evidence or fragmented time sequence, the level will be downgraded, with a "continued observation recommended" prompt.

[0042] After the level determination is completed, a level record is generated, and the "warning level" is strongly bound to the "monitoring segment index" to ensure that the specific area and equipment can be quickly located in subsequent handling steps. The level record uses structured field output, which includes at least: monitoring segment index, warning level (high / medium / low), trigger type identifier, evidence sequence details (listed by time window), verification point summary (e.g., "time consistency: pass; spatial consistency: pass; intensity consistency: gait significant, long press meets standard, deduction exceeds limit"), judgment basis version (corresponding to rule table version number), and generation time. Taking A1 as an example, the hierarchical record can be presented as follows: the segment is A1; the warning level is high; the trigger type is suspicious intrusion; the evidence sequence is [22:30 voiceprint disturbance number A1-01, 22:31 footsteps tracking number A1-02, 22:31 impedance long press number A1-03, 22:32 access control anomaly number A1-04]; the verification point is "a complete link is formed within 180 seconds, and both the strength and timing meet the high threshold"; the rule table version is "V2025-09"; and the generation time is "2025-09-15 22:33". This hierarchical record will serve as the direct input for the subsequent step S6, driving corresponding access control locking, lighting guidance, audio-visual prompts, and communication reporting actions, achieving a smooth connection and closed-loop management from "alarm generation" to "handling execution".

[0043] Example 2 Please see Figure 2 As shown, the present invention also provides an alarm system based on intelligent monitoring of chicken farms, comprising: The baseline modeling module is used to generate corresponding baselines from the daily data of the perimeter conductive ring, access control lock spring sensing unit, footstep pressure array, chicken flock voiceprint array and feed trough impedance bait, and write them into the anti-theft baseline library. The anomaly extraction module is used to perform differential operations on real-time data and the anti-theft baseline library to obtain open circuit events, abnormal access control lock spring unlocking sequences, footstep pressure trajectories, voiceprint disturbances and impedance touch traces, which are aggregated into a set of evidence. The password verification module is used to broadcast the feeding password when the primary triggering condition is met, record the response sequence of the chicken flock's voiceprint array, and perform consistency verification on the voiceprint perturbation and footstep array trajectory in the evidence set based on the response sequence to obtain the feeding password verification result. The rule matching module is used to match intrusion criteria rules based on the evidence set and the feeding password query results, and generate alarm information. The alarm information includes trigger type identifier, trigger location, monitoring segment index and evidence sequence. The early warning classification module is used to determine the high, medium or low warning level according to the classification rule table and in combination with the trigger type identifier and evidence sequence in the alarm information, and bind the monitoring segment index. The linkage execution module is used to drive the fence electric lock, corridor lighting, buzzer and communication dialer according to the warning level and monitoring section index to perform access control locking, lighting guidance, sound and light prompts and communication reporting.

[0044] The foregoing has provided a detailed description of one embodiment of the present invention, but this description is merely a preferred embodiment and should not be construed as limiting the scope of the invention. All equivalent variations and modifications made within the scope of the claims of this invention should still fall within the patent coverage of this invention.

Claims

1. An alarm method based on intelligent monitoring of chicken farms, characterized in that, Includes the following steps: S1: Generate corresponding baselines from the daily data of the perimeter conductive ring, access control spring sensing unit, footstep pressure array, chicken flock voiceprint array and feed trough impedance bait, and write them into the anti-theft baseline library. S2: Perform differential operations on real-time data and the anti-theft baseline database to obtain open circuit events, abnormal access control lock spring unlocking sequences, footstep pressure trajectories, voiceprint disturbances and impedance touch traces, and aggregate them into a set of evidence; S3: When the primary triggering condition is met, the feeding command is broadcast, the response sequence of the chicken flock's voiceprint array is recorded, and the consistency of the voiceprint perturbation and footstep array trajectory in the evidence set is checked accordingly to obtain the feeding command verification result. S4: Based on the evidence set and the feeding password verification result, match the intrusion judgment rule and generate alarm information. The alarm information includes trigger type identifier, trigger location, monitoring segment index and evidence sequence. S5: Determine the high, medium or low warning level according to the hierarchical rule table and in combination with the trigger type identifier and evidence sequence in the alarm information, and bind the monitoring segment index; S6: Based on the warning level and monitoring section index, drive the fence electric lock, corridor lighting, buzzer and communication dialer to perform access control locking, lighting guidance, sound and light prompts and communication reporting.

2. The alarm method for intelligent monitoring of chicken farms according to claim 1, characterized in that, In step S1, the process of generating corresponding baselines from the daily data of the perimeter conductive ring, access control lock spring sensing unit, footstep pressure array, chicken flock voiceprint array, and feed trough impedance bait, and writing them into the anti-theft baseline library, is as follows: During daily feeding and nocturnal roosting periods when no intrusion is confirmed, data from the perimeter conductive ring, access control spring sensor unit, footstep array, chicken flock acoustic array, and feed trough impedance bait are continuously collected and archived to the raw data storage area. Data is extracted from the original data storage area to generate on / off state sequences, access control lock spring release timing sequences, footstep pressure trajectory segments, chicken flock voiceprint response segments, and feed trough impedance touch curves, and a daily feature table is generated according to a unified field format. Based on the quiet period of nighttime roosting, short-term noise segments are removed, and daily features are aggregated according to time period to obtain on / off templates, detangle sequence templates, gait templates, voiceprint templates and impedance templates. Based on the monitoring section index, channel number, and time period label, each template is solidified into a corresponding baseline, written into the anti-theft baseline library, and the baseline version label and effective identifier are recorded.

3. The alarm method for a chicken farm based on intelligent monitoring according to claim 1, characterized in that, In step S2, the process of performing differential operations on real-time data and the anti-theft baseline database to obtain open circuit events, abnormal access control lock spring unlocking sequences, footstep tracking trajectories, voiceprint disturbances, and impedance touch traces, and aggregating them into an evidence set, is as follows: Real-time data is segmented according to the monitoring segment index and the unified time window, and templates with the same segment and the same time window are retrieved in the anti-theft baseline library to form one-to-one aligned data pairs. Perform differential operations on the aligned data pairs to output on / off deviation markers, de-tethering timing deviation markers, array position deviation markers, acoustic frequency band deviation markers, and impedance touch deviation markers. A circuit break event is generated based on the on / off deviation marker; an access control lock spring release sequence anomaly is generated based on the release timing deviation marker; and a foot pressure trajectory is generated based on the pressure position deviation marker. Based on the acoustic frequency band deviation mark, acoustic disturbance is generated; based on the impedance touch deviation mark, impedance touch trace is generated; and the monitoring segment index and time window number are bound together. Using the monitoring section index as the primary key, the evidence set is generated by aggregating circuit breakers, abnormal access control lock spring unlocking sequences, footstep tracking trajectories, voiceprint disturbances, and impedance touch traces in chronological order according to time windows.

4. The alarm method for intelligent monitoring of chicken farms according to claim 1, characterized in that, In step S3, the process of obtaining the feeding command verification result is as follows: After determining that the primary triggering condition is met, the feeding command is broadcast through the loudspeaker according to the monitoring segment index, and the start time of the command is recorded. The chicken flock's voiceprint array records the start time and sequence of responses, while the footstep array synchronously captures the footstep trajectory within the same time window. Based on the start time of the password, the time window of the voiceprint disturbance and the footstep trajectory is aligned to verify the consistency of the monitoring section and generate a consistency judgment result. Based on the consistency determination results and the monitoring segment index, a feeding password verification result is generated and bound to the relevant evidence sequence number.

5. The alarm method for a chicken farm based on intelligent monitoring according to claim 4, characterized in that, The specific details of determining whether the primary triggering condition is met are as follows: When a circuit breaker event or an abnormal access control lock spring release sequence occurs, or when footsteps and voiceprint disturbances, footsteps and impedance touch traces, or voiceprint disturbances and impedance touch traces are present simultaneously in the same monitoring segment and time window, the primary triggering condition is determined to be met.

6. The alarm method for a chicken farm based on intelligent monitoring according to claim 1, characterized in that, In step S4, the process of matching the intrusion criteria rules and generating alarm information based on the evidence set and the feeding password verification results is as follows: Align the evidence set with the feeding command inquiry results according to the monitoring segment index and time window, and merge them into an evidence mapping table; Based on the feeding password verification results, locate the corresponding rule branch and its matching field in the intrusion criterion rules; The rule fields are compared item by item in the evidence mapping table to generate a trigger type identifier, and the trigger location is determined based on the monitoring segment index. Organize the evidence entries for the monitored area in chronological order of the time window, generate an evidence sequence, and associate it with the trigger type identifier; Integrate trigger type identifier, trigger location, monitoring segment index, and evidence sequence to generate structured alarm information.

7. The alarm method for intelligent monitoring of chicken farms according to claim 1, characterized in that, In step S5, the process of determining the high, medium, or low warning level according to the hierarchical rule table and combining the trigger type identifier and evidence sequence in the alarm information, and binding the monitoring segment index, is as follows: Analyze alarm information to extract trigger type identifiers, evidence sequences, and monitoring segment indexes, which serve as inputs for hierarchical judgment. Based on the trigger type identifier, locate the corresponding rule entry in the hierarchical rule table and obtain the evidence verification requirements and warning level mapping for that entry; The evidence sequence is verified item by item according to the rules to confirm the consistency between the evidence and the time window, and to determine the high, medium or low warning level. The warning level is bound to the monitoring segment index, and a graded record is generated together with the trigger type identifier and evidence sequence.

8. An alarm system based on intelligent monitoring of a chicken farm, applied to the alarm method based on intelligent monitoring of a chicken farm as described in any one of claims 1-7, characterized in that, include: The baseline modeling module is used to generate corresponding baselines from the daily data of the perimeter conductive ring, access control lock spring sensing unit, footstep pressure array, chicken flock voiceprint array and feed trough impedance bait, and write them into the anti-theft baseline library. The anomaly extraction module is used to perform differential operations on real-time data and the anti-theft baseline library to obtain open circuit events, abnormal access control lock spring unlocking sequences, footstep pressure trajectories, voiceprint disturbances and impedance touch traces, which are aggregated into a set of evidence. The password verification module is used to broadcast the feeding password when the primary triggering condition is met, record the response sequence of the chicken flock's voiceprint array, and perform consistency verification on the voiceprint perturbation and footstep array trajectory in the evidence set based on the response sequence to obtain the feeding password verification result. The rule matching module is used to match intrusion criteria rules based on the evidence set and the feeding password query results, and generate alarm information. The alarm information includes trigger type identifier, trigger location, monitoring segment index and evidence sequence. The early warning classification module is used to determine the high, medium or low warning level according to the classification rule table and in combination with the trigger type identifier and evidence sequence in the alarm information, and to bind the monitoring segment index. The linkage execution module is used to drive the fence electric lock, corridor lighting, buzzer and communication dialer according to the warning level and monitoring section index to perform access control locking, lighting guidance, sound and light prompts and communication reporting.

Images