Intelligent security and protection integrated system

Abstract

The invention relates to the technical field of closed-loop control of security and protection systems, and discloses an intelligent security and protection integrated system which comprises an AI intelligent analysis module, a response unit state sensing module, a system load calibration module, a threat information speed regulator, an intelligent patrol robot dog, an equipment health management module and a main controller. And the main controller is used for controlling the threat information speed regulator to suppress alarm when the system is overloaded according to a comparison result of a system load level and a security steady-state working point, and scheduling the intelligent patrol robot dog to assist in intervention. Security and protection information pushing is converted into resource regulation and control, and the problem of response saturation or failure caused by manpower bottleneck of a traditional security and protection system is avoided.

Description

Technical Field

[0001] This invention relates to an integrated intelligent security system, belonging to the field of closed-loop control technology for security systems. Background Technology

[0002] Currently, video surveillance systems, perimeter alarm systems, fire monitoring systems, and visitor management systems have been widely deployed, becoming the basic technical means to ensure the security of key areas. These systems collect environmental and personnel information from different dimensions, providing diverse data sources for security and prevention. However, the design of the above-mentioned security systems all implicitly assumes a common operating premise: relying on a timely and effective human security team as the final execution end. But in actual operation, the security industry generally faces objective problems such as an aging workforce, recruitment difficulties, and limited manpower. This makes the actual handling capacity and attention span of the security team a key constraint on the operation of the entire security system. Traditional security systems are designed to maximize information collection and dissemination rather than precisely control response capabilities. Video surveillance systems generate massive amounts of footage, far exceeding the limits of manual monitoring. Perimeter alarm systems suffer from high false alarm rates, frequently interfering with security personnel's handling of invalid information, leading to reduced sensitivity or even equipment shutdown during operation. Data from each subsystem is independent, requiring extensive manual review for post-incident investigations, hindering unified command. This open-loop design, under the reality of limited human resources, generates massive amounts of alarms and redundant data, which essentially constitute a continuous interference with limited security forces, leading to alarm fatigue and response saturation, making it difficult for technical security systems to play a proactive early warning and control role.

[0003] Even though some existing technologies attempt to integrate multiple sensors, the technical approach remains focused on one-way information collection and transmission, failing to address the bottleneck at the response end. For example, the utility model patent with authorization announcement number CN209980422U discloses an intelligent integrated security monitoring system. Although it integrates cameras, smoke and dust sensors, natural gas detection sensors, and pressure sensors, and collects information uniformly to a PLC controller for analysis, the ultimate goal is still to achieve emergency warnings by sharing data with remote devices through a GPRS data transmitter after integrated analysis. This design is essentially an open-loop mode of information transmission, failing to consider, and unable to perceive, the real-time workload and handling capabilities of security personnel as remote receivers. When multiple alarms occur concurrently, the system will indiscriminately push all information to the handling end, which exacerbates the alarm fatigue and response saturation problems mentioned above. Under the reality of limited security personnel, this can easily lead to response failure.

[0004] Therefore, the technical problem to be solved by this invention is how to build an integrated management and control system that avoids simple information aggregation and push, integrates multi-source sensing data, faces the objective constraints of human resources, and realizes automatic analysis, hierarchical early warning and automated control execution of security situation through intelligent analysis. Summary of the Invention

[0005] To address the problems mentioned in the background art, the technical solution of the present invention is as follows: An integrated intelligent security system, comprising: The AI ​​intelligent analysis module is used to analyze multi-source data and generate original security events sorted by their original risk level. The response unit status perception module is used to obtain real-time work status parameters reported by security personnel; The system load calibration module is used to calculate the system load level based on a set of real-time operating status parameters and to preset the security steady-state operating point. Threat information regulator, used to buffer raw security events; Intelligent patrol robot dog; The equipment health management module is used to monitor the working status of the front-end equipment of the security subsystem and generate equipment faults; The main controller connects to the system load calibration module, threat information speed controller, intelligent patrol robot, and equipment health management module. The main controller periodically acquires the system load level and compares it with the security steady-state operating point. When the system load level is determined to be higher than the security steady-state operating point, it executes saturation control: controlling the threat information speed controller to suppress low-priority primary security events and autonomously scheduling the intelligent patrol robot to perform intervention as an auxiliary control to human response. Furthermore, the main controller is also used to adjust the security steady-state operating point when a device fault is received.

[0006] Preferably, the real-time working status parameters include at least: idle and standby, routine patrol, handling low-priority events, and handling high-priority events; the system load calibration module is used to calculate the system load level based on the ratio of the number of security personnel in the handling state to the total number of security personnel, combined with the number of original security events temporarily stored in the threat information speed controller.

[0007] Preferably, the equipment health management module is also used to calculate the integrity rate of the front-end equipment of the security subsystem based on the working status of the front-end equipment of the security subsystem; the main controller is used to use the integrity rate of the front-end equipment of the security subsystem as a product factor to dynamically calculate the security steady-state operating point preset by the system load calibration module, so as to generate an adjusted security steady-state operating point, and use the adjusted security steady-state operating point to replace the security steady-state operating point for comparison.

[0008] Preferably, the threat information speed controller includes a buffer queue; the main controller is also used to perform underload optimization when the system load level is determined to be lower than the security steady-state operating point: control the threat information speed controller to start, extract the highest priority original security events from the buffer queue, and accurately dispatch them to security personnel in an idle standby state; and the main controller is also used to actively generate cross-inspection tasks when the buffer queue is empty, and dispatch them to security personnel.

[0009] Preferably, the AI ​​intelligent analysis module is used to integrate data from video surveillance systems, perimeter security systems, and fire protection systems as multi-source data; and to monitor abnormal human behavior in the multi-source data in real time through visual analysis algorithms, including falls, abnormal running, and prolonged loitering; the AI ​​intelligent analysis module generates original security events based on abnormal human behavior.

[0010] Preferably, the intelligent patrol robot dog includes an infrared thermal imaging camera, a lidar, and a high-definition camera; when the main controller schedules the intelligent patrol robot dog to perform intervention, it is also used to control the intelligent patrol robot dog to perform autonomous navigation and target locking using the infrared thermal imaging camera and lidar, and to perform voice-based expulsion.

[0011] Preferably, the system also includes a standard digital terminal for security personnel; a response unit status perception module for obtaining real-time working status parameters reported by the standard digital terminal for security personnel; and a main controller for determining the original security event as a high-priority event when it receives an original security event generated by the AI ​​intelligent analysis module targeting a blacklisted person, and forcibly releasing the high-priority event to the standard digital terminal for security personnel even if the system load level is higher than the security steady-state working point.

[0012] Preferably, the device health management module is used to monitor the working status of the front-end devices of the security subsystem in real time, and generate a device fault when an abnormality is detected in the working status of the front-end devices of the security subsystem; the abnormality in the working status of the front-end devices of the security subsystem includes camera failure, perimeter system failure, and abnormal fire water pressure.

[0013] Preferably, when the main controller is performing saturation control, it controls the threat information speed regulator to suppress low-priority original security events, including: comparing the original risk level of a newly generated original security event with the highest risk level of an event already temporarily stored in the buffer queue; and only allowing the newly generated original security event to proceed if the original risk level of the newly generated original security event is higher than the highest risk level.

[0014] Preferably, the system also includes infrared thermal imaging monitoring equipment installed on the campus wall and in remote corners; the AI ​​intelligent analysis module is also used to integrate the human heat source signals captured by the infrared thermal imaging monitoring equipment, and combine them with preset algorithms to filter out animal or wind and rain interference in order to generate original security events.

[0015] Compared with the prior art, the beneficial effects of the present invention are: 1. By unifying and integrating multiple isolated subsystems such as video surveillance, perimeter security, and fire protection, and using artificial intelligence analysis technology to centrally analyze multi-source data, the working mode of the security system is transformed from passively pushing scattered information to proactively managing the security situation in a closed loop. It is no longer about piling up alarm information, but about automatically guiding and controlling the subsequent handling process through event classification and intelligent analysis, avoiding the response saturation or failure problems of traditional security systems due to information fragmentation and manpower bottlenecks.

[0016] 2. Construct an automated control and execution mechanism for human-machine collaboration. When the system identifies specific behaviors such as abnormal running or prolonged loitering through visual analysis algorithms, the response is not a single alarm, but rather triggers a collaborative control chain. The system automatically controls the nearest camera to perform continuous tracking and recording, locking onto the target's status. According to the preset control strategy, the intelligent patrol robot dog is autonomously dispatched to move forward along the autonomous navigation path and perform physical intervention actions such as voice warnings. This closed loop of perception, locking, and intervention control, which is automatically completed by the system, replaces the slow-responding manual links and achieves immediate control of emergencies.

[0017] 3. By introducing equipment health management functions, closed-loop control and automatic adjustment of the security system's own operating status are achieved. The system monitors the working status of all front-end devices in real time. When a camera malfunction or perimeter system failure is detected, it no longer relies on time-consuming and labor-intensive manual inspections and troubleshooting, but automatically generates a maintenance work order and pushes it to the maintenance unit. This system-level self-diagnosis and maintenance mechanism ensures the long-term operational reliability of the entire control system and avoids functional gaps in the entire security management system due to the failure of local components. Attached Figure Description

[0018] Figure 1 This is a block diagram of the closed-loop load control system of the present invention; Figure 2 This is a comparison chart of the load control effects of the system of the present invention; Figure 3 This is a diagram illustrating the causes of failure in the traditional security system according to the present invention. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of the present invention clearer, the present invention will be described in further detail below. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0020] This invention provides an integrated intelligent security system, including a main controller that connects and coordinates an AI intelligent analysis module, a response unit status perception module, a system load calibration module, a threat information speed regulator, an equipment health management module, and an intelligent patrol robot. Taking a campus with thousands of cameras as an example, the system performs data integration. The AI ​​intelligent analysis module is used to uniformly access data from the existing video surveillance system, perimeter security system, and fire protection system within the campus, serving as the raw input for the control system. The AI ​​intelligent analysis module utilizes visual analysis algorithms to analyze video data streams in real time, identifying specific abnormal behaviors of individuals, such as falls, abnormal running, prolonged loitering, or integrating human heat source signals captured by infrared thermal imaging monitoring equipment added to the corners of the perimeter walls, filtering animal interference, and identifying intrusion behaviors such as climbing over walls. This module analyzes... The system generates original security events with their original risk level, time, and location, and sends them all to the buffer queue of the threat information regulator. The system establishes real-time status awareness of the security execution end, i.e., the security team, as a feedback path for closed-loop control. The response unit's status awareness module is used to acquire real-time work status parameters reported by each security personnel's standard digital terminal, which can be a dedicated APP or digital handheld device. Real-time work status parameters are well-defined data structures, including at least enumerated values ​​such as idle / standby, routine patrol, handling low-priority events, handling high-priority events, and temporarily off duty. This status reporting is periodic, for example, automatically every 10 seconds or triggering an instantaneous status change when a security personnel accept or close an order, ensuring the main controller has a grasp of the security team's immediate status.

[0021] The system load calibration module receives all real-time operating status parameters collected by the response unit status perception module, quantifies and calculates the overall process variables of the current system, and derives the system load level. This system load level is a key control indicator, and the calculation procedure is predefined. For example, the system load calibration module calculates the system load level using a weighted algorithm: it counts the number of security personnel in both low-priority and high-priority event handling states and calculates their ratio to the total number of security personnel; it obtains the number of pending original security events temporarily stored in the threat information speed controller and calculates their ratio to the maximum length of the preset buffer queue; the system load level... Through formula The calculation shows that, For system load level, The number of people being dealt with, Total number of security personnel This represents the current queue length. This represents the maximum capacity of the queue. and The weighting coefficients and , It can be set to 0.7. It can be set to 0.3; at the same time, this module also presets a security steady-state operating point, this operating point The optimal operating state that the system is expected to maintain, for example This value can be calibrated by offline analysis of historical security data to identify the critical load point that ensures timely response and prevents alarm accumulation in the system; the main controller is the core of the closed-loop control system, periodically executing a control cycle, such as every 30 seconds; the main controller obtains the current real-time system load level from the system load calibration module. Compare it with the preset security steady-state operating point A comparison is made; based on the deviation signal generated by the comparison, the main controller performs adjustment operations on the threat information speed controller; when the main controller determines... When the system load level is lower than the security steady-state operating point, the system is in an underload state. At this time, the underload optimization mode is executed. The main controller controls the valve of the threat information speed controller to open. The speed controller extracts the original security event with the highest original risk level from its buffer queue. Based on the location of the event, the location of the security personnel, and their idle / standby status, it accurately assigns the event to the standard digital terminal of the optimal security personnel. If the buffer queue is empty at this time, the main controller actively generates and assigns cross-inspection tasks to avoid idle security resources, thereby improving the overall utilization rate of the system.

[0022] When the main controller determines When the system load level is higher than or close to the security steady-state operating point, and is in a saturated or overloaded state, saturation control mode is executed. The main controller controls the valve of the threat information speed controller to close slightly, actively suppressing newly generated low-priority original security events. The suppression logic is that the speed controller compares the original risk level of the newly generated original security event with the highest risk level of all events already temporarily stored in the buffer queue. Only when the risk level of the new event is higher than the highest risk level in the current queue is it released. Other events are temporarily stored in the queue to avoid information interference to security personnel who are already under high-load handling. Under this mode, the system only ensures that the highest priority events, such as fires, violent conflicts, or original security events generated by the AI ​​intelligent analysis module targeting blacklisted personnel, are forcibly released. While executing the saturation control mode, the main controller initiates a control mode upgrade and autonomously dispatches intelligent patrol robot dogs to perform intervention. The intelligent patrol robot dogs are equipped with infrared thermal imaging cameras and laser radar. The system includes high-definition cameras and a robot dog. The main controller controls the robot to autonomously navigate to the incident area using lidar, lock onto the target using an infrared thermal imaging camera, and execute voice-based decoys. The robot dog serves as an auxiliary control method when the human response unit is saturated, providing the main controller with additional schedulable control resources to ensure the system still has basic intervention capabilities when human intervention reaches its limit. The system also includes an equipment health management module for managing the reliability of the security control system itself. This module monitors the real-time operating status of all front-end devices in the security subsystems, including camera malfunctions, perimeter system failures, and abnormal fire water pressure. When an abnormal status is detected, an equipment fault is generated, and a repair work order is automatically generated and pushed to maintenance personnel. This module's function is coupled with the control logic of the main controller. Based on the operating status of all front-end devices, the equipment health management module also calculates the front-end device integrity rate of the security subsystem, which is a percentage parameter characterizing the overall perception capability of the system. The main controller acquires this availability rate either upon receiving equipment fault information or periodically. At that time, the security steady-state operating point will be adjusted. The adjustment procedure will improve the availability of front-end equipment by controlling the main controller. As a product factor, it affects the preset security steady-state operating point. Perform dynamic calculations to generate the adjusted security steady-state operating point. The main controller uses this Alternative Subsequent control comparisons were conducted, among which This is the adjusted steady-state operating point for security. This is the preset steady-state operating point for security. This refers to the availability of front-end equipment in the security subsystem; when the front-end sensing equipment malfunctions and its sensing capability decreases, i.e. It automatically lowers its security steady-state operating point, proactively becomes more conservative, makes the system more likely to enter saturation control conditions, prioritizes high-risk events, and ensures that the control system can still operate safely and stably even when its own components are not in ideal conditions.

[0023] Example 1: Total number of security personnel in the system under specific high-load concurrency scenarios For 10 people, the system load calibration module is preset to the security steady-state operating point. The percentage was 70%. When the campus experienced sudden severe convective weather and multiple areas saw gatherings of people, the AI ​​intelligent analysis module generated 35 original security incidents within a short period, including perimeter false alarms and loitering incidents. All of these events flooded into the threat information regulator's buffer queue, and the response unit's status perception module reported that 8 security personnel were in a state of "ongoing response." The system load calibration module calculates the current system load level based on the reported set of status parameters and the buffer queue length. Reaching 82%, higher than 70%; the main controller determines that the system has entered the saturation control condition, the control threat information speed controller executes the suppression logic, the newly generated low-priority perimeter false alarms and loitering events are temporarily stored in the buffer queue, and the standard digital terminals of security personnel do not receive the push.

[0024] The device health management module detected that a camera in a critical area failed due to a lightning strike, generating a device fault report and calculating the availability of the front-end devices in the security subsystem. The load drops to 90%; the main controller receives this status, executes the adjustment procedure, and dynamically adjusts the security steady-state operating point to... Based on this, the steady-state operating threshold is tightened, and the control strategy automatically shifts to a conservative approach. The AI ​​intelligent analysis module integrates data from infrared thermal imaging monitoring equipment to generate a high-priority fence-climbing intrusion event. The threat information speed controller determines that the original risk level of this intrusion event is higher than the highest risk level of events temporarily stored in the buffer queue, and the main controller forcibly allows this high-priority event. The main controller checks the system load level. Higher than the adjusted security steady-state operating point Security personnel The system remains at a high priority level, indicating insufficient human resources for immediate response. The main controller initiates a control mode upgrade, autonomously dispatching the nearest intelligent patrol robot to intervene at the intrusion point. It uses lidar to lock onto the target and executes voice warnings. This high-priority event is precisely assigned to a security personnel who has just completed a task and is now on standby. This personnel rushes to the scene to handle the situation. At the end of the control cycle, the high-priority intrusion event is handled jointly by the intelligent patrol robot and subsequent security personnel. Low-priority events temporarily stored in the buffer queue do not interfere with the critical handling process. Under conditions of decreased perception capability and concurrent input disturbances, the system maintains a stable output of core security response capabilities through closed-loop load control and auxiliary resource scheduling.

[0025] Example 2: This example compares the intelligent security integrated system of the present invention with a traditional open-loop system, demonstrating the control effect of maintaining high-priority event response capability and avoiding security response saturation under different event load pressures. A simulation platform was built to simulate a security area equipped with 10 security personnel. The platform generates two types of raw security event streams, high and low, at a preset rate. The average time for security personnel to handle low-priority events is set to 5 minutes, and the average time to handle high-priority events is set to 15 minutes. A control group and the prototype of this invention are set up. The control group simulates a traditional open-loop system, receiving all raw security events without distinguishing priority or system load, and pushing them to all 10 security personnel. The prototype of this invention deploys a closed-loop control system, including a response unit state perception module, a system load calibration module, and a threat information speed controller. The security steady-state operating point... The system load level is set to 70%. According to the regulations Calculation, weight Set it to 0.7. Set to 0.3; during a 60-minute test period, low, medium, and high event loads were input to the two groups respectively; the test monitoring index was the average load of security personnel, defined as the proportion of personnel in a state of handling. The average response time for high-priority events is defined as the time from the generation of an event to the acceptance of the order by security personnel; the backlog of low-priority events is defined as the number of events that have not been dealt with at the end of the test. The test results are shown in Table 1.

[0026] Table 1: Comparison of System Response Performance under Different Loads

[0027] Table 1 shows that under low load conditions, both groups of security personnel had low average workloads and responded promptly to high-priority events. Under medium load conditions, the control group pushed all 35 events, with the average workload of security personnel reaching 98.3%, and the average response time for high-priority events increasing to 215.4 seconds. The system load level of the sample group in this invention at this time... At security steady-state operating points exceeding 70%, the main controller triggers saturation control mode, the threat information speed regulator suppresses the issuance of low-priority events, the average load of security personnel is controlled at around 72.5%, the average response time for high-priority events is 33.1 seconds, and 24 low-priority events are backlogged. Under high-load conditions, the control group's security personnel average load is 100.0%, and the response time for high-priority events increases to 750.8 seconds; the sample group of this invention has an average security personnel load of 78.6%, an average response time for high-priority events of 48.7 seconds, and 53 low-priority events backlogged. Under high-load conditions, the sample group of this invention was tested at different security steady-state operating points. Performance at set values; three sets of settings Values: 30%, 70% and 95%, see Table 2 for test results.

[0028] Table 2: Different Security Steady-State Operating Points ( Performance table under parameters

[0029] The data in Table 2 shows that, When set to 30%, the system prematurely enters saturation control mode, with security personnel's average workload at 29.5% and a low-priority event handling rate of 3.3%. When the load was set to 95%, the average workload of security personnel was 100.0%, and the response time for high-priority events increased to 741.3 seconds, which is similar to the high-load conditions of the control group. When the threshold is set to 70%, the average response time for high-priority events is 48.7 seconds, the workload of security personnel is 78.6%, and the handling rate for low-priority events is 11.7%.

[0030] Example 3: This example combines Figures 1 to 3 A description of the integrated intelligent security system, such as... Figure 1As shown, the system begins with multi-source data input, including video surveillance, perimeter security, and fire protection systems. This data is sent to the AI ​​intelligent analysis module to analyze the multi-source data and generate raw security events. These events then enter the threat information regulator, serving as a buffer event queue. Based on control commands from the main controller, suppression or release operations are executed. Released events are dispatched to the standard digital terminals of security personnel. Security personnel report real-time work status parameters. The response unit's status perception module acquires the real-time work status of the security personnel and transmits the status information to the system load calibration module. This module calculates the system load level and presets the steady-state operating point. The positioning module receives the equipment availability rate from the equipment health management module, which is calculated by the equipment health management module based on the monitoring of the front-end equipment. The calculated system load level and preset steady-state operating point information are sent to the main controller. The equipment health management module also sends the equipment availability rate to the main controller. The main controller compares the load level with the dynamic steady-state operating point based on this. Based on the comparison results, the main controller issues control commands to the threat information speed controller, performs precise dispatch or forced release to the standard digital terminals of security personnel, and issues dispatch intervention commands to the intelligent patrol robot dog when saturation occurs. The intelligent patrol robot dog then performs auxiliary intervention, such as voice-based deterrence.

[0031] like Figure 2 As shown, under low-load conditions, the average load of security personnel in both the control group and the sample group of this invention was at a low level, at 32.5% and 31.8% respectively. Under medium-load conditions, the average load of the control group reached 98.3%, while that of the sample group of this invention was controlled at 72.5%. Under high-load conditions, the average load of the control group reached 100.0%, while the average load of the sample group of this invention was still controlled at 78.6%. Figure 3 As shown, this reveals the technical dilemmas faced by traditional security systems, specifically manifested in system design flaws such as open-loop design, pursuit of information maximization, failure to address human resource constraints, and lack of closed-loop processing capabilities; information silos leading to independent systems, fragmented data, lack of unified analysis, and inability to coordinate command; data and alarm overload resulting in massive amounts of redundant data, high false alarm rates, alarm fatigue, and massive amounts of patrol footage; and human resource bottlenecks including limited staff energy, an aging workforce, recruitment difficulties, and reliance on manual processing. These factors collectively lead to response saturation or failure in traditional security systems.

[0032] Example 4: This example illustrates the steady-state operating point of the security system in the system load calibration module. The calibration procedure defines the algorithm path for the main controller to perform precise dispatching and proactive generation of cross-inspection tasks under underload optimization conditions, and determines the steady-state operating point of the security system. When the system is deployed for the first time, an offline calibration procedure is executed; all original security event logs and security personnel handling logs from the past 30 days in the security area are retrieved as input data for the simulation model; based on a 10-person security team model and the event handling time setting, the procedure iterates through the data in 5% increments. From 30% to 95% of the set value; for each Set values, simulate historical data streams, and calculate two evaluation indicators: the 95th percentile response time of high-priority events and the average workload of security personnel; plot the two indicators against historical data. The changing performance curve; selecting high-priority events before the 95th percentile response time deteriorates to an inflection point. The value, such as 70%, is used as a preset value; this is the preset security steady-state operating point for the system load calibration module. Adapt to the long-term evolution of security zones, such as increases or decreases in protected areas, structural changes in historical risk heat maps, or the total number of security personnel. When adjustments are made, the main controller is also used to perform periodic or triggered recalibration of the security steady-state operating point, which differs from recalibration based on equipment availability. The real-time product adjustment procedure involves automatically retrieving the most recent historical security event logs and handling logs as update data at the end of the preset period or when changes in the aforementioned boundary conditions are detected. The offline calibration procedure is then re-executed to calculate and generate an updated security steady-state operating point to replace the original one. Default value.

[0033] When the main controller determines that the system is in an underload optimization state and needs to perform precise dispatch, the algorithm path is as follows: The input is the original security event to be dispatched extracted by the threat information speed controller, including the event location coordinates. A list of all available, standby security personnel, including each person's current location coordinates. The main controller iterates through all available personnel and calculates the overall dispatch cost for each person. The cost model is ,in, Calculate the shortest path distance between two people for a geographic information system; The cumulative handling time of the security personnel during past work cycles is used for load balancing; and Preset weighting coefficients; the main controller selects the lowest value. On-duty personnel The system dispatches events to its standard digital terminals. When the main controller determines that the system is in an underload optimization state and the threat information speed regulator buffer queue is empty, the algorithm path for actively generating cross-inspection tasks is as follows: The input is a historical risk heat map of the security area, generated based on the historical event location density statistics and the current real-time location of all security personnel; the main controller calculates the current patrol coverage of all personnel and performs spatial overlay analysis with the historical risk heat map; areas with high historical risk levels, such as the top 20% of the heat map, and areas not currently covered by patrols are identified as inspection target areas; the system selects the nearest idle standby security personnel to the inspection target area, generates an inspection task to go to that area, and dispatches it to their standard digital terminals.

[0034] Example 5: This example illustrates the specific algorithm path for the main controller to perform a control mode upgrade under saturation control conditions, namely, rescheduling security personnel and autonomously scheduling intelligent patrol robots; the security steady-state operating point. The system security personnel are set at 70% of the total number of personnel. For 10 people; at a certain operating moment, the system load level calculated by the system load calibration module. It is 85%, higher than The response unit's status perception module reports that all 10 security personnel are in a state of handling, with 3 handling high-priority events and 7 handling low-priority events. The AI ​​intelligent analysis module generates the original security event, i.e., violent conflict, which is determined to have the highest original risk level. This is confirmed by a combination of cries for help captured by a specific audio sensor and physical conflict behavior identified by a visual analysis algorithm. The threat information speed controller determines that the risk level of this event is higher than all events in the queue, and the main controller forces its passage. The main controller checks the human resources pool and finds that there are 0 idle personnel on standby. The main controller calculates all intelligent patrol robot dogs, such as 2, and sends them to the event location. Estimated time ,choose The shortest robot dog was immediately dispatched to intervene, using a high-definition camera to transmit real-time footage and issue a voice warning to drive it away; the main controller assigned human support, initiated a control mode upgrade, and forcibly reassigned the 7 security personnel handling low-priority incidents; the main controller was for the 7 security personnel. Calculate the rescheduling cost separately The cost model is used to balance the urgency of new events with the sunk costs of existing tasks; rescheduling costs. ,in, For the security personnel from their current location Arrive at the new event location The estimated shortest time, This refers to the time that the security personnel have already invested in their current low-priority tasks. and These are the weighting coefficients. , Prioritizing rapid response to new events; the main controller calculates the response time for all 7 personnel. Value, select the lowest For personnel on duty, their standard digital terminals are forced to switch to high-priority event handling mode, and new event handling instructions are pushed to them; previously handled low-priority events are automatically released and re-enter the threat information speed controller's buffer queue, awaiting the system load level. Falling back to The following will be redistributed.

[0035] System load level Calculation of weighting coefficients in the procedure and Dispatch cost under underload optimized operating conditions Weight coefficients in the model and and the cost of rescheduling under saturation control conditions Weight coefficients in the model and The specific values ​​are determined based on the offline calibration procedure, which includes using historical security event logs and handling logs as simulation inputs, setting multi-dimensional operation evaluation targets, such as the 95th percentile response time of high-priority events not exceeding a specific threshold and the average workload of security personnel remaining near the steady-state working point, using grid search or Bayesian optimization algorithms to iteratively run the simulation within the preset parameter range, and automatically selecting the weight coefficient combination that is closest to the above operation evaluation targets as the initial configuration parameters of the system.

[0036] Example 6: This example illustrates the algorithm path of the AI ​​intelligent analysis module in generating original security events based on prolonged loitering behavior, and the data structure of the buffer queue in the threat information speed controller. To analyze prolonged loitering behavior, the AI ​​intelligent analysis module, during deployment, marks key monitoring areas, such as school walls, secluded corners, or the entrance to the finance office, with electronic fences. The module receives real-time video streams from the video surveillance system and uses the YOLOv5 target detection algorithm or an algorithm of equivalent precision to continuously track each person entering the electronic fence area, assigning a unique target ID. The system establishes a timestamp stack for each target ID, recording the time of entry, continuous presence within the area, and departure. The system presets a loitering time threshold. Based on the sensitivity of the area, such as the area around the school wall. Set to 120 seconds, outside the finance office during non-working hours. Set to 30 seconds; when the AI ​​intelligent analysis module detects the continuous dwell time of the target ID within the electronic fence. Exceed Upon detection of prolonged loitering, a raw security event is immediately generated. This event is then pushed to the threat information regulator's buffer queue. The buffer queue is a priority-based first-in-first-out queue, storing data in the structure {Event_ID, Timestamp, Location_Coord, Event_Type, Risk_Level_Raw}. Here, Event_ID is the unique identifier for the event; Timestamp is the event generation timestamp; Location_Coord is the event's location coordinates; Event_Type is an enumerated value for the event type, such as prolonged loitering, abnormal running, or climbing over a wall; and Risk_Level_Raw is the event's raw risk level, determined by the AI ​​intelligent analysis module based on factors such as event type, location sensitivity, and time period, using a preset risk matrix table. For example, climbing over a wall at night... The calibration is set to 95, and after prolonged lingering in the public area, the calibration is reduced to 30; when the main controller is executing saturation control, it compares the new events... With all events already in the buffer queue The maximum value determines whether to allow passage.

[0037] 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.

[0038] 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 integrated intelligent security system, characterized in that the system... include: The AI ​​intelligent analysis module is used to analyze multi-source data and generate original security events sorted by their original risk level. The response unit status perception module is used to obtain real-time work status parameters reported by security personnel; The system load calibration module is used to calculate the system load level based on a set of real-time operating status parameters and to preset the security steady-state operating point. Threat information regulator, used to buffer raw security events; Intelligent patrol robot dog; The equipment health management module is used to monitor the working status of the front-end equipment of the security subsystem and generate equipment faults; The main controller connects to the system load calibration module, threat information speed controller, intelligent patrol robot dog, and equipment health management module. The main controller is used to periodically acquire the system load level and compare it with the security steady-state operating point. When it is determined that the system load level is higher than the security steady-state operating point, the saturation control mode is executed: the control threat information speed controller suppresses low-priority original security events and autonomously dispatches intelligent patrol robot dogs to perform intervention as an auxiliary control for human response. In addition, the main controller is also used to adjust the security steady-state operating point when a device failure is received.

2. The intelligent integrated security system according to claim 1, characterized in that, Real-time operational status parameters include at least: idle and on standby, routine patrol in progress, handling low-priority events, and handling high-priority events; the system load calibration module is used to calculate the system load level based on the ratio of the number of security personnel in the handling state to the total number of security personnel, combined with the number of original security events temporarily stored in the threat information speed controller.

3. The intelligent integrated security system according to claim 1, characterized in that, The equipment health management module is also used to calculate the integrity rate of the front-end equipment of the security subsystem based on the working status of the front-end equipment of the security subsystem; the main controller is used to use the integrity rate of the front-end equipment of the security subsystem as a product factor to dynamically calculate the security steady-state operating point preset by the system load calibration module, so as to generate an adjusted security steady-state operating point, and use the adjusted security steady-state operating point to replace the security steady-state operating point for comparison.

4. The intelligent integrated security system according to claim 2, characterized in that, The threat information speed controller includes a buffer queue; the main controller is also used to perform underload optimization when the system load level is determined to be lower than the security steady-state operating point: control the threat information speed controller to start, extract the highest priority original security events from the buffer queue, and accurately dispatch them to security personnel in an idle standby state. The main controller is also used to actively generate cross-inspection tasks when the buffer queue is empty and dispatch them to security personnel.

5. The intelligent integrated security system according to claim 1, characterized in that, The AI ​​intelligent analysis module integrates data from video surveillance systems, perimeter security systems, and fire protection systems as multi-source data. It uses visual analysis algorithms to monitor abnormal human behavior in the multi-source data in real time, including falls, abnormal running, and prolonged loitering. Based on the abnormal human behavior, the AI ​​intelligent analysis module generates original security events.

6. The intelligent integrated security system according to claim 1, characterized in that, The intelligent patrol robot dog includes an infrared thermal imaging camera, a lidar, and a high-definition camera. When the main controller dispatches the intelligent patrol robot dog to perform intervention, it is also used to control the intelligent patrol robot dog to use the infrared thermal imaging camera and lidar for autonomous navigation and target locking, and to execute voice-based expulsion.

7. The intelligent integrated security system according to claim 1, characterized in that, The system also includes standard digital terminals for security personnel; a response unit status perception module for obtaining real-time working status parameters reported by the standard digital terminals of security personnel; and a main controller for determining the original security event as a high-priority event when it receives an original security event generated by the AI ​​intelligent analysis module targeting blacklisted personnel, and forcibly releasing the high-priority event to the standard digital terminals of security personnel even if the system load level is higher than the security steady-state operating point.

8. The intelligent integrated security system according to claim 1, characterized in that, The device health management module is used to monitor the working status of the front-end devices of the security subsystem in real time. When an abnormality is detected in the working status of the front-end devices of the security subsystem, a device fault is generated. Abnormalities in the working status of the front-end devices of the security subsystem include camera failure, perimeter system failure, and abnormal fire water pressure.

9. The intelligent integrated security system according to claim 4, characterized in that, When the main controller is executing saturation control mode, it controls the threat information speed regulator to suppress low-priority original security events, including: comparing the original risk level of a newly generated original security event with the highest risk level of events already temporarily stored in the buffer queue; and only allowing the newly generated original security event to proceed if the original risk level of the newly generated original security event is higher than the highest risk level.

10. The intelligent integrated security system according to claim 1, characterized in that, The system also includes infrared thermal imaging monitoring equipment installed on the campus walls and in remote corners; the AI ​​intelligent analysis module is also used to integrate the human heat source signals captured by the infrared thermal imaging monitoring equipment, and combine them with preset algorithms to filter out animal or wind and rain interference in order to generate original security events.

Images