A microwave radar-based active early-warning safety sentry system and early-warning method

Abstract

The application discloses a microwave radar-based active early-warning safety sentry system and an early-warning method, and belongs to the safety early-warning field.The microwave radar-based active early-warning safety sentry system comprises a safety sentry main body, and a self-adaptive millimeter wave radar module, a UWB ad hoc network communication module, an edge computing unit and an early-warning module are arranged in the conical barrel shell of the safety sentry main body.The self-adaptive millimeter wave radar module is integrated with a multi-angle detection antenna array, and can detect and identify vehicles in any direction;the UWB ad hoc network communication module is used for automatically pairing and networking multiple safety sentry main machines, and generates a dynamic electronic fence through triangulation based on UWB positioning data;the edge computing unit completes vehicle target identification and early-warning determination locally, triggers the early-warning module to send an early-warning signal when a vehicle intrudes, and pushes the early-warning signal in the network through the UWB ad hoc network communication module.The application does not need manual calibration of the radar direction, realizes single-soldier rapid deployment and multi-soldier cooperative early-warning, and has complete local protection capability when the network is disconnected.

Description

Technical Field

[0001] This invention belongs to the field of security early warning technology, specifically relating to a microwave radar-based active early warning security sentinel system and early warning method. Background Technology

[0002] The statements in this section are merely background information related to the present invention and do not necessarily constitute prior art.

[0003] In road maintenance, accident handling, and emergency clearing scenarios, warning facilities must be installed behind the construction area to prevent vehicles from entering and causing secondary accidents. Currently, the main technical solutions include traditional reflective traffic cones, simple smart cones, and ordinary radar cones.

[0004] However, the inventors discovered that traditional reflective cones rely solely on reflective film for passive visual identification, lacking electronic sensing and active warning capabilities. In low-visibility environments such as rain, snow, fog, and nighttime, their effective visibility is severely insufficient, leaving vehicles with insufficient braking and avoidance time by the time they detect the construction area, resulting in a very high risk of secondary accidents. While simple smart cones are equipped with GNSS positioning and LED flashing lights, they can only report tipping over and low battery malfunctions, unable to identify whether passing vehicles pose an intrusion threat. Ordinary radar cones integrate millimeter-wave radar, but their radar detection direction is fixed, requiring precise manual calibration of oncoming vehicle directions during deployment. In emergency scenarios such as breakdowns and obstacle clearing, a single person cannot quickly deploy them; furthermore, each cone operates independently without a collaborative mechanism, failing to form a continuous protective area and lacking linkage with personnel-worn terminals, thus unable to monitor the safety of the workers in real time.

[0005] Existing technologies propose a V2X-based intelligent cone warning system, which obtains the cone's location through GNSS positioning, analyzes the data using a high-precision map in the cloud, and then broadcasts the warning information via V2X. However, this system relies entirely on cloud computing, rendering it completely ineffective in tunnels, mountainous areas, and other road sections without 4G / 5G signals. Furthermore, the warnings only cover vehicles equipped with V2X vehicle terminals, leaving many ordinary vehicles uncovered. Additionally, the equipment can only passively report faults after complete failure, failing to anticipate progressive hazards such as radar attenuation and battery degradation, easily leading to hidden protection blind spots. Summary of the Invention

[0006] The purpose of this invention is to overcome the shortcomings of the prior art and provide a microwave radar-based active early warning security sentinel system and early warning method.

[0007] To achieve the above objectives, the present invention is implemented through the following technical solution: In a first aspect, the technical solution of the present invention provides a microwave radar-based active early warning security sentinel system, comprising: a security sentinel body, which includes a cone-shaped shell and an adaptive millimeter-wave radar module, a UWB self-organizing network communication module, an edge computing unit and an early warning module disposed inside the cone-shaped shell; Among them, the adaptive millimeter-wave radar module is used to detect vehicle targets. It integrates a multi-angle detection antenna array, which can detect and identify vehicles within a preset range regardless of the orientation of the antenna. The UWB self-organizing network communication module is used to automatically pair and network multiple security sentinel hosts, and generate dynamic electronic fences through triangulation based on the UWB positioning data of multiple security sentinel hosts. The edge computing unit is configured to receive detection data from the adaptive millimeter-wave radar module and perform vehicle target identification and early warning determination locally; when a vehicle is detected to have entered, the edge computing unit triggers the early warning module to issue an early warning signal and pushes the early warning command to other security sentinel hosts in the network through the UWB self-organizing network communication module.

[0008] In at least one embodiment, a wearable personnel early warning terminal is also included, which has a built-in UWB positioning chip; the UWB positioning chip communicates bidirectionally with the UWB self-organizing network communication module of the security sentry host, and is used to upload the location information of the wearer to the security sentry host in real time. The edge computing unit is also configured to: based on the personnel location information uploaded by the UWB positioning chip, compare the personnel location information with the boundary of the dynamic electronic fence, and trigger the wearable personnel early warning terminal to issue a boundary crossing alarm when the personnel cross the dynamic electronic fence.

[0009] In at least one embodiment, a cloud-based management and control platform is also included. The cloud-based management and control platform communicates with the security sentinel host via a 4G or 5G network to receive early warning event data, device health data, and personnel location data uploaded by the security sentinel host, and performs situation visualization, alarm storage, and maintenance work order generation.

[0010] In at least one embodiment, the adaptive millimeter-wave radar module is specifically a multi-layer fan-shaped staggered antenna array integrated with a 77GHz vehicle-grade millimeter-wave radar chip; the multi-layer fan-shaped staggered antenna array is composed of multiple fan-shaped antenna elements with different orientations arranged in a multi-layer staggered manner, each antenna element having a different beam pointing angle, which, when combined, forms a continuous detection fan-shaped area covering the horizontal direction from 0 to 120°.

[0011] In at least one embodiment, the edge computing unit receives detection data from the adaptive millimeter-wave radar module and performs vehicle target identification and warning determination locally, specifically including: The edge computing unit receives the echo signal from the adaptive millimeter-wave radar module in real time. After amplification, filtering, and analog-to-digital conversion, it runs a target detection algorithm to extract vehicle targets and calculates the vehicle's distance, speed, and azimuth relative to the current security sentry host, generating a vehicle target list. It also monitors the distance and speed of vehicles in the vehicle target list in real time. When the calculated vehicle distance is less than a preset warning distance threshold and the vehicle speed is greater than a preset speed threshold, the edge computing unit determines that a vehicle has entered the warning area in the vehicle target list, immediately generates a vehicle intrusion warning signal, and sends it to the warning module.

[0012] In at least one embodiment, the edge computing unit uses the personnel location information uploaded by the UWB positioning chip to compare the personnel location information with the boundary of the dynamic electronic fence. When a person crosses the dynamic electronic fence, the wearable personnel early warning terminal is triggered to issue a boundary crossing alarm, specifically including: The edge computing unit receives personnel ranging data uploaded by the UWB self-organizing network communication module in real time, calculates the real-time coordinates of each worker in the electronic fence coordinate system through the triangulation algorithm, and performs spatial geometric comparison between the coordinates of each worker and the polygonal boundary of the dynamic electronic fence to determine whether there are any workers outside the boundary. When the coordinates of a worker are outside the boundary of the dynamic electronic fence, an early warning signal for personnel crossing the boundary is immediately generated and sent to the corresponding wearable personnel early warning terminal through the UWB self-organizing network communication module.

[0013] In at least one embodiment, the security sentry body is also provided with a power management module; The edge computing unit is also configured to: collect the signal-to-noise ratio of the adaptive millimeter-wave radar module, the packet loss rate of the UWB self-organizing network communication module, the battery voltage and charge-discharge cycle count of the power management module, and the processor temperature in real time; input the collected multi-dimensional operating parameters into a preset hierarchical threshold-attenuation coupling judgment model to determine whether the device has a failure risk; if the device is determined to have a failure risk, immediately generate a device fault warning signal to the warning module.

[0014] In at least one embodiment, the hierarchical threshold-attenuation coupling judgment model adopts a two-layer parallel judgment structure. The first layer is a static hard threshold judgment layer, which is used to detect sudden failures. The second layer is a dynamic attenuation fitting judgment layer, which is used to predict progressive degradation failures. The static hard threshold judgment layer and the dynamic attenuation fitting judgment layer run in parallel and independently. The output of a failure warning signal by either judgment layer triggers an operation and maintenance warning.

[0015] In at least one embodiment, the warning module is a multi-mode warning module with sound, light, and vibration, which integrates a strobe LED array, a high-decibel buzzer, and a driving circuit. When the warning signal is a vehicle intrusion warning signal, the strobe LED is driven to flash at the highest frequency and the high-decibel buzzer is driven to sound at the maximum volume. When the warning signal is a personnel crossing the boundary warning signal, the strobe LED is driven to flash at a medium frequency and the high-decibel buzzer is driven to sound at a medium volume. When the warning signal is an equipment failure warning signal, the strobe LED is driven to flash at a specific frequency to indicate the fault type.

[0016] Secondly, the technical solution of the present invention also provides an early warning method based on a microwave radar active early warning security sentinel system, including: Deploy one or more security sentinel hosts; the UWB self-organizing network communication module automatically scans surrounding devices and completes network formation; the adaptive millimeter-wave radar module automatically adapts to any placement angle and activates vehicle target detection. When the adaptive millimeter-wave radar module detects a vehicle entering the warning area from behind, the edge computing unit locally triggers the local warning module to issue an audible and visual warning. At the same time, the warning command is pushed to all safety sentinel hosts and all wearable personnel warning terminals in the network through the UWB self-organizing network communication module. Three or more security sentinel hosts generate a dynamic electronic fence through UWB triangulation positioning. Wearable personnel early warning terminals upload personnel location information in real time. The edge computing unit compares the personnel location information with the electronic fence boundary. When a person crosses the boundary, the wearable personnel early warning terminal is triggered to issue a boundary crossing alarm. The edge computing unit collects operating parameters in real time and inputs them into the hierarchical threshold-attenuation coupling judgment model. When the device is determined to have a failure risk, an operation and maintenance warning is generated and uploaded to the cloud management and control platform.

[0017] The beneficial effects of the above-described technical solution of the present invention are as follows: This invention discloses a microwave radar-based active early warning safety sentinel system. Through adaptive millimeter-wave radar multi-angle detection, it enables deployment in any orientation, eliminating manual calibration and significantly improving emergency deployment efficiency. UWB self-organizing networking facilitates multi-machine collaborative linkage and dynamic electronic fence generation. Combined with wearable personnel early warning terminals, it provides two-way, three-dimensional protection against vehicle intrusion and personnel boundary violations. Edge computing units locally perform vehicle identification and early warning determination, ensuring complete protection even in network outage scenarios. Built-in predictive maintenance algorithms monitor equipment operating parameters in real time and predict failure risks, eliminating hidden safety blind spots caused by faulty equipment. This comprehensively solves a series of technical problems in existing technologies, such as passive warning methods, low deployment efficiency, isolated and uncoordinated equipment, reliance on cloud computing, and lack of health prediction, significantly improving the safety protection level of road maintenance work sites.

[0018] The advantages of the present invention, as well as its additional advantages, will be described in detail in the following specific embodiments. Attached Figure Description

[0019] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute a limitation thereof.

[0020] Figure 1 This is a schematic diagram of a microwave radar-based active early warning security sentinel system disclosed in Embodiment 1 of the present invention; Figure 2 This is a schematic diagram of the individual soldier deployment of the active early warning security sentry system based on microwave radar disclosed in Embodiment 1 of the present invention; Figure 3 This is a schematic diagram of the dual-soldier deployment of the active early warning security sentry system based on microwave radar disclosed in Embodiment 1 of the present invention; Figure 4 This is a schematic diagram of the multi-soldier deployment of the active early warning security sentry system based on microwave radar disclosed in Embodiment 1 of the present invention. Detailed Implementation

[0021] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0022] It should be noted that the following detailed descriptions are exemplary and intended to provide further illustration of the invention. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0023] It should be noted that the terminology used herein is for describing particular embodiments only and is not intended to limit the exemplary embodiments of the present invention. As used herein, the singular form is intended to include the plural form as well, unless the context clearly indicates otherwise. Furthermore, it should be understood that when the terms "comprising" and / or "including" are used in this specification, they indicate the presence of features, steps, operations, devices, components, and / or combinations thereof. It should be noted that, without conflict, the various embodiments and features within those embodiments can be combined with each other. The embodiments will now be described in detail with reference to the accompanying drawings.

[0024] Example 1 In a typical embodiment of the present invention, such as Figures 1 to 4 As shown in the figure, this embodiment discloses a microwave radar-based active early warning security sentinel system, including: a security sentinel body, a wearable personnel early warning terminal, and a cloud-based situational control platform.

[0025] The main body of the security sentinel includes a cone-shaped shell and an adaptive millimeter-wave radar module, a UWB self-organizing network communication module, an edge computing unit, and an early warning module installed inside the cone-shaped shell.

[0026] In this embodiment, the cone housing adopts the standard road crash cone shell structure. The exterior of the housing maintains the standard cone shape and can directly replace traditional traffic cones for road construction, maintenance, and emergency operations. The housing has an IP68 waterproof and dustproof rating, can withstand vehicle pressure without damaging internal electronic components, and ensures long-term reliable operation of the equipment in harsh road environments.

[0027] The conical shell has an internal space for accommodating various electronic modules, including an adaptive millimeter-wave radar module, a UWB self-organizing network communication module, an edge computing unit, and an early warning module.

[0028] In this embodiment, the adaptive millimeter-wave radar module is used to detect vehicle targets. It integrates a multi-angle detection antenna array, which can detect and identify vehicles within a preset range regardless of its orientation.

[0029] Specifically, the adaptive millimeter-wave radar module employs an adaptive small-angle millimeter-wave radar array, which is a multi-layered fan-shaped staggered antenna array integrated with a 77GHz automotive-grade millimeter-wave radar chip. This multi-layered fan-shaped staggered antenna array consists of multiple fan-shaped antenna elements with different orientations arranged in a multi-layered staggered manner. Each antenna element has a different beam pointing angle, and when combined, they form a continuous detection fan-shaped area covering 0~120° in the horizontal direction, with a detection range of 0~50m. It can accurately detect the distance, speed, and azimuth information of vehicles approaching from a distance. Furthermore, the 77GHz band millimeter waves have the ability to penetrate rain, snow, and fog, allowing for all-weather operation unaffected by lighting conditions.

[0030] In this embodiment, the adaptive millimeter-wave radar module is communicatively connected to the edge computing unit, transmitting radar echo signals to the edge computing unit for processing. The edge computing unit has a pre-built angle compensation algorithm that calculates the target's azimuth angle relative to the current orientation of the security sentry host based on the intensity distribution and phase difference of the echo signals received by each antenna unit. It also performs coordinate transformation and amplitude compensation on the detection data under different orientations, ensuring that regardless of the horizontal orientation of the security sentry host, a unified vehicle target detection coordinate system can be established with the host as the center. Furthermore, the built-in angle compensation algorithm enables the security sentry host to detect and identify vehicles approaching from behind when deployed in any horizontal orientation from 0° to 120°, without requiring manual adjustment of the radar orientation.

[0031] In this embodiment, the UWB self-organizing network communication module is used to automatically pair and network multiple security sentinel hosts, and generate a dynamic electronic fence based on the UWB positioning data of multiple security sentinel hosts through triangulation.

[0032] Specifically, the UWB self-organizing network communication module uses an ultra-wideband wireless communication chip, operates in the UWB band, and features high speed, low power consumption, and centimeter-level ranging accuracy. This UWB self-organizing network communication module has a built-in keyless automatic pairing algorithm. After multiple security sentinel hosts are powered on, each UWB self-organizing network communication module automatically sends network broadcast signals and receives network broadcast signals from surrounding devices on preset frequencies. The keyless automatic pairing algorithm completes device discovery and pairing without manual operation. After power-on, it scans for surrounding devices and completes network formation within 300ms.

[0033] In this embodiment, the UWB ad hoc network communication module of each security sentinel calculates the distance between devices by measuring the signal flight time between them, and transmits the ranging data to the edge computing unit. When there are three or more security sentinel hosts in the network, the edge computing unit calculates the relative position of each device in a unified coordinate system based on the mutual ranging data between the devices, using a trigonometric positioning algorithm, and generates a continuous dynamic electronic fence. This dynamic electronic fence is a polygonal closed area formed by connecting the positions of each security sentinel host.

[0034] As a further implementation, the UWB self-organizing network communication module and the wearable personnel early warning terminal can receive the UWB positioning signal sent by the wearable personnel early warning terminal, measure the distance between each wearable personnel early warning terminal and multiple safety sentry hosts, and transmit the ranging data to the edge computing unit, which calculates the position coordinates of each worker relative to the dynamic electronic fence based on the triangulation algorithm.

[0035] As a further implementation, a centimeter-level positioning module is also installed inside the conical shell, which includes a GNSS satellite positioning unit and a UWB positioning unit. The GNSS satellite positioning unit includes a multi-band satellite signal receiving antenna and a GNSS baseband processing chip, capable of receiving signals from the Global Navigation Satellite System and outputting the latitude and longitude coordinates of the current security sentinel host. The UWB positioning unit includes a UWB transceiver antenna and a UWB ranging chip. The UWB ranging chip transmits and receives ultra-wideband pulse signals through the UWB transceiver antenna, performing bidirectional ranging with the UWB self-organizing network communication modules of other security sentinel hosts in the network, measuring the straight-line distance between the current security sentinel host and other security sentinel hosts. The centimeter-level positioning module primarily uses satellite positioning data from the GNSS satellite positioning unit, and uses the relative distance between the current security sentinel host and other security sentinel hosts in the network measured by the UWB positioning unit as a correction factor to continuously calculate the precise position of the current security sentinel host and output its device coordinates. Meanwhile, the UWB positioning unit communicates bidirectionally with the UWB positioning chip of the wearable personnel early warning terminal to measure the straight-line distance between the current security sentinel host and each wearable personnel early warning terminal, and outputs the personnel coordinates.

[0036] As a further implementation, a power management module is also installed inside the cone-shaped shell of the safety sentinel main body. Specifically, the power management module includes a lithium battery pack, a charge / discharge management circuit, and a power monitoring circuit. The lithium battery pack is a rechargeable lithium battery that powers all electronic modules within the safety sentinel main unit; the charge / discharge management circuit connects to the lithium battery pack, manages the charging and discharging process of the lithium battery, and provides overcharge protection, over-discharge protection, and short-circuit protection; the power monitoring circuit collects the voltage, current, and temperature parameters of the lithium battery pack in real time, calculates the remaining power and the number of charge / discharge cycles, and transmits the power data to the edge computing unit via a data bus.

[0037] In this embodiment, the edge computing unit is used to receive the detection data from the adaptive millimeter-wave radar module and perform vehicle target identification and early warning determination locally; when it is determined that a vehicle has entered, the edge computing unit triggers the early warning module to issue an early warning signal, and pushes the early warning command to other security sentinel hosts in the network through the UWB self-organizing network communication module.

[0038] Specifically, the edge computing unit employs an embedded processor with built-in non-volatile memory, which stores vehicle target recognition programs, warning determination programs, and offline cache scheduling programs. The edge computing unit connects to the adaptive millimeter-wave radar module, UWB self-organizing network communication module, centimeter-level positioning module, and warning module via a data bus, receiving raw data uploaded by each module and issuing control commands.

[0039] In this embodiment, the edge computing unit receives the echo signal from the adaptive millimeter-wave radar module in real time. After amplification, filtering, and analog-to-digital conversion, it runs a target detection algorithm to analyze vehicle targets in real time. Vehicle targets are extracted using distance-dimensional Fast Fourier Transform (FFT), velocity-dimensional FFT, and angle-dimensional FFT, respectively, and the vehicle's distance, speed, and azimuth relative to the current security sentry host are calculated. After constant false alarm rate (CFAR) detection to filter out noise and clutter, a clustering algorithm groups spatially adjacent targets with similar speeds into the same vehicle target, ultimately generating a vehicle target list containing distance, speed, azimuth, and signal-to-noise ratio. The distance and speed of vehicles in the target list are detected in real time. When the calculated vehicle distance is less than a preset warning distance threshold and the vehicle speed is greater than a preset speed threshold, the edge computing unit determines that a vehicle has entered the warning area and immediately generates a vehicle intrusion warning signal, which is sent to the warning module. This target detection algorithm runs locally in real time on the edge computing unit, with a single-frame processing cycle of no more than 50ms. It does not rely on cloud computing, ensuring complete local vehicle perception and warning capabilities even during network interruptions.

[0040] In this embodiment, the edge computing unit receives personnel ranging data uploaded by the UWB self-organizing network communication module in real time, calculates the real-time coordinates of each worker in the electronic fence coordinate system using a triangulation algorithm, and performs a spatial geometric comparison between the coordinates of each worker and the polygonal boundary of the dynamic electronic fence to determine whether any worker is outside the boundary. When the coordinates of a worker are outside the boundary of the dynamic electronic fence, a worker boundary crossing warning signal is immediately generated and sent to the corresponding wearable worker warning terminal through the UWB self-organizing network communication module.

[0041] In this embodiment, the edge computing unit collects the signal-to-noise ratio of the adaptive millimeter-wave radar module, the packet loss rate of the UWB self-organizing network communication module, the battery voltage and charge-discharge cycle count of the power management module, and the processor temperature in real time. The collected multi-dimensional operating parameters are input into a preset hierarchical threshold-attenuation coupling judgment model to determine whether the device has a failure risk. If the device is determined to have a failure risk, a device fault warning signal is immediately generated and sent to the warning module.

[0042] The layered threshold-attenuation coupling judgment model adopts a two-layer parallel judgment structure. The first layer is a static hard threshold judgment layer, used to detect sudden failures; the second layer is a dynamic attenuation fitting judgment layer, used to predict progressive degradation failures. The two judgment layers operate in parallel and independently. The output of a failure warning signal by either judgment layer triggers an operation and maintenance warning, thereby achieving the dual capabilities of immediate alarm for sudden failures and early prediction of progressive degradation.

[0043] Specifically, the static hard threshold determination layer presets a set of safety thresholds for each operating parameter, including an upper limit threshold and / or a lower limit threshold. It compares the measured value of each operating parameter with the corresponding safety threshold. If the measured value of an operating parameter exceeds its preset safety threshold range, the static hard threshold determination layer immediately outputs a hard threshold over-limit alarm signal for that operating parameter. The alarm signal includes the over-limit parameter number, the measured value, and the degree of over-limit. When multiple operating parameters exceed their limits simultaneously, the alarm signal corresponding to each operating parameter is output separately to ensure that no anomaly is missed. When all operating parameters are within the safety threshold range, the static hard threshold determination layer outputs a normal operating status signal.

[0044] The dynamic degradation fitting judgment layer predicts the gradual degradation of radar sensitivity and battery capacity based on long-term operating curves, and pushes maintenance work orders in advance. Its core logic is to capture the pattern of gradual degradation of equipment performance parameters over time based on historical time-series data of key equipment parameters, such as the gradual decrease of millimeter-wave radar signal-to-noise ratio, continuous degradation of battery capacity, slow increase of UWB communication packet loss rate, and gradual drift of processor temperature baseline. It establishes a degradation trend model to fit its performance degradation trajectory, calculates the remaining margin of current parameters to preset failure thresholds, and outputs gradual degradation warnings in advance before parameters actually exceed the limits. This provides maintenance personnel with sufficient handling window, avoids equipment operating with defects and forming hidden safety blind spots, and makes up for the shortcomings of the static hard threshold judgment layer, which can only detect anomalies that have occurred but cannot predict future failure risks.

[0045] In this embodiment, the dynamic decay fitting decision layer adopts a lightweight statistical algorithm, which does not require a large number of fault samples for offline training, has low computational complexity, is compatible with the limited computing resources of edge computing units, and can run locally in real time without relying on cloud support.

[0046] In this embodiment, the dynamic attenuation fitting decision layer operates according to the process of data buffering, preprocessing, trend fitting, RUL calculation, and graded decision.

[0047] Specifically, firstly, the edge computing unit opens a circular buffer to store historical sampled values ​​of each monitoring parameter at fixed intervals, retaining data from the most recent 30 days by default. To address random fluctuations in the raw data, a sliding window midpoint filtering algorithm is used to remove outliers and eliminate abnormal jumps caused by transient interference, ensuring that the data sequence input to the fitting model conforms to the physical law that performance only deteriorates and does not recover.

[0048] Next, the corresponding degradation model is fitted using the least squares method to the preprocessed time series data sequence, and the degradation slope is calculated. To address the constraints of edge computing power and the real-time requirements of the system, the dynamic decay fitting decision layer uses a classic degradation fitting algorithm that is highly interpretable, computationally inexpensive, and requires no offline training. Different models are matched according to the parameter decay characteristics, as shown in Table 1.

[0049] Table 1. Parameter Attenuation Characteristic Matching Model

[0050] For parameters such as signal-to-noise ratio, packet loss rate, and temperature baseline that exhibit linear decay, a univariate linear regression model is directly used for fitting, and the current degradation slope is solved using the least squares method. and initial baseline value For parameters such as battery voltage / capacity that decay exponentially, a log-linearized exponential model is used. The original data is first converted to a linear sequence by taking the natural logarithm before fitting the model, and then the decay coefficient is solved using the least squares method. and initial baseline value .

[0051] After fitting is complete, calculate the goodness of fit. To determine whether the data sequence exhibits a significant degradation trend. When When the data falls below a preset goodness-of-fit threshold, it is determined that the current data fluctuates significantly and shows no significant degradation trend; therefore, no warning signal is output to avoid false alarms. If the goodness-of-fit threshold is not lower than the preset threshold, the degradation trend is determined to be valid, and the degradation slope is output. Used for subsequent remaining lifetime calculations. As an example, the preset goodness-of-fit threshold is preferably 0.7.

[0052] Subsequently, based on the fitted degradation model, the remaining time for the current parameters to reach the preset failure threshold is calculated, i.e. :

[0053] In the formula, The failure thresholds for the corresponding parameters are as follows: radar signal-to-noise ratio below 10dB, battery voltage below 3.3V, and packet loss rate above 5%. These are the measured values ​​of the current parameters; The slope is the degradation rate, taken as its absolute value.

[0054] Finally, based on the degradation slope and the calculated It outputs a three-level progressive degradation warning. Specifically, when... Exceeding the first remaining time threshold and the degradation slope If the slope is lower than 1.2 times the preset baseline slope, it is considered a normal state, and a normal operation status signal is output; when... Between the first remaining time threshold and the second remaining time threshold, or the degradation slope If the slope exceeds 1.5 times the preset baseline, it is judged as a Level 1 warning state, and maintenance is recommended. At this time, a sub-health warning is output, marked as a predictive maintenance recommendation, and an operation and maintenance work order is generated by uploading it to the cloud. Local audible and visual alarms are not triggered. Less than the second remaining time threshold, or the degradation slope If the slope exceeds 2.5 times the baseline slope, it is judged as a level 2 warning, requiring emergency maintenance. At this time, an imminent failure warning is output, an emergency maintenance work order is generated in the cloud, and the local warning module is triggered to flash at a low frequency to remind maintenance personnel to pay attention.

[0055] The static hard threshold judgment layer and dynamic attenuation fitting judgment layer of the hierarchical threshold-attenuation coupling judgment model operate in parallel and independently. When both the static hard threshold judgment layer and the dynamic attenuation fitting judgment layer output early warning signals, the alarm signal output by the static hard threshold judgment layer is processed with the highest priority, immediately triggering a local audible, visual, and vibration early warning. At the same time, all early warning signals output by the dynamic attenuation fitting judgment layer are automatically invalidated. When the static hard threshold judgment layer does not output an alarm but the dynamic attenuation fitting judgment layer outputs a progressive attenuation early warning, the edge computing unit marks the early warning as a predictive maintenance suggestion and uploads it to the cloud management platform via the UWB self-organizing network and 4G / 5G network to generate an operation and maintenance work order, but does not trigger an alarm in the on-site early warning module to avoid unnecessary interference to on-site personnel. When neither the static hard threshold judgment layer nor the dynamic attenuation fitting judgment layer outputs an alarm, the device outputs a normal operating status signal. Through the coupling and collaborative mechanism of the static hard threshold judgment layer and the dynamic attenuation fitting judgment layer in the hierarchical threshold-attenuation coupling judgment model, the system can respond to sudden failures in real time through the static layer and predict the risk of gradual degradation in advance through the dynamic layer. This achieves dual protection capabilities of immediate alarm for sudden failures and early prediction of gradual degradation, eliminating the hidden safety blind spots caused by equipment operating with defects.

[0056] In this embodiment, the early warning module is specifically a multi-mode sound, light, and vibration early warning module, which integrates a strobe LED array, a high-decibel buzzer, and a driving circuit. It is connected to the edge computing unit and receives early warning driving signals sent by the edge computing unit. When the early warning signal is a vehicle intrusion warning signal, the strobe LEDs are driven to flash at the highest frequency and the high-decibel buzzer sounds at the maximum volume; when the early warning signal is a person crossing the boundary warning signal, the strobe LEDs are driven to flash at a medium frequency and the high-decibel buzzer sounds at a medium volume; when the early warning signal is an equipment failure warning signal, the strobe LEDs are driven to flash at a specific frequency to indicate the type of failure.

[0057] In this embodiment, a 4G / 5G cloud communication unit is also installed inside the cone-shaped shell of the security sentry body. Specifically, the 4G / 5G cloud communication unit uses a cellular mobile communication module, connects to the edge computing unit, and establishes a data connection with the remote cloud management platform through a 4G or 5G mobile communication network. Under normal network conditions, the 4G / 5G cloud communication unit uploads the early warning event data, device health data, and personnel location data generated by the edge computing unit to the cloud management platform in real time; when the edge computing unit detects a network interruption, the 4G / 5G cloud communication unit enters a sleep state, without consuming local computing power and power.

[0058] In this embodiment, the wearable personnel early warning terminal is a portable device for workers to wear, including a UWB positioning chip, a micro vibration motor, and a dual-color sound and light lamp.

[0059] Specifically, the UWB positioning chip uses ultra-wideband communication technology to communicate bidirectionally with the UWB self-organizing network communication module of the security sentinel host. On the one hand, it receives the early warning commands issued by the security sentinel host, and on the other hand, it actively sends UWB positioning signals for the security sentinel host to measure distance and locate.

[0060] The miniature vibration motor is connected to the control output of the UWB positioning chip and can receive vibration drive signals sent by the security sentry host through the UWB self-organizing network communication module to generate vibration tactile alerts.

[0061] The dual-color light and sound system includes a red LED, a yellow LED, and a buzzer, and can receive light and sound drive signals from the security sentinel host.

[0062] In this embodiment, the UWB positioning chip has a built-in early warning command parsing program that can distinguish the type of received early warning command and generate corresponding vibration drive signals. Specifically, when a vehicle enters the construction area from behind, an early warning command is received, driving a micro vibration motor to vibrate at the highest frequency, simultaneously driving a red LED to flash at a high frequency and driving a buzzer to emit a high-frequency buzzing sound, achieving a combined warning of strong vibration and high-frequency red light, reminding personnel to evacuate immediately; when a personnel crosses the boundary of the UWB electronic fence, an early warning command is received, driving a micro vibration motor to vibrate at a low frequency, simultaneously driving a yellow LED to continuously light up and driving a buzzer to emit a low-frequency intermittent alert sound, achieving a low-frequency vibration accompanied by continuous yellow light alarm, reminding personnel that they have exceeded the safe area and should return to the construction area.

[0063] As a further implementation, the wearable personnel early warning terminal has a waterproof shell and a long-lasting battery, making it suitable for all-weather wear during outdoor maintenance operations.

[0064] In this embodiment, the cloud-based situational awareness management platform is a software platform deployed on a cloud server. It establishes a data connection with the 4G / 5G cloud communication units of each security sentinel host through a 4G or 5G network and receives the full amount of data synchronously uploaded by the edge computing unit.

[0065] In this embodiment, the cloud-based management platform integrates a data storage module, a visualization module, a maintenance work order generation module, and a security management module. The data storage module stores all warning event data, equipment health data, personnel trajectory data, and equipment operation logs uploaded by all safety sentinel hosts. The visualization module displays the real-time location and status of safety sentinel hosts at each construction site, as well as vehicle alarm events and personnel location distribution on a map. The maintenance work order generation module automatically generates maintenance work orders based on maintenance warnings generated by the predictive maintenance module from the equipment health data and pushes them to relevant maintenance personnel. The security management module stores and retrieves full-process event data, supporting accident tracing, safety statistics, and management optimization analysis.

[0066] It should be noted that the cloud-based management platform in this embodiment only undertakes the above-mentioned data storage, display, operation and maintenance management and background analysis functions, and does not participate in the on-site real-time early warning calculation. The on-site real-time vehicle intrusion judgment, personnel boundary crossing judgment and early warning triggering are all completed independently by the edge computing unit of the security sentinel host.

[0067] The specific workflow of a microwave radar-based active early warning security sentinel system disclosed in this embodiment is as follows: (1) After the staff places one or more safety sentinel hosts in the predetermined location in the construction area, the safety sentinel hosts will automatically power on and start. The UWB self-organizing network communication module of each safety sentinel host will automatically send a network broadcast signal containing the device identification code on the preset frequency point, and at the same time scan and receive the network broadcast signals sent by similar devices in the vicinity. The device discovery and pairing interconnection will be completed through the keyless automatic pairing algorithm, and the network will be completed without manual operation. At the same time, the adaptive millimeter-wave radar module will automatically start working after power-on. The multi-layer fan-shaped staggered antenna array will receive echo signals from all directions. The angle compensation algorithm in the edge computing unit will automatically calculate the detection coordinate system under the current deployment direction based on the echo signal intensity distribution and phase difference of each antenna unit, and establish a unified vehicle target detection model. The radar orientation will be calibrated manually, and vehicle target detection in the range of 0~50m will be started immediately. It should be noted that a single safety sentinel host will work independently when deployed, and multiple hosts will automatically link together when deployed.

[0068] (2) The adaptive millimeter-wave radar module continuously transmits millimeter waves and receives echoes. The echo signals are transmitted to the edge computing unit. The edge computing unit runs a target detection algorithm to extract vehicle targets from the echo signals and calculates the vehicle distance, speed and azimuth. When the calculated vehicle distance is less than the preset warning distance threshold and the vehicle speed is greater than the preset speed threshold, the edge computing unit determines that a vehicle has entered. It immediately generates a vehicle intrusion warning signal and sends it to the sound, light and vibration multi-mode warning module. The module drives the strobe LED to strobe at the highest frequency and the high-decibel buzzer to sound at the maximum volume to realize the local sound and light warning.

[0069] Meanwhile, the edge computing unit broadcasts the vehicle intrusion warning command to all other safety sentinel hosts and all paired wearable personnel warning terminals within the network via the UWB self-organizing network communication module. All safety sentinel hosts within the network simultaneously trigger audible and visual warnings, and all wearable personnel warning terminals simultaneously trigger strong vibration and high-frequency red light warnings, achieving multi-machine, multi-point synchronization and multiple warning methods including sound, light, and vibration.

[0070] (3) When there are three or more security sentinel hosts in the network, the UWB self-organizing network communication module of each security sentinel host obtains the distance data between devices through mutual ranging and transmits it to the edge computing unit. Based on the mutual ranging data between devices, the edge computing unit uses the trigonometric positioning algorithm to calculate the relative position of each device in a unified coordinate system, and connects the positions of each security sentinel host as vertices to form a polygonal closed area as a dynamic electronic fence.

[0071] The wearable personnel early warning terminal actively sends UWB positioning signals through its UWB positioning chip. The UWB self-organizing network communication module of each safety sentry host receives the positioning signal and measures the distance between itself and the wearable personnel early warning terminal, transmitting the distance measurement data from multiple devices to the edge computing unit. Based on the distance measurement data from multiple safety sentry hosts to the same wearable personnel early warning terminal, the edge computing unit calculates the real-time coordinates of the worker relative to the dynamic electronic fence coordinate system using a triangulation algorithm.

[0072] The edge computing unit performs a spatial geometric comparison between the real-time coordinates of each worker and the polygonal boundary of the dynamic electronic fence. When a worker's real-time coordinates are outside the polygonal boundary of the dynamic electronic fence, the edge computing unit determines that the worker has crossed the boundary and immediately generates a worker boundary crossing alarm command. This command is pushed to the wearable personnel warning terminal worn by the worker through the UWB self-organizing network communication module, triggering a low-frequency vibration and continuous yellow light alarm until the worker retreats back inside the dynamic electronic fence.

[0073] (4) Throughout the system's operation, the edge computing unit continuously collects multi-dimensional operating parameters in parallel. This includes radar signal-to-noise ratio from the adaptive millimeter-wave radar module, communication packet loss rate from the UWB self-organizing network communication module, battery voltage and charge / discharge cycle count from the power management module, and processor operating temperature from the edge computing unit's own temperature sensor. The collected multi-dimensional operating parameters are then processed by a pre-defined hierarchical threshold-attenuation coupling judgment model. Sudden failures are detected based on a static hard threshold judgment layer, while progressive degradation failures are predicted based on a dynamic attenuation fitting judgment layer. The two judgment layers operate independently in parallel. The output of a failure warning signal from either judgment layer triggers an operation and maintenance warning, achieving the dual capability of immediate alarm for sudden failures and early prediction of progressive degradation.

[0074] (5) The edge computing unit continuously monitors the network connection status of the 4G / 5G cloud communication unit. When a network interruption is detected, the edge computing unit independently completes vehicle detection, UWB self-organizing network, early warning judgment and predictive maintenance calculations locally. At the same time, the edge computing unit writes all radar detection data, early warning event data, personnel trajectory data and equipment operating parameter data generated during the network interruption into the local storage cache unit for caching.

[0075] Once the 4G / 5G cloud communication unit detects network recovery, the edge computing unit automatically reads all cached data stored in the local storage cache unit and uploads it in batches to the cloud management platform in chronological order via the 4G / 5G cloud communication unit, enabling breakpoint resume. The cloud management platform then receives the uploaded data and performs storage, display, and security management analysis.

[0076] Example 2 In a typical embodiment of the present invention, this embodiment discloses an early warning method based on a microwave radar active early warning security sentinel system, comprising: Deploy one or more security sentinel hosts; the UWB self-organizing network communication module automatically scans surrounding devices and completes network formation; the adaptive millimeter-wave radar module automatically adapts to any placement angle and activates vehicle target detection. When the adaptive millimeter-wave radar module detects a vehicle entering the warning area from behind, the edge computing unit locally triggers the local warning module to issue an audible and visual warning. At the same time, the warning command is pushed to all safety sentinel hosts and all wearable personnel warning terminals in the network through the UWB self-organizing network communication module. Three or more security sentinel hosts generate a dynamic electronic fence through UWB triangulation positioning. Wearable personnel early warning terminals upload personnel location information in real time. The edge computing unit compares the personnel location information with the electronic fence boundary. When a person crosses the boundary, the wearable personnel early warning terminal is triggered to issue a boundary crossing alarm. The edge computing unit collects operating parameters in real time and inputs them into the hierarchical threshold-attenuation coupling judgment model. When the device is determined to have a failure risk, an operation and maintenance warning is generated and uploaded to the cloud management and control platform.

[0077] While the specific embodiments of the present invention have been described above in conjunction with the accompanying drawings, this is not intended to limit the scope of protection of the present invention. Those skilled in the art should understand that various modifications or variations that can be made by those skilled in the art without creative effort based on the technical solutions of the present invention are still within the scope of protection of the present invention.

Claims

1. A microwave radar-based active early warning security sentinel system, characterized in that, include: The main body of the security sentinel includes a cone-shaped shell and an adaptive millimeter-wave radar module, a UWB self-organizing network communication module, an edge computing unit, and an early warning module disposed inside the cone-shaped shell; Among them, the adaptive millimeter-wave radar module is used to detect vehicle targets. It integrates a multi-angle detection antenna array, which can detect and identify vehicles within a preset range even when deployed in any orientation. The UWB self-organizing network communication module is used to automatically pair and network multiple security sentinel hosts, and generate a dynamic electronic fence through triangulation based on the UWB positioning data of multiple security sentinel hosts. The edge computing unit is configured to receive detection data from the adaptive millimeter-wave radar module and perform vehicle target identification and early warning determination locally; when a vehicle is detected to have entered, the edge computing unit triggers the early warning module to issue an early warning signal and pushes the early warning command to other security sentinel hosts in the network through the UWB self-organizing network communication module.

2. The microwave radar-based active early warning security sentinel system as described in claim 1, characterized in that, It also includes wearable personnel early warning terminals, which have a built-in UWB positioning chip; the UWB positioning chip communicates bidirectionally with the UWB self-organizing network communication module of the security sentinel host to upload the location information of the wearer to the security sentinel host in real time. The edge computing unit is also configured to: based on the personnel location information uploaded by the UWB positioning chip, compare the personnel location information with the boundary of the dynamic electronic fence, and trigger the wearable personnel early warning terminal to issue a boundary crossing alarm when the personnel cross the dynamic electronic fence.

3. The microwave radar-based active early warning security sentinel system as described in claim 1, characterized in that, It also includes a cloud-based management platform; The cloud-based management platform communicates with the security sentinel host via a 4G or 5G network. It receives early warning event data, device health data, and personnel location data uploaded by the security sentinel host, and performs situation visualization, alarm storage, and maintenance work order generation.

4. The microwave radar-based active early warning security sentinel system as described in claim 1, characterized in that, The adaptive millimeter-wave radar module is specifically a multi-layer fan-shaped staggered antenna array integrated with a 77GHz automotive-grade millimeter-wave radar chip. The multi-layer fan-shaped staggered antenna array is composed of multiple fan-shaped antenna elements with different orientations arranged in a multi-layer staggered manner. Each antenna element has a different beam pointing angle, and after combination, it forms a continuous detection fan-shaped area covering the horizontal direction from 0 to 120°.

5. A microwave radar-based active early warning security sentinel system as described in claim 1, characterized in that, The edge computing unit receives detection data from the adaptive millimeter-wave radar module and performs vehicle target identification and warning determination locally, specifically including: The edge computing unit receives the echo signal from the adaptive millimeter-wave radar module in real time. After amplification, filtering, and analog-to-digital conversion, it runs a target detection algorithm to extract vehicle targets and calculates the vehicle's distance, speed, and azimuth relative to the current security sentry host, generating a vehicle target list. It also monitors the distance and speed of vehicles in the vehicle target list in real time. When the calculated vehicle distance is less than a preset warning distance threshold and the vehicle speed is greater than a preset speed threshold, the edge computing unit determines that a vehicle has entered the warning area in the vehicle target list, immediately generates a vehicle intrusion warning signal, and sends it to the warning module.

6. A microwave radar-based active early warning security sentinel system as described in claim 2, characterized in that, The edge computing unit uses the personnel location information uploaded by the UWB positioning chip to compare the personnel location information with the boundary of the dynamic electronic fence. When a person crosses the dynamic electronic fence, the wearable personnel early warning terminal is triggered to issue a boundary crossing alarm, specifically including: The edge computing unit receives personnel ranging data uploaded by the UWB self-organizing network communication module in real time, calculates the real-time coordinates of each worker in the electronic fence coordinate system through the triangulation algorithm, and performs spatial geometric comparison between the coordinates of each worker and the polygonal boundary of the dynamic electronic fence to determine whether there are any workers outside the boundary. When the coordinates of a worker are outside the boundary of the dynamic electronic fence, an early warning signal for personnel crossing the boundary is immediately generated and sent to the corresponding wearable personnel early warning terminal through the UWB self-organizing network communication module.

7. A microwave radar-based active early warning security sentinel system as described in claim 1, characterized in that, The main body of the security sentry is also equipped with a power management module; The edge computing unit is also configured to: collect the signal-to-noise ratio of the adaptive millimeter-wave radar module, the packet loss rate of the UWB self-organizing network communication module, the battery voltage and charge-discharge cycle count of the power management module, and the processor temperature in real time, and input the collected multi-dimensional operating parameters into a preset hierarchical threshold-attenuation coupling judgment model to determine whether the device has a failure risk. If it is determined that there is a risk of equipment failure, an equipment failure warning signal will be generated immediately and sent to the warning module.

8. A microwave radar-based active early warning security sentinel system as described in claim 7, characterized in that, The hierarchical threshold-attenuation coupling judgment model adopts a two-layer parallel judgment structure. The first layer is a static hard threshold judgment layer, which is used to detect sudden failures. The second layer is a dynamic attenuation fitting judgment layer, which is used to predict progressive degradation failures. The static hard threshold judgment layer and the dynamic attenuation fitting judgment layer run in parallel and independently. The output of a failure warning signal by either judgment layer will trigger an operation and maintenance warning.

9. A microwave radar-based active early warning security sentinel system as described in claim 1, characterized in that, The warning module is a multi-mode warning module with sound, light and vibration, which integrates a strobe LED array, a high-decibel buzzer and a drive circuit. When the warning signal is a vehicle intrusion warning signal, the strobe LEDs are driven to flash at the highest frequency and the high-decibel buzzer sounds at the maximum volume. When the warning signal is a personnel crossing the boundary warning signal, the strobe LED light is driven to flash at a medium frequency and the high-decibel buzzer sounds at a medium volume; when the warning signal is an equipment fault warning signal, the strobe LED light is driven to flash at a specific frequency to indicate the fault type.

10. A warning method based on a microwave radar active early warning security sentinel system, characterized in that, include: Deploy one or more security sentinel hosts, and the UWB self-organizing network communication module automatically scans surrounding devices and completes network formation; The adaptive millimeter-wave radar module automatically adapts to any placement angle to enable vehicle target detection; When the adaptive millimeter-wave radar module detects a vehicle entering the warning area from behind, the edge computing unit locally triggers the local warning module to issue an audible and visual warning. At the same time, the warning command is pushed to all safety sentinel hosts and all wearable personnel warning terminals in the network through the UWB self-organizing network communication module. Three or more security sentinel hosts generate a dynamic electronic fence through UWB triangulation positioning. Wearable personnel early warning terminals upload personnel location information in real time. The edge computing unit compares the personnel location information with the electronic fence boundary. When a person crosses the boundary, the wearable personnel early warning terminal is triggered to issue a boundary crossing alarm. The edge computing unit collects operating parameters in real time and inputs them into the hierarchical threshold-attenuation coupling judgment model. When the device is determined to have a failure risk, an operation and maintenance warning is generated and uploaded to the cloud management and control platform.

Images