Edge relay type communication control method and system for strong electromagnetic interference environment

Abstract

The application relates to the technical field of edge relay type communication control, and provides an edge relay type communication control method and system for a strong electromagnetic interference environment. The edge relay type communication control method for the strong electromagnetic interference environment comprises the following steps: receiving all to-be-communicated related data, and forming a unified task queue; performing a task attribute and environment state identification operation according to the unified task queue; adopting a two-layer communication control mechanism with an edge relay as an execution subject to perform path optimization; in the first layer communication, a communication mode, a frequency band / working channel and a sending mode are selected according to the task attribute and the environment state, and a schedulable path set is formed; in the second layer communication, a main path, a backup path and a preset low-overlap self-healing path are found out in a network corresponding to the selected communication mode based on the schedulable path set, so that key alarm loss caused by single link failure can be avoided.

Description

Technical Field

[0001] This invention relates to the field of edge relay communication control technology, and in particular to an edge relay communication control method and system for environments with strong electromagnetic interference. 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 the confined spaces of nuclear power plants, under conditions of strong electromagnetic interference, communication denial, and high safety constraints, such as complex environments like nuclear power plant maintenance tunnels, underground tunnels, and industrial manholes, traditional single communication methods are prone to high latency, high packet loss, and disconnection. Furthermore, existing communication scheduling mainly relies on fixed links, static priorities, or single-path routing, which makes it impossible to reliably transmit critical safety alarms such as SOS, fall alarms, and gas over-limit alarms. When the public network is interrupted or the link deteriorates, the system cannot form a closed loop and it is difficult to provide deterministic transmission guarantees. Summary of the Invention

[0004] To address the aforementioned technical problems, this invention provides an edge relay communication control method and system for environments with strong electromagnetic interference. It can construct a highly reliable communication system with anti-interference and self-healing capabilities, centered on task reliability, network resilience, and edge autonomy. This system uses the edge relay box as the scheduling entity and task classification, communication mode selection, and path optimization under system constraints as its core.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: The first aspect of the present invention provides an edge relay communication control method for environments with strong electromagnetic interference.

[0006] In one or more embodiments, an edge relay communication control method for environments with strong electromagnetic interference is provided, comprising: Receive all data to be communicated and form a unified task queue; Based on a unified task queue, perform task attribute and environment state identification operations; A two-layer communication control mechanism with edge relay as the main execution entity is adopted for path optimization; In the first layer of communication, the communication method, frequency band / working channel and transmission mode are selected according to the task attributes and environmental status to form a set of schedulable paths. In the second layer of communication, based on the set of schedulable paths, the main path, backup path and preset low-overlap self-healing path are found in the network corresponding to the selected communication method to avoid the loss of critical alarms due to the failure of a single link.

[0007] In one implementation, all data related to communication includes: monitoring data uploaded by the monitoring terminal via the communication module; image monitoring data; edge relay environment data; communication link status data; edge relay's own status data; task queue and remote configuration data.

[0008] As one implementation method, all data related to communication is encapsulated into a unified task queue using a custom lightweight binary protocol; the frame structure of the lightweight binary protocol includes a frame header, a length field, a type field, a priority field, a data payload, and a checksum field.

[0009] In one implementation, the task attributes include task categories, which include three levels: safety-critical flow, status monitoring flow, and business data flow.

[0010] As one implementation method, in the first-layer communication, a task-driven heterogeneous link intelligent scheduling algorithm is adopted to form a set of corresponding schedulable paths according to the priority of safety-critical flows, status monitoring flows, and business data flows, which includes: Safety-critical flows take precedence over status monitoring flows and business data flows. After the safety-critical flow is locked, the schedulable path of the status monitoring flow is then solved on the remaining resources; After ensuring the resources for safety-critical flows and status monitoring flows, the schedulable paths for business data are determined.

[0011] As one implementation method, during the process of solving the schedulable path for safety-critical flows, the safety-critical flow tasks are sorted according to their urgency; for each safety-critical flow task, a link that meets the worst-case interference constraint is selected; if a single link cannot meet the reliability requirement and the available communication method meets the redundancy condition, a redundant link is enabled; if there are insufficient available external transmission links, the external transmission link with the highest current reliability is enabled and a local alarm, cache record, and duplicate transmission mechanism are triggered; and safety-critical flow resources are locked to prevent subsequent status monitoring flows / service data flows from preempting them.

[0012] As one implementation method, in the second-layer communication, path optimization is performed based on graph-optimized resilient ad hoc networks and multi-path self-healing algorithms, which includes: Construct a dynamic graph in the network topology corresponding to the communication method; For communication methods with multiple relays, multiple gateways, or multiple exits, the main path, backup path, and preset low-overlap self-healing path are calculated based on the dynamic graph and real-time monitored link information, and based on the task-aware edge weights. For communication methods with only single-hop connections, link quality assessment, switching to alternative communication methods, and buffer retransmission control are performed based on dynamic graphs and real-time monitored link information.

[0013] As one implementation method, optimization is performed layer by layer in the order of safety-critical flow, status monitoring flow, and business data flow. When the main path is detected to be disturbed or the link fails, the backup path takes over first, and then the affected area is partially and incrementally reconstructed.

[0014] As one implementation method, for safety-critical task flows, the main path and backup path are solved simultaneously, and the risk of common failure is reduced by penalties for edge overlap, node overlap and interference area overlap.

[0015] A second aspect of the present invention provides an edge relay communication control system for environments with strong electromagnetic interference.

[0016] In one or more embodiments, an edge relay communication control system for environments with strong electromagnetic interference includes: The queue unification module is used to receive all data to be communicated and form a unified task queue; The information recognition module is used to perform recognition operations on task attributes and environmental status based on a unified task queue. The path optimization module is used to perform path optimization using a two-layer communication control mechanism with edge relay as the main execution body. In the first layer of communication, the communication method, frequency band / working channel and transmission mode are selected according to the task attributes and environmental status to form a set of schedulable paths. In the second layer of communication, based on the set of schedulable paths, the main path, backup path and preset low-overlap self-healing path are found in the network corresponding to the selected communication method to avoid the loss of critical alarms due to the failure of a single link.

[0017] In other embodiments, an edge relay communication control system for environments with strong electromagnetic interference includes: an intelligent monitoring terminal, an edge relay, and a cloud. The intelligent monitoring terminal is used to upload all monitoring information to the edge relay; The edge relay is used to perform the steps in the edge relay communication control method for strong electromagnetic interference environments described above. The cloud and edge relays communicate with each other.

[0018] A third aspect of the present invention provides a computer-readable storage medium.

[0019] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the edge relay communication control method for strong electromagnetic interference environments as described above.

[0020] A fourth aspect of the present invention provides an edge relay device.

[0021] An edge relay device includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it implements the steps of the edge relay communication control method for strong electromagnetic interference environments as described above.

[0022] Compared with the prior art, the beneficial effects of the present invention are: This invention provides an edge relay communication control method and system for environments with strong electromagnetic interference. It employs a two-layer communication control mechanism with edge relay as the main execution entity for path optimization. Based on task attributes and environmental conditions, it selects the communication mode, frequency band / working channel, and transmission mode to form a set of schedulable paths. Then, it finds the main path, backup path, and preset low-overlap self-healing path in the network corresponding to the selected communication mode. This solves the problems of reliable transmission of safety-critical data, edge closed-loop processing of multi-source data, and two-layer communication link control in environments with strong interference. It avoids the loss of critical alarms due to the failure of a single link, reduces the risk of simultaneous breakdown of the main path and backup path in a local strong electromagnetic interference area, and improves the communication self-healing capability in nuclear power plant maintenance corridor scenarios. Attached Figure Description

[0023] 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 an improper limitation of the invention.

[0024] Figure 1 This is a flowchart of an edge relay communication control method for strong electromagnetic interference environments according to an embodiment of the present invention; Figure 2 This is a schematic diagram of an edge relay structure according to an embodiment of the present invention. Detailed Implementation

[0025] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0026] It should be noted that the following detailed description is illustrative and intended to provide further explanation 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.

[0027] It should be noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the scope of exemplary embodiments according to the 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.

[0028] From a hardware structure perspective, an edge relay communication control system for strong electromagnetic interference environments according to an embodiment of the present invention includes: an intelligent monitoring terminal, an edge relay, and a cloud; the intelligent monitoring terminal is used to upload all monitoring information to the edge relay; the edge relay is used to execute the steps in the edge relay communication control method for strong electromagnetic interference environments; the cloud and the edge relay communicate with each other.

[0029] Specifically, intelligent monitoring terminals include, but are not limited to: wearable low-bandwidth personnel monitoring terminals, external video acquisition units, and environmental acquisition units; In this embodiment, edge relay is implemented using an edge relay box.

[0030] Wearable low-bandwidth personnel monitoring terminals are used to collect small amounts of status information such as heart rate, blood oxygen, respiratory rate, posture, fall detection, SOS signal, battery level, and wearing status of workers; external video acquisition units are connected to the edge relay box via Wi-Fi, Ethernet, USB, or industrial video interfaces; the internal environment acquisition unit is used to collect oxygen, carbon monoxide, combustible gas, temperature and humidity, and box temperature; the edge relay box is responsible for data aggregation, task classification, communication method selection, path optimization, local alarms, and cloud data transmission; the cloud monitoring platform is used for remote display, storage, and manual confirmation.

[0031] The wearable low-bandwidth personnel monitoring terminal is a dedicated smart watch with strong electromagnetic interference resistance. This terminal does not handle video acquisition or high-bandwidth data transmission; its primary function is to monitor the vital signs and posture of personnel in a low-power, low-bandwidth, and continuous manner. The wearable low-bandwidth personnel monitoring terminal integrates a heart rate and blood oxygen sensor, an inertial measurement unit, an SOS button, a wear detection module, a low-power communication module, and a linked power management chip. It can collect data on the worker's heart rate, blood oxygen, respiratory rate, body posture, fall status, wear status, terminal battery level, and active distress signals. An optional radiation dose detection module can be added to record the individual's radiation exposure status. Wearable low-bandwidth personnel monitoring terminals upload personnel status data via one or more of Bluetooth BLE, LoRa, NB-IoT, or other narrowband low-power communication methods to low-power receiving modules, NB-IoT communication modules, narrowband private network access gateways, or on-site low-bandwidth access gateways deployed in edge relay boxes or near the site. The edge relay box then receives, parses, and adds the data to the task queue. Under normal circumstances, status monitoring data such as heart rate, blood oxygen, posture, and battery level are uploaded in periodic small data packets. When a fall, prolonged stillness, SOS triggering, or abnormal vital signs are detected, the terminal immediately generates a critical safety alarm packet and uploads it to the edge relay box. The edge relay box can send control commands to the watch via a downlink low-power link, including adjusting sampling frequency, sleep mode, wake-up, vibration alerts, alarm confirmations, and heart rate cycle adjustments.

[0032] To ensure stable operation in the strong electromagnetic interference environment of nuclear power plants, the wearable low-bandwidth personnel monitoring terminal adopts the following industrial-grade anti-interference hardware design: Screen module and auxiliary absorbing material layer: An auxiliary absorbing material layer is attached between the display driving circuit and the sensor acquisition area to effectively absorb and attenuate electromagnetic waves coupled to the display area, and reduce crosstalk between the display driving signal and the sensor signal.

[0033] Radiation-resistant PCB: Utilizing PCB materials and reinforced wiring technology suitable for nuclear power plant maintenance environments, critical signal lines are grounded and impedance-matched to improve device lifespan and signal integrity in complex electromagnetic and radiation environments.

[0034] Electromagnetic reflection suppression structure: Absorbing and reflection suppression layers are set inside the housing and between PCB layers to suppress internal signal reflection and external electromagnetic coupling, thereby reducing the system bit error rate.

[0035] Low-power communication module and linked power management chip: The terminal integrates an industrial-grade low-power communication module, which can use one or more of BLE, LoRa, NB-IoT, or other narrowband low-power communication methods to upload small amounts of personnel status information to the receiving module, NB-IoT communication module, narrowband private network access gateway, or field low-bandwidth access gateway on the edge relay box side. When transmitting small amounts of data and high-reliability tasks such as emergency alarms, the terminal prioritizes generating safety-critical alarm packets and uploading them through a low-power, low-bandwidth link; the device also integrates a linked power management chip, which can dynamically adjust its working status according to the sampling frequency, sleep, wake-up, vibration alert, and heartbeat cycle adjustment instructions issued by the edge relay box, achieving energy-saving control while ensuring communication continuity.

[0036] For example, the main body size of the wearable low-bandwidth personnel monitoring terminal can be adjusted according to the specific structural design, preferably about 36mm~45mm × 43mm~45mm × 14mm, and the weight is controlled within 50g. The shell adopts a composite structure of ABS engineering plastic and metal shielding layer.

[0037] Main control chip: It adopts a low-power microcontroller chip, preferably the ESP32 series dual-core TensilicaXtensa LX6 microprocessor or an industrial-grade compatible control chip, with a maximum frequency of 240 MHz, and provides a variety of memory options, including up to 520 KB of SRAM and optional 4 MB of SPI Flash storage.

[0038] Sensor configuration: Heart rate and blood oxygen: GH3011 high-precision health sensor chip; Body posture: MPU6050 six-axis inertial measurement unit; An optional individual radiation dose sensor is available; environmental parameters such as oxygen concentration, carbon monoxide concentration, combustible gas concentration, and temperature and humidity are preferably collected by gas sensors built into the edge relay box and are not required sensor configurations for the watch.

[0039] Anti-interference design: The PCB has a four-layer structure, with the middle layer being the power layer and the ground layer. The differential signal line impedance is controlled within 100Ω ± 5%. A miniature metal shield (0.1mm thick copper foil, single-point grounding) is installed on the outside of the sensitive area (heart rate sensor, ADC acquisition circuit). The data acquisition circuit adopts a "single-point grounding" design; The antenna features an independent compartment design, with a metal isolation wall separating it from the main circuit.

[0040] Battery and power management: 500mAh lithium polymer battery, NBLP linked power management chip, and the power consumption of the whole machine in sleep mode is ≤2mA.

[0041] The external video acquisition unit is used to collect on-site video, images, or keyframe data in nuclear power plant maintenance corridors, work areas, or confined spaces. This video data is not collected by a wearable watch, but rather by fixed cameras, temporarily deployed cameras, externally mounted cameras, head-mounted cameras, or badge-style cameras, and then transmitted to the edge relay box.

[0042] External video acquisition units can be connected to edge relay boxes via Wi-Fi, Ethernet, USB, or industrial video interfaces. When Wi-Fi, industrial routing, or wired links are disrupted, bandwidth is insufficient, or safety-critical tasks preempt resources, the system reduces the frame rate and bit rate of video data, uploads only key frames, caches locally, or delays uploads to prevent video services from preempting communication resources for safety-critical tasks.

[0043] The environmental acquisition unit is used to collect the safety status of the operating environment of the work area and the edge relay box itself. This unit is preferably integrated inside a strong electromagnetic interference-adaptive edge relay box and communicates with the work environment through gas sampling ports on the box shell, a dustproof, waterproof, and breathable membrane, or an external sampling pipeline. The environmental acquisition unit includes an oxygen concentration sensor, a carbon monoxide sensor, a combustible gas sensor, a temperature and humidity sensor, a box temperature sensor, and an optional radiation dose sensor. These sensors are connected to the main control board via UART, I2C, SPI, ADC, or RS485 interfaces. The main control board performs sampling, filtering, timestamp marking, anomaly threshold judgment, and protocol encapsulation.

[0044] When oxygen concentration, carbon monoxide concentration, combustible gas concentration, or chamber temperature is abnormal, the main control board can directly trigger local buzzers, LED indicators, display screen prompts, or relay outputs without relying on the cloud platform, and mark the event as a safety-critical task to be submitted to the intelligent communication link control module. For safety-critical tasks involving gas exceeding limits, the system prioritizes redundant transmission using low-bandwidth fallback communication modules, industrial routers, or wired fallback channels, while retaining local caching and retransmission mechanisms until confirmation information is received or the manual processing flow is initiated.

[0045] The edge relay in this embodiment undertakes an integrated closed-loop function of "multi-source data access - task classification - communication mode selection - path optimization - hardware execution - local alarm - cloud transmission", and is a field edge control node for strong electromagnetic interference scenarios in nuclear power plant maintenance tunnels.

[0046] like Figure 2 As shown, taking the edge repeater box as an example: The edge repeater box is integrated into an ABS engineering plastic safety waterproof box. The interior is divided and fixed by a customized acrylic plate or metal bracket, which includes a 24V battery compartment, DC input interface, main control board, industrial router / 5G CPE, internal gas sensor, GPU module, Wi-Fi / Mesh repeater module, watch low-power communication module, low-bandwidth fallback communication module, camera interface, cooling fan, buzzer, LED indicator, display screen and external antenna mounting area.

[0047] The power supply / IO main control board serves as the hardware execution and interface management center, responsible for power management, sensor acquisition, gas sensor interface, voltage / current / power consumption acquisition, communication module start / stop control, buzzer / LED / display control, cooling fan control, watchdog reset, protocol encapsulation, and communication with the GPU module. The overall hardware structure of the power supply / IO main control board adopts a modular design, constructing a power tree through a step-down circuit to provide stable power to the ESP32 main controller, display module, and other peripherals. Simultaneously, it incorporates ADC sampling and current detection functions to improve the stability and reliability of the equipment in the complex electromagnetic environment of nuclear power plants. Its key circuit design is as follows: For power input and conversion circuits: 24V DC power input: The system supports 24V battery or 24V power adapter power supply. The input port is equipped with TVS transient suppression diodes and EMI filter circuits, which have strong anti-interference protection capabilities, reduce the impact of nuclear power plant site interference on the main control system and communication links, and can be designed and verified in accordance with the relevant electromagnetic compatibility requirements of GB / T 17626.2~17626.6 Severity Level 4.

[0048] The DC-DC step-down circuit uses the TPS54560DDAR chip to convert 24V to +12V and +5V. The circuit includes input and output filter capacitors to smooth the input voltage and stabilize the output voltage. To handle the instantaneous load surges generated when industrial routers and wireless communication modules operate simultaneously in a nuclear power environment, a feedback network, overcurrent protection, and necessary output margin can be configured. The chip also supports wide input voltage to adapt to voltage fluctuations along long-distance power lines in nuclear power plants, improving the stability of continuous equipment operation. The industrial-grade wide-temperature chip also enables continuous operation at high temperatures.

[0049] The LDO linear regulator circuit uses an LD29150DT33R to convert +5V to +3.3V to power the ESP32 and 3.3V peripherals. Compared to switching power supplies, LDOs have lower output ripple, which can reduce the impact of high-frequency interference on the stability of ADC sampling and wireless communication in the complex electromagnetic environment of nuclear power plants, thereby improving system reliability.

[0050] For data acquisition circuits: Battery voltage acquisition: The battery voltage is divided to the ESP32ADC range (0-3.3V) using a voltage divider resistor network. A filter capacitor is also added to make the voltage acquisition more stable in the complex electromagnetic environment of nuclear power plants.

[0051] Current acquisition: An INA199A2 current sensing amplifier, in conjunction with a shunt resistor, is used to measure the main circuit current. The amplifier output is directly connected to the ESP32ADC channel. This is used to monitor the operating status and power consumption changes of relay equipment at nuclear power plant sites.

[0052] Power calculation: The ESP32 software calculates P=U×I and displays it on the LCD in real time.

[0053] For communication interface circuits: The system adopts a multi-communication link design to improve the stability and reliability of data transmission in the complex environment of nuclear power plants.

[0054] USB interface: USB2 is a USB-to-serial debugging interface used for firmware download and serial port printing. This facilitates on-site equipment maintenance.

[0055] Wi-Fi / Mesh, watch low-power communication interface and NB-IoT communication module or narrowband gateway interface: The Wi-Fi / Mesh module is used for local area communication, camera access and multi-cabinet relay; the watch low-power communication module is used to receive small data packets such as heart rate, blood oxygen, posture, SOS, battery level and wearing status uploaded by wearable monitoring terminals, and performs protocol parsing and task classification through the main control board; the NB-IoT communication module, narrowband private network access gateway or field low-bandwidth access gateway is used to connect the data of field low-power terminals or narrowband nodes to the edge relay box.

[0056] Low-bandwidth fallback communication module and industrial router interface: Reserved interfaces for NB-IoT, LoRa, 4G / 5G CPE, or other narrowband / cellular communication modules. The low-bandwidth fallback communication module is used for SOS, gas over-limit, heartbeat, and critical status fallback transmission in weak network environments, sending critical alarms to upper-level relay nodes, monitoring centers, field gateways, or cloud backhaul links with coverage. The industrial router / 5G CPE is used to transmit data, alarm events, keyframes, and logs aggregated at the edge relay box back to the cloud monitoring platform or monitoring center.

[0057] For user interaction circuits: LCD display screen: connected via SPI interface (CS, DC, RST, BL, etc. pins).

[0058] LED indicator lights: Multiple LEDs (LED2-LED5) indicate power status and communication status, making it easy for staff to quickly determine the equipment status.

[0059] Buzzer: Controlled by GPIO, used for alarm prompts.

[0060] Dial encoder: Used for menu navigation; the corresponding pins can be seen in the circuit diagram.

[0061] To meet the requirements for stable operation in environments with strong electromagnetic interference, the PCB layout emphasizes power integrity and EMC immunity. The 24V power input is located near the board edge, with widened high-current traces (≥2mm) to reduce voltage drop and prevent overheating. The DC-DC conversion circuit area has separate copper plating, and input / output capacitors are placed close to the chip pins. Analog signal acquisition circuits (voltage divider, current detection) are located away from high-frequency switching nodes and use single-point grounding to improve ADC sampling stability. Communication interfaces (RS485, USB) are located near the board edge and equipped with ESD protection devices, making them less susceptible to electrostatic discharge and voltage surges during long-distance communication. Extensive copper plating and dense via connections reduce signal return paths, improving overall EMC performance and stability. The GPU core is located away from the power and communication RF areas, with a separate power supply and a filtering ferrite bead; short signal traces, partial grounding shielding, and grounded heat dissipation vias at the bottom enhance interference immunity and stability.

[0062] The internal layout of the enclosure is as follows: Left side: 24V battery compartment, DC input interface and external antenna mounting area.

[0063] Upper middle layer: Acrylic board or metal bracket to fix the main control board, industrial router / 5G CPE, internal gas sensor, camera interface and GPU module.

[0064] Lower middle layer: Wi-Fi / Mesh relay module, watch low-power communication module, low-bandwidth fallback communication module and its interface board.

[0065] Right side: Cooling fan, voltage display module, power switch, network port, DC output interface, buzzer, LED indicator and display screen.

[0066] All devices are centrally powered by a 24V battery, which is converted to 12V, 5V, and 3.3V by the power / IO main control board to supply power to each module. The GPU module is preferably powered by the +5V branch output from the power / IO main control board and is not directly connected to a 24V or 12V power supply. Industrial routers / 5G CPEs can be powered by 12V or 24V depending on the model. Wi-Fi / Mesh modules are powered by 5V or 12V depending on the specific device specifications. Low-power communication modules for watches, NB-IoT communication modules, narrowband private network access gateways or low-bandwidth access gateways in the field, and low-bandwidth fallback communication modules are powered by 3.3V or 5V depending on the device specifications. The exterior of the enclosure features a waterproof interface, a camera interface, and an antenna lead-out interface. The antenna is led out via an SMA or other industrial connector.

[0067] The GPU module serves as an edge computing auxiliary unit, used for task classification, communication method selection, link path calculation, and data processing requiring high computing power. It can also optionally perform keyframe extraction, compression, or caching strategy calculations on video or image data input from external cameras. The GPU module outputs scheduling decisions, which are then translated into specific hardware control actions by the power / IO main control board.

[0068] Industrial routers / 5G CPEs are used to transmit data, alarm events, keyframes, and logs aggregated in the edge relay box back to the cloud monitoring platform or monitoring center; Wi-Fi / Mesh relay modules are used for local area communication, camera access, local high-speed data transmission, and multi-box Mesh relay; watch low-power communication modules are used to receive small data packets uploaded by watches, such as heart rate, blood oxygen, posture, SOS, battery level, and wearing status; NB-IoT communication modules, narrowband private network access gateways, or field low-bandwidth access gateways are used to connect narrowband data from wearable terminals or other low-power nodes to the edge relay box; low-bandwidth fallback communication modules are used for critical alarm fallback transmission when the network is weak, the public network is interfered with, or the main link is unavailable.

[0069] To adapt to the strong electromagnetic interference and complex shielding environment of nuclear power plants, the edge relay box has the following industrial-grade electromagnetic compatibility and core functions.

[0070] Industrial-grade EMC electromagnetic compatibility design: Adopting a shielded cavity and filter interface design, it meets the electromagnetic compatibility standard of GB / T 17626.2~.6 Severity Level 4, achieving stable signal relay and data transmission under extreme conditions such as strong magnetic fields, shielding, and multipath reflection.

[0071] The intelligent link control function of the edge relay box relies on task hierarchies and two-layer intelligent communication link control modules deployed within the box. The algorithm's execution is primarily driven by the edge computing module within the edge relay box; wearable terminals, external cameras, and in-box environmental sensors serve only as data sources or access nodes and do not undertake global link optimization tasks. Reliable algorithm operation is based on standardized data input, hierarchical processing, and executable hardware output procedures.

[0072] The algorithm inputs for edge relay boxes include six categories: The first category consists of personnel status data uploaded by wearable monitoring terminals via the watch's low-power communication module, NB-IoT communication module, narrowband private network access gateway, or on-site low-bandwidth access gateway. This data includes heart rate, blood oxygen, respiratory rate, posture, fall flag, prolonged stillness flag, SOS flag, wearing status, terminal battery level, and watch link status.

[0073] The second category is video and image data input from external cameras, including live video streams, image frames, keyframes, and event screenshots.

[0074] The third category is data collected from the environment inside the chamber, including oxygen concentration, carbon monoxide concentration, combustible gas concentration, temperature and humidity, chamber temperature, optional radiation dose, and gas over-limit indicators.

[0075] The fourth category is communication link status data, including RSSI, SINR, available bandwidth, end-to-end latency, packet loss rate, ACK success rate, link availability status, queue length, and communication module operating status for Bluetooth Low Energy links, NB-IoT communication module / narrowband gateway links, low-bandwidth fallback links, Wi-Fi / Mesh links, industrial routing links, and wired links.

[0076] The fifth category is the edge relay box's own status data, including battery voltage, current, remaining power, box temperature, GPU temperature, main control board temperature, fan status, cache usage, and watchdog status.

[0077] The sixth category is task queues and remote configuration data, including task type, task level, data size, maximum allowable latency, minimum reliability requirements, whether caching is allowed, whether preemption is allowed, remote configuration version, parameter weights, and manual confirmation results.

[0078] Data Protocol and Encapsulation: To ensure reliable transmission and parsing in environments with strong interference, all data is encapsulated using a custom lightweight binary protocol. The protocol frame structure includes a frame header, length field, type field, priority field, data payload, and checksum field. The data payload can be encoded in TLV format, and the checksum field can use CRC-16 verification.

[0079] Algorithm processing: The edge relay box performs task identification, task classification, link status fusion, communication mode selection, and path optimization under system constraints based on the above inputs. For safety-critical tasks, priority is given to low latency, high reliability, and acknowledgment-based backhaul; for status monitoring tasks, priority is given to stable continuity and low power consumption; for business data tasks, priority is given to high bandwidth and allows for caching, degradation, and delayed upload.

[0080] The algorithm output includes: communication mode selection, frequency band or working channel selection, transmission mode selection, primary / backup path selection, path switching, redundant transmission, time slot allocation, power or working level adjustment, video bitrate reduction, keyframe upload, local caching, watch sampling period adjustment, watch sleep / wake-up, buzzer / LED / display alarm, fan start / stop, industrial routing backhaul strategy, and cache retransmission strategy under public network interruption. These outputs are generated by the edge computing module and converted by the power / IO main control board into specific control actions for the communication module, gateway, audible and visual alarms, power supply branches, and heat dissipation components.

[0081] The GPU module is an auxiliary computing unit within the edge relay box, responsible for handling more demanding algorithmic computation tasks than the main control board, such as task classification, communication mode selection, link path calculation, candidate path ranking, and video or image data processing. The main control board remains the hardware interface and execution control center, responsible for sensor acquisition, communication module control, power management, local alarms, and security fallback.

[0082] Hardware Configuration and Power Supply Design: Based on a Raspberry Pi or other embedded computing board, and equipped with a matching low-power GPU expansion module, parallel accelerated computing is achieved to meet the real-time computing requirements of communication algorithms and optional video keyframe processing. The module is preferably powered by the +5V branch output from the chassis power supply / IO main control board, and is not directly connected to a 24V or 12V power supply. The power supply can be dynamically adjusted according to the computing load. At the same time, it interacts with the main control board, industrial router, Wi-Fi / Mesh repeater module, NB-IoT communication module, narrowband private network access gateway and camera interface through Gigabit Ethernet, UART, USB or inter-board bus, receives link status and task data, and outputs scheduling decision instructions.

[0083] Low power consumption and computing power adaptation: The module adopts a dynamic power management strategy, with normal operating power consumption controlled within 5W and power consumption not exceeding 10W in full-load computing mode, adapting to limited power supply conditions on site; it can enter a low-power sleep state during idle periods and quickly wake up after receiving scheduling task instructions; the computing power of the GPU module can meet the real-time or near-real-time computing requirements of algorithms such as candidate link pruning, graph optimization path calculation, multi-path weight update, and lexicographical order task sorting, ensuring that on-site scheduling decisions can be output within an acceptable timeframe for the project.

[0084] Sensing and Data Standardization: Receives data from the watch packaged on the main control board, the internal environment data, link status, edge box status, and data input from the external camera, and completes data cleaning, protocol parsing, and format conversion to form standardized algorithm input.

[0085] Decision-making and computation: The system performs task classification, communication mode selection, graph optimization ad hoc network path algorithm and lexicographical task scheduling algorithm, and completes primary and backup path weight update, task priority sorting, candidate path calculation and cache resending strategy generation.

[0086] Control and Execution: The GPU module makes decisions such as communication mode selection, path switching, time slot allocation, transmission power or working level adjustment, video bitrate reduction, buffer retransmission, watch sampling period adjustment, and local alarms. Based on the above decisions, the main control board controls the watch's low-power communication module, NB-IoT communication module, narrowband private network access gateway or field low-bandwidth access gateway, low-bandwidth fallback communication module, Wi-Fi / Mesh module, industrial routing module, wired interface, buzzer, LED, display screen, and fan.

[0087] Local persistence and disaster recovery: The algorithm configuration parameters and historical link status are cached locally. Even when the public network is interrupted, the system can still make self-organizing network scheduling decisions based on local data to ensure that critical system communications are not interrupted.

[0088] Under the constraints of strong electromagnetic interference, communication denial, and high security levels in the confined space of nuclear power plants, traditional fixed routing, static scheduling, and single-objective optimization methods are insufficient to provide deterministic transmission guarantees. This invention focuses on task reliability, network resilience, and edge autonomy, constructing a highly reliable, self-healing communication system with the edge relay box as the scheduling entity and task hierarchical management, communication mode selection, and path optimization under system constraints as its core. Wearable terminals, cameras, and in-box sensors are only responsible for collecting, encapsulating, or uploading data; NB-IoT communication modules, narrowband private network access gateways or low-bandwidth field access gateways, Wi-Fi / Mesh, industrial routers, and wired interfaces serve as schedulable communication resources for the edge relay box, uniformly calculated by the edge computing module and uniformly executed by the power / IO main control board.

[0089] Figure 1A flowchart of an edge relay communication control method for environments with strong electromagnetic interference, according to an embodiment of the present invention, is provided. Figure 1 The edge relay communication control method for strong electromagnetic interference environments in this embodiment may include the following steps S1 to S3.

[0090] The specific implementation process of steps S1 to S3 is as follows: Step S1: Receive all data to be communicated and form a unified task queue.

[0091] In the specific implementation process, all data related to communication includes: monitoring data uploaded by the monitoring terminal through the communication module; image monitoring data; edge relay environment data; communication link status data; edge relay's own status data; task queue and remote configuration data.

[0092] Specifically, all data to be communicated is encapsulated into a unified task queue using a custom lightweight binary protocol; the frame structure of the lightweight binary protocol includes a frame header, length field, type field, priority field, data payload, and checksum field.

[0093] Step S2: Based on the unified task queue, perform the identification operation of task attributes and environment status.

[0094] The task attributes include task categories, which include three levels: safety-critical flow, status monitoring flow, and business data flow.

[0095] Step S3: Path optimization is performed using a two-layer communication control mechanism with edge relay as the main execution body.

[0096] In the first-layer communication, the communication method, frequency band / working channel and transmission mode are selected according to the task attributes and environmental status to form a set of schedulable paths. In the second-layer communication, based on the set of schedulable paths, the main path, backup path and preset low-overlap self-healing path are found in the network corresponding to the selected communication method to avoid the loss of critical alarms due to the failure of a single link.

[0097] In the first-layer communication, a task-driven heterogeneous link intelligent scheduling algorithm is adopted to form a set of corresponding schedulable paths according to the priority of safety-critical flows, status monitoring flows, and business data flows, which includes: Safety-critical flows take precedence over status monitoring flows and business data flows. After the safety-critical flow is locked, the schedulable path of the status monitoring flow is then solved on the remaining resources; After ensuring the resources for safety-critical flows and status monitoring flows, the schedulable paths for business data are determined.

[0098] The specific process of forming a set of schedulable paths according to the priority of safety-critical flows, status monitoring flows, and business data flows using a task-driven heterogeneous link intelligent scheduling algorithm is as follows: S311: Construct a schedulable heterogeneous communication resource pool for edge relay boxes; Edge relay boxes model their internal or field-side controllable communication resources into a unified schedulable resource pool. This includes low-power access links for smartwatches, NB-IoT communication modules / narrowband gateway links, LoRa / Sub-GHz low-bandwidth links, Wi-Fi / Mesh local area communication links, industrial routers / 5G CPE backhaul links, and wired fallback links, mathematically expressed as: (1) In the formula, For Bluetooth BLE or other watch low-power access links, For NB-IoT communication modules / narrowband gateway links, For low-bandwidth backup outbound links, For Wi-Fi / Mesh local area communication links, For industrial routing / 5G CPE backhaul links, This is a wired backup link; this formula is used to uniformly describe the schedulable communication resources of the edge relay box, forming a standardized resource set.

[0099] (2) C represents the set of communication methods that can be scheduled by the edge relay box; This indicates Bluetooth BLE or other low-power short-range reception methods for watches; This indicates the NB-IoT communication module or narrowband private network access method; This indicates LoRa / Sub-GHz, low-bandwidth fallback transmission, or other low-speed, high-reliability fallback methods; Indicates Wi-Fi / Mesh communication method; Indicates the industrial routing / 5G CPE backhaul method; This indicates a wired backup communication method; Let K(t) represent the set of available links corresponding to communication mode c at time t; K(t) represents the set of all available links that can be scheduled by the edge relay box at time t. This formula is used to divide the communication resources in the system into a two-layer structure: a "set of communication modes" and a "set of specific links". The first layer is used for the edge relay box to select the communication mode, and the second layer is used for path optimization in the link subgraph corresponding to the selected communication mode.

[0100]

[0101] (3)

[0102] m represents the communication task number; This represents the set of tasks within the current scheduling period; t represents the scheduling time or scheduling period number. The binary decision variable represents whether the edge relay box enables communication mode c or its corresponding communication module for task m at time t; This represents the task activation variable. If task m enters the immediate scheduling queue of the edge relay box at time t, then... = 1; If task m does not enter the immediate scheduling queue at time t, then = 0; This indicates the maximum number of communication methods that task m is allowed to use simultaneously within a scheduling cycle. This indicates the number of communication methods enabled for task m. This formula guarantees that no communication method is assigned when the task is not active, and that at least one method is selected when the task is active, but the maximum redundancy is not exceeded.

[0103] This formula describes the first-level communication method selection decision on the edge relay box side. When a task is not active, no communication method is assigned to it; when a task is active, the system selects at least one communication method for it, and the number of selected communication methods does not exceed the maximum allowed redundancy for that task.

[0104] S312: Collect link status parameters; The edge relay box collects data in real time on each communication method and its available links, including interference power, signal-to-interference-plus-noise ratio, available bandwidth, end-to-end delay, packet loss rate, ACK success rate, energy consumption per unit bit, communication module operating status, and link availability flags, forming a dynamic state vector.

[0105] (4) (5) Γ represents the task-communication compatibility matrix; This flag indicates the compatibility between the task type and the communication method. When the compatibility flag is 1, it means that the task type is allowed to use this communication method; when the compatibility flag is 0, it means that the task type is not allowed to use this communication method. This indicates the task type or task level to which task m belongs. ; L1 is for safety-critical tasks (including SOS alarms, fall alarms, gas over-limit alarms, alarms for serious abnormal vital signs, and critical control commands); L2 is for status monitoring tasks (including heart rate, blood oxygen, respiratory rate, posture, battery level, wearing status, enclosure temperature, and equipment status); L3 is for business data tasks (including video streams, images, logs, bulk data, and inspection records). constraint This means that the system can only enable communication mode c for task m if the task type and communication mode are compatible.

[0106] The compatibility matrix can be set according to the following rules: L1 safety-critical tasks are allowed to use low-bandwidth fallback communication, NB-IoT communication modules, narrowband private network access gateways or field low-bandwidth access gateways, Wi-Fi / Mesh, industrial routers / 5G CPEs, and wired fallback links; L2 status monitoring tasks are allowed to use smartwatch low-power access, NB-IoT communication modules, narrowband private network access gateways or field low-bandwidth access gateways, low-bandwidth fallback communication, and industrial router links; L3 service data tasks prioritize the use of Wi-Fi / Mesh, industrial routers / 5G CPEs, and wired fallback links, while narrowband links are only used for event summaries, keyframe indexes, or alarm association information.

[0107] This formula constrains the communication methods selectable for different task types. Safety-critical tasks and status monitoring tasks involve small to medium data volumes, allowing for the selection of multiple communication methods based on link status. Business data tasks typically involve large data volumes, prohibiting the use of BLE, NB-IoT communication modules, narrowband gateways, or low-bandwidth fallback communication methods to transmit complete video streams, thus avoiding the consumption of low-speed critical communication resources by video, images, and large data packets. For L3 services, event summaries, keyframe indexes, or alarm association information can be transmitted via narrowband links only when necessary.

[0108]

[0109] (6)

[0110] A binary variable indicating whether the communication method is enabled by the edge relay box at time t; it is 1 when enabled and 0 otherwise. The binary decision variable representing whether the edge relay box enables communication mode c for task m at time t; A matrix representing the concurrency feasibility of communication methods; A flag indicating whether communication mode c and communication mode c′ are allowed to be concurrently enabled: when When =1, it means that both can be enabled simultaneously; when When =0, it means that the two cannot be enabled at the same time (or it is not recommended to enable them at the same time due to hardware, power supply, antenna, frequency band constraints).

[0111] This formula describes the hardware concurrency constraints when multiple communication methods are redundantly transmitted. If two communication methods correspond to independent communication modules, independent antennas, or RF resources that can operate in parallel, then both are allowed to be enabled concurrently. If two communication methods share the RF front-end, antenna, power channel, or there is significant mutual interference, then both are not allowed to be enabled simultaneously. For example, when the low-power communication module and the NB-IoT communication module, narrowband private network access gateway, or low-bandwidth fallback communication module are independent of each other, the corresponding concurrency flag can be set to 1 to indicate that they can run concurrently.

[0112] (7) :link At any moment The link state vector. Electromagnetic interference intensity of link k at time t; Signal-to-interference-to-noise ratio (SINR); Available bandwidth; End-to-end basic latency; Packet loss rate; Energy consumption per unit of data transmission; Link availability status (1 indicates availability, 0 indicates unavailability); this formula digitizes the real-time quality of the link, providing a calculable and comparable quantitative basis for scheduling decisions.

[0113] S3: Construct a hierarchical task model; In nuclear power plant maintenance corridor scenarios, different communication services exhibit significant differences in safety levels, transmission requirements, and resource consumption. Adopting a unified scheduling strategy could easily lead to high-bandwidth services preempting critical communication resources. Therefore, based on service attributes and security levels, all communication tasks within the system are divided into three categories: a safety-critical layer, a status monitoring layer, and a service data layer, forming a differentiated scheduling model. (8) L1 safety critical layer: = {SOS active alarm, fall alarm, prolonged inactivity alarm, abnormal vital signs alarm, gas over-limit alarm, critical control command};

[0114] This represents the L1 safety-critical task set (including SOS alarms, fall alarms, gas over-limit alarms, serious abnormal vital signs alarms, and critical control commands). (t) represents a binary decision variable for whether the edge relay box enables communication mode c for task m at time t; This represents the minimum number of redundant communication methods required for a safety-critical task m. It represents a set of outward communication methods that can transmit alarm events from the edge relay box to the cloud monitoring platform, monitoring center, upper-level relay node, or on-site receiving gateway; This indicates the NB-IoT communication module or narrowband private network access method; This indicates a low-bandwidth fallback communication method; Indicates the Wi-Fi / Mesh communication method; Indicates the industrial routing / 5G CPE backhaul method; This indicates a wired backup communication method.

[0115] It should be noted that, This indicates the dual-path redundancy requirement under conditions of sufficient external communication resources. When the number of available external communication methods is insufficient to meet this redundancy requirement, the system does not force the maintenance of dual-path external transmission, but instead activates the currently most reliable available external transmission link and simultaneously triggers local audible and visual alarms, cache recording, and retransmission mechanisms. When all external communication methods are unavailable, the edge relay box enters local autonomous mode and performs cache retransmission after communication is restored.

[0116] This formula is used to ensure minimum redundant transmission capacity for safety-critical tasks. For L1 safety-critical tasks such as SOS, fall detection, and gas over-limit detection, the system must select at least [a specific number of] external transmission links, provided the number of available external transmission links meets the requirements. The system employs multiple communication methods for transmission, including at least one method capable of transmitting back to external platforms, upstream nodes, or on-site receiving gateways, to prevent safety-critical alarms from remaining solely on the local short-range access link. When available external transmission links are insufficient, the system prioritizes the external transmission link with the highest reliability and simultaneously triggers local audible and visual alarms, cache recording, and retransmission mechanisms. When all external communication methods are unavailable, the edge relay box enters local autonomous mode, triggering alarms via buzzer, LED indicator, and display screen, and writing the alarm data to the local cache. Cache retransmission is performed after communication is restored.

[0117] This layer of tasks is characterized by low latency, high reliability, non-degradability, preemptibility, and redundancy. For L1 tasks, the system goal is not to save bandwidth or improve video quality, but to ensure speed, stability, and reliability as much as possible. L1 tasks should have a significantly higher scheduling priority than other tasks.

[0118] L2 Status Monitoring Layer: = {Heart rate, blood oxygen, respiratory rate, posture, wearing status, terminal battery level, sensor status, device operating status, periodic environmental monitoring data}; This level of service prioritizes continuous stability, low power consumption, moderate latency, and minimal transmission fluctuations. Data does not require millisecond-level ultra-fast response, but it must maintain long-term uninterrupted transmission to avoid transmission interruptions or severe latency fluctuations, ensuring the continuity and reliability of the monitoring process.

[0119] L3 Business Data Layer: = {External camera video, images, keyframes, logs, large packet data, inspection records}; This layer of business data has a large volume and high bandwidth consumption, and has flexible adaptability capabilities such as bitrate reduction, caching, delayed transmission, and fragmented transmission. In the event of strong electromagnetic interference or network resource constraints, L3 services proactively yield resources to prioritize the normal transmission of L1 and L2 services, without affecting the core security and monitoring functions of the system.

[0120] S314: Construct a nonlinear transmission model; Based on this state vector, the scheduler performs resource mapping within the allowed operating slots, channels, transmit power, and bandwidth of the communication module according to task priorities. Its core relies on a nonlinear link rate model to describe the coupling relationship between power, bandwidth, and external interference. (10) Let m be the actual effective transmission rate of task m on link k; : The bandwidth allocated to task m; : Allocated transmission power; : Channel gain; N0: Noise power spectral density constant; : Link interference intensity; This formula is used to characterize the impact of power, bandwidth and external interference on the effective transmission rate under strong electromagnetic interference environment, avoiding scheduling deviations caused by relying entirely on the ideal channel assumption.

[0121] The end-to-end transmission delay of task m is defined as (11) : Waiting time in queues; Link switching cost. Indicates task At any moment The amount of data to be transmitted; Indicates task The data quality factor, compression factor, or bitrate level is used to characterize the effective amount of data after bitrate reduction, compression, or quality adjustment.

[0122] Since the link rate is non-linear and the delay includes a 1 / R term, the task delay is also non-linear.

[0123] If a task is redundantly sent via multiple links or various communication methods, the task is considered successfully transmitted as long as any available link successfully delivers the message and receives a valid acknowledgment. The task success rate is defined as: (12) : Overall transmission success rate of task m; : Decision variables for selecting the link on the edge relay box side (1 indicates selection, 0 indicates non-selection); Link availability status; : Link packet loss rate; This formula characterizes the effect of redundant transmission across multiple links on reducing the overall transmission failure risk. Under conditions where the correlation between failures of different links or communication methods is low, a product form can be approximated. For scenarios with common causes of failure such as strong local electromagnetic interference, shared node failures, or shared area obstruction, the system further reduces the risk of joint failure through path design with low edge overlap, low node overlap, and low interference area overlap. If a communication method corresponds only to a single-hop link, this term serves as the basis for link quality assessment and switching to backup communication methods.

[0124] The communication energy consumption of task m is defined as follows: (13) Indicates task Total energy consumption for communication; Indicates link At any moment Standby power consumption or power consumption to keep connected.

[0125] S315: Construct a lexicographically robust nonlinear programming model; (14) : Lexicographical order optimization; , , The objective function of L1 safety critical layer, L2 status monitoring layer, and L3 business data layer. : Set of scheduling decision variables; This formula ensures that L1 tasks have the best priority over L2 and L3 tasks, achieving absolute guarantee of safety level.

[0126] (15) These represent the weight coefficients of latency, failure risk, link switching cost, and redundancy loss penalty in the L1 objective, respectively. Indicates task At any moment Enable redundant transmission? thus: This indicates that the latency has been reduced. This indicates a reduction in the risk of failure; Used to avoid frequent switching; This is used to encourage the use of redundant transmissions for critical missions.

[0127] (16) in These represent the weighting coefficients for latency, failure risk, energy consumption, and latency fluctuation in the L2 objective, respectively.

[0128] =|Tm(t) -Tm(t -1)|(17) This indicates transmission latency fluctuations. This item is used to avoid issues such as inconsistent speeds or interruptions in vital signs and status monitoring data.

[0129] (18) in, These represent the weighting coefficients for latency, failure risk, energy consumption, and service quality benefits in the L3 objective, respectively. This indicates gains in video clarity, image quality, or log integrity.

[0130] The goal of L3 is to improve the quality of business data as much as possible without affecting L1 and L2.

[0131] The strict three-level lexicographical order solution can be written as follows: first, find the optimal solution for the L1 safety-critical task among all feasible scheduling variables; then, find the optimal solution for the L2 status monitoring task without significantly violating the L1 objective; and finally, find the optimal solution for the L3 business data task without significantly violating the L1 and L2 objectives.

[0132] ; ; (19) It represents the entire feasible scheduling decision space that meets the requirements of communication mode compatibility, concurrency constraints, link availability, bandwidth and power limitations, and task latency and reliability.

[0133] The first layer solution yields... To achieve L1 safety-critical mission objectives To achieve the optimal; The second layer solution yields... ,Require No more than With safety margin The sum of these values ​​simultaneously enables the L2 state monitoring target to be monitored. The optimal solution is obtained from the third layer. ,Require and Neither is significantly damaged, while ensuring that L3 business data targets are... To achieve optimal results.

[0134] The final output is: u*(t) = .in, and It represents a safety margin, indicating that subsequent level optimizations cannot significantly compromise the optimal guarantees of previous levels.

[0135] Simultaneously, robust constraints for strong interference are applied to the L1 task, ensuring that reliability and latency hard limits are met even within the range of interference fluctuations. (20) (twenty one) In the formula, Let K be the set of interference uncertainties for link k. This represents the minimum success rate required for L1 tasks. The maximum allowable latency for L1 tasks. This is the L1 safety-critical task set; this set of formulas ensures that critical alarms still meet the reliability and delay constraints of nuclear power safety requirements under worst-case interference conditions.

[0136] S316: Perform hierarchical fast solution; A method employing link candidate set pruning, priority locking of safety-critical tasks, rolling allocation of remaining resources, and local approximation of continuous variables is used to quickly obtain scheduling results. A complete robust nonlinear programming model is used to describe the objectives and constraints, and in field operation, candidate link pruning, heuristic path search, local linear approximation, and rolling updates are used for fast approximate solutions.

[0137] (twenty two) Indicates task Maximum permissible link base latency; Indicates task Maximum allowable packet loss rate on the link. Indicates task At any moment The candidate link set obtained after filtering by link availability, latency, and packet loss rate.

[0138] This reduces the task from enumerating all communication methods and links to simply selecting from a limited number of compatible communication methods and then optimizing the path in the corresponding subgraph. A typical pruning strategy is as follows: L1 prioritizes NB-IoT communication modules, narrowband private network access gateways or low-bandwidth field access gateways, low-bandwidth fallback communication, industrial routing, wired fallback, and necessary local auxiliary alarm outputs; L2 retains watch low-power communication links, NB-IoT communication module / narrowband gateway links, low-bandwidth fallback links, and industrial routing links; L3 retains Wi-Fi / Mesh, wired, and high-bandwidth industrial routing links.

[0139] Prioritize and lock in L1 safety-critical tasks for resource allocation; First, solve: (twenty three); This represents the feasible scheduling decision space corresponding to L1 safety-critical tasks.

[0140] The execution strategy is as follows: sort tasks according to their urgency; select a link that meets the worst-case interference constraint for each L1 task; if a single link cannot meet the reliability requirement and the available communication methods meet the redundancy requirements, then enable the redundant link; if there are insufficient available external transmission links, then enable the external transmission link with the highest current reliability and trigger local alarms, cache records, and duplicate transmission mechanisms; lock the L1 resources and do not allow subsequent L2 or L3 preemption.

[0141] The set of secure resource locks can be written as: (twenty four) This represents the set of resources after an L1 safety-critical task has been locked. , , These represent the link selection, bandwidth allocation, and power allocation results locked after the L1 solution is obtained, respectively.

[0142] L2 optimizes for remaining resources: After L1 locking, the remaining bandwidth of the link is: (25) in: Indicates the link after L1 resource locking. The remaining bandwidth; Indicates link At any moment Total allocatable bandwidth capacity.

[0143] The remaining power resources are: (26) This indicates that the link or communication module has been locked after L1 resource locking. The remaining available transmission power; This indicates its maximum allocatable transmit power.

[0144] Then solve on the remaining resources: (27) And maintain: J1(u) ≤ J1(u1*) + ε1(28) This represents the feasible scheduling space for L2 state monitoring tasks under L1 resource guarantee constraints.

[0145] L3 Elastic Service Degradation and Optimization: L3 services are optimized after L1 and L2 resource guarantees are in place: (29) This represents the feasible scheduling space for L3 business data tasks under the constraints of L1 and L2 resource guarantees.

[0146] When strong interference increases or remaining resources are insufficient, the system will reduce the video bitrate, compress the image, delay the log, and upload large data packets in chunks.

[0147] Local approximation and rolling updates for continuous variables: For fixed link selection The rest and A fast solution can be obtained through local approximation. The solution from the previous time step is taken as the linearized point: (30) , , These represent the reference rate, reference bandwidth, and reference transmit power at the previous scheduling cycle or the current iteration point, respectively. , These represent the partial derivatives of the rate function with respect to bandwidth and transmit power, respectively.

[0148] Instead of starting from scratch, each moment is updated in a rolling fashion: u(t) ←u(t-1) + u(t)(31) This indicates the scheduling decision relative to the previous time step. The incremental update volume.

[0149] This mechanism transforms complex mixed-integer nonlinear programming into small-scale candidate search and continuous variable approximation updates, making it more suitable for real-time on-site scheduling.

[0150] S317: Perform link scheduling and feedback updates; The edge computing module generates decisions based on the solution results, including communication mode selection, frequency band or working channel selection, transmission mode selection, path switching, redundant transmission, video bitrate reduction, key frame upload, data fragmentation, and buffered delayed transmission. The power / IO main control board converts these decisions into specific control actions for the communication module, NB-IoT communication module, narrowband private network access gateway or field low bandwidth access gateway, industrial router, Wi-Fi / Mesh module, wired interface, and local alarm module, and feeds back the actual transmission results to the link state estimation and optimization model for the next moment.

[0151] Unlike conventional fixed communication modes or simple priority queue mapping, this algorithm is essentially a robust scheduling mechanism with strong coupling between task, communication method, and path, with edge relay boxes as the execution entity. Its core lies in first determining the appropriate communication method, frequency band, or working channel and transmission mode for the task, then optimizing the path within the network subgraph corresponding to the selected communication method, and using the success rate, latency, and acknowledgment return of L1 tasks as necessary constraints to maintain priority for safety-critical tasks within a certain range of interference fluctuations.

[0152] In the second-layer communication, path optimization is performed based on graph-optimized resilient self-organizing networks and multi-path self-healing algorithms, including: Construct a dynamic graph in the network topology corresponding to the communication method; For communication methods with multiple relays, multiple gateways, or multiple exits, the main path, backup path, and preset low-overlap self-healing path are calculated based on the dynamic graph and real-time monitored link information, and based on the task-aware edge weights. For communication methods with only single-hop connections, link quality assessment, switching to alternative communication methods, and buffer retransmission control are performed based on dynamic graphs and real-time monitored link information.

[0153] The optimization process proceeds layer by layer in the order of safety-critical flow, status monitoring flow, and business data flow. When a disruption or link failure is detected on the main path, the backup path takes over first, and then the affected area undergoes local incremental reconstruction. For safety-critical task flows, both the main path and the backup path are solved simultaneously, and the risk of common failure is reduced through penalties for overlapping edges, nodes, and interference areas.

[0154] The graph-optimized resilient self-organizing network and multi-path self-healing algorithm belong to the second layer of a two-layer communication control mechanism. The first layer has already determined the communication methods or sets of communication methods that can be used by the task by the edge relay box; the second layer constructs a dynamic graph in the network topology corresponding to the communication methods, calculates the primary path, backup path, and low-overlap self-healing path, and is used to solve the problem of "how to resist interference, how to switch between primary and backup paths, and how to self-heal and reconstruct when multiple nodes, multiple links, and multiple tasks are concurrent" in the network. When the communication method only corresponds to single-hop access or single-exit backhaul, this algorithm degenerates into link quality assessment, backup communication method switching, ACK retransmission, and buffer retransmission control.

[0155] Specifically, the process of graph-optimized resilient self-organizing networks and multi-path self-healing algorithms is as follows: S321 dynamic graph construction; The wearable monitoring terminals, in-cabinet gas sensors, external camera access nodes, edge relay boxes, NB-IoT communication modules, narrowband private network access gateways or field low-bandwidth access gateways, relay nodes, and gateways in the nuclear power plant maintenance tunnel are modeled as node sets. The available connections formed by the selected communication methods in the first layer are modeled as edge sets, resulting in a dynamic graph: (32) G(t) represents the value generated by the currently available communication nodes at time t. and links The resulting dynamic image.

[0156] The node set is: (33) The terminals for maintenance personnel are listed in order. Sensor nodes as well as (NB-IoT communication module, narrowband private network access gateway or low-bandwidth access gateway in the field, relay node), edge computing node With gateway node .

[0157] The edge set E(t) represents the links that can communicate at the current time. Different communication methods correspond to different subgraphs, such as the Wi-Fi / Mesh subgraph, the NB-IoT communication module / narrowband gateway subgraph, the low-bandwidth fallback communication subgraph, the industrial routing backhaul subgraph, and the wired fallback subgraph. Path optimization is performed within the corresponding subgraph to avoid confusing communication method selection with path selection.

[0158] S322 State Acquisition and Edge Weight Generation; Collect latency, packet loss rate, bandwidth, RSSI, SINR, interference intensity, energy consumption, availability, ACK success rate, and congestion level for each edge to construct an edge state vector. .

[0159] (34) : Link-based transmission delay; : Packet loss rate; Available bandwidth of the link; The electromagnetic interference intensity of the environment in which the link is located; Energy consumption per unit of data transmission or power cost per node; ∈{0,1}: Link availability indicator, 1 indicates availability, 0 indicates unavailability; : Current load and congestion level of the link.

[0160] S323 task flow layering; Based on the task's security level, latency limit, reliability limit, and bandwidth requirements, the task flow is divided into a security-critical layer, a status monitoring layer, and a business data layer.

[0161] S324 robust edge weight calculation; Based on the uncertainty set We compute task-aware robust edge weights to assess link risks under worst-case interference. Different tasks have different sensitivities to the same edge. Security alerts prioritize reliability and latency, while video processing prioritizes bandwidth. Therefore, we define task-aware edge costs for the task flow f. : (35) Represents task flow Regarding latency Packet loss ,interference ,bandwidth Energy consumption and congestion items The non-negative weighting coefficients. Indicates the task stream to be transmitted; These represent the sets of safety-critical flows, status monitoring flows, and business data flows, respectively.

[0162] Typical configuration: in f ∈ In medium reliability, latency, and low overlap, the penalty weights are larger, and in f ∈ Medium bandwidth and caching costs are given greater weight. This makes graph optimization truly "task-specific," rather than using the same shortest path rule for all tasks.

[0163] S325 primary / backup low overlap path optimization; The main path and backup path are solved for the safety-critical task flow, and the risk of common failure is reduced by penalties for edge overlap, node overlap and interference area overlap.

[0164] For safety-critical flows f ∈ At the same time, plan the main path and backup path .

[0165] If the two paths can be taken over independently, the overall success rate is higher. for: (36) Represents task flow Overall success rate after adopting a primary / backup path; , These represent the primary path and backup path at time [time]. The transmission success rate. This indicates that the failure event for critical tasks has changed from "primary path failure" to "simultaneous failure of both primary and backup paths".

[0166] Edge overlap penalty: ; Node overlap penalty: ; in, Represents task flow The main path Represents task flow Backup path; Representing a path The set of edges or links contained therein; Representing a path The set of communication nodes traversed; This indicates the number of elements in the set. Therefore, This indicates the number of links shared between the primary path and the backup path. This indicates the number of nodes shared between the primary and backup paths. A larger shared number indicates a higher risk of both primary and backup paths failing simultaneously. Interference area overlap penalty: Divide the utility tunnel into several areas: Z = {Z1,Z2, . . .,ZH}, (37) This indicates the number of interference regions obtained from the division. Indicates the first Interference area.

[0167] Each edge belongs to a region z(i,j)∈ Z. The set of regions traversed by the path is... Z(P) ={z(i,j):(i,j)∈P} (38) (39) This indicates the number of common interference areas traversed by the primary and backup paths.

[0168] For safety-critical flows, TRGO (Task-aware Robust Graph Optimization) is used to solve the problem: (40) Represents task flow At any moment Select path The total cost of the path; These represent the penalty weights for edge overlap, node overlap, and interference region overlap, respectively. This indicates the cost weight for path switching or reconstruction. : Cost of path switching or reconstruction, used to avoid frequent jittery switching.

[0169] TRGO's "redundancy" is not simply about selecting an extra path, but rather a joint optimization of primary and backup paths with low edge overlap, low node overlap, and low interference area overlap. This avoids a localized area of ​​strong interference simultaneously breaking down both the primary and backup paths, which is more in line with the local fault propagation characteristics of nuclear power plant maintenance tunnels.

[0170] S326 dictionary order diagram optimization; Optimize layer by layer in the order of safety-critical flow, status monitoring flow, and business data flow to ensure that ordinary business operations do not compromise critical business operations.

[0171] L1 safety-critical flow objectives: (41) The objective function for graph optimization represents the L1 safety-critical flow; This represents the set of decision variables for path selection and primary / backup paths during the graph optimization phase.

[0172] L2 Status Monitoring Flow Target: The state monitoring stream prioritizes stability and low power consumption. Defining latency fluctuations: (42) but: (43) Let represent the graph optimization objective function for the L2 state monitoring flow. Represents task flow At any moment End-to-end path delay for the selected path. Represents task flow Select path The path energy consumption or node power cost at that time. and These are weight parameters; L3 business data flow objectives: Business data flows allow for caching, degradation, and latency. Let... For video or image quality gains, As for the cost of cache latency, then: (44) and The weight parameters represent the business quality benefits and cache latency costs.

[0173] S327 self-healing switching and incremental reconstruction; When a disruption to the main path or a link failure is detected, the backup path takes over first, and then the affected area is partially reconstructed incrementally, instead of performing a full-map brute-force recalculation.

[0174] If the L1 main path is disturbed: (45) Indicates time A set of edges that are disturbed, out of service, or have significantly degraded link quality was detected; Then immediately switch to the backup path, and recalculate the new backup path in the background: (46) Represents task flow A new backup path recalculated outside the current primary path or host path; Indicates the candidate path.

[0175] In this way, the system always maintains a secure structure of "current carrying path + one backup path".

[0176] S328 Feedback Update; The actual transmission delay, packet loss, number of handovers, and link recovery results are fed back into the edge weight estimation and path optimization of the next cycle.

[0177] In one or more embodiments, another edge relay communication control system for strong electromagnetic interference environments can be implemented in software. The edge relay communication control system for strong electromagnetic interference environments includes the following software modules: The queue unification module is used to receive all data to be communicated and form a unified task queue; The information recognition module is used to perform recognition operations on task attributes and environmental status based on a unified task queue. The path optimization module is used to perform path optimization using a two-layer communication control mechanism with edge relay as the main execution body. In the first layer of communication, the communication method, frequency band / working channel and transmission mode are selected according to the task attributes and environmental status to form a set of schedulable paths. In the second layer of communication, based on the set of schedulable paths, the main path, backup path and preset low-overlap self-healing path are found in the network corresponding to the selected communication method to avoid the loss of critical alarms due to the failure of a single link.

[0178] It should be noted that each module in the edge relay communication control system for strong electromagnetic interference environment in this embodiment corresponds one-to-one with each step in the edge relay communication control method for strong electromagnetic interference environment in the above embodiment, and their specific implementation process is the same, so it will not be repeated here.

[0179] This invention employs a two-layer communication control mechanism with edge relay as the execution entity for path optimization. Based on task attributes and environmental conditions, it selects the communication method, frequency band / working channel, and transmission mode to form a set of schedulable paths. Then, within the network corresponding to the selected communication method, it identifies the primary path, backup path, and preset low-overlap self-healing path. This solves the problems of reliable transmission of safety-critical data, edge closed-loop processing of multi-source data, and two-layer communication link control under strong interference environments. It avoids the loss of critical alarms due to single link failure, reduces the risk of simultaneous breakdown of the primary and backup paths in localized strong electromagnetic interference areas, and improves the communication self-healing capability in nuclear power plant maintenance corridor scenarios.

[0180] Based on the above Figure 1 The control method for the edge relay communication monitoring and control system oriented towards the strong electromagnetic interference environment of nuclear power plants, as shown, includes the following steps: Step 1: Low-bandwidth personnel status access via wearable terminal. The wearable low-bandwidth personnel monitoring terminal collects low-bandwidth personnel status data such as heart rate, blood oxygen, respiratory rate, posture, fall status, prolonged static state, SOS signal, battery level, and wearing status of the personnel. This data is uploaded via BLE, LoRa, NB-IoT, or other narrowband low-power communication methods to the low-power receiving module, NB-IoT communication module, narrowband private network access gateway, or on-site low-bandwidth access gateway at the edge relay box. After receiving the data, the edge relay box timestamps, parses the protocol, performs CRC verification, and initially identifies the task level. The watch does not collect video data, nor does it handle high-bandwidth service transmission or global link scheduling.

[0181] Step 2: External Video and Internal Environment Data Access. The external camera uploads live video streams, image frames, keyframes, or event screenshots to the edge relay box via Wi-Fi, Ethernet, USB, or an industrial video interface. The internal gas sensor connects to the main control board via UART, I2C, SPI, ADC, or RS485 to collect environmental data such as oxygen concentration, carbon monoxide concentration, combustible gas concentration, temperature and humidity, and box temperature. Neither video nor environmental data is collected by the wearable terminal.

[0182] Step 3: The edge relay box establishes a unified task queue. The main control board within the edge relay box aggregates, filters, timestamps, serializes, performs CRC checks, and encapsulates local alarm data, including personnel status data, internal environment data, power status data, communication module status data, and local alarm status, before submitting it to the edge computing module. The edge computing module combines this data with input data from external cameras and communication link status to form a unified task queue.

[0183] Step 4: The edge computing module performs task classification and scheduling decisions. The GPU module or edge computing module receives structured data packaged on the main control board, input data from external cameras, and link status data. It performs task identification, link status fusion, task classification, and scheduling decisions, dividing the tasks into L1 safety-critical tasks, L2 status monitoring tasks, and L3 business data tasks.

[0184] Step 5: Selection of communication method, frequency band, and transmission mode at the first layer of the edge computing module. The edge computing module selects the communication method, frequency band or working channel, and transmission mode based on task level, data size, maximum allowable latency, minimum reliability requirements, whether caching is allowed, current electromagnetic interference status, availability of each communication module, and remaining power in the edge box. For L1 tasks such as SOS, fall detection, gas over-limit detection, and critical control commands, the system prioritizes one or more of the following: low-bandwidth fallback link, NB-IoT communication module, narrowband private network access gateway or on-site low-bandwidth access gateway, industrial router / 5G CPE, wired fallback link, or Wi-Fi / Mesh upstream repeater, and can simultaneously trigger local buzzer, LED, and display alarms. For L2 status monitoring tasks, the system prioritizes low-power, low-bandwidth, stable, and continuous access links. For L3 service data tasks, the system prioritizes high-bandwidth links such as Wi-Fi / Mesh, wired interface, or industrial router, and performs bitrate reduction, key frame upload, local caching, or delayed upload when the link deteriorates.

[0185] Step 6: Path optimization and self-healing under the second-layer standard constraints at the box end. After selecting the communication mode at the first layer, the edge computing module constructs a dynamic graph in the corresponding communication network, and calculates edge weights based on RSSI, SINR, available bandwidth, end-to-end latency, packet loss rate, ACK success rate, node load, electromagnetic interference level, and energy consumption. For communication modes with multiple relays, multiple gateways, or multiple exits, the system calculates the primary path, backup path, and low-overlap self-healing path; for communication modes with only single-hop connections, the system performs link quality assessment, backup communication mode switching, and buffer retransmission control.

[0186] Step 7: The main control board executes hardware control. The edge computing module sends decisions to the main control board regarding communication mode selection, path switching, redundant transmission, time slot allocation, transmission power or operating level adjustment, video bitrate downsampling, buffer retransmission, watch sampling frequency adjustment, local alarms, and fan control. Based on these decisions, the main control board controls the NB-IoT communication module, narrowband private network access gateway or field low-bandwidth access gateway, watch low-power communication module, low-bandwidth fallback communication module, Wi-Fi / Mesh module, industrial routing module, wired interface, buzzer, LED, display screen, fan, power supply branch, and local cache module. It also sends control commands such as sampling frequency adjustment, sleep mode, wake-up, vibration alert, alarm confirmation, and heartbeat cycle adjustment to the wearable terminal via a low-power downlink.

[0187] Step 8: ACK Confirmation, Cache Retransmission, and Anomaly Rollback. For L1 safety-critical tasks, the system requires the receiving end to return an ACK confirmation. The ACK can come from the cloud platform, monitoring center, upstream relay node, field receiving gateway, or local edge box confirmation module. If no ACK is received within the confirmation time, the system retransmits via backup communication methods and triggers a local audible and visual alarm or manual confirmation process. When the external transmission link is unavailable, the system records the non-transmission status and triggers an alarm and cache retransmission locally. For L2 status monitoring tasks, the system can use periodic or batch confirmation. For L3 business data tasks, the system can use local caching, breakpoint resumption, and retransmission after recovery.

[0188] Step 9: Edge Autonomy Under Public Network Interruption. When industrial routing, cloud links, or some external transmission links are unavailable, the edge repeater box enters local autonomy mode. In this mode, the edge repeater box continues to receive data from the watch, external camera, and internal gas sensor, and continues to perform local risk assessment, local audible and visual alarms, low-bandwidth backup transmission for critical tasks, video caching, and path reconstruction. After the public network or upstream receiving link is restored, the system performs cache retransmission according to task level, timestamp, and event type, prioritizing the retransmission of data required for safety-critical events and manual confirmation.

[0189] Step 10: Cloud Reception, Remote Display, and Configuration Updates. The cloud monitoring platform receives alarm events, status monitoring data, keyframes, video clips, and log data uploaded by the edge relay boxes, and performs remote display, storage, manual confirmation, and configuration updates. The cloud can send updated task thresholds, communication weights, risk level rules, and configuration parameters to the edge relay boxes, but it does not undertake real-time intelligent identification or real-time communication scheduling on-site. Real-time security closure on-site is primarily completed locally by the edge relay boxes.

[0190] In one or more embodiments, an edge relay device is also provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the steps of the edge relay communication control method for strong electromagnetic interference environments as described above.

[0191] Specifically, according to embodiments of this application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of this application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program including functions for executing... Figure 1 The program code for the method shown. In such an embodiment, the computer program can be downloaded and installed from a network via a communication component, and / or installed from a removable medium. When the computer program is executed by the central processing unit, it performs the various functions defined in the apparatus of this application.

[0192] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, as well as combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0193] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. An edge relay communication control method for environments with strong electromagnetic interference, characterized in that, Executed within the edge relay, including: Receive all data to be communicated and form a unified task queue; Based on a unified task queue, perform task attribute and environment state identification operations; A two-layer communication control mechanism with edge relay as the main execution entity is adopted for path optimization; In the first layer of communication, the communication method, frequency band / working channel and transmission mode are selected according to the task attributes and environmental status to form a set of schedulable paths. In the second layer of communication, based on the set of schedulable paths, the main path, backup path and preset low-overlap self-healing path are found in the network corresponding to the selected communication method to avoid the loss of critical alarms due to the failure of a single link.

2. The edge relay communication control method for strong electromagnetic interference environments as described in claim 1, characterized in that, All data to be communicated includes: monitoring data uploaded by the monitoring terminal via the communication module; image monitoring data; edge relay environment data; communication link status data; edge relay's own status data; task queue and remote configuration data.

3. The edge relay communication control method for strong electromagnetic interference environments as described in claim 2, characterized in that, All data to be communicated is encapsulated into a unified task queue using a custom lightweight binary protocol; the frame structure of the lightweight binary protocol includes a frame header, length field, type field, priority field, data payload, and checksum field.

4. The edge relay communication control method for strong electromagnetic interference environments as described in claim 1, characterized in that, The task attributes include task categories, which include three levels: safety-critical flow, status monitoring flow, and business data flow. In the first-layer communication, a task-driven heterogeneous link intelligent scheduling algorithm is adopted to form a set of corresponding schedulable paths according to the priority of safety-critical flows, status monitoring flows, and business data flows, which includes: Safety-critical flows take precedence over status monitoring flows and business data flows. After the safety-critical flow is locked, the schedulable path of the status monitoring flow is then solved on the remaining resources; After ensuring the resources for safety-critical flows and status monitoring flows, the schedulable paths for business data are determined.

5. The edge relay communication control method for strong electromagnetic interference environments as described in claim 4, characterized in that, During the process of finding schedulable paths for safety-critical flows, the safety-critical flow tasks are sorted according to their urgency. For each safety-critical flow task, a link that meets the worst-case interference constraint is selected. If a single link cannot meet the reliability requirements and the available communication methods meet the redundancy requirements, a redundant link is enabled. If there are insufficient available external transmission links, the external transmission link with the highest reliability is enabled, and local alarms, cache records, and duplicate transmission mechanisms are triggered. Safety-critical flow resources are locked to prevent subsequent status monitoring flows / business data flows from preempting them.

6. The edge relay communication control method for strong electromagnetic interference environments as described in claim 1, characterized in that, In the second-layer communication, path optimization is performed based on graph-optimized resilient self-organizing networks and multi-path self-healing algorithms, including: Construct a dynamic graph in the network topology corresponding to the communication method; For communication methods with multiple relays, multiple gateways, or multiple exits, the main path, backup path, and preset low-overlap self-healing path are calculated based on the dynamic graph and real-time monitored link information, and based on the task-aware edge weights. For communication methods with only single-hop connections, link quality assessment, switching to alternative communication methods, and buffer retransmission control are performed based on dynamic graphs and real-time monitored link information.

7. The edge relay communication control method for strong electromagnetic interference environments as described in claim 6, characterized in that, The system optimizes the flow layer by layer in the order of safety-critical flow, status monitoring flow, and business data flow. When the main path is detected to be disturbed or the link fails, the backup path takes over first, and then the affected area is partially and incrementally reconstructed. For safety-critical task flows, both the main path and backup path are solved simultaneously, and the risk of common failure is reduced by penalties for edge overlap, node overlap, and interference region overlap.

8. An edge relay communication control system for environments with strong electromagnetic interference, characterized in that, The edge relay communication control method for strong electromagnetic interference environments based on any one of claims 1-7 includes: The queue unification module is used to receive all data to be communicated and form a unified task queue; The information recognition module is used to perform recognition operations on task attributes and environmental status based on a unified task queue. The path optimization module is used to perform path optimization using a two-layer communication control mechanism with edge relay as the main execution body. In the first layer of communication, the communication method, frequency band / working channel and transmission mode are selected according to the task attributes and environmental status to form a set of schedulable paths. In the second layer of communication, based on the set of schedulable paths, the main path, backup path and preset low-overlap self-healing path are found in the network corresponding to the selected communication method to avoid the loss of critical alarms due to the failure of a single link.

9. An edge relay communication control system for environments with strong electromagnetic interference, characterized in that, include: Intelligent monitoring terminals, edge relays, and cloud computing; The intelligent monitoring terminal is used to upload all monitoring information to the edge relay; The edge relay is used to perform the steps in the edge relay communication control method for strong electromagnetic interference environment as described in any one of claims 1-7; The cloud and edge relays communicate with each other.

10. An edge relay device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the steps in the edge relay communication control method for strong electromagnetic interference environments as described in any one of claims 1-7.

Images