Multi-protocol compatible industrial wireless gateway integrated device and its data conversion method

The industrial wireless gateway integration device, which integrates hardware modularization and software intelligent scheduling, solves the problems of incomplete protocol integration, poor adaptability, and low conversion efficiency of industrial wireless gateways. It achieves efficient and stable multi-protocol data transmission and industrial scenario adaptation, supports dynamic expansion and remote upgrade of the protocol stack, and meets the high real-time and stability requirements of industrial sites.

CN123123480APending Publication Date: 2026-11-17JINAN BODAXUN COMM TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611304261.1
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-08-26
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

Existing industrial wireless gateways suffer from technical defects such as incomplete protocol integration and coverage, poor adaptability, low conversion efficiency, insufficient customization for industrial scenarios, insufficient support for domestic protocols, weak scalability, and low transmission reliability, which fail to meet the high real-time and stability requirements of industrial sites.

Method used

It adopts a technical architecture of modular hardware support, intelligent software scheduling, and a dedicated protocol conversion engine. It designs a protocol scheduling algorithm and a link state evaluation model based on data priority. Combined with multi-protocol wireless transceiver modules, industrial protocol interface modules, protocol conversion acceleration modules, and data caching modules, it realizes end-to-end protocol integration, hardware-level conversion, and redundancy fault tolerance mechanisms, and supports dynamic hot loading and remote upgrade of the protocol stack.

Benefits of technology

It achieves end-to-end integrated management of industrial equipment-specific protocols, general network communication protocols, and cloud platform standardized protocols, with protocol conversion latency controlled within 10ms, packet loss rate reduced to below 0.1%, ensuring priority transmission of high-priority data, link status monitoring and automatic fault switching, supporting dynamic expansion and remote upgrades of the protocol stack, and guaranteeing stable transmission 24/7.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123123480A_ABST
    Figure CN123123480A_ABST
Patent Text Reader

Abstract

This invention provides an integrated industrial wireless gateway device compatible with multiple protocols and its data conversion method. The device includes a core control module, a multi-protocol wireless transceiver module, an industrial protocol interface module, a protocol conversion acceleration module, a data buffer module, and a power supply and protection module. These modules interact via a bus. The core control module and the protocol conversion acceleration module constitute an intelligent protocol conversion engine, employing a hardware and software collaborative conversion architecture. Through a conversion pipeline of "feature recognition—protocol parsing—data mapping—format encapsulation," it achieves low-latency, packet-loss-free conversion between protocols. The data conversion method includes: automatic protocol identification, unified data mapping, preemptive scheduling based on four-dimensional priority, link quality evaluation and selection, triple-threshold fault-tolerant switching, and supports hot loading of protocol plugins, breakpoint resumption, and hierarchical congestion control.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of industrial wireless gateway technology, and more particularly to an integrated device for an industrial wireless gateway compatible with multiple protocols and its data conversion method. Background Technology

[0002] The rapid development of the Industrial Internet of Things (IIoT) has driven the intelligent upgrading of industrial production. As the core hub for interconnecting industrial field devices, transmitting data, and connecting to cloud platforms, the industrial wireless gateway's protocol compatibility directly determines the interoperability, data transmission efficiency, and scenario adaptability of industrial systems. While current industrial wireless gateway products have achieved partial protocol integration, in practical industrial applications, they still suffer from numerous technical shortcomings, including incomplete protocol integration coverage, poor protocol compatibility, insufficient customization for industrial scenarios, weak support for domestic protocols, and low data conversion and transmission efficiency. These shortcomings are specifically reflected in the following aspects: 1. Incomplete protocol integration system and lack of end-to-end compatibility. Existing industrial wireless gateways mostly adopt an integration mode of "primarily using a single type of protocol and supplementing with other protocols", which makes it difficult to achieve end-to-end protocol compatibility for native interaction of industrial equipment, multi-link data transmission, and standardized docking with cloud platforms. Additional protocol conversion middleware needs to be configured, which increases system deployment costs and the number of failure points.

[0003] 2. Poor compatibility between industrial-specific protocols and general network protocols, resulting in low conversion efficiency. Existing gateways mostly adopt a "simple software parsing" mode for protocol conversion, lacking dedicated protocol conversion engines and hardware support. This leads to problems such as high latency, high packet loss rate, and data format corruption during the data conversion process, failing to meet the high real-time requirements of scenarios such as industrial motion control and real-time monitoring.

[0004] 3. Protocol integration lacks customization for industrial scenarios, resulting in insufficient anti-interference and stability. Industrial sites are characterized by strong electromagnetic interference and dense equipment deployment. Existing gateway protocol integration has not been specifically optimized for industrial scenarios. Multi-protocol data transmission adopts a "first-come, first-served" scheduling method, and high-priority data such as industrial control commands are easily blocked by low-priority data, affecting production stability.

[0005] 4. Insufficient support for domestic industrial protocols and industry-specific protocols, resulting in poor localization adaptability. Most mainstream gateways are developed by foreign manufacturers and lack sufficient support for domestic industrial wireless standards such as WIA-PA, the HJ212 environmental protection standard, and the DL / T 645 power standard, failing to meet the specific needs of industry-specific device interconnection and regulatory reporting.

[0006] 5. Weak protocol scalability, unable to adapt to the needs of industrial technology upgrades. Existing gateways mostly have fixed protocol stacks, with hardware modules tightly bound to the protocol stack, making it impossible to dynamically enable / disable protocols or perform remote upgrades. Technology upgrades require replacing the entire device, increasing enterprise upgrade costs.

[0007] 6. Multi-protocol links lack redundancy and fault tolerance mechanisms, resulting in low transmission reliability. Existing gateways' multi-protocol links mostly operate in independent mode, lacking a unified link status monitoring and fault tolerance mechanism. When a protocol link fails, data cannot automatically switch to a backup link, leading to interruptions in industrial data transmission. Summary of the Invention

[0008] To address the aforementioned technical deficiencies in existing industrial wireless gateways, this invention provides a multi-protocol compatible integrated industrial wireless gateway device and its data conversion method. By constructing a comprehensive, layered protocol integration system and designing a technical architecture of modular hardware support, intelligent software scheduling, and a dedicated protocol conversion engine, the following specific technical problems are solved: 1. To address the incompleteness of existing gateway protocol integration systems, achieve full-link, integrated integration of three major categories of protocols: industrial equipment-specific protocols, general network communication protocols, and cloud platform standardized protocols, covering more than 95% of industrial scenarios.

[0009] 2. To address the issues of poor compatibility and low conversion efficiency between industrial-specific protocols and general network protocols, a hardware-level + software-level collaborative intelligent protocol conversion engine was designed to control the protocol conversion latency to within 10ms and reduce the packet loss rate to below 0.1%.

[0010] 3. To address the lack of industrial-scenario customization in protocol integration, a protocol scheduling algorithm and link state evaluation model based on data priority are designed to ensure the priority transmission of high-priority data such as industrial control commands.

[0011] 4. Addresses the issue of insufficient support for domestic industrial protocols and industry-specific protocols, natively integrating localized protocols such as WIA-PA (domestic industrial wireless standard), HJ212 (environmental protection standard), and DL / T 645 (power standard).

[0012] 5. To address the issue of weak protocol extensibility, a pluggable protocol design and a decoupled hardware / software architecture are adopted, supporting dynamic hot reloading of the protocol stack and remote version upgrades.

[0013] 6. To address the lack of redundancy and fault tolerance mechanisms in multi-protocol links, a unified protocol link status monitoring and automatic fault switching mechanism is designed to achieve redundant backup of multi-protocol links and ensure stable 24 / 7 transmission of industrial data.

[0014] This invention is achieved through the following technical solution: a multi-protocol compatible industrial wireless gateway integrated device, comprising: a core control module for running a dual system of real-time operating system and embedded Linux, performing protocol scheduling, link management, and data processing; a multi-protocol wireless transceiver module, including multiple pluggable radio frequency units, each of which is physically isolated through independent radio frequency isolation circuits and shielded cavities, for supporting signal transmission and reception of multiple wireless protocols; an industrial protocol interface module, integrating multiple industrial interfaces for adapting to data access of multiple wired industrial protocols; a protocol conversion acceleration module, employing an FPGA chip or programmable logic device, pre-programmed with hardware logic for protocol parsing and data format conversion, for implementing hardware-level conversion of data packets between different protocols; a data caching module, employing a dual-layer caching design of memory and solid-state storage, for dynamically caching data according to data priority; and a power supply and protection module for providing redundant power supply and circuit protection; the core control module, multi-protocol wireless transceiver module, industrial protocol interface module, protocol conversion acceleration module, and data caching module interact with each other via a high-speed serial bus and an internal data bus.

[0015] As a preferred solution, the core control module and the protocol conversion acceleration module constitute an intelligent protocol conversion engine, which adopts a hardware and software collaborative conversion architecture. The protocol conversion acceleration module is used for hardware-level acceleration of data packet decapsulation, verification calculation and format conversion, while the core control module is used for software processing of instruction mapping, semantic parsing and logical judgment. The intelligent protocol conversion engine adopts a conversion pipeline of "feature recognition - protocol parsing - data mapping - format encapsulation", which controls the protocol conversion latency to within 10ms and reduces the packet loss rate to below 0.1%.

[0016] As a preferred solution, the multi-protocol wireless transceiver module includes: a low-power RF unit to support LoRa / LoRaWAN, ZigBee, and BLE short-range low-power protocols; a high-speed WiFi RF unit to support WiFi 6 and WiFi 7 (with reservations); a cellular communication unit to support 5G / 4G / NB-IoT wide-area protocols; and an industrial wireless bus RF unit to support WIA-PA / WirelessHART industrial wireless protocols. The industrial protocol interface module includes: multiple RS485 / RS232 serial ports for Modbus RTU serial protocol compatibility; multiple CAN / CAN FD bus interfaces for CANopen bus protocol compatibility; and multiple Gigabit Ethernet ports for Profinet and EtherNet / IP industrial Ethernet protocols. Each interface employs independent protection circuitry, supporting simultaneous access from multiple devices. In the data caching module, the memory cache adopts a circular buffer design for high-speed temporary caching of P1 to P3 level data, with read / write latency of less than 1ms; solid-state storage is used for persistent storage of P4 to P5 level data and data to be retransmitted in case of link interruption; high-priority data resides in memory, while low-priority data is flushed to solid-state storage when memory reaches the water level threshold. The power supply and protection module supports wide-voltage DC input and PoE power supply, and adopts a dual-power redundancy design, automatically switching when one power supply fails; it has built-in overvoltage, overcurrent, overtemperature, and reverse connection protection circuits; the device casing is made of aluminum alloy, which has dustproof and vibration-proof characteristics, and the internal circuit board is coated with conformal coating.

[0017] The present invention also provides a data conversion method based on the above-mentioned device, comprising the following steps: Automatic Protocol Identification: A two-layer identification strategy, employing feature-based matching and weighted similarity as a fallback, is used to identify the protocol type of received data packets. Multi-dimensional feature vectors are extracted for each protocol, including frame header feature codes, frame length distribution features, verification algorithm features, address field length features, and function code value range features, with each feature dimension assigned a different weight coefficient. If a feature code is not matched, the weighted similarity between the data packet and each protocol in the protocol feature library is calculated. If the similarity exceeds a threshold, the packet is identified as belonging to the corresponding protocol. If all similarities are below the threshold, the density-based DBSCAN clustering algorithm is used to automatically cluster the data packets, generating temporary protocol rule templates for each cluster. Incremental learning mechanisms are used to continuously optimize the rules, enabling the self-generation of rules for unknown protocols.

[0018] Unified Data Mapping: Based on the unified data model of "device-measurement point-attribute" triple, data from different protocols is converted into a standardized structure according to a configurable mapping table. The mapping table defines the mapping relationship between the register address, data type, byte order, coefficient offset, unit and unified model of each protocol through XML or JSON configuration files, and supports dynamic loading and modification at runtime. The mapping process includes four steps: address translation, data type conversion, engineering quantity conversion and unit unification.

[0019] Priority scheduling: The priority score of data frames is calculated based on four dimensions: time sensitivity, packet loss tolerance, latency requirements and business importance. Priority levels are divided and a preemptive scheduling strategy is used for data transmission.

[0020] Link evaluation and selection: Real-time acquisition of signal strength, packet loss rate, round-trip time, available bandwidth and bit error rate of each link, calculation of comprehensive link quality score, and allocation of transmission links to data according to priority and link quality matching strategy.

[0021] Fault tolerance and switching: When the link score is lower than the fault threshold, the candidate backup link score is higher than the normal threshold, and the state duration exceeds the stable duration threshold, link fault switching is performed.

[0022] As a preferred solution, the weighted similarity calculation formula is: Similarity(P) = Σ(wi × match_i) / Σwi, where wi is the weight of the i-th feature dimension and match_i is the matching degree of that dimension; the frame header feature code accurate matching in the fast feature code matching takes less than 1ms, achieving fast recognition of more than 90% of common protocols; the similarity threshold is set to 0.85 by default; data type conversion covers conversion between bit, byte, word, double word and floating point types.

[0023] As the preferred option, the priority score is calculated using the following formula: Priority = α × TimeSensitivity + β × LossTolerance + γ ×DelayRequirement + δ × BusinessImportance Among them, TimeSensitivity is the time sensitivity, reflecting the sensitivity of data to transmission latency, with a value range of 0-1; LossTolerance is the reciprocal of packet loss tolerance, with a higher score for lower packet loss tolerance, with a value range of 0-1; DelayRequirement is the urgency of latency requirements, with hard real-time data scoring the highest, with a value range of 0-1; BusinessImportance is the level of business importance, preset by the business type, with a value range of 0-1; α, β, γ, and δ are the weight coefficients of each dimension and satisfy α+β+γ+δ=1, which can be configured according to the scenario.

[0024] The data is divided into five levels based on priority scores: Level P1 is emergency control data with a score ≥ 0.9, Level P2 is routine control data with a score ≤ 0.7 and < 0.9, Level P3 is real-time monitoring data with a score ≤ 0.5 and < 0.7, Level P4 is routine monitoring data with a score ≤ 0.3 and < 0.5, and Level P5 is log diagnostic data with a score < 0.3.

[0025] The preemptive scheduling strategy is as follows: when a high-priority data frame arrives, if low-priority data is currently being transmitted, the current transmission is interrupted, the already transmitted portion is recorded as the offset, and high-priority data is transmitted first. After the high-priority data is transmitted, the transmission of low-priority data is resumed from the interruption point. P1 level data has the highest preemption right and can preempt all low-priority data. P2 level data can preempt P3 to P5 level data. Data of the same priority is scheduled using a first-in-first-out queue.

[0026] As the preferred option, the formula for calculating the overall link quality score is: LinkScore = w1 × norm(RSSI) + w2 × (1 - norm(PLR)) + w3 × (1 -norm(RTT)) + w4 × norm(BW) + w5 × (1 - norm(BER)) Among them, RSSI is the signal strength, PLR is the packet loss rate, RTT is the round-trip time, BW is the available bandwidth, BER is the bit error rate, w1 to w5 are the weight coefficients of each parameter, with default values ​​of 0.25, 0.3, 0.25, 0.1 and 0.1 respectively, which can be adjusted according to the business scenario, norm() is the normalization function, which maps each parameter to the 0-1 range, and LinkScore ranges from 0 to 100.

[0027] The link status is divided into three levels based on LinkScore: Normal status is LinkScore ≥ 80, the link quality is good and can carry all priority data; Warning status is 50 ≤ LinkScore < 80, the link quality is degraded, it can only carry P3 level and above data, and low priority data is suspended or switched; Fault status is LinkScore < 50, the link is unavailable and triggers the fault switching process.

[0028] The priority and link quality matching strategy is as follows: P1 and P2 level data are forcibly assigned to the link with the highest score to ensure low latency and high reliability; P3 level data is assigned to the link with the second-highest score; P4 and P5 level data are assigned to the link with lower cost or lighter current load to achieve multi-link load balancing; when the quality of the primary link degrades, high priority data is switched to the backup link first.

[0029] As a preferred solution, failover must simultaneously meet the following three threshold conditions: Condition 1, the current primary link score is lower than the fault threshold, i.e., LinkScore_current<50; Condition 2, the candidate backup link score is higher than the normal threshold, i.e., LinkScore_backup>80; Condition 3, the duration of the above conditions exceeds the stable duration threshold, which is set to 500ms by default; After the failover is completed, a cooldown time is set, during which reverse failover is not triggered, and the cooldown time is set to 2s by default.

[0030] As a preferred option, a weighted adaptive learning step is also included: a reinforcement learning mechanism is introduced, with data transmission success rate and average latency as reward signals, and the weight coefficients of each parameter in the link scoring model are dynamically optimized through the Q-Learning algorithm, so that the scoring model gradually adapts to the current network environment.

[0031] As a preferred solution, the following steps are also included: receiving the protocol plugin package, which contains the protocol stack binary file and configuration file; verifying the plugin signature to ensure its legitimacy; creating an independent protocol runtime context in memory; registering the plugin to the global protocol status table; initializing the protocol stack and starting to listen to the corresponding interface; hot loading is completed in the background without affecting the normal operation of other protocols, with a loading time of less than 2 seconds. The global protocol status table records the protocol ID, protocol name, plugin version, bound interface, running status, data statistics, and configuration parameters of all loaded protocols. The running status includes three states: running, paused, and faulty. It supports querying and modification at runtime, allowing users to dynamically enable or disable specified protocols and modify protocol parameters without restarting the device.

[0032] As a preferred solution, a breakpoint resumption step is also included: each data packet is assigned a globally unique incremental sequence number, the sender records the acknowledged sequence number and the sent sequence number, and the receiver returns the acknowledged sequence number; when the link is interrupted, the data is cached to solid-state storage; after the link is restored, the breakpoint is located by comparing the sent and received sequence numbers, the sender retransmits the data packets that have not been acknowledged in order, and the acknowledged data packets are not retransmitted. The breakpoint location accuracy reaches the single packet level, and the resumption efficiency is close to 100%.

[0033] As a preferred solution, it also includes a tiered congestion control step: when the cache occupancy rate reaches the first-level warning threshold of 70%, the receiving rate of low-priority data is reduced, and the sender is notified to limit the flow; when the cache occupancy rate reaches the second-level warning threshold of 85%, P5 level log data is discarded, and all P4 level data is transferred to solid-state storage; when the cache occupancy rate reaches the third-level alarm threshold of 95%, P4 level non-critical monitoring data is discarded, and only P1 to P3 level core data is retained to ensure that high-priority data always has sufficient cache space.

[0034] As a preferred option, the following steps are also included: receiving downlink commands from the cloud platform or industrial control system, verifying their legitimacy after decryption and authentication by the security module; mapping standardized commands to the target device's proprietary protocol format and checking the parameters' legitimacy; assigning downlink control commands with high priority by default and preemptively inserting them into the head of the transmission queue; selecting the downlink channel to send commands based on the interface bound to the target device, waiting for the device's response, and automatically retrying if the timeout occurs.

[0035] As a preferred option, the protocols natively integrated into the device fall into three main categories: industrial equipment-specific protocols, including Modbus RTU / ASCII / TCP, OPC DA / UA / UA over MQTT general industrial basic protocols; Siemens Profinet / Profibus-DP, Rockwell EtherNet / IP / CIP, Mitsubishi CC-Link IE / MELSEC, Omron FINS / HostLink, and Delta ASDA vendor-specific industrial protocols; CAN / CAN FD, RS485 / RS232, WirelessHART, ISA100.11a, and WIA-PA industrial bus protocols; and DL / T 645, IEC 101 / 103 / 104, HJ 212, SL 651, CJ / T188, and MBUS domestic and industry-specific protocols. General network communication protocols, including LoRa / LoRaWAN, ZigBee / ZigBeePRO, BLE5.0 / BLE Mesh short-range low-power wireless protocols; WiFi 6 and WiFi 5 high-speed wireless LAN protocols, with a reserved WiFi 7 interface; NB-IoT, 4G-LTE (Cat.1 / Cat.4 / Cat.12), and 5G NR wide-area IoT wireless protocols; and TCP / IP protocol suite, VPN (IPsec / OpenVPN), and DHCP / DDNS basic network transmission protocols. Standardized cloud platform protocols, including MQTT, MQTT-SN, and CoAP lightweight IoT cloud protocols; HTTP / HTTPS and FTP / FTPS general network cloud protocols; and MQTT SparkPlug B and OPC UA over MQTT industrial-specific cloud protocols.

[0036] By employing the above technical solutions, this invention has the following beneficial effects compared to existing technologies: (1) High protocol conversion efficiency and strong real-time performance. Through the intelligent protocol conversion engine with hardware and software collaboration, combined with the automatic recognition mechanism of feature vector weighted matching, the protocol conversion latency is ≤10ms and the packet loss rate is ≤0.1%, which meets the requirements of high real-time scenarios such as industrial motion control.

[0037] (2) The scheduling mechanism is scientific and the transmission is highly deterministic. Based on the four-dimensional priority scoring formula and preemptive scheduling algorithm, combined with the priority-link quality matching strategy, the deterministic transmission of high-priority data such as industrial control instructions is ensured, with latency fluctuations of less than ±2ms.

[0038] (3) The link is fault-tolerant and reliable, and the transmission stability is high. The multi-dimensional weighted scoring model + triple threshold switching mechanism + weight adaptive learning realizes accurate evaluation and reliable switching of the link status. The switching time is ≤200ms, avoiding the ping-pong effect and ensuring stable transmission of industrial data 24 / 7.

[0039] (4) The protocol is highly scalable and adaptable to technology upgrades. Standardized plug-in interface + hot loading mechanism + protocol status table management realize hardware and software decoupling, support dynamic protocol expansion and remote upgrades, and adapt to new technologies such as Wi-Fi 7 and 6G without replacing the whole machine, thus extending the device life cycle.

[0040] (5) Intelligent caching strategy and good data integrity. The dual-layer caching of memory and solid-state, hierarchical congestion control and precise breakpoint resume ensure that high-priority data is not lost, and the breakpoint resume efficiency is close to 100%, which greatly improves data integrity in harsh network environments.

[0041] (6) Wide protocol compatibility and comprehensive coverage of scenarios. It natively supports more than 30 protocols in three major categories: industrial equipment protocols, general network protocols, and cloud platform protocols, covering mainstream domestic and international standards and domestic industry-specific protocols, covering more than 95% of industrial scenarios.

[0042] Additional aspects and advantages of the invention will become apparent in the following description or may be learned by practice of the invention. Attached Figure Description

[0043] The above and / or additional aspects and advantages of the present invention will become apparent and readily understood from the description of the embodiments taken in conjunction with the following drawings, in which: Figure 1 Overall architecture diagram of an industrial wireless gateway integration device compatible with multiple protocols; Figure 2 A flowchart of the intelligent protocol conversion engine; Figure 3 This is a schematic diagram of a priority-based real-time scheduling mechanism. Figure 4 Flowchart for link status evaluation and fault switching; Figure 5 This is a schematic diagram of the protocol plug-in architecture; Figure 6 This is a diagram illustrating the breakpoint resume and caching mechanism; Figure 7 This is a hardware structure module diagram. Detailed Implementation

[0044] To better understand the above-mentioned objectives, features, and advantages of the present invention, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be noted that, unless otherwise specified, the embodiments and features described in these embodiments can be combined with each other.

[0045] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and therefore the scope of protection of the invention is not limited to the specific embodiments disclosed below.

[0046] The following is combined Figures 1 to 6 The present invention provides a detailed description of an integrated industrial wireless gateway device compatible with multiple protocols and its data conversion method.

[0047] This invention addresses the technical shortcomings of existing industrial wireless gateways, such as incomplete protocol integration, low conversion efficiency, poor adaptability to industrial scenarios, insufficient support for domestic protocols, weak scalability, and low transmission reliability. It provides a multi-protocol compatible industrial wireless gateway integration device and its data conversion method. Through five core technological innovations—an intelligent protocol conversion engine, a priority-based scheduling module, a link fault tolerance module, a protocol plug-in management module, and a breakpoint resume caching module—this invention achieves efficient parsing, rapid conversion, stable transmission, and intelligent scheduling of multiple protocols, solving the end-to-end protocol compatibility issues in industrial field interconnection of multi-brand equipment, multi-link data transmission, and seamless integration with industrial cloud platforms.

[0048] like Figure 1 , Figure 2 As shown, this invention proposes an industrial wireless gateway integrated device compatible with multiple protocols and its data conversion method. The industrial wireless gateway integrated device of this invention adopts a compact aluminum alloy shell, and its internal structure is divided into a core control layer, a protocol transceiver layer, an interface conversion layer, a data buffer layer, and a power supply protection layer according to function. The layers communicate with each other through a high-speed serial bus and an internal data bus.

[0049] The core control module uses an ARM Cortex-A53 quad-core industrial-grade processor with a clock speed of 1.5GHz, supporting a wide operating temperature range of -40℃ to 85℃. This module simultaneously runs a FreeRTOS real-time operating system and an embedded Linux dual-system, with FreeRTOS handling high real-time tasks and Linux handling non-real-time management tasks. The processor incorporates a hardware encryption engine and a floating-point unit to meet the requirements of data encryption and high-speed computation.

[0050] The multi-protocol wireless transceiver module adopts a pluggable modular design, with each RF unit physically isolated through independent shielding cavities and isolation circuits. Specifically, the LoRa unit uses the SX1278 chip (470MHz, 20dBm output, -148dBm sensitivity), the ZigBee unit uses the CC2530 (supporting ZigBee PRO), the Bluetooth unit uses the nRF52832 (BLE 5.0 and Mesh), the WiFi 6 unit uses the IPQ6000 (dual-band, up to 1.8Gbps), and the 5G cellular unit uses the Quectel RG500Q (supporting SA / NSA, 4G, NB-IoT, dual-SIM redundancy). In addition, an industrial wireless bus RF unit is also included for protocols such as WIA-PA and WirelessHART.

[0051] The industrial protocol interface module provides multiple RS485 / RS232 serial ports (compatible with Modbus RTU), multiple CAN / CANFD interfaces (compatible with CANopen), and multiple Gigabit Ethernet ports (compatible with Profinet and EtherNet / IP). All interfaces are equipped with independent protection circuits (TVS, resettable fuses, etc.) and support simultaneous connection of multiple devices.

[0052] The protocol conversion acceleration module uses a Xilinx Artix-7 FPGA (XC7A35T) with a dedicated ASSP chip, pre-programmed with hardware logic for protocol parsing and format conversion, to achieve hardware-level acceleration of time-consuming operations such as data packet decapsulation, verification, and format conversion.

[0053] The data caching module consists of 4GB DDR4 memory and a 128GB industrial-grade SSD. DDR4 serves as a high-speed circular buffer for temporarily storing high-priority data (P1-P3 levels), with read / write latency below 1ms. The SSD is used for persistent storage of P4-P5 level data and data to be retransmitted in case of link interruption. The cache management strategy automatically schedules data based on priority: high-priority data resides in memory, while low-priority data is transferred to the SSD when memory usage reaches a threshold.

[0054] The power supply and protection module adopts dual-redundancy with DC 9-36V wide voltage input and PoE power supply, and the MP2307 power management chip realizes automatic switching (switching time <1ms). It has built-in overvoltage (40V), overcurrent (3A), overtemperature (85℃) and reverse connection protection. The whole machine shell is made of aluminum alloy with sandblasting and oxidation, with an IP40 protection rating. The internal circuit board is coated with conformal coating to adapt to harsh environments such as high humidity and dust.

[0055] In this embodiment, the gateway natively integrates three major categories of protocols, covering the entire link of industrial equipment interconnection, network transmission, and cloud platform docking.

[0056] Industrial equipment-specific protocols include: Modbus RTU / ASCII / TCP, OPC DA / UA / UA over MQTT; Siemens Profinet / Profibus-DP, Rockwell EtherNet / IP / CIP, Mitsubishi CC-Link IE / MELSEC, Omron FINS / HostLink, Delta ASDA; CAN / CAN FD, RS485 / RS232, WirelessHART, ISA100.11a, WIA-PA; and domestic industry protocols such as DL / T 645, IEC 101 / 103 / 104, HJ 212, SL 651, CJ / T 188, and MBUS.

[0057] Common network communication protocols include: LoRa / LoRaWAN, ZigBee / ZigBee PRO, BLE 5.0 / Mesh; WiFi 6 (802.11ax), WiFi 5 (802.11ac), with a reserved WiFi 7 interface; NB-IoT, 4G LTE (Cat.1 / 4 / 12), 5G NR; TCP / IP, VPN (IPsec / OpenVPN), DHCP / DDNS, etc.

[0058] The cloud platform's standardized protocols cover: MQTT, MQTT-SN, CoAP; HTTP / HTTPS, FTP / FTPS; and industrial cloud-specific protocols such as MQTTSparkplug B and OPC UA over MQTT.

[0059] Each protocol is encapsulated as an independent plugin, interacting with the core system through standard interfaces, and supporting dynamic loading and unloading.

[0060] The engine is composed of a main processor (software) and an FPGA (hardware), and adopts a four-step pipeline of "feature recognition → protocol parsing → data mapping → format encapsulation".

[0061] The automatic protocol identification employs a two-layer strategy: "signature-first, AI as a fallback." The system pre-builds a feature vector library for each protocol, with dimensions including fixed frame header codes, frame length distribution, verification algorithm type, address field length, and function code range, each with different weights. When a data packet arrives, it first attempts to match the frame header feature code. If a match is found, the protocol type is directly determined, taking less than 1ms and covering over 90% of common protocols. If no match is found, the system calculates the weighted similarity between the unknown packet and each protocol in the library. Similarity(P) = Σ(wi × match_i) / Σwi Where wi is the weight of the i-th dimension, and match_i is the matching degree of that dimension. If the highest similarity exceeds the threshold (default 0.85), it is determined to be the protocol; if all similarities are below the threshold, the DBSCAN clustering algorithm is started to automatically cluster the unknown packets and generate a temporary protocol rule template. At the same time, the template is continuously optimized through incremental learning to achieve the adaptation of the new protocol.

[0062] Data mapping is based on a unified "device-measurement point-attribute" triple model. The mapping table uses an XML / JSON configuration file, which can be dynamically modified without restarting. The mapping process includes: address translation (such as Modbus register address mapping), data type conversion (bit / byte / word / double word / floating point), engineering quantity conversion (multiplication by a coefficient plus offset), and unit unification, ensuring semantic consistency of data from different sources.

[0063] In terms of hardware-software collaboration, the FPGA handles simple depackaging, verification, and format conversion with a latency of <1ms; the main processor is responsible for instruction mapping, semantic parsing, and logical judgment. The two interact through a high-speed internal bus, with an overall conversion latency of ≤10ms and a packet loss rate of ≤0.1%.

[0064] The scheduling module calculates a four-dimensional priority score for each data frame: Priority = α×TimeSensitivity + β×LossTolerance + γ×DelayRequirement + δ×BusinessImportance Among them, TimeSensitivity is the time sensitivity (0-1), LossTolerance is the reciprocal of the packet loss tolerance (0-1), DelayRequirement is the delay urgency (highest for hard real-time), and BusinessImportance is the business importance level. The sum of the weight coefficients α, β, γ, and δ is 1, with default values ​​of 0.3, 0.25, 0.25, and 0.2, which can be adjusted according to the scenario.

[0065] The data is divided into five levels based on the score: P1 (≥0.9, emergency control), P2 (0.7-0.9, routine control), P3 (0.5-0.7, real-time monitoring), P4 (0.3-0.5, routine monitoring), and P5 (<0.3, log diagnosis).

[0066] The scheduling algorithm is a priority-based preemptive scheduling. When a high-priority frame arrives, if a low-priority frame is currently being transmitted, the transmission is immediately interrupted, the transmitted offset is recorded, and the high-priority frame is transmitted first. After the high-priority frame is completed, the low-priority frame is resumed from the interruption point. P1 can preempt all frames, P2 can preempt P3-P5, and so on. Frames of the same priority are queued in a FIFO manner.

[0067] Meanwhile, the scheduling module, based on link status, allocates optimal links for different priorities: P1 / P2 are forced to use the link with the highest score; P3 uses the second-best link; and P4 / P5 are assigned to links with lighter loads or lower costs. When the quality of the primary link degrades, high-priority data is preferentially switched to backup links.

[0068] The system collects the signal strength (RSSI), packet loss rate (PLR), round-trip time (RTT), available bandwidth (BW), and bit error rate (BER) of each link in real time, and calculates a comprehensive score: LinkScore = w1×norm(RSSI) + w2×(1-norm(PLR)) + w3×(1-norm(RTT)) +w4×norm(BW) + w5×(1-norm(BER)) The default values ​​for w1-w5 are 0.25, 0.3, 0.25, 0.1, and 0.1 respectively. norm() is a normalization function that maps each parameter to the 0-1 interval, with a total score of 0-100.

[0069] Based on the score, the link is divided into three states: normal (≥80 points, can carry all data), warning (50-79 points, only carries P3 and above, low priority pause or switchover), and fault (<50 points, triggers switchover).

[0070] Fault switching adopts a "triple threshold" condition, which must be met simultaneously to perform the switch: (1) primary link score < 50; (2) backup link score > 80; (3) the above conditions last for more than 500ms. A 2-second cooldown period is set after the switch to prevent the ping-pong effect.

[0071] In addition, the system introduces a Q-Learning reinforcement learning mechanism, using transmission success rate and average latency as reward signals, and dynamically adjusts the weight coefficients w1-w5 to enable the scoring model to adapt to changes in the network environment.

[0072] This invention employs a plug-in architecture that decouples software and hardware. Each protocol plug-in must implement standardized interfaces: init(), parse(), encode(), get_feature(), and destroy(), and communicate with the core system through a unified bus.

[0073] The plugin hot-loading process is as follows: The system receives the plugin package (including binary data and configuration files), verifies the validity of the digital signature, then creates an independent runtime context in memory, registers the plugin information (ID, name, version, bound interface, etc.) with the global protocol state table, and finally calls init() to initialize and start listening. The entire process is completed asynchronously in the background, without affecting the operation of existing protocols, and the loading time is less than 2 seconds.

[0074] The global protocol status table is stored in the form of a hash table, recording the running status (running / paused / faulty), packet sending and receiving statistics, and configuration parameters of each plugin. It allows users to dynamically enable / disable the protocol or modify parameters via Web or CLI without restarting the device.

[0075] The cache employs a two-tier architecture of memory and SSD. A circular buffer in memory is used for high-speed caching of data at levels P1-P3, while the SSD is used for persistent storage of data at levels P4-P5 and data awaiting retransmission in case of link interruption. Data priority determines its scheduling between the two cache levels: high-priority data resides in memory, while low-priority data is automatically migrated to the SSD when memory usage exceeds a threshold.

[0076] Congestion control is divided into three levels: when the occupancy rate is ≥70%, the low-priority receiving rate is reduced and the sender is notified to limit the rate; when the occupancy rate is ≥85%, the P5 log is discarded and the P4 log is transferred to SSD; when the occupancy rate is ≥95%, the non-critical data of P4 is discarded and only the core data of P1-P3 is retained.

[0077] The breakpoint resumption mechanism assigns a globally unique incrementing sequence number (SeqNum) to each data packet. The sender records the sent and acknowledged sequence numbers, and the receiver returns an ACK. After the link is restored, the breakpoint is accurately located by comparing the sequence numbers. Only unacknowledged data packets are retransmitted, while acknowledged packets are not retransmitted, thus avoiding bandwidth waste and achieving a resumption efficiency of nearly 100%.

[0078] Uplink process (device → cloud platform): 1. Device data enters the gateway via wired / wireless interface. The protocol automatic identification module uses feature vector matching to quickly determine the protocol type. For unknown protocols, clustering is used to generate temporary rules.

[0079] 2. The conversion engine calls the corresponding plugin to parse the raw data, extract valid information, and map it to a unified data model.

[0080] 3. The scheduling module calculates priority scores, classifies them into levels P1-P5, and places them into the transmission queue according to the preemption strategy. 4. The link evaluation module calculates the score for each link in real time and selects the best link based on the priority-link matching strategy.

[0081] 5. After the data is encapsulated according to the target cloud platform protocol (such as MQTT), it is uploaded through the selected link.

[0082] 5. If the link is interrupted during transmission, the breakpoint resume module will cache unacknowledged data to the SSD, and only retransmit the data after the breakpoint after the link is restored.

[0083] Downlink process (cloud platform → device): 1. Receive downlink commands, and verify their legitimacy through secure decryption and identity authentication.

[0084] 2. The conversion engine maps standardized commands to device-specific protocol formats and verifies the validity of parameters.

[0085] 3. Downlink control commands are assigned priority to P1 by default, preempting the insertion head.

[0086] 4. Select the downlink channel to send the command according to the interface bound to the target device, and start the timeout retry mechanism (retry 3 times by default).

[0087] In this embodiment, the key parameters are preferably set as follows: priority weights: α=0.3, β=0.25, γ=0.25, δ=0.2; link scoring weights: w1=0.25, w2=0.3, w3=0.25, w4=0.1, w5=0.1; link status thresholds: normal ≥80, warning 50-79, fault <50; switching stability time 500ms, cooling time 2s; cache warning 70%, level 2 85%, level 3 95%; protocol identification similarity threshold 0.85.

[0088] In the description of this specification, the terms "one embodiment," "some embodiments," "specific embodiment," etc., refer to a specific feature, structure, material, or characteristic described in connection with that embodiment or example, which is included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0089] 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. A multi-protocol compatible industrial wireless gateway integration device, comprising: include: The core control module is used to run a dual system of real-time operating system and embedded Linux, and to perform protocol scheduling, link management and data processing. The multi-protocol wireless transceiver module includes multiple pluggable radio frequency units. Each radio frequency unit is physically isolated through an independent radio frequency isolation circuit and a shielding cavity, which is used to support the signal transmission and reception of multiple wireless protocols. The industrial protocol interface module integrates multiple industrial interfaces to adapt to data access for various wired industrial protocols. The protocol conversion acceleration module uses an FPGA chip or programmable logic device and pre-programs hardware logic for protocol parsing and data format conversion to achieve hardware-level conversion of data packets between different protocols. The data caching module adopts a dual-layer caching design of memory and solid-state storage to dynamically cache data according to data priority; In addition, a power supply and protection module is provided to offer redundant power supply and circuit protection; The core control module, multi-protocol wireless transceiver module, industrial protocol interface module, protocol conversion acceleration module, and data cache module interact with each other via a high-speed serial bus and an internal data bus.

2. The compatible multi-protocol industrial wireless gateway integration device of claim 1, wherein, The core control module and the protocol conversion acceleration module constitute an intelligent protocol conversion engine, which adopts a hardware and software collaborative conversion architecture. The protocol conversion acceleration module is used to perform hardware-level acceleration for data packet decapsulation, verification calculation, and format conversion, while the core control module is used to perform software processing for instruction mapping, semantic parsing, and logical judgment. The intelligent protocol conversion engine adopts a conversion pipeline of "feature recognition - protocol parsing - data mapping - format encapsulation" to control the protocol conversion latency to within 10ms and reduce the packet loss rate to below 0.1%.

3. The compatible multi-protocol industrial wireless gateway integration device of claim 1, wherein, The multi-protocol wireless transceiver module includes: a low-power radio frequency unit for supporting LoRa / LoRaWAN, ZigBee, and BLE short-range low-power protocols; a high-speed WiFi radio frequency unit for supporting WiFi 6 and WiFi 7 reserved; a cellular communication unit for supporting 5G / 4G / NB-IoT wide-area protocols; and an industrial wireless bus radio frequency unit for supporting WIA-PA / WirelessHART industrial wireless protocols. The industrial protocol interface module includes: multiple RS485 / RS232 serial ports for adapting to the Modbus RTU serial port protocol; multiple CAN / CAN FD bus interfaces for adapting to the CANopen bus protocol; and multiple Gigabit Ethernet ports for adapting to Profinet and EtherNet / IP industrial Ethernet protocols; each interface adopts an independent protection circuit and supports multiple devices to connect simultaneously. In the data caching module, the memory cache adopts a circular buffer design for high-speed temporary caching of P1 to P3 level data, with a read / write latency of less than 1ms; solid-state storage is used for persistent storage of P4 to P5 level data and data to be retransmitted when the link is interrupted; high-priority data resides in memory, and low-priority data is flushed to solid-state storage when the memory reaches the water level. The power supply and protection module supports wide voltage DC input and POE power supply, and adopts a dual power supply redundancy design, which automatically switches when one power supply fails; it has built-in overvoltage, overcurrent, overtemperature and reverse connection protection circuits; the device shell is made of aluminum alloy, which has dustproof and vibration-proof characteristics, and the internal circuit board is coated with conformal coating.

4. The data conversion method of a compatible multi-protocol industrial wireless gateway integrated device according to any one of claims 1 to 3, characterized in that, Specifically, the following steps are included: Automatic Protocol Identification: A two-layer identification strategy, prioritizing feature code matching and using weighted similarity as a fallback, is employed to identify the protocol type of received data packets. Multi-dimensional feature vectors are extracted for each protocol, including frame header feature codes, frame length distribution features, verification algorithm features, address field length features, and function code value range features, with each feature dimension assigned a different weight coefficient. If a feature code is not matched, the weighted similarity between the data packet and each protocol in the protocol feature library is calculated. If the similarity exceeds a threshold, the packet is identified as belonging to the corresponding protocol. If all similarities are below the threshold, the density-based DBSCAN clustering algorithm is used to automatically cluster the data packets, generating temporary protocol rule templates for each cluster. Incremental learning mechanisms are used to continuously optimize the rules, enabling the self-generation of rules for unknown protocols. Unified Data Mapping: Based on the unified data model of "device-measurement point-attribute" triple, data from different protocols are converted into a standardized structure according to a configurable mapping table. The mapping table defines the mapping relationship between the register address, data type, byte order, coefficient offset, unit and unified model of each protocol through XML or JSON configuration files, and supports dynamic loading and modification at runtime. The mapping process includes four steps: address translation, data type conversion, quantity conversion, and unit unification. Priority scheduling: Calculate the priority score of data frames based on four dimensions: time sensitivity, packet loss tolerance, latency requirements, and business importance, classify priority levels, and use a preemptive scheduling strategy for data transmission; Link evaluation and selection: Real-time acquisition of signal strength, packet loss rate, round-trip time, available bandwidth and bit error rate of each link, calculation of comprehensive link quality score, and allocation of transmission links to data according to priority and link quality matching strategy; Fault tolerance and switching: When the link score is lower than the fault threshold, the candidate backup link score is higher than the normal threshold, and the state duration exceeds the stable duration threshold, link fault switching is performed.

5. The data conversion method of claim 4, wherein, The weighted similarity calculation formula is: Similarity(P) = Σ(wi × match_i) / Σwi, where wi is the weight of the i-th feature dimension and match_i is the matching degree of that dimension; the frame header feature code accurate matching in the fast feature code matching takes less than 1ms, achieving fast recognition of more than 90% of common protocols; the similarity threshold is set to 0.85 by default; data type conversion covers the conversion between bit, byte, word, double word and floating point.

6. The data conversion method of claim 4, wherein, The formula for calculating the priority score is as follows: Priority = α × TimeSensitivity + β × LossTolerance + γ ×DelayRequirement + δ × BusinessImportance Among them, TimeSensitivity is the time sensitivity, reflecting the sensitivity of data to transmission latency, with a value range of 0-1; LossTolerance is the reciprocal of packet loss tolerance, with a higher score for lower packet loss tolerance, and a value range of 0-1; DelayRequirement is the urgency of latency requirements, with hard real-time data scoring the highest, and a value range of 0-1; BusinessImportance is the level of business importance, preset by the business type, and a value range of 0-1; α, β, γ, and δ are the weight coefficients of each dimension, satisfying α+β+γ+δ=1, and can be configured according to the scenario; The data is divided into five levels based on priority scores: Level P1 is emergency control data with a score ≥ 0.9, Level P2 is routine control data with a score 0.7 ≤ score < 0.9, Level P3 is real-time monitoring data with a score 0.5 ≤ score < 0.7, Level P4 is routine monitoring data with a score 0.3 ≤ score < 0.5, and Level P5 is log diagnostic data with a score < 0.

3. The preemptive scheduling strategy is as follows: when a high-priority data frame arrives, if low-priority data is currently being transmitted, the current transmission is interrupted, the already transmitted portion is recorded as an offset, and high-priority data is transmitted first. After the high-priority data is transmitted, the transmission of low-priority data is resumed from the interruption point. P1 level data has the highest preemption right and can preempt all low-priority data. P2 level data can preempt P3 to P5 level data. Data of the same priority is scheduled using a first-in-first-out queue.

7. The data conversion method of claim 4, wherein the data conversion method is characterized by: The formula for calculating the overall link quality score is as follows: LinkScore = w1 × norm(RSSI) + w2 × (1 - norm(PLR)) + w3 × (1 - norm(RTT)) + w4 × norm(BW) + w5 × (1 - norm(BER)) Where RSSI is signal strength, PLR is packet loss rate, RTT is round-trip time, BW is available bandwidth, BER is bit error rate, w1 to w5 are weight coefficients of each parameter, with default values ​​of 0.25, 0.3, 0.25, 0.1 and 0.1 respectively, which can be adjusted according to business scenarios, norm() is a normalization function that maps each parameter to the 0-1 range, and LinkScore ranges from 0 to 100. The link status is divided into three levels based on the LinkScore: Normal status is LinkScore ≥ 80, the link quality is good and can carry all priority data; Warning status is 50 ≤ LinkScore < 80, the link quality is degraded, it can only carry P3 level and above data, and low priority data is suspended or switched; Fault status is LinkScore < 50, the link is unavailable and the fault switching process is triggered. The priority and link quality matching strategy is as follows: P1 and P2 level data are forcibly assigned to the link with the highest score; P3 level data is assigned to the link with the second-highest score; P4 and P5 level data are assigned to the link with lower cost or lighter current load to achieve multi-link load balancing; when the quality of the main link degrades, high priority data is preferentially switched to the backup link.

8. The data conversion method for a multi-protocol compatible industrial wireless gateway integrated device according to claim 7, characterized in that, The failover must simultaneously meet the following three threshold conditions: Condition 1, the current primary link score is lower than the fault threshold, i.e., LinkScore_current < 50; Condition 2, the candidate backup link score is higher than the normal threshold, i.e., LinkScore_backup > 80; Condition 3, the duration of the above conditions exceeds the stable duration threshold, which is set to 500ms by default; After the failover is completed, a cooldown period is set, during which reverse failover is not triggered, and the cooldown period is set to 2s by default. It also includes a weighted adaptive learning step: introducing a reinforcement learning mechanism, using data transmission success rate and average latency as reward signals, and dynamically optimizing the weight coefficients of each parameter in the link scoring model through the Q-Learning algorithm, so that the scoring model gradually adapts to the current network environment.

9. The data conversion method for a multi-protocol compatible industrial wireless gateway integrated device according to claim 4, characterized in that, It also includes a protocol plugin hot-loading step: receiving a protocol plugin package, the plugin package containing a protocol stack binary file and a configuration file; Verify the plugin signature to ensure its legitimacy; Create a separate protocol runtime context in memory; Register the plugin to the global protocol state table; initialize the protocol stack and start listening to the corresponding interface; The hot loading is completed in the background, without affecting the normal operation of other protocols, and the loading time is less than 2 seconds; The global protocol status table records the protocol ID, protocol name, plugin version, bound interface, running status, data statistics and configuration parameters of all loaded protocols. The running status includes three states: running, paused and faulty. It supports querying and modification at runtime. Users can dynamically enable or disable specified protocols and modify protocol parameters without restarting the device. It also includes a breakpoint resumption step: each data packet is assigned a globally unique incrementing sequence number, the sender records the acknowledged sequence number and the sent sequence number, and the receiver returns the acknowledged sequence number; when the link is interrupted, the data is cached to solid-state storage; after the link is restored, the breakpoint is located by comparing the sent and received sequence numbers, the sender retransmits the data packets that have not been acknowledged in order, and the acknowledged data packets are not retransmitted. The breakpoint location accuracy reaches the single packet level, and the resumption efficiency is close to 100%. Tiered congestion control steps: When the cache occupancy rate reaches the first-level warning threshold of 70%, reduce the receiving rate of low-priority data and notify the sender to limit the flow; when the cache occupancy rate reaches the second-level warning threshold of 85%, discard P5 level log data and transfer all P4 level data to solid-state storage; when the cache occupancy rate reaches the third-level alarm threshold of 95%, discard P4 level non-critical monitoring data and retain only P1 to P3 level core data. Downlink command steps: Receive downlink commands from the cloud platform or industrial control system, and verify their legitimacy after decryption and identity authentication by the security module; map standardized commands to the target device's proprietary protocol format and check parameter legitimacy; assign high priority to downlink control commands by default and preemptively insert them at the head of the transmission queue; select the downlink channel to send commands according to the interface bound to the target device, wait for device response, and automatically retry if timeout occurs.

10. The data conversion method for a multi-protocol compatible industrial wireless gateway integrated device according to claim 4, characterized in that, The protocols natively integrated into the device include the following three main categories: Industrial equipment-specific protocols: including Modbus RTU / ASCII / TCP, OPC DA / UA / UA over MQTT general industrial basic protocols; Siemens Profinet / Profibus-DP, Rockwell EtherNet / IP / CIP, Mitsubishi CC-Link IE / MELSEC, Omron FINS / HostLink, Delta ASDA manufacturer-specific industrial protocols; CAN / CAN FD, RS485 / RS232, WirelessHART, ISA100.11a, WIA-PA industrial bus protocols; and DL / T 645, IEC 101 / 103 / 104, HJ212, SL 651, CJ / T 188, MBUS domestic and industry-specific protocols; General network communication protocols: including LoRa / LoRaWAN, ZigBee / ZigBee PRO, BLE5.0 / BLE Mesh short-range low-power wireless protocols; WiFi 6 and WiFi 5 high-speed wireless LAN protocols, with a reserved WiFi 7 interface; NB-IoT, 4G-LTE (Cat.1 / Cat.4 / Cat.12), 5G NR wide-area IoT wireless protocols; and TCP / IP protocol suite, VPN (IPsec / OpenVPN), DHCP / DDNS basic network transmission protocols; Standardized cloud platform protocols include MQTT, MQTT-SN, and CoAP lightweight IoT cloud protocols; and HTTP / HTTPS and FTP / FTPS universal network cloud protocols. And MQTT SparkPlug B, OPC UA over MQTT industrial cloud protocol.