A deterministic centralized scheduling method and system for ultra-low latency industrial wireless network
Patent Information
- Application Number
- CN202611247779.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-08-18
- Publication Date
- 2026-09-22
AI Technical Summary
[0006]1.多用户并发重传具有较强的整体耦合性，难以基于独立队列进行新旧数据的细粒度拆解与混合重组，局部节点出错容易导致整个调度组重传或正常节点被阻塞
[0038]关闭了媒体接入控制(MAC)层的随机退避机制，采用中心节点基于正交频分多址(OFDMA)技术进行绝对中心化联合调度。这不仅在物理层实现了绝对零碰撞，还将时延从概率性波动变为了绝对确定性。同时，利用多用户并发帧(下行多用户帧和触发帧)，在保证时延确定性的前提下，极大提升了工业高频控制场景下的网络并发容量。 打破并发重传整体耦合，消除多用户调度的木桶效应
Smart Images
Figure CN122803066A_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of industrial wireless communication technology, specifically relating to a deterministic centralized scheduling method and system for ultra-low latency industrial wireless networks. Background Technology
[0002] With the profound evolution of the Industrial Internet of Things (IIoT) and smart manufacturing, industrial wireless control networks (such as robotic arm collaboration and high-frequency sensing monitoring applications) place extremely stringent demands on the underlying communication links. Unlike traditional consumer-grade wireless networks, data transmission in industrial control scenarios must possess extremely high determinism and ultra-low latency, typically requiring microsecond to millisecond-level response times. In such scenarios, any uncontrollable delays caused by channel contention or latency jitter due to inefficient retransmission mechanisms can potentially lead to serious production accidents or system paralysis. Therefore, achieving zero channel collisions and maximally compressing the processing latency of the underlying protocol stack are core issues that urgently need to be addressed in current industrial wireless communication.
[0003] Existing wireless LAN and industrial wireless underlying access mechanisms still have obvious inherent defects, making it difficult to perfectly balance collision-free operation and extremely low latency: Under random access mechanisms, nodes need to compete for channels. For example, the carrier sense multiple access / collision avoidance (CSMA / CA) mechanism widely used in traditional Wi-Fi can easily cause collisions when multiple nodes are accessing concurrently, leading to an exponential increase in the backoff window and resulting in serious latency uncertainty.
[0004] To achieve zero collisions, some industrial networks have adopted traditional deterministic access mechanisms (such as simple Time Division Multiple Access (TDMA) or master-slave polling). However, these mechanisms typically use serial communication, resulting in high overhead for single-point interactions and a lack of multi-user high concurrency capabilities. When bit errors occur, retransmissions often need to span the entire scheduling cycle, leading to extremely high latency penalties.
[0005] Even in the latest standards that have introduced concurrent scheduling technologies such as Orthogonal Frequency Division Multiple Access (OFDMA), their retransmission and status acknowledgment (ACK) mechanisms remain inefficient.
[0006] 1. Multi-user concurrent retransmission has strong overall coupling, making it difficult to perform fine-grained decomposition and mixed recombination of new and old data based on independent queues. Errors in local nodes can easily lead to retransmission of the entire scheduling group or blockage of normal nodes.
[0007] 2. Confirmation information is usually buried deep within the complete Media Access Control (MAC) layer data payload. The central node must wait for the complete data packet to be received and verified at the physical layer before it can be extracted. This results in excessive cross-layer resolution latency, which greatly limits the scheduling flexibility and response speed in ultra-low latency scenarios.
[0008] In summary, existing wireless communication underlying access and status confirmation mechanisms can no longer meet the dual requirements of industrial deterministic networks for "absolute collision-free" and "extremely low latency". Summary of the Invention
[0009] The purpose of this invention is to provide a deterministic centralized scheduling method and system for ultra-low latency industrial wireless networks, in order to solve the problems existing in the prior art.
[0010] The above-mentioned technical objective of the present invention is achieved through the following technical solution:
[0011] A deterministic centralized scheduling method for ultra-low latency industrial wireless networks is provided, applicable to industrial wireless networks comprising a central node and multiple terminal nodes. The method includes the following steps:
[0012] Downlink control steps: The media access control layer of both the central node and the terminal node disables the conflict avoidance mechanism, and the central node obtains absolute control of the wireless channel; during the downlink transmission cycle, the central node uses orthogonal frequency division multiple access technology to encapsulate the downlink data into multi-user frames and send them concurrently to each terminal node;
[0013] Uplink scheduling steps: During the uplink transmission cycle, the central node sends a trigger frame to allocate the corresponding resource unit, so as to schedule the triggered terminal node to concurrently upload the trigger-based frame on the specified resource unit;
[0014] Status confirmation step: The central node and the terminal node implicitly transmit status confirmation information for the previous transmission cycle in the transmitted data frames or control frames through a cross-layer carrying mechanism, and the central node performs a retransmission decision in the next transmission cycle based on the status confirmation information.
[0015] Furthermore, the retransmission decision includes a decoupling mechanism based on independent queues:
[0016] The central node maintains independent transmission queues for each terminal node at the media access control layer; when assembling the multi-user frame, it extracts the head data frame from each transmission queue and sends them together.
[0017] Furthermore, the decoupling mechanism based on independent queues also includes independent pointer operations and hybrid framing steps:
[0018] After the multi-user frame is sent, the head pointer of each sending queue is operated independently according to the acknowledgment status of the previous cycle: if a successful acknowledgment status is received, the head frame of the corresponding sending queue is released and the next new frame is extracted; if no acknowledgment status is received or a timeout occurs, the head frame of the corresponding sending queue is retained.
[0019] In the next scheduling cycle, the newly extracted frames and the retained retransmitted old frames are mapped to different resource units using orthogonal frequency division multiple access technology, and then mixed and packaged into the same new multi-user frame for multiple terminal nodes and sent concurrently.
[0020] Furthermore, a signaling-free frequency hopping enhancement mechanism is introduced into the retransmission decision:
[0021] Both the central node and the terminal node have multiple sets of orthogonal frequency division multiple access resource unit frequency domain mapping tables pre-stored locally, and the system maintains a global scheduling sequence number;
[0022] Each time a retransmission scheduling is initiated, the central node and the terminal node synchronously execute the frequency hopping algorithm based on the global scheduling sequence number, so that the retransmitted old frame is transmitted on a different resource unit than the previous transmission.
[0023] Furthermore, the cross-layer payload mechanism includes a downlink physical layer payload mechanism:
[0024] During downlink transmission, the central node generates a global confirmation bitmap from the status confirmation information of all terminal nodes in the previous uplink cycle, and broadcasts the global confirmation bitmap in the physical layer preamble common signaling field of the current downlink multi-user frame or the trigger frame.
[0025] Furthermore, the downlink physical layer carrying mechanism also includes a low-power sleep mechanism:
[0026] Terminal nodes that have not been allocated downlink data resources obtain their own acknowledgment status by listening to the physical layer preamble common signaling field, and then directly enter sleep mode after obtaining the status.
[0027] Furthermore, the downlink physical layer payload mechanism also includes an empty squad broadcast confirmation mechanism:
[0028] If the central node's current sending queue is empty and there are no multi-user frames or trigger frames to be sent, the central node independently sends a broadcast confirmation frame and includes the global confirmation bitmap in the broadcast confirmation frame.
[0029] Furthermore, the cross-layer mounting mechanism includes an implicit mounting mechanism for the uplink media access control header:
[0030] During uplink transmission, the terminal node will include the acknowledgment status for downlink data in the extended control field of the Media Access Control header of the transmitted trigger-based frame;
[0031] If the terminal node currently has no uplink data, it constructs and sends an idle trigger-based frame with zero data payload, and transmits the acknowledgment status in the extended control field of its Media Access Control header.
[0032] Furthermore, the state confirmation step also includes a state streaming resolution mechanism:
[0033] When the central node receives an uplink trigger-based frame, it utilizes the streaming processing capabilities of the baseband hardware to extract the confirmation status of each terminal node from the demodulated media access control header within the physical time window before the data load has been received or verified.
[0034] A deterministic centralized scheduling system for ultra-low latency industrial wireless networks includes:
[0035] A central node, whose media access control layer disables the collision avoidance mechanism to obtain absolute control of the wireless channel; the central node includes a queue management module, a hybrid framing module, a preamble carrying module, and a streaming parsing module, used to execute the method steps on the central node side;
[0036] Multiple terminal nodes, whose media access control layer is configured as a purely passive state machine that does not respond to triggers; the terminal nodes include a preamble listening module and a header mounting module, used to execute the method steps on the terminal node side.
[0037] In summary, the present invention has the following beneficial effects:
[0038] The random backoff mechanism at the Media Access Control (MAC) layer was disabled, and a centralized, jointly scheduled system based on Orthogonal Frequency Division Multiple Access (OFDMA) technology was adopted. This not only achieved absolute zero collisions at the physical layer but also transformed latency from probabilistic fluctuations to absolute determinism. Simultaneously, by utilizing multi-user concurrent frames (downlink multi-user frames and trigger frames), the network concurrency capacity in industrial high-frequency control scenarios was significantly improved while maintaining latency determinism. This also broke down the overall coupling of concurrent retransmissions and eliminated the bottleneck effect in multi-user scheduling.
[0039] Within a single scheduling cycle, old frames requiring retransmission are seamlessly mixed and packaged with brand-new frames extracted from other successful queues in the frequency domain. This mechanism completely breaks the retransmission bottleneck effect in multi-user concurrent transmission, ensuring that the throughput of normal terminal nodes is no longer dragged down by local packet loss, and maximizing channel utilization under extremely low latency.
[0040] This application introduces a signaling-free deterministic frequency hopping mechanism based on a global scheduling sequence number. When retransmitting an old frame, it automatically switches to a completely new frequency domain resource unit different from the previous transmission during the next scheduling initiation. This significantly enhances the system's reliability against narrowband interference in harsh industrial electromagnetic environments with zero signaling negotiation overhead and zero additional latency.
[0041] This application completely eliminates the overhead of sending independent acknowledgment (ACK) control frames. In the downlink direction, the global acknowledgment bitmap is embedded in the common signaling field of the physical layer preamble; terminal nodes without allocated data resources only need to listen to the physical layer preamble to quickly obtain the acknowledgment status and directly enter sleep mode, reducing node power consumption. In the uplink direction, the acknowledgment status is embedded in the extended control field of the trigger-based Media Access Control (MAC) header; when a terminal node has no uplink data, it carries the acknowledgment status by replying with an empty frame with zero data load. Combined with the streaming processing characteristics of the baseband hardware on the central node side, decisions are made in advance before the data load is fully received, pushing the retransmission response speed to the hardware limit. Attached Figure Description
[0042] Figure 1 This is a flowchart of the deterministic joint scheduling and streaming parsing process on the central node (AD) side of the present invention.
[0043] Figure 2 This is a flowchart of the passive response and cross-layer implicit mounting process on the terminal node (FD) side as described in this invention. Detailed Implementation
[0044] To make the technical means, creative features, objectives and effects of this invention easier to understand, the invention will be further described below with reference to the figures and specific embodiments.
[0045] This invention provides a deterministic centralized scheduling method and system for ultra-low latency industrial wireless networks. The industrial wireless network adopts a star topology, consisting of a central node with high computing power and multiple identical or heterogeneous terminal nodes. To completely eliminate latency jitter in high-frequency control loops at the millisecond or even microsecond level, the system undergoes a deep reconstruction of the traditional underlying wireless communication protocol stack.
[0046] I. Protocol Stack Architecture and Static Time-Frequency Resource Configuration
[0047] This scheme deploys a deterministic media access control (D-MAC) layer and an orthogonal frequency division multiple access physical (OFDMA-PHY) layer in both the central node and each of the terminal nodes.
[0048] Deterministic control architecture: After system startup, the D-MAC layers of the central node and all terminal nodes forcibly disable the Distributed Coordination Function (DCF) and random backoff algorithm components in the Carrier Sense Multiple Access / Collision Avoidance (CSMA / CA) mechanism via firmware instructions. The central node obtains absolute control over the current wireless spectrum channel through exclusive channel measurement and beacon announcement; the local access state machine of each terminal node is forcibly configured to a purely passive follow-up state of "no trigger, no response".
[0049] Multi-user queue management architecture: Within the D-MAC layer, the central node constructs and maintains an asymmetric multi-independent transmission queue addressing matrix. This matrix allocates and maps a separate, isolated, and non-interfering first-in-first-out (FIFO) transmission queue for each legitimate terminal node's physical address (MAC Address) in the network. The head pointer and read / write control lines of each queue are independently and dynamically addressed and controlled by the central node's scheduling core.
[0050] Discretization of Time-Frequency Resources: The OFDMA-PHY layer divides the continuous wireless spectrum into multiple discrete orthogonal subcarriers, and aggregates these subcarriers into multiple independent frequency domain resource units (RUs). The central node and each terminal node synchronously pre-store multiple sets of frequency domain mapping tables for the discretized resource units in their local non-volatile memory (NVM). Simultaneously, the system maintains a fixed-length global scheduling sequence number (SN) in the underlying hardware timer, which increments with the joint scheduling period.
[0051] II. Downlink Control Flow and Asymmetric Cross-Layer Physical Carrying Mechanism
[0052] Combination Figure 1 and Figure 2 Within each defined joint scheduling cycle, the system downlink executes high-concurrency multi-user transport streams and macro-state declaration logic:
[0053] Dynamic offloading addressing and decoupled framing: The scheduling operator inside the central node is activated. For the current downlink transmission cycle, the central node independently searches the independent transmission queues corresponding to each terminal node. If a queue is detected to be non-empty, the data frame currently at the head of the queue is extracted from each non-empty transmission queue through a multiplexer (MUX). Subsequently, using orthogonal frequency division multiple access (OFDMA) technology, the data frames at the head of the queues extracted from different queues and oriented to different physical addresses are mapped to different physical resource units in the frequency domain, and finally aligned and combined in the time domain to encapsulate into a single downlink multi-user frame (MU frame).
[0054] The global acknowledgment status is implemented at the physical layer: the central node extracts and summarizes all media access control layer status acknowledgment information received from each terminal node in the previous uplink cycle. The scheduling core compresses and packages this non-aligned status information to generate a high-density global acknowledgment bitmap (global ACK bitmap). When constructing the Physical Layer Protocol Data Unit (PPDU) for the current cycle, the central node determines whether there is a downlink MU frame or the trigger frame to be sent. If the determination is "yes" (i.e., there is a frame to be sent), the baseband modulation chip directly writes the global acknowledgment bitmap into the physical layer preamble common signaling field of the downlink multi-user frame or the trigger frame to be sent through the cross-layer hardware pass-through bus, and sends it down along with the physical layer broadcast signal. If the determination is "no" (i.e., all independent transmission queues of the central node are empty and there is no downlink data or control frame to be sent), in order to avoid the terminal node causing a timeout retransmission failure due to not receiving acknowledgment information, the central node activates a dedicated triggering mechanism, independently constructs and sends a dedicated broadcast acknowledgment frame with a fixed physical layer length, and loads the global acknowledgment bitmap into the physical layer signaling space of the broadcast acknowledgment frame for forced transmission.
[0055] Terminal-side rapid interception and selective low-power sleep: On the downlink, the physical layer RF front-end of each terminal node continuously monitors the channel. When a downlink PPDU arrives, the baseband processor of the terminal node immediately intercepts and extracts the global acknowledgment bitmap the instant it demodulates the physical layer preamble common signaling field. Terminal nodes that have not been allocated downlink data resources in the current cycle and have determined by retrieving the global acknowledgment bitmap that their uplink data in the previous cycle has been successfully received by the central node (corresponding bit is 1) do not need to continue activating the subsequent physical layer baseband demodulator and data buffer. Their access state machine directly issues an interrupt command to the power management chip (PMIC), controlling the RF front-end and baseband core to directly enter a microsecond-level low-power sleep state. This fundamentally removes the constraint that uplink and downlink scheduling objects must maintain strict symmetrical interaction. Terminal nodes that have been allocated downlink data resources or need to perform uplink backhaul release their locally confirmed uplink historical buffer frames rapidly according to the bitmap state and continue decoding their own multi-user frame data payload area on the corresponding resource unit.
[0056] III. Implicit Embedding Mechanism of Uplink Forced Scheduling Stream and Media Access Control Header
[0057] See Figure 2 During the uplink transmission cycle, the network uses a time-frequency strictly aligned triggering mechanism to ensure that the uplink channel completely avoids the risk of random collisions.
[0058] Deterministic frequency domain macro-scheduling: After sending a multi-user frame, the central node immediately sends a timing-precisely aligned trigger frame. The trigger frame contains an explicit physical layer identifier for a specific terminal node, as well as the specified uplink frequency domain resource unit (RU) size and index.
[0059] Implicit cross-layer uplink control domain integration: After receiving the trigger frame, each of the triggered terminal nodes triggers its corresponding hardware timer to initiate uplink transmission within strictly defined timing boundaries. The terminal node directly writes the verification and reception acknowledgment (ACK / NACK) status of the currently received downlink multi-user frame into the extended control domain of the Media Access Control (MAC) header of the currently assembled trigger-based frame (TB frame), completing the implicit integration in the uplink direction.
[0060] Zero-load uplink feedback idle frame mechanism: If the local transmission queue of the scheduled terminal node is empty in the current time slot, i.e., there is no substantial industrial high-frequency sensor data to be transmitted, in order to ensure the integrity of the central node's closed-loop control algorithm, the terminal node's D-MAC layer will automatically truncate data generation in memory and construct an idle trigger-based frame with a data payload length field encapsulated as zero. This idle frame only retains the physical layer preamble and the necessary MAC header, and continues to retain and transmit the reception acknowledgment status for the current downlink data in its Media Access Control header extended control field, and then sends a concurrent response uplink.
[0061] IV. Hardware Pipeline Streaming Analysis and Decoupling Frequency Hopping Retransmission Mechanism
[0062] See Figure 1 After receiving multiple uplink TB frame signals concurrently, the central node, through deep cross-layer coupling at the baseband level, maximizes the computational latency of scheduling decisions:
[0063] Parallel pipelined streaming state extraction: When the physical layer baseband chip of the central node concurrently receives trigger-based frames (or idle trigger-based frames) uploaded by multiple terminal nodes at the RF front end, it initiates a hardware-level streaming processing pipeline. The baseband chip does not need to wait for the entire data payload of the TB frame to be fully received, nor does it need to wait for the Cyclic Redundancy Check (CRC) circuit to complete the entire packet calculation. Instead, at the instant the physical layer preamble and MAC header demodulation are completed, it directly extracts the state confirmation information of each terminal node from the extended control field of the MAC header through a high-speed hardware pipeline.
[0064] Asynchronous decoupling control of the queue head pointer: Based on the confirmation status extracted from the streaming above, the scheduling core of the central node performs asynchronous decoupling control on each sending queue in the asymmetric multi-independent sending queue addressing matrix.
[0065] For the sending queue that has successfully retrieved the acknowledgment status (ACK=1) from the streaming, the scheduling core directly sends a release pulse to the queue, pops and destroys the current head frame, controls the read pointer to move one position forward, and extracts the next frame of new data from the FIFO memory and moves it to the head of the queue.
[0066] For transmission queues that have not received an acknowledgment or whose demodulation determination has timed out due to interference, the scheduling core freezes the pointer of the queue, strictly preserves the current first frame of the queue, marks it as a retransmission of an old frame, and leaves it in its original position to wait for reactivation.
[0067] Seamless frequency-domain hybrid reassembly: At the start of the next scheduling cycle, the hybrid framing module of the central node performs deep frequency-domain reassembly. This module, through multiplexing, horizontally merges newly extracted frames from the successful queue with retransmitted old frames remaining at the head of the failed queue. In the orthogonal frequency division multiple access modulation mapping circuit, these new frames and retransmitted old frames are allocated and mapped in parallel to different resource units (RUs), uniformly compressed and packaged into a single multi-user frame for multiple terminal nodes for concurrent transmission. Thus, the throughput of the normal queue in the time domain is not affected by the retransmission cycle of local packet loss nodes, fundamentally eliminating the "weakest link" effect in concurrent transmission.
[0068] Overhead-free noncoherent deterministic frequency hopping: During the retransmission and framing process described above, the system's global scheduling sequence number increments (SN = SN + 1). To prevent continuous packet loss at failed nodes due to sudden narrowband electromagnetic interference in industrial settings, the system initiates a signaling-free deterministic frequency hopping mechanism. The D-MAC layer scheduling subsystems of the central node and each terminal node directly use the newly incremented global scheduling sequence number as an independent variable and substitute it into the locally pre-stored frequency hopping mapping algorithm. Without consuming any air interface control signaling or introducing any protocol interaction delay, the system automatically calculates and switches the resource unit mapping configuration locally, so that the retransmitted old frame is forcibly allocated to a completely new spectrum frequency domain resource unit that is entirely different from the previous transmission for modulation and transmission in the next transmission, thereby giving the network extremely high-frequency channel anti-interference resilience and high determinism.
[0069] To more clearly illustrate the operational logic of this invention, the following example, using an industrial wireless communication system comprising one central node (AD) and eight terminal nodes (FD1-FD8), with two consecutive scheduling cycles as examples, details the "decoupling retransmission" and "rapid confirmation" processes of this system when facing partial transmission failures. Specifically, the media access control layer of the central node disables the collision avoidance mechanism to obtain absolute control of the wireless channel; the central node includes a queue management module, a hybrid framing module, a preamble mounting module, and a streaming parsing module; the media access control layers of the eight terminal nodes are configured as purely passive state machines that do not respond to triggers; each terminal node includes a preamble listening module and a header mounting module.
[0070] I. Initial State Setup: AD maintains 8 independent transmit queues (Q1 to Q8) for FD1 to FD8 at the Media Access Control (MAC) layer. Assume the current global scheduling sequence number is SN=1.
[0071] II. First Scheduling Cycle (SN=1): Initial Concurrency and Local Failure
[0072] AD sends out multi-user data and trigger frames (at time T0): AD extracts a brand new head-of-line data frame from Q1 to Q8, combines them into a multi-user (MU) frame using OFDMA technology, and sends it concurrently to FD1 to FD8 on 8 resource units (RU). Immediately following the MU frame, AD sends out a trigger frame to schedule the specified RU and trigger each FD to reply with a TB frame.
[0073] FD reception and uplink confirmation (time T1):
[0074] Successful receiving terminal (taking FD1 and FD3 to FD8 as examples): Successfully parses the data belonging to itself in the MU frame and updates the ack information locally. Subsequently, it sends a TB frame on the uplink RU assigned by the Trigger frame, and sets the ACK flag in its MAC header control field to 1 (if the terminal's transmit queue is empty at this time, it sends an empty TB frame to carry the ack information uplink).
[0075] Terminal that fails to receive data (taking FD2 as an example): Due to local interference, FD2 failed to successfully parse the MU frame, and therefore could not perform normal ACK and data reply.
[0076] AD streaming parsing and queue state machine update (time T2): AD concurrently receives uplink signals, parses the MAC header, and obtains ACK information:
[0077] For Q1 and Q3~Q8, once the AD detects that an ACK has been received, the first frame of these queues is released, the pointer is moved forward, and the new data of the next frame is obtained.
[0078] For Q2, AD determines that no ACK has been received (timeout / NACK), does not release the first frame of Q2, and prepares to retransmit it in the next cycle.
[0079] III. Second Scheduling Cycle (SN=2): Hybrid Frame Retransmission
[0080] The global scheduling sequence number is incremented to SN=2.
[0081] AD hybrid framing and downlink global physical layer confirmation (time T3):
[0082] Independent queue decoupling framing: AD extracts brand new data frames from Q1 and Q3 to Q8, and seamlessly mixes them with the retransmitted old data frames retained in Q2 to generate new MU frames.
[0083] Deterministic frequency hopping without signaling: To avoid prolonged interference on the same frequency band causing FD2 to continue losing packets, the AD and all FDs synchronously switch the RU mapping configuration based on the current sequence number SN=2. The AD maps the retransmitted data of Q2 to a completely new RU frequency band that is entirely different from that when SN=1, without the need for additional signaling negotiation throughout the process.
[0084] Equipped with a global ACK bitmap: The AD generates an ACK bitmap from the uplink reception status of the 8 FDs in the previous cycle (SN=1), and directly embeds it into the physical layer preamble common signaling field (PHYCommonInfo) of the MU frame. Then, the MU frame and Trigger frame are sent out according to the logic at time T0.
[0085] FD Reception and ACK Status Update (Time T4): All terminals (FD1 to FD8) listen for the physical layer preamble of the MU frame and read the global ACK bitmap. Based on the ACK bitmap status, FD1 to FD8 rapidly update their own uplink queue status (releasing successfully transmitted uplink frames). Subsequently, all eight terminals successfully parsed the contents of the data payload area (including FD2 successfully receiving the retransmitted frame after frequency hopping, and the other terminals receiving the new frame).
[0086] Uplink Acknowledgment (Time T5): Assume that all uplink industrial data from all FDs has been transmitted (i.e., the terminal transmission queue is empty). Eight terminals (FD1~FD8) concurrently transmit empty TB frames (i.e., frames containing only the physical layer preamble and MAC header, with a data payload length of 0) on their respective RUs, and set the ACK flag to 1 in the MAC header of each terminal. After the AD receives the signal and stream-parses the eight ACK states from the MAC header, it releases all the first frames of queues Q1~Q8. The system then perfectly enters the next new scheduling cycle with SN=3 with extremely low latency and extremely high channel utilization.
[0087] In this document, the terms "upper," "lower," "front," "back," "left," "right," "top," "bottom," "inner," "outer," "vertical," and "horizontal," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only used for the clarity of expressing the technical solution and for the convenience of description, and therefore should not be construed as limiting the present invention.
[0088] In this document, the terms “including,” “comprising,” or any other variations thereof are intended to cover non-exclusive inclusion, which includes not only the elements listed but also other elements not expressly listed.
[0089] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of this invention is defined by the appended claims and their equivalents.
Claims
1. A deterministic centralized scheduling method for ultra-low latency industrial wireless networks, characterized in that, Applied to an industrial wireless network, which includes a central node and multiple terminal nodes, the method includes the following steps: Downlink control steps: The media access control layer of both the central node and the terminal node disables the conflict avoidance mechanism, and the central node obtains absolute control of the wireless channel; during the downlink transmission cycle, the central node uses orthogonal frequency division multiple access technology to encapsulate the downlink data into multi-user frames and send them concurrently to each terminal node; Uplink scheduling steps: During the uplink transmission cycle, the central node sends a trigger frame to allocate the corresponding resource unit, so as to schedule the triggered terminal node to concurrently upload the trigger-based frame on the specified resource unit; Status confirmation step: The central node and the terminal node implicitly transmit status confirmation information for the previous transmission cycle in the transmitted data frames or control frames through a cross-layer carrying mechanism, and the central node performs a retransmission decision in the next transmission cycle based on the status confirmation information.
2. The deterministic centralized scheduling method for ultra-low latency industrial wireless networks according to claim 1, characterized in that, The retransmission decision includes a decoupling mechanism based on independent queues: The central node maintains independent transmission queues for each terminal node at the media access control layer; when assembling the multi-user frame, it extracts the head data frame from each transmission queue and sends them together.
3. The deterministic centralized scheduling method for ultra-low latency industrial wireless networks according to claim 2, characterized in that, The decoupling mechanism based on independent queues also includes independent pointer operations and hybrid framing steps: After the multi-user frame is sent, the head pointer of each sending queue is operated independently according to the acknowledgment status of the previous cycle: if a successful acknowledgment status is received, the head frame of the corresponding sending queue is released and the next new frame is extracted; if no acknowledgment status is received or a timeout occurs, the head frame of the corresponding sending queue is retained. In the next scheduling cycle, the newly extracted frames and the retained retransmitted old frames are mapped to different resource units using orthogonal frequency division multiple access technology, and then mixed and packaged into the same new multi-user frame for multiple terminal nodes and sent concurrently.
4. The deterministic centralized scheduling method for ultra-low latency industrial wireless networks according to claim 1, characterized in that, The retransmission decision also incorporates a signaling-free frequency hopping enhancement mechanism: Both the central node and the terminal node have multiple sets of orthogonal frequency division multiple access resource unit frequency domain mapping tables pre-stored locally, and the system maintains a global scheduling sequence number; Each time a retransmission scheduling is initiated, the central node and the terminal node synchronously execute the frequency hopping algorithm based on the global scheduling sequence number, so that the retransmitted old frame is transmitted on a different resource unit than the previous transmission.
5. The deterministic centralized scheduling method for ultra-low latency industrial wireless networks according to claim 1, characterized in that, The cross-layer payload mechanism includes a downlink physical layer payload mechanism: During downlink transmission, the central node generates a global confirmation bitmap from the status confirmation information of all terminal nodes in the previous uplink cycle, and broadcasts the global confirmation bitmap in the physical layer preamble common signaling field of the current downlink multi-user frame or the trigger frame.
6. The deterministic centralized scheduling method for ultra-low latency industrial wireless networks according to claim 5, characterized in that, The downlink physical layer carrying mechanism also includes a low-power sleep mechanism: Terminal nodes that have not been allocated downlink data resources obtain their own acknowledgment status by listening to the physical layer preamble common signaling field, and then directly enter sleep mode after obtaining the status.
7. The deterministic centralized scheduling method for ultra-low latency industrial wireless networks according to claim 5, characterized in that, The downlink physical layer payload mechanism also includes an empty squad broadcast confirmation mechanism: If the central node's current sending queue is empty and there are no multi-user frames or trigger frames to be sent, the central node independently sends a broadcast confirmation frame and includes the global confirmation bitmap in the broadcast confirmation frame.
8. The deterministic centralized scheduling method for ultra-low latency industrial wireless networks according to claim 1, characterized in that, The cross-layer mounting mechanism includes an implicit mounting mechanism for the uplink media access control header: During uplink transmission, the terminal node will include the acknowledgment status for downlink data in the extended control field of the Media Access Control header of the transmitted trigger-based frame; If the terminal node currently has no uplink data, it constructs and sends an idle trigger-based frame with zero data payload, and transmits the acknowledgment status in the extended control field of its Media Access Control header.
9. The deterministic centralized scheduling method for ultra-low latency industrial wireless networks according to claim 1, characterized in that, The state confirmation step also includes a state streaming resolution mechanism: When the central node receives an uplink trigger-based frame, it utilizes the streaming processing capabilities of the baseband hardware to extract the confirmation status of each terminal node from the demodulated media access control header within the physical time window before the data load has been received or verified.
10. A deterministic centralized scheduling system for ultra-low latency industrial wireless networks, characterized in that, include: A central node whose media access control layer disables the collision avoidance mechanism to obtain absolute control of the wireless channel; the central node includes a queue management module, a hybrid framing module, a preamble carrying module, and a streaming parsing module, for executing the method steps of the central node side as described in any one of claims 1 to 9; Multiple terminal nodes, whose media access control layer is configured as a purely passive state machine that does not respond to triggers; the terminal nodes include a preamble listening module and a header mounting module, used to execute the method steps on the terminal node side of any one of claims 1 to 9.