An autonomous routing method and system for massive internet of things

Abstract

The application discloses an autonomous routing method and system for large-scale Internet of Things, and belongs to the technical field of low-power lossy network routing, which comprises the following steps: allocating a 3-byte unique identifier containing a subnet number and a hardware serial number to a node; nodes in a subnet mutually notify the identifier, and a cluster head is dynamically elected based on signal quality; the cluster head runs an RPL protocol to build a DODAG in the cluster; the cluster heads run a BATMAN protocol to obtain link metrics and establish an optimal path; the cluster head generates a service message containing DODAG information of the cluster, and sends the message to all reachable other cluster heads according to the BATMAN optimal path, so as to realize routing synchronization; a target cluster head analyzes the message and sets a BATMAN gateway, and publishes external routing to the cluster through an RPL DIO; and cluster head failure is periodically detected to trigger reselection and routing reconstruction. The method realizes efficient, reliable and self-healing routing in a large-scale Internet of Things environment through innovative network layering and protocol fusion, and is suitable for Internet of Things applications in a complex dynamic environment.

Description

Technical Field

[0001] This invention relates to the field of low-power lossy network routing technology, and more specifically to an autonomous routing method and system for large-scale Internet of Things (IoT). Background Technology

[0002] With the rapid development of IoT technology, numerous new application scenarios have emerged in fields such as intelligent transportation and industrial automation, including vehicle-to-everything (V2X) networks, emergency response mobile base stations, and industrial sensor networks. These scenarios typically feature large node scales, dynamically changing network topologies, and harsh communication environments, such as high interference and high loss. Traditional single routing protocols, such as Low-Power Lossy Network Routing Protocol (RPL), while performing well in static low-power networks, are prone to routing detours and single-point failures due to their inherent tree structure and reliance on root nodes in large-scale mobile networks. This makes it difficult to meet the low-latency, high-reliability communication requirements of intelligent transportation and industrial control.

[0003] While some existing technologies have attempted to improve RPL to support mobility or implement hierarchical and partitioned management within the RPL framework, these solutions are mostly limited to optimizations within a single protocol stack and fail to fundamentally solve the dynamic routing problem across regions and subnets. In complex applications, communication networks often need to integrate fixed and mobile nodes to form a hybrid communication system that spans regions and involves multiple hops. This heterogeneous and highly dynamic network environment poses unprecedented challenges to the scalability and autonomy of routing protocols. Summary of the Invention

[0004] To address the aforementioned technical problems, this invention provides an autonomous routing method and system for large-scale Internet of Things (IoT) systems, which can adapt to changes in network size and topology, combining local efficiency with global flexibility to meet the complex communication infrastructure requirements of modern IoT systems.

[0005] To achieve the above objectives, the present invention adopts the following technical solution:

[0006] In a first aspect, embodiments of the present invention provide an autonomous routing method for large-scale Internet of Things (IoT), comprising the following steps: Node Identifier: Each node in the Internet of Things network is assigned a unique identifier, which includes the subnet number and the node hardware serial number; Cluster head election: The entire large-scale Internet of Things network is divided into multiple subnets. Each node in the subnet announces its own identification number to each other. The receiving node dynamically elects a cluster head based on comprehensive signal quality indicators. Intra-cluster route establishment: The cluster head node is used as the subnet center point to run the RPL protocol and construct a Destination Oriented Directed Acyclic Graph (DODAG) within the subnet to form intra-cluster routes; Inter-cluster routing establishment: The cluster head nodes form a backbone network and run a better mobile hotspot network mode (BATMAN protocol). They obtain the metric value of the inter-cluster link by exchanging source message (OGM) messages and establish the optimal path between cluster heads. Route synchronization and propagation: The cluster head node generates a service packet containing the DODAG routing information of its own cluster, and sends the service packet to all other reachable cluster heads in the BATMAN routing table according to the optimal inter-cluster path determined by the BATMAN protocol, thereby realizing the interaction of inter-cluster routing information and the synchronization of the entire network routing table; External cluster route injection: After receiving the service packet, the target cluster head parses the DODAG information, sets the next hop to the BATMAN gateway to the source cluster head, and publishes the external route to the cluster node through RPL DIO messages; Dynamic route self-healing: Periodically detect the cluster head's operating status. When a cluster head failure is detected, trigger the corresponding subnet to re-elect a cluster head. The new cluster head rebuilds intra-cluster RPL routes and inter-cluster BATMAN routes, completing the full network route recovery.

[0007] Furthermore, the method for generating the node hardware serial number includes: extracting the last 16 bits of the node hardware serial number; if the hardware serial number is less than 16 bits, padding with zeros; if it exceeds 16 bits, using the CRC-16 check algorithm to calculate and taking the lower 16 bits as the node hardware serial number.

[0008] Furthermore, subnet numbers are assigned based on physical distance or usage scenario, with nodes within the same physical area or functional domain being assigned the same subnet number.

[0009] Furthermore, the receiving node dynamically elects the cluster head based on comprehensive signal quality indicators, including: when there is only one node in the subnet, that node automatically becomes the cluster head; When there are multiple nodes in a subnet, each node in the subnet broadcasts its own node number and simultaneously collects the RSSI signal strength, SNR signal-to-noise ratio, LQI link quality indicator, and SOC node power parameters of adjacent nodes. For benefit-type indicators and cost-type indicators, corresponding normalization formulas are used for normalization. The signal quality score is obtained by weighting and summing the normalized parameters using preset weighting coefficients, and the node with the highest signal quality score is selected as the cluster head node of the current subnet.

[0010] Furthermore, in the inter-cluster routing establishment step, the cluster head obtains the link delay, link stability, and peer cluster head power through OGM messages; for paths spanning multiple hops, the sum of the delays of each hop is taken as the path delay, the minimum stability of each hop is taken as the path stability, and the minimum power of each hop is taken as the path power; after normalizing the path delay, path stability, and path power respectively, they are weighted and summed with preset weight coefficients to calculate the path metric value, and the path with the optimal metric value is selected as the optimal path between cluster heads.

[0011] Furthermore, in the routing synchronization and diffusion step, the service message format includes a message header and a data segment. The message header contains a 3-byte source cluster head number, a 2-byte sequence number, and a 1-byte CRC-8 checksum.

[0012] Furthermore, the CRC-8 checksum is calculated using the CRC-8 algorithm on the source cluster head number, sequence number, and intra-cluster DODAG routing data in the data segment, generating a polynomial x. 8 +x 5 +x 4 +1, the initial value is 0xFF, the input and output are not flipped, and the result XOR value is 0x00; the receiving end recalculates the same data range with the same parameters, compares the calculation result with the check bit, and if they match, the message is determined to be complete and error-free, otherwise the message is discarded.

[0013] Furthermore, in the routing synchronization and propagation step, the service packets are sent to all other reachable cluster heads in the BATMAN routing table. Specifically, the IP destination of the service packets is set to the destination IP address in the BATMAN routing table, and then forwarded according to the next hop in the BATMAN routing table.

[0014] Furthermore, in the external cluster route injection step, the external route is advertised to the ordinary nodes in the cluster through the DIO message of the RPL protocol, so that the ordinary nodes can obtain the external route to the external cluster; the ordinary nodes add the external route according to the DIO message, and use the best next-hop node to reach the cluster head of the cluster determined by the RPL protocol as the gateway.

[0015] In a second aspect, embodiments of the present invention provide an autonomous routing system for large-scale Internet of Things (IoT), employing the method described in the first aspect, the system comprising: The node identification module is used to assign a numbered identifier to each node in the Internet of Things network. The numbered identifier includes a subnet number and a node hardware serial number. The cluster head election module is used to divide the entire large-scale Internet of Things network into multiple subnets. Each node in the subnet announces its own identification number to each other, and the receiving node dynamically elects a cluster head based on comprehensive signal quality indicators. The intra-cluster routing establishment module is used to run the RPL protocol with the cluster head node as the subnet center point to construct the DODAG within the subnet and form intra-cluster routes. The inter-cluster routing establishment module is used to form a backbone network between the cluster head nodes, run the BATMAN protocol, obtain the metric value of the inter-cluster link by exchanging OGM messages, and establish the optimal path between cluster heads. The routing synchronization and diffusion module is used by the cluster head node to generate a service packet containing the DODAG routing information of the cluster, and send the service packet to the target cluster head according to the optimal inter-cluster path determined by the BATMAN protocol, so as to achieve global routing synchronization. The external cluster routing injection module is used to parse the DODAG information after the target cluster head receives the service message, set the next hop to the BATMAN gateway to the source cluster head, and publish the external route to the local cluster node through RPL DIO messages. The dynamic routing self-healing module is used to periodically detect the cluster head's operating status. When a cluster head failure is detected, it triggers the corresponding subnet to re-elect a cluster head. The new cluster head rebuilds the intra-cluster RPL routes and inter-cluster BATMAN routes, completing the full network routing recovery.

[0016] The beneficial effects of the above-described technical solutions provided in the embodiments of the present invention include at least the following: This invention adopts a hierarchical network architecture of intranet RPL and inter-cluster BATMAN, localizing routing control and decoupling intranet routing from inter-cluster routing. This allows the network to scale up to tens of thousands of nodes without incurring significant control overhead, fundamentally eliminating the risk of single point of failure and freeing the network size from being limited by the processing capacity of a single root node.

[0017] RPL and BATMAN protocols operate independently and decoupled. RPL focuses on routing management for low-power lossy networks within a cluster, while BATMAN focuses on link quality awareness and optimal path calculation for inter-cluster backbone networks. Synchronization of intra-cluster routing information does not depend on the convergence process of the BATMAN protocol; instead, it is achieved through controlled forwarding of independent service packets, avoiding negative interference and redundant control overhead between protocols and significantly accelerating route convergence. BATMAN offers superior packet delivery rate and end-to-end latency compared to RPL in dense networks and mobile scenarios. This invention's protocol fusion scheme fully leverages the respective advantages of both protocols.

[0018] The dynamic cluster head election mechanism ensures that the subnet can quickly elect a new cluster head after a single point of failure, while the multi-dimensional link evaluation mechanism enables the backbone network to automatically avoid unstable paths, ensuring the reliability of end-to-end communication. The directional routing synchronization mechanism effectively isolates global routing oscillations caused by changes in a single cluster head.

[0019] Using node energy consumption (SOC) as one of the core indicators for cluster head election and path selection balances network energy consumption and prevents network fragmentation caused by some nodes running out of energy too early. The directional routing synchronization mechanism avoids global flooding of routing information exchange, significantly reducing communication overhead and improving bandwidth utilization.

[0020] Physical subnetting simplifies network topology management; the transparent global routing mechanism allows ordinary nodes to access all network resources without running complex protocols. Ordinary nodes only need to maintain a simple default route pointing to the cluster head, which greatly reduces the complexity of application development and operation. Attached Figure Description

[0021] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0022] Figure 1 This is a flowchart of an autonomous routing method for large-scale Internet of Things provided in an embodiment of the present invention.

[0023] Figure 2 This is a protocol stack architecture diagram of an autonomous routing method for large-scale Internet of Things provided in an embodiment of the present invention.

[0024] Figure 3 This is a network topology diagram of an autonomous routing method for large-scale Internet of Things provided in an embodiment of the present invention.

[0025] Figure 4 This is a diagram illustrating the service message format of the autonomous routing method for large-scale Internet of Things provided in this embodiment of the invention.

[0026] Figure 5 This is a schematic diagram of routing synchronization for an autonomous routing method for large-scale Internet of Things provided in an embodiment of the present invention. Detailed Implementation

[0027] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0028] Example 1: like Figure 1As shown, this embodiment of the invention discloses an autonomous routing method for large-scale Internet of Things (IoT). Node Identifier: Assign a unique 3-byte identifier to each node in the IoT network, including the subnet number and the node hardware serial number, to ensure the uniqueness and traceability of the identity of all nodes in the network.

[0029] Cluster head election: The entire large-scale Internet of Things network is divided into multiple subnets. Each node in the subnet announces its own identification number to each other. The receiving node dynamically elects the cluster head based on comprehensive signal quality indicators. Through a comprehensive weighted evaluation of multi-dimensional link quality and node power, the cluster head is always in the optimal communication center position in the subnet.

[0030] Intra-cluster route establishment: The cluster head node is used as the subnet center point to run the RPL protocol and construct the DODAG within the subnet to form intra-cluster routes; this mechanism ensures the reliable establishment of uplink paths from any node in the cluster to the cluster head and downlink paths from the cluster head to any node.

[0031] Inter-cluster routing establishment: Cluster head nodes form a backbone network, run the BATMAN protocol, and obtain the metric value of inter-cluster links by exchanging OGM messages to establish the optimal path between cluster heads; this mechanism enables the backbone network to perceive changes in link quality in real time and dynamically select the optimal inter-cluster forwarding path.

[0032] Route synchronization and diffusion: The cluster head node generates a service packet containing the DODAG routing information of its own cluster, and sends the service packet to all other reachable cluster heads in the BATMAN routing table according to the optimal inter-cluster path determined by the BATMAN protocol, thereby realizing the interaction of routing information between clusters and the synchronization of the routing table of the entire network. This directional diffusion mechanism avoids global flooding and significantly reduces network control overhead while ensuring the consistency of routing information.

[0033] External cluster route injection: After receiving the service message, the target cluster head parses the DODAG information, sets the next hop to the BATMAN gateway leading to the source cluster head, and publishes the external route to the local cluster node through RPL DIO messages. This mechanism allows ordinary nodes to transparently access external cluster resources without running the BATMAN protocol, simplifying the protocol stack complexity of the terminal node.

[0034] Dynamic route self-healing: Periodically check the cluster head's operating status. When a cluster head failure is detected, trigger the corresponding subnet to re-elect a cluster head. The new cluster head rebuilds intra-cluster RPL routes and inter-cluster BATMAN routes, completing the full network route recovery. This mechanism gives the network the ability to self-heal from node failures, ensuring that a single cluster head failure will not cause the entire network's service to be interrupted.

[0035] like Figure 2As shown, this invention adopts a hierarchical network architecture, dividing the entire large-scale network into multiple subnets (clusters). Within each cluster, a cluster head is dynamically elected based on comprehensive link quality, ensuring the optimality of the cluster head node. The cluster head runs the RPL protocol to manage intra-cluster routing, forming a stable and efficient access network. Between cluster heads, a backbone network is formed, running the BATMAN protocol. Based on this architecture, this invention achieves deep decoupling of the protocol. The synchronization of intra-cluster routing information does not depend on the convergence process of the BATMAN protocol, but is achieved through independent, controlled forwarding of service packets. This allows the BATMAN protocol to focus on inter-cluster link quality awareness and optimal path calculation, without carrying heavy global routing information. Regarding route synchronization, cluster heads do not engage in global flooding of routing information exchange, but instead perform point-to-point directed diffusion through the optimal path determined by BATMAN, effectively avoiding broadcast storms and isolating global routing oscillations caused by changes in a single cluster head. Simultaneously, the cluster head naturally integrates the learned external routes into the intra-cluster routing through the RPL protocol, allowing ordinary nodes to access network resources without running complex protocols. In addition, to ensure that the contributions of measurement indicators with different dimensions and value ranges can be fairly and accurately reflected in the weighted calculation, this invention normalizes all original indicators used for weighted calculation and maps them to a unified, dimensionless interval of [0,1].

[0036] The following example illustrates the autonomous routing method of the present invention for large-scale Internet of Things: Assume a smart factory IoT system comprising three workshops (subnets), numbered 0x01, 0x02, and 0x03. Each workshop is equipped with dozens of sensor nodes. The network topology in this embodiment is as follows: Figure 3 As shown.

[0037] 1. Initialization and Identification: Each sensor node generates a 2-byte node serial number based on its hardware ID, which is then combined with the subnet number (0x01 / 0x02 / 0x03) of its workshop to form a 3-byte identifier. For example, a node in workshop A might be numbered 0x01 0x11 0x11. Subnet numbers are assigned based on physical distance or usage scenario; nodes within the same physical area or functional domain are assigned the same subnet number.

[0038] The specific parameters of the CRC-16 checksum algorithm used in this step are as follows: the generator polynomial is... + + +1 corresponds to the hexadecimal identifier 0x1021; the initial value of the CRC register is 0x0000; each byte of input data is first reversed bitwise, and then participates in the shift or operation; after all byte operations are completed, the 16-bit register result is reversed bitwise as a whole, and the final 16-bit value is the node hardware serial number.

[0039] 2. Election of Cluster Leader: Within workshop A (subnet 0x01), each node begins broadcasting its own identifier and its RSSI (Signal Strength Indicator), SNR (Signal-to-Noise Ratio), LQI (Link Quality Indicator), and SOC (State of Charge) parameters. After receiving the broadcast from nodes 0x01, 0x11, and 0x12, nodes 0x01, 0x11, and 0x12 calculate a signal quality score based on the current RSSI = -65dBm (historical maximum -50dBm, minimum -100dBm), SNR = 25dB (historical maximum 30dB, minimum 5dB), LQI = 110 (historical maximum 255, minimum 0), and SOC = 80%, using preset weights (α = 0.3, β = 0.3, γ = 0.2, δ = 0.2).

[0040] The RSSI, SNR, LQI, and SOC mentioned above are all benefit-related indicators and must be processed using a normalization function. The formula is as follows:

[0041] Where NR(X) is the normalized value of index X, with a value range of [0,1], and X is the original collected index value. min X is the historical minimum sample value of this indicator. max This is the historical maximum sample value for this indicator.

[0042] NR(RSSI)=(-65-(-100)) / (-50-(-100))=0.7 NR(SNR) = (25-5) / (30-5) = 0.8 NR(LQI) = 110 / 255 = 0.43 Q=α×NR(RSSI)+β×NR(SNR)+γ×NR(LQI)+δ×NR(SOC)=0.3×0.7+0.3×0.8+0.2×0.43+0.2×0.8=0.696 Where NR() is the normalized value of the corresponding index, Q is the node's overall signal quality score, and α, β, γ, and δ are the weighting coefficients of RSSI, SNR, LQI, and SOC, respectively, satisfying α+β+γ+δ=1.

[0043] After a round of evaluation, workshop A elected the nodes 0x01, 0x11, and 0x12 with the highest scores as the cluster head.

[0044] Among them, network interference intensity refers to the background electromagnetic interference power in the node's working channel, and the unit is dBm. The larger the value, the stronger the interference.

[0045] The measurement method is as follows: During the data transmission idle period, the node performs multiple background RSSI samples on the current working channel and takes the average value as the current interference intensity value; or directly reads the channel noise floor value output by the MAC layer Channel Clearance Assessment (CCA) module. When the measured value is higher than -70dBm, it is judged as a strong interference scenario.

[0046] In actual deployments, the aforementioned weighting coefficients α, β, γ, and δ are not fixed but dynamically configured based on network interference intensity and the average remaining power of nodes to adapt to different environments. Specifically: When the network interference intensity is higher than -70dBm (strong interference scenario), anti-interference capability should be given priority, and α=0.3, β=0.3, γ=0.2, δ=0.2 should be set. When the average remaining power of a node is less than 30% (low power scenario), node survival time is taken into priority, and α=0.2, β=0.2, γ=0.2, δ=0.4 are set. When both of the above scenarios are satisfied, set α=0.25, β=0.25, γ=0.2, and δ=0.3.

[0047] The sum of all the above weighting coefficients is always equal to 1, and all are within the interval [0,1].

[0048] 3. Route establishment: Within the cluster: Starting from cluster head 0x01, 0x11, and 0x12, DIO messages are sent every 100ms to build an RPLDODAG with itself as the root. All nodes in workshop A join this DODAG.

[0049] Cluster heads: The cluster heads of workshops A, B, and C are A0, B0, and C0 respectively. They are connected by a wireless link and start running the BATMAN protocol, periodically exchanging OGM messages.

[0050] Link stability (stab) quantifies the sustained reliable transmission capability of wireless links between cluster heads, ranging from 0% to 100%. A higher value indicates a lower risk of packet loss and link interruption. It is obtained through OGM message sliding window statistics: cluster heads broadcast OGM messages with incrementing sequence numbers at fixed intervals. The receiving end counts the actual number of messages received and the theoretically expected number of messages received within the sliding window, and calculates the result using the following formula:

[0051] Where Stab represents link stability; N recv N represents the actual number of OGM messages received within the sliding window. total This represents the theoretical total number of OGM messages that should be received within the sliding window.

[0052] For inter-cluster path metrics, this invention uses link delay as a cost-based indicator and link stability (stab) and cluster head remaining power (SOC) as benefit-based indicators, specifically: Total delay of multi-hop path:

[0053] Among them, delay path The total delay for the entire multi-hop path; delay i Let be the single-link delay of the i-th hop; n is the total number of hops in the path.

[0054] Multi-hop path stability:

[0055] Among them, stab path For the overall stability of multi-hop paths; stab i Let represent the stability of the i-th hop single link.

[0056] Minimum cluster head power for multi-hop paths:

[0057] Among them, SOC path The minimum remaining charge of all cluster heads along the path; SOC i This represents the remaining battery power of the i-th hop cluster head.

[0058] Cost-related metrics, specifically link latency normalization:

[0059] Where NR (delay) is the normalized value of the path delay; delay path Total delay for multi-hop paths; delay min delay max These represent the minimum and maximum network link latency, respectively.

[0060] Path stability normalization:

[0061] Where NR(stab) is the path stability normalized value; stab path For multi-hop path stability; stab min stab max These represent the minimum and maximum values ​​for overall network link stability, respectively.

[0062] Path minimum charge normalization:

[0063] Among them, NR(SOC) pathThe value is the normalized minimum energy level for the path; SOC path Minimum cluster head charge for the path; SOC min SOC max These represent the minimum and maximum charge levels of the cluster head, respectively.

[0064] Inter-cluster path comprehensive metric:

[0065] Where NR() is the normalized value of the corresponding index; Metric is the comprehensive metric value of inter-cluster paths; W1, W2, and W3 are the delay, stability, and power weight coefficients, respectively, satisfying W1+W2+W3=1.

[0066] Assume the link latency from A0 to B0 is 10ms, stability is 95%, and B0's battery level is 70%; the link latency from A0 to C0 is 50ms, stability is 90%, and C0's battery level is 50%; the link latency from B0 to C0 is 30ms, and stability is 93%. Calculate link metrics using preset weights (W1=0.4, W2=0.3, W3=0.3).

[0067] The system link latency has historically ranged from a maximum of 100ms to a minimum of 2ms. Link latency is a cost-related metric. Link A0→B0→C0 delay = + =10+30=40 NR( = 1 - [(40 - 2) / (100 - 2)] = 0.612 Link A0→C0 delay =50, NR( ) = 1 - [(50 - 2) / (100 - 2)] = 0.51 The system link stability historically ranged from a maximum of 99% to a minimum of 70%. Link stability is a benefit-oriented indicator. Link A0→B0→C0 stability =MIN( , ) = 93 NR( ) = (93-70) / (99-70) = 0.793; Link A0→C0 stability NR ( = (90-70) / (99-70) = 0.69 Power consumption of link A0→B0→C0 =MIN( , )=0.5 Link A0→C0 power =0.5 After weighted calculation, the quality of the A0→B0→C0 link =(0.612)×0.4+(0.793)×0.3+0.5×0.3=0.633, A0→C0 link quality =(0.51)×0.4+(0.69)×0.3+(0.5)×0.3=0.561, > For the route from A0 to C0, the path A0→B0→C0 is more optimal. Therefore, in the routing table of cluster head A0, the best next hop to reach both B0 and C0 is B0.

[0068] 4. Inter-cluster route propagation: Cluster head A0 generates a service message, the data segment of which contains DODAG routing information for all nodes in workshop A. This service message does not depend on the convergence process of the BATMAN protocol; instead, it is sent as an independent message through B0 according to the BATMAN routing table, achieving directional diffusion and synchronization of routing information between clusters. For the specific format definition of the above service message, please refer to [link to relevant documentation]. Figure 4 Its message header contains a 3-byte source cluster head number, a 2-byte sequence number, and a 1-byte CRC-8 checksum, while the data segment carries intra-cluster DODAG routing information.

[0069] The service message generated in this step contains a 1-byte CRC-8 checksum. The CRC-8 algorithm is used to calculate the entire service message, including the 3-byte source cluster head number in the message header, the 2-byte sequence number, and the intra-cluster DODAG routing data in the data segment, but excluding the checksum itself. The generator polynomial is 0x31. + + 1) The initial value is 0xFF, the input and output are not flipped, and the result XOR value is 0x00; the sending end obtains an 8-bit check value by modulo 2 division and fills it into the check bit field. The receiving end recalculates the same data range with the same parameters and compares the calculation result with the check bit. If they match, the message is determined to be complete and error-free; otherwise, the message is discarded.

[0070] 5. External cluster route injection: Cluster head C0 learns the routing information to subnet 0x01, sets the next hop to the BATMAN gateway leading to source cluster head A0, and distributes it to all nodes within workshop C via RPL DIO messages, naturally integrating external routing into the intra-cluster RPL routing system. Subsequently, if any node in workshop C wants to communicate with a node in workshop A, the data packet will first be routed to cluster head C0, then via BATMAN routing (through cluster head B) to finally reach cluster head A0, and finally be distributed to the target node by cluster head A0. This process achieves full network routing transparency for ordinary nodes, allowing them to access external cluster resources without running complex protocols like BATMAN. The complete routing synchronization process is as follows: Figure 5 As shown in Table 1 and Table 2, the routing table structure formed by each node after the process is completed.

[0071] Table 1: Cluster Head AO Routing Table

[0072] Table 2: Routing Table for Ordinary Node A12

[0073] 6. Fault Recovery: This embodiment uses a tiered periodic adjustment of the RPL interval to detect cluster head failures. The total number of nodes is used as an indicator of network size, and the mapping relationship between the tiers and parameters is as follows: Small-scale (≤50 nodes): RPL interval period 50ms; Medium scale (≤200 nodes): RPL interval period 100ms; Large-scale (>200 nodes): RPL interval period 200ms; Taking a medium-sized network as an example, if cluster head A0 goes offline due to a fault, its neighboring nodes will trigger a re-election after not receiving a DIO for 300ms (3 RPL cycles of 100ms each). After the new cluster head is generated, it will re-establish the intra-cluster RPL DODAG and re-establish BATMAN routes with other cluster heads. At the same time, as the new representative, it will announce the new routing information to other cluster heads through service packets, thereby completing the self-repair of the entire network.

[0074] Thus, through the above steps, the three workshops in the smart factory IoT have achieved hierarchical collaboration, decoupled operation, and efficient synchronization of intra-cluster RPL routing and inter-cluster BATMAN routing, building a scalable, highly reliable, and self-healing large-scale IoT autonomous routing system.

[0075] Example 2: Based on the same inventive concept, embodiments of the present invention also provide an autonomous routing system for large-scale Internet of Things (IoT), the system comprising: The node identification module is used to assign a unique identifier to each node in the Internet of Things network, which includes a subnet number and a node hardware serial number. The cluster head election module is used to divide the entire large-scale Internet of Things network into multiple subnets. Each node in the subnet announces its own identification number to each other, and the receiving node dynamically elects a cluster head based on comprehensive signal quality indicators. The intra-cluster routing establishment module is used to run the RPL protocol with the cluster head node as the subnet center point to construct the DODAG within the subnet and form intra-cluster routes. The inter-cluster routing establishment module is used to form a backbone network between the cluster head nodes, run the BATMAN protocol, obtain the metric value of the inter-cluster link by exchanging OGM messages, and establish the optimal path between cluster heads. The routing synchronization and diffusion module is used by the cluster head node to generate a service packet containing the DODAG routing information of the cluster, and send the service packet to the target cluster head according to the optimal inter-cluster path determined by the BATMAN protocol, so as to achieve global routing synchronization. The external cluster routing injection module is used to parse the DODAG information after the target cluster head receives the service message, set the next hop to the BATMAN gateway to the source cluster head, and publish the external route to the local cluster node through RPL DIO messages. The dynamic routing self-healing module is used to periodically detect the cluster head's operating status. When a cluster head failure is detected, it triggers the corresponding subnet to re-elect a cluster head. The new cluster head rebuilds the intra-cluster RPL routes and inter-cluster BATMAN routes, completing the full network routing recovery.

[0076] Since these systems and the principles they address are similar to the aforementioned autonomous routing method for large-scale Internet of Things, the implementation of this system can be found in the implementation of the aforementioned method, and the repetitions will not be repeated.

[0077] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the systems disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the descriptions are relatively simple; relevant parts can be referred to the method section.

[0078] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in its embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. An autonomous routing method for large-scale Internet of Things (IoT), characterized in that, Includes the following steps: Node Identifier: Each node in the Internet of Things network is assigned a unique identifier, which includes the subnet number and the node hardware serial number; Cluster head election: The entire large-scale Internet of Things network is divided into multiple subnets. Each node in the subnet announces its own identification number to each other. The receiving node dynamically elects a cluster head based on comprehensive signal quality indicators. Intra-cluster routing establishment: The cluster head node is used as the subnet center point to run the RPL protocol, construct the DODAG within the subnet, and form intra-cluster routes; Inter-cluster routing establishment: The cluster head nodes form a backbone network, run the BATMAN protocol, obtain the metric value of inter-cluster links by exchanging OGM messages, and establish the optimal path between cluster heads; Route synchronization and propagation: The cluster head node generates a service packet containing the DODAG routing information of its own cluster, and sends the service packet to all other reachable cluster heads in the BATMAN routing table according to the optimal inter-cluster path determined by the BATMAN protocol, thereby realizing the interaction of inter-cluster routing information and the synchronization of the entire network routing table; External cluster route injection: After receiving the service packet, the target cluster head parses the DODAG information, sets the next hop to the BATMAN gateway to the source cluster head, and publishes the external route to the cluster node through RPL DIO messages; Dynamic route self-healing: Periodically detect the cluster head's operating status. When a cluster head failure is detected, trigger the corresponding subnet to re-elect a cluster head. The new cluster head rebuilds intra-cluster RPL routes and inter-cluster BATMAN routes, completing the full network route recovery.

2. The method as described in claim 1, characterized in that, The method for generating the node hardware serial number includes: extracting the last 16 bits of the node hardware serial number; if the hardware serial number is less than 16 bits, padding with zeros; if it exceeds 16 bits, using the CRC-16 check algorithm to calculate and taking the lower 16 bits as the node hardware serial number.

3. The method as described in claim 1, characterized in that, The subnet numbers are assigned based on physical distance or usage scenario, and nodes within the same physical area or functional domain are assigned the same subnet number.

4. The method as described in claim 1, characterized in that, The receiving node dynamically elects the cluster head based on comprehensive signal quality indicators, including: when there is only one node in the subnet, that node automatically becomes the cluster head; When there are multiple nodes in a subnet, each node in the subnet broadcasts its own node number and simultaneously collects the RSSI signal strength, SNR signal-to-noise ratio, LQI link quality indicator, and SOC node power parameters of adjacent nodes. For benefit-type indicators and cost-type indicators, corresponding normalization formulas are used for normalization processing respectively; The signal quality score is obtained by weighting and summing the normalized parameters using preset weighting coefficients, and the node with the highest signal quality score is selected as the cluster head node of the current subnet.

5. The method as described in claim 1, characterized in that, In the inter-cluster routing establishment step: The cluster head obtains link latency, link stability, and peer cluster head power through OGM messages; For paths spanning multiple hops, the sum of the delays of each hop is the path delay, the minimum stability of each hop is the path stability, and the minimum energy of each hop is the path energy. After normalizing the path delay, path stability, and path power consumption, the paths are then weighted and summed using preset weighting coefficients to calculate the path metric. The path with the optimal metric is then selected as the optimal path between cluster heads.

6. The method as described in claim 1, characterized in that, In the routing synchronization and diffusion step, the service message format includes a message header and a data segment. The message header contains a 3-byte source cluster head number, a 2-byte sequence number, and a 1-byte CRC-8 checksum.

7. The method as described in claim 6, characterized in that, The CRC-8 checksum is calculated using the CRC-8 algorithm on the source cluster head number, sequence number, and intra-cluster DODAG routing data in the data segment, generating a polynomial of x. 8 +x 5 +x 4 +1, the initial value is 0xFF, the input and output are not flipped, and the result XOR value is 0x00; the receiving end recalculates the same data range with the same parameters, compares the calculation result with the check bit, and if they match, the message is determined to be complete and error-free; otherwise, the message is discarded.

8. The method as described in claim 1, characterized in that, In the route synchronization and propagation step, the service packets are sent to all other reachable cluster heads in the BATMAN routing table. Specifically, the IP destination of the service packets is set to the destination IP address in the BATMAN routing table, and then forwarded according to the next hop in the BATMAN routing table.

9. The method as described in claim 1, characterized in that, In the external cluster route injection step, the external route is advertised to the ordinary nodes in the cluster through the DIO message of the RPL protocol, so that the ordinary nodes can obtain the external route to the external cluster. Ordinary nodes add external routes based on the DIO message and use the best next-hop node to reach the cluster head, as determined by the RPL protocol, as the gateway.

10. An autonomous routing system for large-scale Internet of Things (IoT), characterized in that, The system comprises: The node identification module is used to assign a unique identifier to each node in the Internet of Things network, which includes a subnet number and a node hardware serial number. The cluster head election module is used to divide the entire large-scale Internet of Things network into multiple subnets. Each node in the subnet announces its own identification number to each other, and the receiving node dynamically elects a cluster head based on a comprehensive signal quality index. The intra-cluster routing establishment module is used to run the RPL protocol with the cluster head node as the subnet center point to construct the DODAG within the subnet and form intra-cluster routes. The inter-cluster routing establishment module is used to form a backbone network between the cluster head nodes, run the BATMAN protocol, obtain the metric value of the inter-cluster link by exchanging OGM messages, and establish the optimal path between cluster heads. The routing synchronization and diffusion module is used by the cluster head node to generate a service packet containing the DODAG routing information of the cluster, and send the service packet to the target cluster head according to the optimal inter-cluster path determined by the BATMAN protocol, so as to achieve global routing synchronization. The external cluster routing injection module is used to parse the DODAG information after the target cluster head receives the service message, set the next hop to the BATMAN gateway to the source cluster head, and publish the external route to the local cluster node through RPL DIO messages. The dynamic routing self-healing module is used to periodically detect the cluster head's operating status. When a cluster head failure is detected, it triggers the corresponding subnet to re-elect a cluster head. The new cluster head rebuilds the intra-cluster RPL routes and inter-cluster BATMAN routes, completing the full network routing recovery.

Images