A method, system, and device for selecting bidirectional relay nodes for multiple robots.

By selecting the optimal relay node using a multi-factor weighted approach, the problems of missing hop count processing logic and single evaluation dimension in relay node selection in multi-robot systems are solved, enabling bidirectional closed-loop self-healing communication and improving communication efficiency and recovery capability after disconnection.

CN123075595APending Publication Date: 2026-11-10福建汉特云智能科技有限公司
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611219033.4
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-08-12
Publication Date
2026-11-10

AI Technical Summary

Technical Problem

In existing multi-robot collaborative operation systems, the relay node selection mechanism suffers from problems such as missing hop count processing logic, single evaluation dimension, and fragmented uplink and downlink decision-making. These issues lead to increased communication latency, higher packet loss rate, uneven cluster load, and low overall communication efficiency, thus restricting the reliability and efficiency of connection recovery.

Method used

A multi-factor weighted selection method is adopted to select the optimal relay node. A hierarchical strategy is used to screen neighbor nodes that meet the preset hop count condition. The method comprehensively considers indicators such as WAN communication quality, load, remaining power, distance and self-organizing network signal quality to construct a closed-loop self-healing communication mechanism with uplink and downlink linkage.

Benefits of technology

Effectively avoids invalid forwarding, improves the availability of communication links, ensures the shortest link and lowest latency, balances cluster load, shortens the recovery time after disconnection, and improves communication success rate and reliability.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123075595A_ABST
    Figure CN123075595A_ABST
Patent Text Reader

Abstract

This invention discloses a method, system, and apparatus for selecting bidirectional relay nodes for multiple robots, belonging to the field of robot communication technology. The method includes: a target robot monitoring its own wide area network communication quality; if the quality falls below a trigger condition, a hierarchical strategy is used to select neighboring nodes that meet the hop count requirement to form an uplink candidate node set; relay performance indicators of each candidate node are obtained, and the optimal uplink relay node is selected through normalized weighted scoring, and uplink data is sent to the platform via it and simultaneously identified; when the platform detects a robot disconnection, it prioritizes acquiring the already synchronized optimal uplink relay node, and if the reuse condition is met, it directly issues commands through it; otherwise, the optimal downlink relay node is selected based on the same hierarchical strategy and weighted scoring for command issuance. This invention achieves globally optimal selection of relay nodes through bidirectional linkage between uplink and downlink, hierarchical hop count constraints, and multi-factor comprehensive scoring, ensuring bidirectional link topology consistency and significantly reducing disconnection recovery latency.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This application relates to the field of multi-robot communication technology, specifically to a method, system, and apparatus for selecting bidirectional relay nodes for multi-robot communication. Background Technology

[0002] In a multi-robot collaborative operation system, when a robot (source node) needs to forward data to the cloud control platform via a short-range ad hoc network (such as WiFi Mesh, Bluetooth Mesh, etc.) through neighboring robots due to poor wide area network signal, there are usually multiple available neighboring robots. In the existing technology, the source node often adopts the following simple strategies to select a relay: (1) Random selection: randomly select a robot from the neighbor list as a relay. (2) Strongest signal priority: select the neighboring device with the highest ad hoc network signal strength (RSSI). (3) Broadcast to all neighbors: broadcast the data to all neighboring devices, and let the neighboring devices decide whether to forward it themselves, and then the cloud control platform performs deduplication.

[0003] However, the existing relay node selection mechanisms generally suffer from the following drawbacks: First, the hop count processing logic is missing, making link quality and latency uncontrollable. Most solutions use a mixed candidate and unified sorting method for nodes with different hop counts, without setting a hierarchical screening mechanism that prioritizes low-hop-count nodes or clearly defining a maximum hop count limit. This leads to the easy selection of high-hop-count nodes as relays in practical applications, resulting not only in an increase in forwarding nodes and packet loss rate but also a significant increase in communication latency, posing a serious threat to the timeliness of critical services such as emergency stop commands and path changes. Second, the evaluation dimension is singular, and relay selection lacks global optimality. Most existing technologies rely solely on a single indicator (such as RSSI or distance) for relay selection or directly use random forwarding, without comprehensively considering key factors such as node load, remaining power, and the WAN communication quality of the relay node itself. This makes the selected relay nodes prone to being overloaded or running out of power and about to go offline, leading to uneven cluster load, low overall communication efficiency, and even invalid forwarding situations where the self-organizing network is connected but data cannot be uploaded. Third, uplink and downlink decisions are disconnected, and bidirectional links lack consistency guarantees. Existing solutions typically focus only on optimizing unidirectional communication paths, lacking bidirectional collaborative designs for how to use self-organizing networks to relay data when uplink fails, and how the platform can proactively recover lost robots via self-organizing networks. The uplink relay selection result is not synchronized to the cloud control platform, requiring independent re-decision during downlink recovery. This leads to inconsistent bidirectional link topologies, increased recovery latency, and an inability to form a complete, self-consistent closed-loop self-healing system, thus limiting the reliability and efficiency of loss-of-connection recovery.

[0004] Therefore, there is an urgent need for a relay method that can achieve closed-loop self-healing and recovery from disconnection. Summary of the Invention

[0005] In view of the above problems, this application provides a method, system and device for selecting bidirectional relay nodes for multiple robots. When the wide-area signal of multiple robots is poor, the optimal node is selected by weighting multiple factors using adjacent robots as relays, thereby realizing bidirectional closed-loop self-healing communication in dual-mode communication.

[0006] To achieve the above objectives, the inventors provide a multi-robot bidirectional relay node selection method, applicable to a system composed of a cloud control platform and multiple mobile robots, each equipped with a wide area network communication module and a short-range wireless self-organizing network module, comprising the following steps:

[0007] The target robot monitors its own wide area network communication quality. When the communication quality is detected to be lower than the preset triggering condition, it obtains neighbor node information through a short-distance self-organizing network and selects neighbor nodes that meet the preset hop count condition based on a hierarchical strategy to form an uplink candidate relay node set.

[0008] Obtain the relay performance index of each candidate node in the uplink candidate relay node set, normalize each relay performance index, calculate the comprehensive score by weighting according to the preset weight, and select the candidate node with the highest score as the optimal uplink relay node.

[0009] The target robot sends uplink data to the cloud control platform through the optimal uplink relay node and synchronizes the identification information of the optimal uplink relay node to the cloud control platform.

[0010] When the cloud control platform detects that a target robot has lost contact, it obtains the identifier of the optimal uplink relay node synchronized with the target robot and determines whether the corresponding relay node meets the preset reuse conditions. If it does, it sends instructions to the target robot through the corresponding relay node.

[0011] If the preset reuse conditions are not met, the cloud control platform obtains the neighbor node information of the target robot, and filters the neighbor nodes that meet the preset hop count conditions based on the hierarchical strategy to form a set of downlink candidate relay nodes; the cloud control platform obtains the relay performance index of each candidate node in the downlink candidate relay node set; normalizes each relay performance index and calculates a comprehensive score by weighting according to preset weights, selects the candidate node with the highest score as the optimal downlink relay node, and issues instructions to the target robot through the optimal downlink relay node.

[0012] In some embodiments, the preset hop count condition is: the data forwarding hop count between the neighbor node and the target robot is 1 hop or 2 hops. In these embodiments, the relay hop count is strictly limited to 1 hop or 2 hops, which effectively controls transmission delay and forwarding overhead, and avoids the increased packet loss rate and uncontrollable path problems caused by multiple hops.

[0013] In some embodiments, relay performance metrics include the WAN communication quality, load, hop count, remaining battery power, distance between the candidate node and the robot, and ad hoc network signal quality for each candidate node. These embodiments achieve multi-dimensional global evaluation by comprehensively incorporating six metrics: WAN communication quality, load, hop count, remaining battery power, distance, and ad hoc network signal quality. This avoids ineffective forwarding or node overload issues caused by relying on a single metric, thus improving the comprehensiveness and accuracy of relay selection.

[0014] In some embodiments, the normalization of various relay performance indicators and the calculation of a comprehensive score based on preset weights include:

[0015] The calculated value of the hop factor H is 1 for one hop, 0.5 for two hops, and 0 for three hops or more.

[0016] Subtracting the node's current combined CPU and bandwidth utilization from 1 yields the load factor L.

[0017] Get the current remaining battery percentage of the node as the battery factor P;

[0018] The distance factor D is calculated based on the ratio of the actual distance between nodes to the maximum communication distance of the ad hoc network, and the value is larger the closer the distance.

[0019] The value of the self-organizing network signal quality factor S is obtained by linearly mapping the received signal strength indication value to the interval [0, 1].

[0020] The WAN communication quality factor of the relay node is calculated by linearly mapping the WAN signal strength, preset full-strength signal strength, and minimum acceptable signal threshold of the relay node to the interval [0, 1]. .

[0021] The overall score is calculated using a weighted average based on preset weights, and the corresponding calculation formula is as follows:

[0022] ,

[0023] in, Here, is the WAN communication quality factor for relay nodes, D is the distance factor, S is the ad hoc network signal quality factor, H is the hop count factor, L is the load factor, and P is the power factor. , , , , , These are the weights of the corresponding items, and .

[0024] These embodiments achieve the scientific and computable nature of the comprehensive scoring mechanism by uniformly mapping various indicators to a comparable quantification range and highlighting the core position of wide area network upload capability and distance through reasonable weight allocation, while taking into account link quality, latency control and node endurance.

[0025] In some embodiments, the method further includes: the cloud control platform or the target robot dynamically adjusting preset weights according to the current task scenario; the task scenario includes at least preset power priority mode, speed priority mode, and load balancing mode. This design in these embodiments supports dynamic weight adjustment based on different task scenarios such as power priority, speed priority, or load balancing, enhancing the method's adaptability to various operating conditions such as high-speed movement and static clustering, and improving flexibility.

[0026] In some embodiments, a hierarchical strategy is employed to select neighbor nodes that meet a preset hop count condition, forming an uplink candidate relay node set. This strategy prioritizes selecting one-hop neighbor nodes, and when no one-hop neighbor node is available, selects two-hop neighbor nodes and removes nodes with three or more hops. Alternatively, a hierarchical strategy is employed to select neighbor nodes that meet a preset hop count condition, forming a downlink candidate relay node set. This approach extends coverage while ensuring low latency for relay nodes, achieving a balance between communication performance and coverage capability.

[0027] In some embodiments, the preset reuse conditions include the uplink optimal relay node being online, the uplink optimal relay node having normal WAN communication, and the number of hops to the target robot not exceeding two. In these embodiments, by setting a triple reuse threshold of online status, normal WAN communication, and no more than two hops, it is ensured that the uplink relay node prioritized for reuse still has effective forwarding capabilities, guaranteeing the success rate of downlink recovery.

[0028] Furthermore, the present invention also provides a multi-robot bidirectional relay node selection system, which applies the aforementioned multi-robot bidirectional relay node selection method. The system includes a cloud control platform and at least two robots that are communicatively connected to the cloud control platform.

[0029] The cloud control platform includes:

[0030] A connection status monitoring unit is used to detect robot disconnection via a heartbeat mechanism;

[0031] The robot disconnection query unit is used to broadcast a query command to the online robot after a disconnection is detected;

[0032] The downlink recovery unit is used to, when a target robot is detected to be out of contact, obtain the identifier of the uplink optimal relay node synchronized with the target robot, and determine whether the corresponding relay node meets the preset reuse conditions. If it does, the unit sends instructions to the target robot through the corresponding relay node. If it does not, the unit obtains the neighbor node information of the target robot, and selects neighbor nodes that meet the preset hop count conditions based on a hierarchical strategy to form a downlink candidate relay node set. The unit obtains the relay performance index of each candidate node in the downlink candidate relay node set, normalizes each relay performance index, calculates a comprehensive score by weighting according to preset weights, selects the candidate node with the highest score as the downlink optimal relay node, and sends instructions to the target robot through the downlink optimal relay node.

[0033] Each robot includes:

[0034] Wide area network communication module, used for direct communication with the cloud control platform;

[0035] The short-range wireless ad hoc network module is used to establish multi-hop ad hoc networks with other robots and maintain a neighbor table for its own short-range ad hoc network.

[0036] A signal quality monitoring unit is used to monitor the signal quality of the wide area network communication module in real time.

[0037] The communication decision unit is used to obtain neighbor node information through the short-range ad hoc network communication module when its own wide area network communication quality is lower than the preset triggering conditions. Based on the hierarchical strategy, it selects neighbor nodes that meet the preset hop count conditions to form an uplink candidate relay node set; obtains the relay performance index of each candidate node in the uplink candidate relay node set, normalizes each relay performance index and calculates a comprehensive score according to the preset weights, selects the candidate node with the highest score as the optimal uplink relay node; sends uplink data to the cloud control platform through the optimal uplink relay node, and synchronizes the identification information of the optimal uplink relay node to the cloud control platform.

[0038] The relay performance indicators include the WAN communication quality, load, hop count, remaining power of each candidate node, as well as the distance between the candidate node and the robot, and the signal quality of the ad hoc network.

[0039] In some embodiments, the instruction downlink recovery unit forms a downlink candidate relay node set according to a hierarchical strategy that prioritizes one-hop neighbor nodes, selects two-hop neighbor nodes when there are no one-hop neighbor nodes, and removes nodes with three or more hops.

[0040] In some embodiments, the communication decision unit forms an uplink candidate relay node set according to a hierarchical strategy that prioritizes one-hop neighbor nodes, selects two-hop neighbor nodes when there are no one-hop neighbor nodes, and eliminates nodes with three or more hops.

[0041] Furthermore, the present invention also provides a multi-robot bidirectional relay node selection device, including a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the program, it implements the steps of the multi-robot bidirectional relay node selection method.

[0042] The present invention adopts the above technical solution, which differs from the prior art. The above technical solution has the following beneficial effects:

[0043] (1) The present invention integrates the signal quality of the wide area network of the candidate robot relay and prioritizes the selection of nodes with reliable upload capability, effectively avoiding the problem of invalid forwarding when the self-organizing network is connected but the data cannot be uploaded, and greatly improving the overall availability of the communication link.

[0044] (2) The present invention adopts a hierarchical hop count constraint strategy of "priority 1 hop, secondary 2 hops, maximum 2 hops". Prioritizing 1 hop node can ensure the shortest link and the lowest latency, meeting the timeliness requirements of key services such as emergency stop instructions and path changes; when 1 hop node is unavailable, 2 hop node is allowed as an alternative to expand the coverage area; 3 hop and above nodes are strictly prohibited to control the transmission delay within an acceptable range, avoid the lag risk and increased packet loss rate caused by long paths, and ensure the security and controllability of task execution.

[0045] (3) This invention is the first to incorporate multiple dimensions of indicators such as self-organizing network signal quality, relay node WAN signal quality, distance, load, hop count, and remaining power into the evaluation model. It achieves comprehensive evaluation through normalized weighted summation. On the basis of taking into account upload capability, link quality, node load and remaining power, it achieves the global perspective of end-to-end link to select the optimal relay node, effectively balance the cluster load, improve the overall resource utilization efficiency, and significantly improve the communication success rate and reliability.

[0046] (4) This invention constructs a closed-loop mechanism for bidirectional linkage between uplink and downlink: the optimal relay ID selected by the uplink robot is synchronized to the platform in real time, and the platform prioritizes reusing this relay when recovering a lost robot. Through the linkage and reuse of uplink and downlink relay decisions, the consistency of bidirectional link topology is ensured, effectively solving the link breakage problem. The platform does not need to perform complex routing calculations again, significantly improving the recovery capability under wide area network failure scenarios and greatly shortening the recovery time under loss scenarios.

[0047] The above description of the invention is merely an overview of the technical solution of this application. In order to enable those skilled in the art to better understand the technical solution of this application and to implement it based on the description and drawings, and to make the above-mentioned objectives and other objectives, features and advantages of this application easier to understand, the following description is provided in conjunction with the specific embodiments and drawings of this application. Attached Figure Description

[0048] The accompanying drawings are only used to illustrate the principles, implementation methods, applications, features, and effects of specific embodiments of the present invention and other related contents, and should not be considered as limitations on this application.

[0049] In the accompanying drawings of the instruction manual:

[0050] Figure 1 This is a schematic diagram of the structure of a multi-robot bidirectional relay node selection system according to the present invention;

[0051] Figure 2 This is a schematic diagram illustrating the working principle of a multi-robot bidirectional relay node selection method according to the present invention.

[0052] Figure 3 This is a flowchart of the uplink optimal relay selection process in a specific implementation.

[0053] Figure 4 This is a flowchart of the downlink optimal relay selection process in a specific implementation. Detailed Implementation

[0054] To illustrate the possible application scenarios, technical principles, implementable specific solutions, and achievable objectives and effects of this application in detail, the following description, in conjunction with the listed specific embodiments and accompanying drawings, provides a detailed explanation. The embodiments described herein are merely illustrative of the technical solutions of this application and are therefore intended to limit the scope of protection of this application.

[0055] Please see Figure 1 or Figure 2 This embodiment specifically discloses a multi-robot bidirectional relay node selection method, applied to a system composed of a cloud control platform and multiple mobile robots. Each mobile robot is equipped with a wide area network communication module and a short-range wireless self-organizing network module, including the following steps:

[0056] Step 1: The target robot monitors its own wide area network communication quality. When the communication quality is detected to be lower than the preset triggering condition, it obtains neighbor node information through a short-distance ad hoc network and selects neighbor nodes that meet the preset hop count condition based on a hierarchical strategy to form an uplink candidate relay node set.

[0057] Furthermore, it should be noted that the target robot's signal quality factor in the wide area network is... (Triggering factor) characterizes the robot's own WAN link quality; as a feasible implementation, WAN signal strength (RSSI) is used as a parameter characterizing communication quality, and the corresponding normalized expression is: ;in, For full signal strength, The lowest acceptable signal threshold; the on-board WAN signal quality factor. ∈[0,1]. As a feasible implementation, the preset trigger condition may be selected as: when the local wide area network signal factor <W is triggered (preset threshold, such as 0.5) or no platform response is received for N consecutive heartbeat cycles, the relay selection algorithm is started; otherwise, the wide area network direct connection is used directly.

[0058] In some embodiments, such as Figure 3 or Figure 4 shown, the preset hop count condition is: the number of data forwarding hops between the neighbor node and the target robot is 1 hop or 2 hops. In these embodiments, the number of relay hops is strictly limited to 1 hop or 2 hops, which effectively controls transmission delay and forwarding overhead, and avoids the problems of increased packet loss rate and uncontrollable path caused by multi-hop transmission.

[0059] Specifically, in some embodiments, neighbor nodes that meet the preset hop count condition are screened according to a hierarchical strategy that preferentially screens one-hop neighbor nodes, screens two-hop neighbor nodes when there are no one-hop neighbor nodes, and eliminates nodes with three or more hops, to form an uplink candidate relay node set; or, neighbor nodes that meet the preset hop count condition are screened according to a hierarchical strategy that preferentially selects one-hop neighbor nodes, selects two-hop neighbor nodes when there are no one-hop neighbor nodes, and eliminates nodes with three or more hops, to form a downlink candidate relay node set. While ensuring low latency of relay nodes, the coverage range is expanded, achieving a balance between communication performance and coverage capability.

[0060] As Figure 3 shown, step 2: obtain the relay performance indicator of each candidate node in the uplink candidate relay node set, perform normalization processing on each relay performance indicator, calculate a comprehensive score by weighting according to preset weights, and select the candidate node with the highest score as the uplink optimal relay node;

[0061] In some embodiments, the above relay performance indicators include the wide area network communication quality, load, hop count, remaining power of each candidate node, as well as the distance between the candidate node and the robot, and the ad hoc network signal quality. As a feasible implementation, each robot carries real-time status information such as its own wide area network quality, load rate, and remaining power in the ad hoc network heartbeat beacon. As another feasible technical solution, a "status query request" is sent to the candidate nodes once before relay selection to obtain the latest parameters.

[0062] In these embodiments, by comprehensively incorporating the six indicators of wide area network communication quality, load, hop count, remaining power, distance, and ad hoc network signal quality, a multi-dimensional global evaluation is realized, the problems of invalid forwarding or node overload caused by relying on a single indicator are avoided, and the comprehensiveness and accuracy of relay selection are improved.

[0063] In some embodiments, the implementation of performing normalization processing on each relay performance indicator and calculating the comprehensive score by weighting according to preset weights includes:

[0064] The hop count factor H is calculated, which represents the number of hops from the candidate node to the target robot. The value of H is 1 if the candidate node has one hop to the target robot, 0.5 if it has two hops, and 0 if it has three or more hops. The corresponding normalized expression is: 1 hop, H=1; 2 hops, H=0.5; 3 or more hops, H=0, which means that the corresponding candidate relay node is directly eliminated.

[0065] The load factor L is obtained by subtracting the current combined CPU and bandwidth utilization of the node from 1. Specifically, the load factor L represents the CPU and bandwidth utilization of the candidate relay node, and the corresponding normalized expression is: L = 1 - LoadRate; where LoadRate ∈ [0,1] is the current combined CPU or bandwidth utilization of the node; the load factor L ∈ [0,1], and the larger the value of the load factor L, the lower the node load.

[0066] Obtain the current remaining battery percentage of the node as the battery factor P; specifically, the battery factor P: represents the remaining battery percentage of the candidate node, and the corresponding normalized expression is: Where Battery%∈[0,100] is the current remaining percentage of the node's battery; the battery factor P∈[0,1], the larger the value of the battery factor P, the more sufficient the node's battery.

[0067] The distance factor D is calculated based on the ratio of the actual distance between nodes to the maximum communication distance of the ad hoc network. Specifically, the distance factor D represents the physical distance between the candidate node and the target robot, and the corresponding normalized expression is: Where d is the current distance between the candidate node and the target robot; The maximum communication distance of the ad hoc network module is represented by the distance factor D ∈ [0,1]. A larger distance factor D indicates a closer distance. In some implementations, the current distance d between the candidate node and the target robot is obtained based on existing conventional ultra-wideband (UWB) module ranging, or by estimating the distance using a path loss model based on the ad hoc network signal strength (RSSI), or by ranging based on time-of-flight (ToF).

[0068] The value of the ad hoc network signal quality factor S is obtained by linearly mapping the received signal strength indication value to the interval [0, 1]. Specifically, the ad hoc network signal quality factor S represents the ad hoc network signal strength (RSSI) between the candidate node and the target robot, and the corresponding normalized expression is: Where RSSI represents the current signal strength of the ad hoc network; This represents the strongest self-organizing network signal strength value. The minimum acceptable threshold for self-organizing network signal strength; the self-organizing network signal quality factor S∈[0,1], the larger the value of the self-organizing network signal quality factor S, the better the quality of the self-organizing network link.

[0069] The WAN communication quality factor of the relay node is calculated by linearly mapping the WAN signal strength, preset full-strength signal strength, and minimum acceptable signal threshold of the relay node to the interval [0, 1]. Specifically, the relay node's WAN signal quality factor The evaluation factor characterizes the WAN link quality of the candidate relay node, and its corresponding normalized expression is: ;in, For full signal strength, The lowest acceptable signal threshold; relay node WAN signal quality factor ∈[0,1], The larger the value, the stronger the node's ability to upload data to the platform.

[0070] As a feasible implementation method, the signal quality of the relay robot's wide area network is evaluated according to preset weights. The comprehensive score is obtained by weighting the relay robot distance D, the self-organizing network signal quality factor S, the hop count factor H, the load factor L, and the power factor P. The corresponding calculation formula is as follows:

[0071] ,

[0072] in, Here, is the WAN communication quality factor for relay nodes, D is the distance factor, S is the ad hoc network signal quality factor, H is the hop count factor, L is the load factor, and P is the power factor. , , , , , These are the weights of the corresponding items, and .

[0073] These embodiments achieve the scientific and computable nature of the comprehensive scoring mechanism by uniformly mapping various indicators to a comparable quantification range and highlighting the core position of wide area network upload capability and distance through reasonable weight allocation, while taking into account link quality, latency control and node endurance.

[0074] As a default and feasible configuration, the weights of the above indicators are as follows: The formula for calculating the overall score is as follows: D: 25%, S: 20%, H: 15%, L: 10%, P: 5%.

[0075] .

[0076] It should be noted that the signal quality of the relay robot's wide area network... The weight of distance D for the relay robot is 25%, which is beneficial for ensuring core data upload capabilities and minimizing the link; specifically, the above weights ( The default configuration (D: 25%, S: 20%, H: 15%, L: 10%, P: 5%) is suitable for balanced modes in typical scenarios. The system supports dynamic adjustment of weight allocation through configuration files, cloud control platform commands, or local robot policies. In practical applications, weight allocation can be dynamically adjusted according to task requirements, including but not limited to the following methods:

[0077] Scenario preset modes: The system presets "Power Priority Mode" (increases P weight), "Speed ​​Priority Mode" (increases S and H weight), "Load Balancing Mode" (increases L weight), etc., which are automatically switched by the cloud control platform according to the global task or by the robot according to its own status.

[0078] Step 3: The target robot sends the uplink data to the cloud control platform through the uplink optimal relay node, and synchronizes the identification information of the uplink optimal relay node to the cloud control platform.

[0079] like Figure 4 As shown, in step 4, when the cloud control platform detects that a target robot has lost contact, it obtains the identifier of the uplink optimal relay node synchronized by the target robot and determines whether the corresponding relay node meets the preset reuse conditions; if it does, it sends instructions to the target robot through the corresponding relay node.

[0080] If the preset reuse conditions are not met, the cloud control platform obtains the neighbor node information of the target robot, and filters neighbor nodes that meet the preset hop count conditions based on a hierarchical strategy to form a downlink candidate relay node set. The cloud control platform obtains the relay performance index of each candidate node in the downlink candidate relay node set; it normalizes each relay performance index and calculates a comprehensive score by weighting it according to preset weights, and selects the candidate node with the highest score as the optimal downlink relay node, and issues instructions to the target robot through the optimal downlink relay node. It should be noted that, as an optional implementation, the cloud control platform maintains a global topology table and continuously reports its own neighbor table information to the cloud control platform when each robot is online, so that the cloud control platform can infer the possible neighbor nodes of the target robot; as another feasible implementation, the cloud control platform broadcasts a "neighbor query instruction" to online robots near the last reported location of the target robot, requiring them to scan and report the neighbor information of the target robot through the ad hoc network in order to obtain the neighbor node information of the target robot.

[0081] In some embodiments, the preset reuse conditions include the uplink optimal relay node being online, the uplink optimal relay node having normal WAN communication, and the number of hops to the target robot not exceeding two. In these embodiments, by setting a triple reuse threshold of online status, normal WAN communication, and no more than two hops, it is ensured that the uplink relay node prioritized for reuse still has effective forwarding capabilities, guaranteeing the success rate of downlink recovery.

[0082] As a further extension of the technical solution of this invention, in some embodiments, the method further includes: the cloud control platform or the target robot dynamically adjusting preset weights according to the current task scenario; the task scenario includes at least preset power priority mode, speed priority mode, and load balancing mode. This design in these embodiments supports dynamic adjustment of weights according to different task scenarios such as power priority, speed priority, or load balancing, enhancing the method's adaptability to various working conditions such as high-speed movement and static clustering, and improving flexibility.

[0083] Furthermore, the present invention also provides a multi-robot bidirectional relay node selection system, which applies the aforementioned multi-robot bidirectional relay node selection method. The system includes a cloud control platform and at least two robots that are communicatively connected to the cloud control platform.

[0084] The cloud control platform includes:

[0085] A connection status monitoring unit is used to detect robot disconnection via a heartbeat mechanism;

[0086] The robot disconnection query unit is used to broadcast a query command to the online robot after a disconnection is detected;

[0087] The downlink recovery unit is used to, when a target robot is detected to be out of contact, obtain the identifier of the uplink optimal relay node synchronized with the target robot, and determine whether the corresponding relay node meets the preset reuse conditions. If it does, the unit sends instructions to the target robot through the corresponding relay node. If it does not, the unit obtains the neighbor node information of the target robot, and selects neighbor nodes that meet the preset hop count conditions based on a hierarchical strategy to form a downlink candidate relay node set. The unit obtains the relay performance index of each candidate node in the downlink candidate relay node set, normalizes each relay performance index, calculates a comprehensive score by weighting according to preset weights, selects the candidate node with the highest score as the downlink optimal relay node, and sends instructions to the target robot through the downlink optimal relay node.

[0088] Each robot includes:

[0089] A wide area network communication module is used to communicate directly with the cloud control platform; the wide area network communication includes a 4G / 5G / WiFi network that can be directly connected to the cloud platform.

[0090] The short-range wireless ad hoc network module is used to establish multi-hop ad hoc networks with other robots and maintain a neighbor table for its own short-range ad hoc network. Specifically, the short-range wireless ad hoc network module supports WiFi Mesh or Bluetooth Mesh to enable multi-hop networking among multiple robots.

[0091] A signal quality monitoring unit is used to monitor the signal quality of the wide area network communication module in real time.

[0092] The communication decision unit is used to obtain neighbor node information through the short-range ad hoc network communication module when its own wide area network communication quality is lower than the preset triggering conditions. Based on the hierarchical strategy, it selects neighbor nodes that meet the preset hop count conditions to form an uplink candidate relay node set; obtains the relay performance index of each candidate node in the uplink candidate relay node set, normalizes each relay performance index and calculates a comprehensive score according to the preset weights, selects the candidate node with the highest score as the optimal uplink relay node; sends uplink data to the cloud control platform through the optimal uplink relay node, and synchronizes the identification information of the optimal uplink relay node to the cloud control platform.

[0093] The relay performance metrics include the WAN communication quality, load, hop count, remaining battery power, distance between the candidate node and the robot, and ad hoc network signal quality for each candidate node. As a feasible implementation, each robot carries its real-time status information, such as WAN quality, load rate, and remaining battery power, in its ad hoc network heartbeat beacon. Alternatively, a "status query request" is sent to the candidate node before relay selection to obtain the latest parameters.

[0094] Specifically, in some implementations, each robot further includes a robot main control module. The robot main control module includes a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, it implements the corresponding functional units of this application. Specifically: the signal quality monitoring unit reads the RSSI value reported by the wide area network communication module through the processor to achieve signal quality monitoring; the communication decision unit executes the relay node selection algorithm program through the processor, calls the short-range wireless ad hoc network module to obtain neighbor table information, and calculates a comprehensive score to achieve relay decision function; the wide area network communication module and the short-range wireless ad hoc network module are respectively connected to the robot main control module through a bus or interface, and complete data transmission and reception under the unified scheduling of the main control module. Furthermore, as a feasible technical supplement, the cloud control platform includes a platform server. The platform server includes a memory, a processor, and a computer program stored in the memory and running on the processor. The connection status monitoring unit, the robot disconnection query unit, and the command downlink recovery unit are all implemented by the processor executing the computer program.

[0095] In some embodiments, the instruction downlink recovery unit forms a downlink candidate relay node set according to a hierarchical strategy that prioritizes one-hop neighbor nodes, selects two-hop neighbor nodes when there are no one-hop neighbor nodes, and removes nodes with three or more hops.

[0096] In some embodiments, the communication decision unit forms an uplink candidate relay node set according to a hierarchical strategy that prioritizes one-hop neighbor nodes, selects two-hop neighbor nodes when there are no one-hop neighbor nodes, and eliminates nodes with three or more hops.

[0097] The core algorithm of this invention involves only neighbor table maintenance, simple arithmetic operations, and sorting, resulting in low computational complexity. It requires no global topology information and does not rely on GPS or maps. With a maximum hop count limited to 2 and a significantly fewer control message count than traditional Ad Hoc routing protocols, it is ideally suited for battery-powered mobile robot platforms with limited computing resources, achieving a balance between high performance and low resource consumption.

[0098] The triggering logic of the aforementioned multi-robot bidirectional relay node selection system is as follows: relay selection is initiated only when the signal strength of the target robot's wide area network (WAN) is poor; priority is given to selecting 1-hop nodes among the target robot's neighbor nodes; if there are no 1-hop neighbor nodes, then 2-hop neighbor nodes are selected as candidate relay nodes; the maximum number of hops for a relay node is 2 hops, and any neighbor node with ≥3 hops is directly eliminated and not considered as a candidate relay node; for each candidate relay node, its performance is evaluated using six key indicators: WAN signal strength, self-organizing network signal strength, distance, load, hop count, and power consumption. The results are obtained through weighted fusion calculation. The optimal relay node is selected based on the comprehensive score of each candidate relay node. When the target robot's communication quality on the wide area network is poor and it cannot directly connect to the cloud control platform, the target robot communicates with the cloud control platform based on the determined optimal uplink relay node and synchronizes the ID of the uplink relay node to the cloud control platform. When the cloud control platform communicates with the target robot downlink, it prioritizes reusing the synchronized relay node. If it is available, the consistency of the uplink and downlink topology is maintained. If it is unavailable, an optimal downlink relay node is re-determined from the target robot's neighboring nodes using the same method and steps as the uplink relay node for communication.

[0099] Furthermore, the present invention also provides a multi-robot bidirectional relay node selection device, including a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the program, it implements the steps of the multi-robot bidirectional relay node selection method.

[0100] The present invention adopts the above technical solution, which differs from the prior art. The above technical solution has the following beneficial effects:

[0101] (1) The present invention integrates the signal quality of the wide area network of the candidate robot relay and prioritizes the selection of nodes with reliable upload capability, effectively avoiding the problem of invalid forwarding when the self-organizing network is connected but the data cannot be uploaded, and greatly improving the overall availability of the communication link.

[0102] (2) The present invention adopts a hierarchical hop count constraint strategy of "priority 1 hop, secondary 2 hops, maximum 2 hops". Prioritizing 1 hop node can ensure the shortest link and the lowest latency, meeting the timeliness requirements of key services such as emergency stop instructions and path changes; when 1 hop node is unavailable, 2 hop node is allowed as an alternative to expand the coverage area; 3 hop and above nodes are strictly prohibited to control the transmission delay within an acceptable range, avoid the lag risk and increased packet loss rate caused by long paths, and ensure the security and controllability of task execution.

[0103] (3) This invention is the first to incorporate multiple dimensions of indicators such as self-organizing network signal quality, relay node WAN signal quality, distance, load, hop count, and remaining power into the evaluation model. It achieves comprehensive evaluation through normalized weighted summation. On the basis of taking into account upload capability, link quality, node load and remaining power, it achieves the global perspective of end-to-end link to select the optimal relay node, effectively balance the cluster load, improve the overall resource utilization efficiency, and significantly improve the communication success rate and reliability.

[0104] (4) This invention constructs a closed-loop mechanism for bidirectional linkage between uplink and downlink: the optimal relay ID selected by the uplink robot is synchronized to the platform in real time, and the platform prioritizes reusing this relay when recovering a lost robot. Through the linkage and reuse of uplink and downlink relay decisions, the consistency of bidirectional link topology is ensured, effectively solving the link breakage problem. The platform does not need to perform complex routing calculations again, significantly improving the recovery capability under wide area network failure scenarios and greatly shortening the recovery time under loss scenarios.

[0105] Finally, it should be noted that although the above embodiments have been described in the text and drawings of this application, this should not limit the scope of patent protection of this application. Any technical solutions that are based on the essential concept of this application and utilize the content described in the text and drawings of this application, resulting in equivalent structural or procedural substitutions or modifications, as well as the direct or indirect application of the technical solutions of the above embodiments to other related technical fields, are all included within the scope of patent protection of this application.

Claims

1. A method for selecting bidirectional relay nodes for multiple robots, applied to a system consisting of a cloud control platform and multiple mobile robots, each mobile robot being equipped with a wide area network communication module and a short-range wireless self-organizing network module, characterized in that... The method includes the following steps: The target robot monitors its own wide area network communication quality. When the communication quality is detected to be lower than the preset triggering condition, it obtains neighbor node information through a short-distance self-organizing network and selects neighbor nodes that meet the preset hop count condition based on a hierarchical strategy to form an uplink candidate relay node set. Obtain the relay performance index of each candidate node in the uplink candidate relay node set, normalize each relay performance index, calculate the comprehensive score by weighting according to the preset weight, and select the candidate node with the highest score as the optimal uplink relay node. The target robot sends uplink data to the cloud control platform through the optimal uplink relay node and synchronizes the identification information of the optimal uplink relay node to the cloud control platform. When the cloud control platform detects that a target robot has lost contact, it obtains the identifier of the optimal uplink relay node synchronized with the target robot and determines whether the corresponding relay node meets the preset reuse conditions. If it does, it sends instructions to the target robot through the corresponding relay node. If the preset reuse conditions are not met, the cloud control platform obtains the neighbor node information of the target robot and selects neighbor nodes that meet the preset hop count conditions based on the hierarchical strategy to form a set of downlink candidate relay nodes. The cloud control platform obtains the relay performance index of each candidate node in the downlink candidate relay node set; normalizes each relay performance index and calculates a comprehensive score by weighting according to a preset weight; selects the candidate node with the highest score as the optimal downlink relay node; and issues instructions to the target robot through the optimal downlink relay node.

2. The method for selecting bidirectional relay nodes for multiple robots according to claim 1, characterized in that, The relay performance metrics include the WAN communication quality, load, hop count, remaining power of each candidate node, as well as the distance between the candidate node and the robot, and the signal quality of the ad hoc network.

3. The method for selecting bidirectional relay nodes for multiple robots according to claim 2, characterized in that, The implementation of normalizing various relay performance indicators and calculating a comprehensive score based on preset weights includes: The calculated value of the hop factor H is 1 for one hop, 0.5 for two hops, and 0 for three hops or more. Subtracting the node's current combined CPU and bandwidth utilization from 1 yields the load factor L. Get the current remaining battery percentage of the node as the battery factor P; The distance factor D is calculated based on the ratio of the actual distance between nodes to the maximum communication distance of the ad hoc network, and the value is larger the closer the distance. The value of the self-organizing network signal quality factor S is obtained by linearly mapping the received signal strength indication value to the interval [0, 1]. The WAN communication quality factor of the relay node is calculated by linearly mapping the WAN signal strength, preset full-strength signal strength, and minimum acceptable signal threshold of the relay node to the interval [0, 1]. ; The overall score is calculated using a weighted average based on preset weights, and the corresponding calculation formula is as follows: , in, Here, is the WAN communication quality factor for relay nodes, D is the distance factor, S is the ad hoc network signal quality factor, H is the hop count factor, L is the load factor, and P is the power factor. , , , , , These are the weights of the corresponding items, and .

4. The method for selecting bidirectional relay nodes for multiple robots according to claim 3, characterized in that, The method also includes: the cloud control platform or the target robot dynamically adjusts the preset weights according to the current task scenario; the task scenario includes at least the preset power priority mode, speed priority mode and load balancing mode.

5. The method for selecting bidirectional relay nodes for multiple robots according to claim 1, characterized in that, The preset hop count condition is: the data forwarding hop count between the neighbor node and the target robot is 1 hop or 2 hops.

6. The method for selecting bidirectional relay nodes for multiple robots according to claim 5, characterized in that, The uplink candidate relay node set is formed by selecting neighbor nodes that meet the preset number of hops by using a hierarchical strategy that prioritizes one-hop neighbor nodes, selects two-hop neighbor nodes when there are no one-hop neighbor nodes, and removes nodes with more than or equal to three hops. Alternatively, a hierarchical strategy can be adopted to select neighbor nodes that meet the preset number of hops, forming a set of downlink candidate relay nodes. This strategy prioritizes one-hop neighbor nodes, selects two-hop neighbor nodes when no one-hop neighbor node is available, and removes nodes with three or more hops.

7. The method for selecting bidirectional relay nodes for multiple robots according to claim 1, characterized in that, The preset reuse conditions include that the uplink optimal relay node is online, the WAN communication of the uplink optimal relay node is normal, and the number of hops to the target robot does not exceed two.

8. A multi-robot bidirectional relay node selection system, employing the multi-robot bidirectional relay node selection method according to any one of claims 1 to 7, characterized in that, The system includes a cloud control platform and at least two robots that communicate with the cloud control platform; The cloud control platform includes: A connection status monitoring unit is used to detect robot disconnection via a heartbeat mechanism; The robot disconnection query unit is used to broadcast a query command to the online robot after a disconnection is detected; The downlink recovery unit is used to, when a target robot is detected to be out of contact, obtain the identifier of the uplink optimal relay node synchronized with the target robot, and determine whether the corresponding relay node meets the preset reuse conditions. If it does, the unit sends instructions to the target robot through the corresponding relay node. If it does not, the unit obtains the neighbor node information of the target robot, and selects neighbor nodes that meet the preset hop count conditions based on a hierarchical strategy to form a downlink candidate relay node set. The unit obtains the relay performance index of each candidate node in the downlink candidate relay node set, normalizes each relay performance index, calculates a comprehensive score by weighting according to preset weights, selects the candidate node with the highest score as the downlink optimal relay node, and sends instructions to the target robot through the downlink optimal relay node. Each robot includes: Wide area network communication module, used for direct communication with the cloud control platform; The short-range wireless ad hoc network module is used to establish multi-hop ad hoc networks with other robots and maintain a neighbor table for its own short-range ad hoc network. A signal quality monitoring unit is used to monitor the signal quality of the wide area network communication module in real time. The communication decision unit is used to obtain neighbor node information through the short-range ad hoc network communication module when its own wide area network communication quality is lower than the preset triggering conditions. Based on the hierarchical strategy, it selects neighbor nodes that meet the preset hop count conditions to form an uplink candidate relay node set; obtains the relay performance index of each candidate node in the uplink candidate relay node set, normalizes each relay performance index and calculates a comprehensive score according to the preset weights, selects the candidate node with the highest score as the optimal uplink relay node; sends uplink data to the cloud control platform through the optimal uplink relay node, and synchronizes the identification information of the optimal uplink relay node to the cloud control platform. The relay performance indicators include the WAN communication quality, load, hop count, remaining power of each candidate node, as well as the distance between the candidate node and the robot, and the signal quality of the ad hoc network.

9. A multi-robot bidirectional relay node selection system according to claim 8, characterized in that, The downlink recovery unit forms a set of downlink candidate relay nodes according to a hierarchical strategy that prioritizes one-hop neighbor nodes, selects two-hop neighbor nodes when there are no one-hop neighbor nodes, and removes nodes with three or more hops. The communication decision unit forms an uplink candidate relay node set according to a hierarchical strategy that prioritizes one-hop neighbor nodes, selects two-hop neighbor nodes when there are no one-hop neighbor nodes, and removes nodes with three or more hops.

10. A multi-robot bidirectional relay node selection device, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the program, it implements the steps of the multi-robot bidirectional relay node selection method according to any one of claims 1 to 8.