A robot cluster knowledge sharing method driven by body perception
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- SMART DYNAMICS CO LTD
- Filing Date
- 2026-01-30
- Publication Date
- 2026-06-02
Smart Images
Figure REF-OBJ-1769759873620-000006
Abstract
Description
Technical Field
[0001] This invention relates to the field of computer data processing technology, and in particular to a knowledge-sharing method for robot swarms driven by embodied perception. Background Technology
[0002] A robot swarm refers to a system composed of multiple robots that collaborate to accomplish complex tasks that are difficult for a single robot to perform. In swarm collaborative operations, knowledge sharing is the foundation for achieving efficient collaboration. It involves how each robot node effectively integrates and distributes the local environmental information it perceives through its own sensors to other members of the swarm to form a unified, global situational awareness. This is essentially a complex distributed data processing and knowledge management process.
[0003] In existing technologies, knowledge sharing among robot swarms typically employs a centralized data fusion and distribution architecture. In this architecture, each robot acts as a data acquisition terminal, reporting its raw sensory data or pre-processed information to a central server or a designated leader robot. The central node is responsible for collecting data from all sources, aligning timestamps, unifying coordinate systems, and using pre-defined rules or algorithms to resolve conflicts and redundancies between different data sources.
[0004] Existing technical solutions have several shortcomings in practical applications. Centralized data processing nodes are prone to becoming the performance bottleneck and single point of failure of the entire system. If the central node fails or communication is blocked, the knowledge sharing of the entire cluster will be paralyzed. The entire process from data collection to final distribution back to robot members has an inherent time delay, which may cause the global knowledge acquired by the robot to lag behind the rapidly changing dynamic environment, indicating room for improvement. Summary of the Invention
[0005] This invention provides a knowledge-sharing method for robot swarms driven by embodied perception, which ensures the quality of shared knowledge from the source and achieves efficient utilization of system resources and improved overall robustness.
[0006] To achieve the above objectives, the present invention adopts the following technical solution: Firstly, an embodied perception-driven method for knowledge sharing in robot swarms is provided, specifically including the following steps: The robot generates verifiable knowledge units with sensor fingerprints containing sensor modalities and state parameters based on its perception data and task status. Based on the verifiable knowledge unit and its associated perceptual fingerprint, a consensus verification is initiated to a dynamically selected verification group, and a verified knowledge unit is generated based on the verification feedback received after orientation and re-perception by robot members within the verification group. Based on the verified knowledge unit and the verification feedback, a trusted knowledge group is constructed that encapsulates the trust chain information that records the verification history and the initial trust energy. Based on the robot's own task requirements, it actively obtains target trusted knowledge groups from the cluster shared network, and performs metabolic fusion and update of the local knowledge graph according to the trust chain information carried by the target trusted knowledge groups.
[0007] Optionally, the generation of verifiable knowledge units with perceptual fingerprints specifically includes: Entity descriptions and spatiotemporal information are extracted from the perceived data to generate structured knowledge assertions; Obtain the set of sensor types and their key state parameters that are relied upon when generating the structured knowledge assertion, and encapsulate them as sensor modal features; Obtain the local task context when the robot performs a perception action, and generate task context labels; The structured knowledge assertions, sensor modal features, and task context labels are bound together to generate the verifiable knowledge unit and its associated perceptual fingerprint.
[0008] Optionally, the generation of verified knowledge units specifically includes: Send the verifiable knowledge unit and the perceptual fingerprint to all robot members of the verification group; Receive verification feedback entries from each robot member, which include their re-perception results and their own perception fingerprints; Based on the consistency of the content of the verification feedback items and the quality of the corresponding perceptual fingerprints, the verifiable knowledge units are confirmed or corrected through a consensus algorithm to generate the verified knowledge units.
[0009] Optionally, the process of forming the dynamically selected verification group specifically includes: Obtain real-time status information of robots in the cluster, including geographical location, task list and list of available sensors; Based on the spatial coordinates and semantic type of the verifiable knowledge unit, and combined with the real-time status information, the spatial proximity, task relevance and sensor complementarity of each robot are calculated to obtain a comprehensive suitability score. Based on the comprehensive suitability score, the robots with the highest scores are selected to form the dynamically selected verification group.
[0010] Optionally, the construction of a trusted knowledge group specifically includes: Extract the identifiers of all robot members participating in the verification and the core sensor modes they use from the verification feedback to form the trust chain information; Obtain the historical reliability scores of each robot member participating in the verification; The initial credibility energy is calculated based on the number of verifiers, sensor modality diversity, and historical reliability scores of each verifier contained in the trust chain information. The verified knowledge unit, the trust chain information, and the initial credibility energy are encapsulated to generate the trusted knowledge group.
[0011] Optionally, when the trusted knowledge group propagates in a cluster-shared network, it further includes: The credibility energy carried by the trusted knowledge group is attenuated based on the number of propagation hops it has undergone. Obtain the real-time reliability score of the nodes through which the trusted knowledge group is forwarded; Based on the real-time reliability scores of the nodes along the route, the attenuated credibility energy is corrected to obtain the updated credibility energy.
[0012] Optionally, the active acquisition of the target trusted knowledge group specifically includes: Based on the robot's task execution progress and the status of the local knowledge graph, a knowledge requirement vector is generated; Using the knowledge demand vector as a query condition, a matching candidate trusted knowledge group is retrieved in the cluster shared network; Based on the current credibility energy and the strength of the trust chain information of each candidate trusted knowledge group, the target trusted knowledge group is selected from the candidate trusted knowledge groups.
[0013] Optionally, the metabolic fusion update of the local knowledge graph specifically includes: The content of the target trusted knowledge group is compared with the existing knowledge in the local knowledge graph to trigger conflict detection; When a conflict is detected, the strength and time freshness of the trust chain information on which the knowledge of the conflicting parties is attached are compared to generate an arbitration result. Based on the arbitration result, a replacement or retention operation is performed on the local knowledge graph to complete the metabolic fusion update.
[0014] Optionally, after initiating consensus verification to the dynamically selected verification group, the method further includes: Within a preset time, it is determined whether the received verification feedback has reached the legally required quantity. If not, a consensus failure signal is generated. Upon receiving the consensus failure signal, the verification priority of the verifiable knowledge unit is increased. Based on the improved verification priority, a new verification group containing different robot members is selected, and the consensus verification is re-initiated.
[0015] Secondly, an embodied perception-driven robot swarm knowledge-sharing system is provided, including: The knowledge unit generation module is configured to generate verifiable knowledge units with sensor fingerprints containing sensor modalities and state parameters by the robot based on its perception data and task status. The consensus verification module is configured to initiate consensus verification to a dynamically selected verification group based on the verifiable knowledge unit and its associated perceptual fingerprint, and generate a verified knowledge unit based on the received verification feedback. The group building and propagation module is configured to build a trusted knowledge group encapsulating trust chain information and initial trust energy based on the verified knowledge unit and the verification feedback. The knowledge metabolism module is configured to actively acquire target trusted knowledge groups from the cluster shared network based on the robot's own task requirements, and perform metabolic fusion and update of the local knowledge graph based on the trust chain information carried by the target trusted knowledge groups. Thirdly, an electronic device is provided, comprising: a processor and a memory; the memory is used to store a computer program, which, when executed by the processor, causes the electronic device to perform the embodied perception-driven robot swarm knowledge-sharing method described in the first aspect.
[0016] In one possible design, the electronic device described in the third aspect may further include a transceiver. This transceiver may be a transceiver circuit or an interface circuit. The transceiver can be used for communication between the electronic device described in the third aspect and other electronic devices.
[0017] In the embodiments of the present invention, the electronic device described in the third aspect may be a terminal, or a chip (system) or other component or assembly disposed in the terminal, or a system containing the terminal.
[0018] Fourthly, a computer-readable storage medium is provided, comprising: a computer program or instructions; when the computer program or instructions are executed on a computer, the computer causes the computer to perform the embodied perception-driven robot swarm knowledge-sharing method described in the first aspect.
[0019] In summary, the above methods and systems have the following technical effects: This invention improves the quality and reliability of the collective knowledge base through a knowledge-sharing process that combines pre-verification and post-evolutionary metabolism. By introducing a consensus verification mechanism executed by a dynamic verification group at the source of knowledge generation, it ensures that only information that has been cross-confirmed by multiple parties can enter the shared network, effectively preventing initial contamination by erroneous or low-quality data. This systematically improves the accuracy and consistency of the entire cluster knowledge base. By encapsulating knowledge with traceable trust chain information and endowing it with dynamically decaying credibility energy, the credibility of knowledge is no longer a static label, but a flowing value that evolves with propagation and time. It can autonomously identify and prioritize the adoption of high-credibility knowledge, while naturally eliminating outdated information. It exhibits stronger adaptability in the face of individual failures or environmental interference. Through precise on-demand allocation of communication and computing resources, it greatly optimizes the overall operating efficiency of the robot cluster. By transforming passive global information broadcasting into proactive knowledge acquisition by individual robots based on task requirements, and combining a metabolic fusion mechanism based on trust chains, it ensures that each robot only processes the most critical and credible information for its current task, avoiding information overload and significantly reducing network bandwidth and individual processor load. This provides support for the deployment and application of large-scale robot clusters. Attached Figure Description
[0020] Figure 1 This is a flowchart illustrating the embodied perception-driven robot swarm knowledge sharing method provided in an embodiment of the present invention. Detailed Implementation
[0021] The technical solution of the present invention will now be described with reference to the accompanying drawings.
[0022] In this embodiment of the invention, "instruction" can include direct and indirect instructions, as well as explicit and implicit instructions. The information indicated by a certain piece of information is called the information to be instructed. In specific implementation, there are many ways to instruct the information to be instructed, such as, but not limited to, directly instructing the information to be instructed, such as the information to be instructed itself or its index. It can also indirectly instruct the information to be instructed by instructing other information, where there is a correlation between the other information and the information to be instructed. It can also instruct only a part of the information to be instructed, while the other parts are known or pre-agreed upon. For example, the instruction of specific information can be achieved by using a pre-agreed (e.g., protocol-defined) arrangement of various pieces of information, thereby reducing instruction overhead to some extent. Simultaneously, common parts of various pieces of information can be identified and uniformly indicated to reduce the instruction overhead caused by individually indicating the same information.
[0023] Furthermore, the specific indication method can also be any existing indication method, such as, but not limited to, the above-mentioned indication methods and their various combinations. Specific details of various indication methods can be found in existing technologies, and will not be elaborated upon here. As described above, for example, when multiple pieces of information of the same type need to be indicated, the indication methods for different pieces of information may differ. In specific implementation, the required indication method can be selected according to specific needs. This embodiment of the invention does not limit the selected indication method; therefore, the indication methods involved in this embodiment of the invention should be understood to cover various methods that enable the party to be indicated to obtain the information to be indicated.
[0024] It should be understood that the information to be indicated can be sent as a whole or divided into multiple sub-information messages sent separately, and the sending period and / or timing of these sub-information messages can be the same or different. The specific sending method is not limited in this embodiment of the invention. The sending period and / or timing of these sub-information messages can be predefined, for example, according to a protocol, or configured by the sending device by sending configuration information to the receiving device.
[0025] "Predefined" or "pre-configured" can be achieved by pre-saving corresponding codes, tables, or other means that can be used to indicate relevant information in the device. This embodiment of the invention does not limit the specific implementation method. "Saving" can refer to saving in one or more memories. These memories can be separate installations or integrated into the encoder, decoder, processor, or electronic device. Alternatively, some memories can be separately installed, while others are integrated into the decoder, processor, or electronic device. The type of memory can be any form of storage medium, and this embodiment of the invention does not limit this.
[0026] In the embodiments of this invention, "protocol" may refer to a protocol family in the field of communication, a standard protocol with a similar protocol family frame structure, or a related protocol applied to a future embodied perception-driven robot swarm knowledge sharing method system. The embodiments of this invention do not specifically limit this.
[0027] In this embodiment of the invention, descriptions such as "when," "under the circumstances," "if," and "if" all refer to the device making corresponding processing under certain objective circumstances, and are not limited to a specific time. They do not require the device to make a judgment action during implementation, nor do they imply any other limitations.
[0028] In the description of the embodiments of the present invention, unless otherwise stated, " / " indicates that the objects before and after are in an "or" relationship. For example, A / B can represent A or B. "And / or" in the embodiments of the present invention is merely a description of the relationship between the related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, and B alone, where A and B can be singular or plural. Furthermore, in the description of the embodiments of the present invention, unless otherwise stated, "multiple" refers to two or more. "At least one of the following" or similar expressions refer to any combination of these items, including any combination of single or plural items. For example, at least one of a, b, or c can represent: a, b, c, ab, ac, bc, or abc, where a, b, and c can be single or multiple. Additionally, to facilitate a clear description of the technical solutions of the embodiments of the present invention, the terms "first" and "second" are used in the embodiments of the present invention to distinguish identical or similar items with essentially the same function and effect. Those skilled in the art will understand that the terms "first," "second," etc., do not limit the quantity or order of execution, and that "first," "second," etc., are not necessarily different. Furthermore, in the embodiments of this invention, words such as "exemplary" or "for example" are used to indicate that something is being described as an example, illustration, or description. Any embodiment or design scheme described as "exemplary" or "for example" in the embodiments of this invention should not be construed as being more preferred or advantageous than other embodiments or design schemes. Specifically, the use of words such as "exemplary" or "for example" is intended to present the relevant concepts in a concrete manner for ease of understanding.
[0029] The network architecture and business scenarios described in the embodiments of this invention are for the purpose of more clearly illustrating the technical solutions of the embodiments of this invention, and do not constitute a limitation on the technical solutions provided by the embodiments of this invention. As those skilled in the art will know, with the evolution of network architecture and the emergence of new business scenarios, the technical solutions provided by the embodiments of this invention are also applicable to similar technical problems.
[0030] Figure 1 This is a flowchart illustrating the method provided in an embodiment of the present invention. The specific process of this embodied perception-driven robot swarm knowledge sharing method is as follows: S1. The robot generates verifiable knowledge units with sensor fingerprints containing sensor modalities and state parameters based on its perception data and task status. S2. Based on the verifiable knowledge unit and its associated perceptual fingerprint, initiate consensus verification to the dynamically selected verification group, and generate a verified knowledge unit based on the verification feedback received by the robot members in the verification group after orientation re-perception. S3. Based on the verified knowledge unit and the verification feedback, construct a trusted knowledge group that encapsulates the trust chain information that records the verification history and the initial trust energy. S4. Based on the robot's own task requirements, actively obtain the target trusted knowledge group from the cluster shared network, and perform metabolic fusion and update of the local knowledge graph according to the trust chain information carried by the target trusted knowledge group.
[0031] Optionally, the generation of verifiable knowledge units with perceptual fingerprints specifically includes: Entity descriptions and spatiotemporal information are extracted from the perceived data to generate structured knowledge assertions; Obtain the set of sensor types and their key state parameters that are relied upon when generating the structured knowledge assertion, and encapsulate them as sensor modal features; Obtain the local task context when the robot performs a perception action, and generate task context labels; The structured knowledge assertions, sensor modal features, and task context labels are bound together to generate the verifiable knowledge unit and its associated perceptual fingerprint.
[0032] Specifically, this method transforms the robot's raw, unstructured perceptual data into standardized information units that can be verified through cluster consensus. Its engineering purpose is to provide a data foundation with clear content, traceable origin, and assessable quality for subsequent distributed verification. This process is automatically executed by the robot's local knowledge processing unit, whose input is real-time data streams from sensors such as cameras and LiDAR, and whose output is verifiable knowledge units bound to perceptual fingerprints.
[0033] The first step in the system's execution is to generate structured knowledge assertions. The engineering objective of this step is to slice and parse the continuous raw perceptual data stream into discrete, meaningful descriptions of events or entities. For example, when the robot's vision processing module, a deep learning-based target detection algorithm, identifies a hot signal target in a video stream, the module outputs the target's category, confidence level, and pixel coordinates in the image. The system then calls a coordinate transformation service, combining its own inertial navigation unit (IMU) and GPS data, to convert the pixel coordinates into three-dimensional spatial coordinates in the world coordinate system. Simultaneously, the system records the precise timestamp of the current operating system. This information, including entity descriptions such as "suspected vital signs," spatial coordinates such as (118.342, 32.091, 1.5), and the timestamp, is encapsulated into a standardized data structure, forming a structured knowledge assertion.
[0034] Following this, the system synchronously generates sensor modal features. Its engineering purpose is to quantify and record the physical perception conditions at the time of generating the aforementioned structured knowledge assertion, providing an objective basis for subsequent evaluation of the assertion's credibility. The system queries and records in real time the internal state parameters of the key sensors upon which the assertion depends. For example, for an infrared thermal imaging sensor detection, the system records the ambient temperature, e.g., 15 degrees Celsius, the sensor's detection distance setting, e.g., 50 meters, and the temperature difference between the target and the background, e.g., 10 Kelvin. Simultaneously, confidence parameters directly obtained from the target detection algorithm, e.g., 0.87, are also incorporated. These parameters collectively constitute the sensor modal features, which directly reflect the physical quality of this perception.
[0035] The system then generates a task context label. The engineering purpose of this step is to associate the perception with the robot's current intent. The robot's internal task scheduler maintains an identifier for the currently executing task. The system reads this identifier, such as "ZoneB-SurvivorSearch-Grid03," and uses it as the task context label. This label indicates that the perception did not occur randomly or by chance, but rather was the result of the robot actively acquiring information to perform a specific, sophisticated search task, which typically implies higher focus and resource investment.
[0036] Finally, the system binds the three data volumes generated in the preceding steps to form the final output. Structured knowledge assertions, as core data, are defined as verifiable knowledge units. Sensor modal features and task context labels are jointly encapsulated into a meta-data package, serving as a perceptual fingerprint attached to this knowledge unit. These two parts are logically strongly bound to ensure inseparability during network transmission. The system also calculates an initial fingerprint quality score (Qfp) based on the parameters of the perceptual fingerprint for internal ranking or preliminary screening; its calculation method can be expressed as the formula... ; Here, Cm represents the model confidence score directly output by the perception algorithm, a normalized value between 0 and 1. Sq represents the normalized environmental quality factor calculated based on the sensor's key state parameters, for example, obtained by mapping the relationship between the current environmental noise level and the sensor's theoretical optimal operating range to a range of 0 to 1. w1 and w2 are preset weighting coefficients, summing to 1, used to adjust the importance of model confidence and environmental quality in the final score. This verifiable knowledge unit and its perceptual fingerprint are then submitted to the next stage of consensus verification module for processing.
[0037] Optionally, the generation of verified knowledge units specifically includes: Send the verifiable knowledge unit and the perceptual fingerprint to all robot members of the verification group; Receive verification feedback entries from each robot member, which include their re-perception results and their own perception fingerprints; Based on the consistency of the content of the verification feedback items and the quality of the corresponding perceptual fingerprints, the verifiable knowledge units are confirmed or corrected through a consensus algorithm to generate the verified knowledge units.
[0038] The newly generated verifiable knowledge units are collaboratively verified on a small scale and with high efficiency. The initiating robot acts as a temporary coordinator, or the distributed consensus service designated in the network schedules the verification. The input is the verifiable knowledge unit with a perceptual fingerprint, and the output is the verified knowledge unit after confirmation or correction by multiple parties.
[0039] The first step is to dynamically select robots to join the dynamic verification group. This intelligently filters out the few most suitable peers for verifying the specific knowledge, avoiding ineffective communication and wasted perception resources. First, the system parses the input verifiable knowledge unit, extracting its core spatial coordinates and semantic type. It then broadcasts a neighbor node query request to the cluster network, containing the spatial coordinates. Upon receiving the request, all robots in the cluster calculate the Euclidean distance between their current position and these coordinates, selecting robots within a preset threshold (e.g., 100 meters) as candidate verifiers. Next, the system further filters these candidate verifiers, requiring that their current or upcoming tasks are highly relevant to the semantic type of the verifiable knowledge unit. For example, for a knowledge unit about "building structural damage," robots currently performing "safe path exploration" or "structural assessment" tasks will be prioritized. This step ensures that verifiers are not only physically close but also aligned in their task intent. Finally, the system selects 2 to 4 robots from the qualified candidates to form the dynamic verification group for this task.
[0040] Next, a verification request is sent to all robot members of the dynamic verification group. The engineering purpose of this is to clearly communicate the verification task and provide sufficient contextual information to guide the perceptual behavior of the verifier. Optionally, the process of forming the dynamically selected verification group specifically includes: Obtain real-time status information of robots in the cluster, including geographical location, task list and list of available sensors; Based on the spatial coordinates and semantic type of the verifiable knowledge unit, and combined with the real-time status information, the spatial proximity, task relevance and sensor complementarity of each robot are calculated to obtain a comprehensive suitability score. Based on the comprehensive suitability score, the robots with the highest scores are selected to form the dynamically selected verification group.
[0041] The system acquires real-time status information of relevant robots in the cluster. Its engineering purpose is to establish an immediate, global resource view for subsequent decision-making. The robot initiating verification broadcasts a status query request to neighboring nodes within its communication range. This request contains the spatial coordinates of verifiable knowledge units. The receiving robot replies with a data packet containing its own real-time status information. This packet is structured to include its precise geographical location, a task list containing current and pending task IDs, and a list of all available sensors and their current status. For example, a data packet might contain location (118.345, 32.093, 0.5), a task list [TaskA, TaskB], and a sensor list. The acquired candidate robots undergo multi-dimensional quantitative evaluation. The engineering objective is to calculate a comprehensive suitability score for each potential validator. This score is composed of a weighted average of three independent sub-scores. First, the system calculates spatial distance—the Euclidean distance between the real-time position of each candidate robot and the spatial coordinates of the verifiable knowledge unit—and converts it into a normalized spatial proximity score Ss, which decreases as distance increases. Second, the system calculates task relevance by querying and matching the candidate robot's task list with the semantic type of the verifiable knowledge unit in a pre-defined semantic association matrix, obtaining a task relevance score Sr between 0 and 1. Finally, the system evaluates sensor complementarity. The system extracts sensor modalities from the raw perceptual fingerprint and compares them with the candidate robot's available sensor list. Based on a pre-defined complementarity scoring table (e.g., high visual-to-infrared complementarity, low visual-to-visual complementarity), each candidate robot is assigned the highest possible sensor complementarity score Sc.
[0042] The third step involves the system making the final selection based on a comprehensive score. Its engineering objective is to make the final team member selection decision based on the quantitative evaluation in the previous step. The system uses a comprehensive suitability scoring formula to calculate the final score S for each candidate robot: ; Where Ss, Sr, and Sc are the normalized scores of spatial proximity, task relevance, and sensor complementarity calculated in the previous step, respectively; ks, kr, and kc are preset weight coefficients, which sum to 1. These coefficients are dynamically adjusted according to the nature of the current overall task, such as whether speed or accuracy is prioritized. For example, in an emergency rescue mission, the weight of ks may be increased to prioritize the nearest robot. All candidate robots are sorted from high to low according to their final score S, and the top 2 to 4 robots are selected. Their ID list is used as the final output to formally form the dynamic verification group. Optionally, the construction of a trusted knowledge group specifically includes: Extract the identifiers of all robot members participating in the verification and the core sensor modes they use from the verification feedback to form the trust chain information; Obtain the historical reliability scores of each robot member participating in the verification; The initial credibility energy is calculated based on the number of verifiers, sensor modality diversity, and historical reliability scores of each verifier contained in the trust chain information. The verified knowledge unit, the trust chain information, and the initial credibility energy are encapsulated to generate the trusted knowledge group.
[0043] The first step is to extract and integrate the trust chain information. The engineering purpose is to establish a transparent source and verification history for the verified knowledge unit. The system parses the verification feedback entries returned by all members within the dynamic verification group, extracting two key fields from each entry: the robot member identifier, a globally unique ID, and the core sensor modality used by the member during the verification, such as LiDAR or a thermal imaging camera. The system aggregates these binary pairs composed of member identifier and sensor modality to form a structured list, which constitutes the trust chain information. This list fully records the supporters of the knowledge unit and their verification methods.
[0044] Next, the system obtains the historical reliability scores of each robot member participating in the verification. The engineering purpose is to introduce the verifier's long-term reputation as an important basis for assessing the credibility of the current verification behavior. Based on the robot member identifier in the trust chain information, the system initiates a query request to the cluster's distributed reputation management service to obtain the historical reliability score corresponding to each member. This score is a floating-point number between 0 and 1, and is periodically calculated and updated based on the robot's performance in all past verification tasks—that is, the number of times its provided verification feedback has been ultimately confirmed as correct or incorrect.
[0045] The system calculates the initial trust energy based on the aforementioned information. Its engineering objective is to integrate multi-dimensional verification strength indicators, including the number of verifiers, the diversity of verification methods, and the verifiers' historical reputation, into a single, standardized initial trust score. This score is calculated according to the formula... ; Where E represents the final calculated initial credibility energy, Nv is the number of unique robot member identifiers in the trust chain information, f(Nv) is a function that normalizes the number of verifiers, such as a logarithmic function, to reflect the diminishing marginal effect of increasing the number of verifiers. Dm is the number of unique core sensor modalities in the trust chain information, f(Dm) is a function that normalizes modal diversity, Ravg is the arithmetic mean of the historical reliability scores of all participating robot members, and wn and wd are preset weight coefficients that sum to 1, used to adjust the relative importance of the number of verifiers and modal diversity in the calculation.
[0046] Ultimately, the system encapsulates all the aforementioned elements into the trusted knowledge group. Its engineering objective is to create an atomic, information-complete knowledge propagation unit. The system creates a new data structure, using the consensus-verified knowledge unit as its core data payload, and the trust chain information generated in the first step and the initial trust energy calculated in the third step as its metadata. This integrated data package, containing data, verification history, and initial trust score, is defined as the trusted knowledge group and submitted to the network propagation module, ready for distribution to other members of the cluster.
[0047] Optionally, when the trusted knowledge group propagates in a cluster-shared network, it further includes: The credibility energy carried by the trusted knowledge group is attenuated based on the number of propagation hops it has undergone. Obtain real-time reliability scores for the nodes that forward the trusted knowledge group; Based on the real-time reliability scores of the nodes along the route, the attenuated credibility energy is corrected to obtain the updated credibility energy.
[0048] Specifically, this method dynamically adjusts the credibility of trusted knowledge groups during their propagation within the cluster network. Its engineering purpose is to simulate the natural decay of information in the transmission chain, ensuring that the timeliness and scope of knowledge dissemination are constrained by its inherent credibility, and preventing outdated or low-quality information from spreading indefinitely within the network. This process is triggered each time a trusted knowledge group is forwarded by a robot node after it is injected into the cluster's shared network. The input is the trusted knowledge group to be forwarded, and the output is the same trusted knowledge group with updated credibility energy.
[0049] The system's dynamic decay processing of the credibility energy of trusted knowledge groups is mainly based on two core factors: the number of propagation hops and the real-time reliability score of the nodes along the way.
[0050] The first step is attenuation based on the number of propagation hops. The engineering purpose is to introduce temporal and spatial attenuation into the credibility of knowledge; that is, the more intermediate links information passes through, the greater its uncertainty. Each trusted knowledge group is initialized with a propagation hop count segment upon creation, with an initial value of 0. When a robot node receives a knowledge spore, F = E0 * D_h. Here, F is the credibility energy after attenuation by the number of propagation hops, E0 is the credibility energy value of the trusted knowledge group at the previous hop node, and D_h is an attenuation factor related to the number of propagation hops h, calculated as D_h = a^h, where a is a preset constant less than 1, such as 0.95, representing the attenuation rate per hop. This exponential attenuation model ensures that information does not propagate indefinitely.
[0051] The second step involves correction based on the real-time reliability scores of the nodes along the path. The engineering goal is to ensure that the quality of the information's propagation path also affects its final credibility; that is, information forwarded by high-reputation nodes should be more credible than information forwarded by low-reputation nodes. When a robot node is about to forward a trusted knowledge group, it first queries its own real-time reliability score, which is dynamically calculated by the reputation management service based on its recent behavior. Then, the system adjusts the decayed credibility energy based on the node's score. If the node's score is higher than the cluster average, the decay is reduced; conversely, if it is lower than the average, the decay is increased. This correction process can be represented as... ; Here, E_next is the trustworthiness energy ultimately carried by the trusted knowledge group when the node forwards it. F is the energy value after the first propagation hop count decay. R_node is the real-time reliability score of the current forwarding node, and R_avg is the average reliability score of all nodes in the cluster, which is a global parameter that can be obtained periodically through network broadcast. k is an adjustment coefficient used to control the sensitivity of node reliability to the decay effect, and its value is usually between 0.1 and 0.5.
[0052] Through the combined processing of the above two steps, when each robot node receives a trusted knowledge group and decides to forward it, it first executes this set of decay calculation procedures to update the trustworthiness energy field inside the spore before sending the updated trusted knowledge group to the next node. If the calculated E_next is lower than a preset propagation threshold, such as 0.05, the robot will stop forwarding the trusted knowledge group, thus achieving the natural termination of propagation.
[0053] Optionally, the active acquisition of the target trusted knowledge group specifically includes: Based on the robot's task execution progress and the status of the local knowledge graph, a knowledge requirement vector is generated; Using the knowledge demand vector as a query condition, a matching candidate trusted knowledge group is retrieved in the cluster shared network; Based on the current credibility energy and the strength of the trust chain information of each candidate trusted knowledge group, the target trusted knowledge group is selected from the candidate trusted knowledge groups.
[0054] Optionally, the metabolic fusion update of the local knowledge graph specifically includes: The content of the target trusted knowledge group is compared with the existing knowledge in the local knowledge graph to trigger conflict detection; When a conflict is detected, the strength and time freshness of the trust chain information on which the knowledge of the conflicting parties is attached are compared to generate an arbitration result. Based on the arbitration result, a replacement or retention operation is performed on the local knowledge graph to complete the metabolic fusion update.
[0055] Optionally, after initiating consensus verification to the dynamically selected verification group, the method further includes: Within a preset time, it is determined whether the received verification feedback has reached the legally required quantity. If not, a consensus failure signal is generated. Upon receiving the consensus failure signal, the verification priority of the verifiable knowledge unit is increased. Based on the improved verification priority, a new verification group containing different robot members is selected, and the consensus verification is re-initiated.
[0056] An embodied perception-driven robot swarm knowledge-sharing system, comprising: The knowledge unit generation module is configured to generate verifiable knowledge units with sensor fingerprints containing sensor modalities and state parameters by the robot based on its perception data and task status. The consensus verification module is configured to initiate consensus verification to a dynamically selected verification group based on the verifiable knowledge unit and its associated perceptual fingerprint, and generate a verified knowledge unit based on the received verification feedback. The group building and propagation module is configured to build a trusted knowledge group encapsulating trust chain information and initial trust energy based on the verified knowledge unit and the verification feedback. The knowledge metabolism module is configured to actively acquire target trusted knowledge groups from the cluster shared network based on the robot's own task requirements, and perform metabolic fusion and update of the local knowledge graph according to the trust chain information carried by the target trusted knowledge groups.
[0057] The electronic device provided in this embodiment of the invention, exemplarily, can be a network device, or a chip (system) or other component or assembly that can be disposed in a network device. The electronic device may include a processor. Optionally, the electronic device may also include a memory and / or a transceiver. The processor is coupled to the memory and transceiver, for example, by means of a communication bus connection.
[0058] The following is a detailed introduction to the various components of the electronic device: In this context, the processor is the control center of the electronic device. It can be a single processor or a collective term for multiple processing elements. For example, a processor can be one or more central processing units (CPUs), an application-specific integrated circuit (ASIC), or one or more integrated circuits configured to implement embodiments of the present invention, such as one or more digital signal processors (DSPs), or one or more field-programmable gate arrays (FPGAs).
[0059] Alternatively, the processor can perform various functions of the electronic device, such as the methods described above, by running or executing software programs stored in memory and by calling data stored in memory.
[0060] In a specific implementation, as one example, the processor may include one or more CPUs, such as CPU0 and CPU1.
[0061] In a specific implementation, as one example, the electronic device may also include multiple processors. Each of these processors may be a single-core processor (single-CPU) or a multi-core processor (multi-CPU). Here, a processor may refer to one or more devices, circuits, and / or processing cores used to process data (e.g., computer program instructions).
[0062] The memory is used to store the software program that executes the solution of the present invention, and the execution is controlled by the processor. The specific implementation method can be referred to the above method embodiment, and will not be repeated here.
[0063] Optionally, the memory can be read-only memory (ROM) or other types of static storage devices capable of storing static information and instructions, random access memory (RAM) or other types of dynamic storage devices capable of storing information and instructions, or electrically erasable programmable read-only memory (EEPROM), compact disc read-only memory (CD-ROM) or other optical disc storage, optical disc storage (including compressed optical discs, laser discs, optical discs, digital universal optical discs, Blu-ray discs, etc.), magnetic disk storage media or other magnetic storage devices, or any other medium capable of carrying or storing desired program code in the form of instructions or data structures and accessible by a computer, but not limited thereto. The memory can be integrated with the processor or exist independently and coupled to the processor through an interface circuit of an electronic device; the embodiments of the present invention do not specifically limit this.
[0064] A transceiver is used for communication with other electronic devices. For example, if the electronic device is a terminal, the transceiver can be used to communicate with a network device or with another terminal device. Similarly, if the electronic device is a network device, the transceiver can be used to communicate with a terminal or with another network device.
[0065] Optionally, the transceiver may include a receiver and a transmitter. The receiver is used to implement the receiving function, and the transmitter is used to implement the sending function.
[0066] Optionally, the transceiver can be integrated with the processor or exist independently and coupled to the processor through the interface circuit of the electronic device. This embodiment of the invention does not specifically limit this.
[0067] It is understandable that the structure of an electronic device does not constitute a limitation on the electronic device. An actual electronic device may include more or fewer components, or combine certain components, or have different component arrangements.
[0068] Furthermore, the technical effects of the electronic device can be referred to the technical effects of the method described in the above method embodiments, and will not be repeated here.
[0069] It should be understood that the processor in the embodiments of the present invention can be a central processing unit (CPU), or it can be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor.
[0070] It should also be understood that the memory in the embodiments of the present invention can be volatile memory or non-volatile memory, or may include both volatile and non-volatile memory. The non-volatile memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. The volatile memory can be random access memory (RAM), which is used as an external cache. By way of example, but not limitation, many forms of random access memory (RAM) are available, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate synchronous DRAM (DDRSDRAM), enhanced synchronous DRAM (ESDRAM), synchronous linked DRAM (SLDRAM), and direct rambus RAM (DRRAM).
[0071] The above embodiments can be implemented, in whole or in part, by software, hardware (such as circuits), firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of the present invention are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more sets of available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. A semiconductor medium can be a solid-state drive.
[0072] It should be understood that the term "and / or" in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. A and B can be singular or plural. Additionally, the character " / " in this article generally indicates an "or" relationship between the preceding and following related objects, but it can also represent an "and / or" relationship. Please refer to the context for a more accurate understanding.
[0073] In this invention, "at least one" means one or more, and "more than one" means two or more. "At least one of the following" or similar expressions refer to any combination of these items, including any combination of a single item or a plurality of items. For example, at least one of a, b, or c can represent: a, b, c, ab, ac, bc, or abc, where a, b, and c can be a single item or multiple items.
[0074] It should be understood that, in various embodiments of the present invention, the sequence number of each process does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.
[0075] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.
[0076] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.
[0077] In the embodiments provided by this invention, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.
[0078] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.
[0079] In addition, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.
[0080] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.
[0081] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.
Claims
1. A method for knowledge sharing in robot swarms driven by embodied perception, characterized in that, Specifically, the following steps are included: The robot generates verifiable knowledge units with sensor fingerprints containing sensor modalities and state parameters based on its perception data and task status. Based on the verifiable knowledge unit and its associated perceptual fingerprint, a consensus verification is initiated to a dynamically selected verification group, and a verified knowledge unit is generated based on the verification feedback received after orientation and re-perception by robot members within the verification group. Based on the verified knowledge unit and the verification feedback, a trusted knowledge group is constructed that encapsulates the trust chain information that records the verification history and the initial trust energy. Based on the robot's own task requirements, it actively obtains target trusted knowledge groups from the cluster shared network, and performs metabolic fusion and update of the local knowledge graph according to the trust chain information carried by the target trusted knowledge groups.
2. The embodied perception-driven robot swarm knowledge sharing method according to claim 1, characterized in that, The generation of verifiable knowledge units with perceptual fingerprints specifically includes: Entity descriptions and spatiotemporal information are extracted from the perceived data to generate structured knowledge assertions; Obtain the set of sensor types and their key state parameters that are relied upon when generating the structured knowledge assertion, and encapsulate them as sensor modal features; Obtain the local task context when the robot performs a perception action, and generate task context labels; The structured knowledge assertions, sensor modal features, and task context labels are bound together to generate the verifiable knowledge unit and its associated perceptual fingerprint.
3. The embodied perception-driven robot swarm knowledge sharing method according to claim 1, characterized in that, The generated verified knowledge unit specifically includes: Send the verifiable knowledge unit and the perceptual fingerprint to all robot members of the verification group; Receive verification feedback entries from each robot member, which include their re-perception results and their own perception fingerprints; Based on the consistency of the content of the verification feedback items and the quality of the corresponding perceptual fingerprints, the verifiable knowledge units are confirmed or corrected through a consensus algorithm to generate the verified knowledge units.
4. The embodied perception-driven robot swarm knowledge sharing method according to claim 3, characterized in that, The process of forming the dynamically selected verification group specifically includes: Obtain real-time status information of robots in the cluster, including geographical location, task list and list of available sensors; Based on the spatial coordinates and semantic type of the verifiable knowledge unit, and combined with the real-time status information, the spatial proximity, task relevance and sensor complementarity of each robot are calculated to obtain a comprehensive suitability score. Based on the comprehensive suitability score, the robots with the highest scores are selected to form the dynamically selected verification group.
5. The embodied perception-driven robot swarm knowledge sharing method according to claim 1, characterized in that, The construction of the trusted knowledge group specifically includes: Extract the identifiers of all robot members participating in the verification and the core sensor modes they use from the verification feedback to form the trust chain information; Obtain the historical reliability scores of each robot member participating in the verification; The initial credibility energy is calculated based on the number of verifiers, sensor modality diversity, and historical reliability scores of each verifier contained in the trust chain information. The verified knowledge unit, the trust chain information, and the initial credibility energy are encapsulated to generate the trusted knowledge group.
6. The embodied perception-driven robot swarm knowledge sharing method according to claim 5, characterized in that, When the trusted knowledge group is propagated in the cluster-shared network, it also includes: The credibility energy carried by the trusted knowledge group is attenuated based on the number of propagation hops it has undergone. Obtain the real-time reliability score of the nodes through which the trusted knowledge group is forwarded; Based on the real-time reliability scores of the nodes along the route, the attenuated credibility energy is corrected to obtain the updated credibility energy.
7. The embodied perception-driven robot swarm knowledge sharing method according to claim 1, characterized in that, The proactive acquisition of the target trusted knowledge group specifically includes: Based on the robot's task execution progress and the status of the local knowledge graph, a knowledge requirement vector is generated; Using the knowledge demand vector as a query condition, a matching candidate trusted knowledge group is retrieved in the cluster shared network; Based on the current credibility energy and the strength of the trust chain information of each candidate trusted knowledge group, the target trusted knowledge group is selected from the candidate trusted knowledge groups.
8. The embodied perception-driven robot swarm knowledge sharing method according to claim 1, characterized in that, The metabolic fusion update of the local knowledge graph specifically includes: The content of the target trusted knowledge group is compared with the existing knowledge in the local knowledge graph to trigger conflict detection; When a conflict is detected, the strength and time freshness of the trust chain information on which the knowledge of the conflicting parties is attached are compared to generate an arbitration result. Based on the arbitration result, a replacement or retention operation is performed on the local knowledge graph to complete the metabolic fusion update.
9. The embodied perception-driven robot swarm knowledge sharing method according to claim 7, characterized in that, After initiating consensus verification to the dynamically selected verification group, the process also includes: Within a preset time, it is determined whether the received verification feedback has reached the legally required quantity. If not, a consensus failure signal is generated. Upon receiving the consensus failure signal, the verification priority of the verifiable knowledge unit is increased. Based on the improved verification priority, a new verification group containing different robot members is selected, and the consensus verification is re-initiated.
10. A knowledge-sharing system for robot swarms driven by embodied perception, characterized in that, include: The knowledge unit generation module is configured to generate verifiable knowledge units with sensor fingerprints containing sensor modalities and state parameters by the robot based on its perception data and task status. The consensus verification module is configured to initiate consensus verification to a dynamically selected verification group based on the verifiable knowledge unit and its associated perceptual fingerprint, and generate a verified knowledge unit based on the received verification feedback. The group building and propagation module is configured to build a trusted knowledge group encapsulating trust chain information and initial trust energy based on the verified knowledge unit and the verification feedback. The knowledge metabolism module is configured to actively acquire target trusted knowledge groups from the cluster shared network based on the robot's own task requirements, and perform metabolic fusion and update of the local knowledge graph according to the trust chain information carried by the target trusted knowledge groups.