Data processing method, electronic device, storage medium and program product

Abstract

The invention provides a data processing method, electronic equipment, a storage medium and a program product. The method comprises the following steps: when an information state of a first sensor node meets a triggering condition of data transmission, mutually transmitting respective sensor data between the first sensor node and a neighbor sensor node; performing iterative operation on the information state of the first sensor node based on the sensor data of the first sensor node and the neighbor sensor node; and when the difference value of the information states obtained by two adjacent iterative operations is less than or equal to a specified threshold value, determining that the iterative operation process of the information state of the first sensor node is ended, and determining the forecast state of the first sensor node at the future moment based on the latest iterative operation result of the information state of the first sensor node.

Description

Technical Field

[0001] This application relates to the technical field of transmission bearers, and more particularly to a data processing method, electronic device, storage medium, and program product. Background Technology

[0002] In scenarios such as cloud-based autonomous driving systems, the collaborative work of multiple sensor nodes is required to achieve a large amount of data transmission and computation. However, in related technologies, redundant data transmission and high computational complexity are prone to occur, which is not conducive to quickly and accurately determining the predicted status of sensor nodes. Summary of the Invention

[0003] This application provides a data processing method, an electronic device, a storage medium, and a program product.

[0004] The technical solution of this application embodiment is implemented as follows: This application provides a data processing method, the method comprising: When the information state of the first sensor node meets the triggering condition for data transmission, the first sensor node and its neighboring sensor nodes transmit their respective sensor data to each other. Based on the sensor data of the first sensor node and its neighboring sensor nodes, the information state of the first sensor node is iteratively calculated. When the difference between the information states obtained from two adjacent iterations is less than or equal to a specified threshold, the iterative calculation process for determining the information state of the first sensor node ends. Based on the latest iterative calculation result of the information state of the first sensor node, the predicted state of the first sensor node at a future time is determined.

[0005] This application also provides an electronic device, which includes a processor and a memory for storing a computer program that can run on the processor; wherein the processor is used to run the computer program to perform any of the above-described data processing methods.

[0006] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements any of the above-described data processing methods.

[0007] This application also provides a computer program product, including a computer program or computer executable instructions, which, when executed by a processor, implements any of the above-described data processing methods.

[0008] As can be seen, the embodiments of this application can only transmit sensor data between the first sensor node and its neighboring sensor nodes when the information state of the first sensor node meets the triggering condition for data transmission. This helps to avoid the transmission of redundant data to a certain extent. Furthermore, it is not necessary to determine the predicted state of the first sensor node in the future based on the result of each iteration of the information state of the first sensor node. Instead, the predicted state of the first sensor node in the future can be determined when the difference between the information states obtained from two adjacent iterations is less than or equal to a specified threshold. This can reduce the computational complexity of the predicted state of the first sensor node to a certain extent, thereby facilitating the rapid and accurate determination of the predicted state of the sensor node. Attached Figure Description

[0009] Figure 1 This is a flowchart of a data processing method according to an embodiment of this application; Figure 2 A schematic diagram illustrating the connectivity of multiple sensor nodes provided in an embodiment of this application; Figure 3 A comparative schematic diagram of various state estimation curves provided for embodiments of this application and related technologies; Figure 4 A comparative diagram showing the number of iterations provided in the embodiments of this application; Figure 5 This is a schematic diagram of the structure of the data processing apparatus according to an embodiment of this application; Figure 6 This is a schematic diagram of the composition structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0010] Looking back at the history of nonlinear system filtering, the Extended Kalman Filter (EKF) algorithm was the first to be proposed. This algorithm mainly utilizes Taylor series expansion to linearize the nonlinear system and achieve the filtering process. However, this algorithm has a large error, and the Jacobian matrix of the system is difficult to solve. To overcome the shortcomings of the EKF algorithm, the Unscented Kalman Filter (UKF) algorithm was proposed. This algorithm uses the idea of ​​Unscented Transformation (UT) to overcome the limitations of the EKF algorithm. As research has deepened, the study of nonlinear filtering has expanded from low-dimensional systems to high-dimensional systems. The increase in dimensionality has brought difficulties to nonlinear filtering. The calculation of the covariance matrix in the UKF algorithm is prone to nondefinite cases, thus failing to guarantee the stability of the filtering system.

[0011] In related technologies, the Cubature Kalman Filter (CKF) algorithm has been proposed. It utilizes a third-order spherical radial criterion, using volume points to obtain the state mean and covariance matrix. In high-dimensional systems, this algorithm can maintain system stability while maintaining high filtering accuracy; it is currently widely used. The Cubature Information Kalman Filter (CIF) algorithm is an information form of CKF, possessing equivalent accuracy. In the information fusion process, CIF, being an information form, omits certain calculations, making it simpler and more convenient to use.

[0012] In scenarios such as cloud-based autonomous driving systems, the collaborative operation of multiple sensor nodes places extremely high demands on network bandwidth, especially when transmitting large amounts of data, which can easily lead to network congestion or anomalies. These issues pose challenges to the design of traditional estimation algorithms. By analyzing knowledge related to distributed filtering estimation, consensus filtering algorithms, and event-triggered mechanisms in nonlinear systems, it can be concluded that the consensus-based distributed CIF algorithm performs exceptionally well in improving estimation accuracy. However, despite its high accuracy, its computational complexity and network transmission burden also increase accordingly.

[0013] The first technical solution in this study proposes an event-triggered heading tracking control scheme for unmanned surface vessels (USVs) based on adaptive dynamic programming. This scheme aims to address the issues of wear and excessive energy consumption during heading control by incorporating an event-triggered mechanism. It introduces an Actor-Critic fuzzy logic system and establishes a weight update law to achieve adaptive dynamic programming heading tracking control under event triggering. However, this scheme has some limitations. For example, the computational complexity of the algorithm increases after introducing the Actor-Critic fuzzy logic system. For USV control systems with high real-time requirements, this may lead to high computational resource demands, especially in resource-constrained embedded environments, affecting the system's response speed.

[0014] In the second technical solution related to this technology, an event-triggered predictive control trajectory tracking method and system for unmanned aerial vehicles (UAVs) are proposed. This solution aims to address the problems of high computational complexity and excessive data transmission frequency in UAV trajectory tracking. This solution combines event triggering with a mechanism-based predictive control algorithm. By utilizing particle swarm optimization (PSO) to optimize the control parameters of the triggering mechanism, the system resource consumption is reduced. However, PSO is prone to getting trapped in local optima when dealing with complex high-dimensional problems, especially when multiple local extrema exist in the search space, making it difficult to find the global optimum.

[0015] In applications such as autonomous driving, drones, and ships, event-triggered mechanisms using related technologies can reduce data transmission. However, in some scenarios, this can still lead to the transmission of a large amount of redundant data, especially under complex network conditions. Frequent transmissions can unnecessarily consume bandwidth resources, increasing reliance on the cloud and exacerbating network bandwidth consumption, making it ineffective in situations with limited bandwidth. To reduce communication volume, event-triggered mechanisms typically rely on preset trigger thresholds. However, choosing an appropriate threshold is challenging; a threshold that is too low can lead to unnecessary communication, while a threshold that is too high may affect data timeliness and thus estimation accuracy.

[0016] In summary, related technologies are prone to problems such as redundant data transmission and high computational complexity, which are not conducive to quickly and accurately determining the forecast status of sensor nodes.

[0017] To address the technical problems existing in related technologies, this application proposes a technical solution based on embodiments. In this application embodiment, a dual event triggering mechanism can be introduced. This mechanism effectively reduces redundant data transmission and lowers network bandwidth consumption by setting innovative triggering conditions. Combining this mechanism with consensus algorithms and the CIF algorithm forms a highly efficient cloud-based nonlinear distributed estimation algorithm for autonomous driving. This algorithm can maintain high estimation accuracy while significantly reducing network burden by minimizing unnecessary data transmission, making it more suitable for environments with limited bandwidth and requiring real-time response (such as autonomous driving environments).

[0018] The embodiments of this application can be applied to scenarios such as smart cities, data transmission, the Internet of Things (IoT), and cloud computing. For example, the embodiments of this application can be applied to systems such as cloud computer internal software integration, intelligent transportation system platforms, smart cities, and mobile computing, as well as edge artificial intelligence (AI) systems, IoT systems, and industrial automation systems, and have broad market prospects.

[0019] In one exemplary application scenario, cloud-based data processing capabilities can be integrated with intelligent transportation systems to achieve accident warnings and route planning, improving urban traffic conditions. The solution adopted in this application embodiment can build a flexible business model, providing users with a pay-as-you-go resource usage model by offering route planning services, reducing the maintenance costs of information technology (IT) infrastructure, and further improving system performance and user experience, providing users with more efficient, flexible, and intelligent services. In another exemplary application scenario, cloud computers can provide an autonomous driving simulation environment. Using the technical solution of this application embodiment, a large amount of autonomous driving-related data processing and training tasks can be transferred to the cloud, supporting strict data isolation and encryption to ensure the data security of the autonomous driving system. Autonomous driving involves a large amount of user data and vehicle data; cloud-based security mechanisms can effectively prevent data leakage and attacks. The embodiments of this application will be further described in detail below with reference to the accompanying drawings and examples. It should be understood that the embodiments provided herein are merely illustrative of the embodiments of this application and are not intended to limit the embodiments of this application. Furthermore, the embodiments provided below are some embodiments for implementing this application, and not all embodiments for implementing this application. Unless otherwise specified, the technical solutions described in the embodiments of this application can be implemented in any combination.

[0020] It should be noted that, in the embodiments of this application, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a method or apparatus that includes a list of elements includes not only the elements expressly described, but also other elements not expressly listed, or elements inherent to implementing the method or apparatus. Without further limitations, an element defined by the phrase "comprising a..." does not exclude the presence of other related elements (e.g., steps in the method or units in the apparatus, such as portions of circuitry, processors, programs, or software, etc.) in the method or apparatus that includes that element.

[0021] The data processing method provided in this application includes a series of steps, but the data processing method provided in this application is not limited to the steps described. Similarly, the data processing apparatus provided in this application includes a series of modules, but the apparatus provided in this application is not limited to the modules explicitly described, and may also include modules that need to be set up for obtaining relevant information or processing based on information.

[0022] This application proposes a data processing method that can be applied to multiple sensor nodes, enabling data interaction between the multiple sensor nodes and uploading data to the cloud.

[0023] Figure 1 This is a flowchart of a data processing method according to an embodiment of this application, such as... Figure 1 As shown, the process includes: Step 101: When the information status of the first sensor node meets the triggering condition for data transmission, the first sensor node and its neighboring sensor nodes transmit their respective sensor data to each other.

[0024] Here, the first sensor node can be any one of multiple sensor nodes. For example, the first sensor node can be a sensor installed on the vehicle in an autonomous driving scenario, and it can be used to collect data such as position and speed. Neighbor sensor nodes are those nodes that are adjacent to the first sensor node on the communication link among the multiple sensor nodes. The first sensor node and neighbor sensor nodes can directly interact with each other without needing to go through other sensor nodes as intermediaries.

[0025] The information state of the first sensor node at the current moment can be determined based on the prediction state and prediction error covariance of the first sensor node at the previous moment. When the information state of the first sensor node meets the triggering condition for data transmission, the first sensor node can transmit its own sensor data to neighboring sensor nodes, and neighboring sensor nodes can also transmit their own sensor data to the first sensor node.

[0026] In this embodiment of the application, the sensor data can be data used to estimate the predicted state of the sensor node, and the predicted state of the sensor node can represent the prediction information of the sensor node for data to be collected at future times.

[0027] In this embodiment, the triggering condition for data transmission can be set and adjusted according to the actual situation. If the information status of the first sensor node does not meet the triggering condition for data transmission, then the first sensor node and its neighboring sensor nodes will not transmit data, which helps to avoid the transmission of redundant data to a certain extent.

[0028] Step 102: Based on the sensor data of the first sensor node and its neighboring sensor nodes, perform iterative calculations on the information state of the first sensor node.

[0029] Step 103: When the difference between the information states obtained from two adjacent iterations is less than or equal to a specified threshold, the iterative calculation process for determining the information state of the first sensor node ends. Based on the latest iterative calculation result of the information state of the first sensor node, the predicted state of the first sensor node at a future time is determined.

[0030] Here, the specified threshold can be a fixed threshold parameter. In some embodiments, if the difference between the information states obtained from two adjacent iterations is greater than the specified threshold, step 102 is repeated until the difference between the information states obtained from two adjacent iterations is less than or equal to the specified threshold. In this way, the information state of the first sensor node can become more stable, and based on the latest iteration result of the information state of the first sensor node, the predicted state of the first sensor node at future times can be determined more reliably.

[0031] In practical applications, steps 101 to 103 can be implemented based on a processor and a communication interface. The processor can be at least one of the following: Application Specific Integrated Circuit (ASIC), Digital Signal Processor (DSP), Digital Signal Processing Device (DSPD), Programmable Logic Device (PLD), Field Programmable Gate Array (FPGA), Central Processing Unit (CPU), controller, microcontroller, and microprocessor.

[0032] As can be seen, the embodiments of this application can only transmit sensor data between the first sensor node and its neighboring sensor nodes when the information state of the first sensor node meets the triggering condition for data transmission. This helps to avoid the transmission of redundant data to a certain extent. Furthermore, it is not necessary to determine the predicted state of the first sensor node in the future based on the result of each iteration of the information state of the first sensor node. Instead, the predicted state of the first sensor node in the future can be determined when the difference between the information states obtained from two adjacent iterations is less than or equal to a specified threshold. This can reduce the computational complexity of the predicted state of the first sensor node to a certain extent, thereby facilitating the rapid and accurate determination of the predicted state of the sensor node.

[0033] The embodiments of this application can be applied to various scenarios such as cloud-based autonomous driving scenarios where sensor nodes are deployed. The following uses formulas (1) and (2) to perform model analysis on cloud-based autonomous driving scenarios.

[0034] (1) (2) in, It is discrete time; Let be the state vector at time t; for Time of the first The observation vector of each sensor node. Indicates the relationship with the first The parameters corresponding to each sensor node An integer greater than or equal to 1; , This indicates the total number of sensor nodes. and They represent The nonlinear state function at time t and the t The observation function of each sensor node For integers greater than or equal to 1, It is an integer greater than or equal to 1. yes Time-based process noise, It is an integer greater than or equal to 1. for Time of the first The observation noise of each sensor node.

[0035] In the context of cloud-based autonomous driving, hypotheses 1 and 2 can be proposed. Hypothesis 1 includes: Gaussian white noise. and noise Uncorrelated, zero mean, Gaussian white noise and noise The variances can be expressed as follows: and Gaussian white noise and noise It satisfies formula (3).

[0036] (3) in, yes Time-based process noise, express Time of the first The observation noise of each sensor node. To represent the spatial independence of observation noise, in the context of autonomous driving in the cloud, we can assume that the observation noise of different sensor nodes is independent of each other. To represent the time independence of observation noise, in the context of autonomous driving in the cloud, we can assume that the observation noise at different times is uncorrelated.

[0037] Assumption 2 includes: the observation vector of each sensor node at the initial time. irrelevant and ,and It satisfies formula (4).

[0038] (4) in, Indicates and The corresponding expectation, Indicates and The corresponding covariance matrix.

[0039] In this embodiment of the application, when the information state of the first sensor node meets the triggering condition for data transmission, a consensus mechanism can be used to transmit the respective sensor data between the first sensor node and its neighboring sensor nodes.

[0040] Consensus mechanisms originated in blockchain architecture and aim to achieve state consistency among multiple nodes in a distributed system through specific rules. In the research of distributed systems, asynchronous communication and fault-tolerant protocols are used to ensure information transmission and state synchronization between nodes. The introduction of consensus mechanisms not only improves system efficiency but also ensures data security.

[0041] In this embodiment, the wireless sensor network constructed through sensor nodes can be represented as an undirected graph. An undirected graph controls the set of nodes. and edge set ,in, This represents the total number of nodes in the undirected graph. Each node in the undirected graph has its neighboring nodes, which together form the set of its neighbors. When using a consensus mechanism to transmit data, each node broadcasts the information it acquires to its neighboring nodes at each sampling time, thus achieving state consistency. It can be stipulated that information exchange only occurs between a node and its neighboring nodes.

[0042] In distributed network estimation systems, consensus mechanisms are used to determine rules for data consistency, process node data according to specific rules, and ensure that the data in the entire system is unified. In some embodiments, a weighted average consensus mechanism can be introduced to reduce the transmission of duplicate data. The weighted average consensus mechanism can be described by formulas (5) and (6).

[0043] (5) (6) in, Indicates the cycle time. It is the first Each sensor node in The state at any given moment (e.g., position, velocity, etc.). It is the first Each sensor node in The state at any given moment, It is the first Each sensor node in The state at any given moment. for Time and the The sensor node and the first The consensus weight matrix corresponding to each sensor node. For the first The set of neighboring nodes of a sensor node. For the first Each sensor node in The degree of time, For the first Each sensor node in The degree of time. The elements in the consensus weight matrix are the original data without modification. The rows of the consensus weight matrix are random. For example, the sum of the elements in each row of the consensus weight matrix is ​​1, but the value of each element in each row of the consensus weight matrix is ​​a random number.

[0044] In this embodiment, the triggering condition for data transmission can be set according to actual conditions. In some embodiments, the triggering condition for data transmission can be determined based on the information state error of the first sensor node and the sum of the information state errors of its neighboring nodes. Here, the information state error represents the difference between the information state of the first sensor node at the current moment and the information state of the first sensor node at the time of the last data transmission; the sum of the information state errors represents the sum of the differences between the information states of each of the first sensor node's neighboring sensor nodes and the first sensor node at the time of the last data transmission.

[0045] For example, under the same conditions, the dynamic event triggering mechanism can achieve a lower communication frequency than the static event triggering mechanism while ensuring estimation performance, thereby saving resources more effectively.

[0046] During the data transmission process of sensor nodes, the dynamic triggering conditions for data transmission can be set using the state information of the local sensor node and neighboring sensor nodes in the CIF algorithm. The dynamic triggering conditions for real-time data transmission can be based on parameters. Settings, parameters The calculation formula is formula (7).

[0047] (7) in, , For example, the weight matrix, It can be the consensus weight matrix described above. for The latest data transmission trigger time before the current time. For the first Each sensor node in Information status at any given moment For the first Each sensor node in Information status at any given moment. For the first Each neighboring sensor node of the first sensor node and the second sensor node Each sensor node in The sum of the differences in information state at any given time. , For the first Each sensor node in Information status at any given moment For the first Each sensor node in Dynamic threshold parameters at any given time.

[0048] Here, when the parameter When it equals 1, it can be considered that the first... Each sensor node in The information state at any given time satisfies the triggering condition for data transmission; when the parameter When it equals 0, it can be considered that the first... Each sensor node in The information state at any given time does not meet the triggering conditions for data transmission.

[0049] It can be seen that the information state error of the first sensor node is used to represent the variation range of the information state of the first sensor node, and the information state error of the neighboring nodes represents the total difference between each neighboring sensor node and the first sensor node. Therefore, the embodiments of this application can reasonably determine the triggering conditions for data transmission based on considering the variation range of the information state of the first sensor node and the total difference between each neighboring sensor node and the first sensor node.

[0050] In some embodiments, the triggering conditions for data transmission can be adjusted based on the difference between the information state of the first sensor node at the current moment and the information state of the first sensor node during the last data transmission.

[0051] It can be seen that the difference between the information state of the first sensor node at the current moment and the information state of the first sensor node at the last data transmission can, to a certain extent, reflect the current change range of the information state of the first sensor node. Therefore, in this embodiment, based on considering the current change range of the information state of the first sensor node, the triggering conditions for data transmission corresponding to the first sensor node at the current moment can be reasonably adjusted to achieve a reasonable adjustment of the triggering frequency of data transmission.

[0052] For example, it can be based on This allows for adjustments to the triggering conditions for data transmission. The pattern of change can be represented by formula (8).

[0053] (8) in, This is the preset value. It can be seen that... The value of the threshold parameter directly affects the convergence speed of data transmission. The triggering mechanism for data transmission can be as follows: when the difference between the latest transmitted data and the current estimation error is large, the dynamic threshold parameter should be decreased to increase the triggering frequency of data transmission events; conversely, when the difference between the newly transmitted data and the current estimation error is small, the dynamic threshold parameter needs to be increased to reduce the triggering frequency of data transmission events. By designing the dynamic threshold parameter in this way, the amount of data transmission in the consensus algorithm can be reduced, avoiding resource waste.

[0054] If there are no other disturbances, the triggering of data transmission will remain in a steady state; that is, the dynamic threshold parameter will eventually converge to a stable value. For example, when... When equal to 0, The value of is a constant between 0 and 1, and the data transmission triggering mechanism gradually approaches the periodic event triggering mechanism. In this embodiment, by using the correlation between the information state error of the first sensor node and the sum of the information state errors of neighboring nodes, the triggering condition for data transmission can be set without additional calculation. The advantage of dynamic event triggering is that it minimizes data transmission.

[0055] In some embodiments, the iterative calculation of the information state of the first sensor node can be performed based on sensor data from the first sensor node and its neighboring sensor nodes, using a volumetric information weighted consistency filtering method. By employing this method, the efficiency of the iterative calculation of the information state can be improved, thereby enhancing the real-time data processing capabilities of the sensor node.

[0056] In this embodiment, the convergence performance and number of iterations of the volume information weighted consistency filtering method are closely related to the differences between the sensors.

[0057] In one example, the state information and related information matrix errors from different sensors are processed using the same parameters. In another example, a consistency weight matrix can be designed to replace these parameters, allowing the consistency system to converge faster. This weight matrix can be obtained using the Metropolis-Hastings weight calculation method. The consistency weight matrix can be as described above. In one example, let The number of iterations representing the information state of a sensor node, in As the value of gradually increases, the iterative calculation process of the information state of the sensor node can be represented by formulas (9) and (10).

[0058] (9) (10) in, Indicates the first Each sensor node in Time passes The information matrix obtained from the next iteration. Indicates the first Each sensor node in Time passes The information state vector obtained from the iteration operation.

[0059] In some embodiments, when the information state of the first sensor node is the trace of a matrix, the information state of the first sensor node can be iteratively calculated based on the sensor data of the first sensor node and its neighboring sensor nodes.

[0060] It can be seen from formulas (9) and (10) that when As the expression approaches infinity, the information state vector and information matrix of each sensor node will converge to a linear minimum variance estimate. However, a major drawback of this filtering algorithm is that during the consensus iteration process, sensor nodes need to transfer data in matrix form, and computation is required. The inverse matrix of the information matrix. When the dimensionality and complexity of the iterative computation system are high, it will significantly increase the burden of computation and communication.

[0061] To reduce computation and communication costs and improve the feasibility of real-time applications, in some embodiments, the trace of the information matrix can be used instead of the information matrix for consistency iteration. For example, the sensor data of the first sensor node and the neighboring sensor nodes can be the information matrix and information state vector described above. When the trace of the information matrix is ​​used for consistency iteration, the iterative calculation process of the information state of the sensor node can be represented by formulas (11) to (14).

[0062] (11) (12) (13) (14) in, Indicates the first Each sensor node in Information matrix at time, Indicates the first Each sensor node in The information state vector at any given time. and For the information state of the sensor node, in the initial calculation and Afterwards, it can be made , Then, formulas (13) and (14) can be used to... and Perform iterative calculations.

[0063] As the number of consensus iterations increases, the effectiveness of the iterations gradually improves, and the average cumulative mean square error decreases. Without constraints, a significant number of iterations are typically required to achieve data consistency across all sensor nodes. By introducing an event-triggered mechanism into the consensus iteration algorithm, an appropriate number of iterations can be selected at each moment, thereby reducing the computational burden. In some embodiments, the termination condition for the iterative operation of the information state can be determined using formula (15).

[0064] (15) in, To specify a threshold, For the first Each sensor node in Time passes The information state obtained from the next iteration. , . and It can be calculated by referring to formula (14).

[0065] here, When the value equals 0, it can be considered that the iterative calculation process of the sensor node's information state needs to be terminated. When the value is 1, it can be assumed that the iterative calculation process of the sensor node's information state needs to continue.

[0066] In real-world scenarios, based on the collaborative work of sensor nodes, the information state values ​​of different sensor nodes gradually converge during the iterative process of information state updates. Therefore, iterative calculations can be stopped when the error value of the information state during the iteration process falls below a specified threshold. This information state iterative calculation mechanism is suitable for complex systems and, compared to consistency iterations without event triggering mechanisms, can reduce the number of iterations while maintaining performance.

[0067] As can be seen, the embodiments of this application can perform iterative operations on the trace of a matrix, which can reduce computational and communication costs to a certain extent compared with the scheme of performing iterative operations on a matrix.

[0068] In some embodiments, to determine the information state of the first sensor node, before iteratively calculating its information state, the prediction error covariance of the first sensor node can be determined based on the latest determined prediction state; and the information state of the first sensor node can be determined based on the prediction error covariance. It can be seen that by using the latest determined prediction state of the first sensor node, the prediction error covariance of the first sensor node can be determined relatively accurately, and thus, based on the prediction error covariance, the information state of the first sensor node can be determined more accurately.

[0069] In this embodiment, the target state can be determined by fusion state estimation, that is, by combining the information of multiple sensor nodes in time or space, a more accurate target state estimate can be obtained. In some embodiments, a central fusion method can be used to estimate the information state of sensor nodes. When using the central fusion method to estimate the information state of sensor nodes, each sensor node is treated as a central node of its neighbors, and then the information matrix and information state vector of the central node are added together. For example, the target state can be determined by formulas (16) and (17). The information matrix and information state vector of each sensor node.

[0070] (16) (17) in, Indicates the first Each sensor node in The forecast covariance error at time, Indicates the first Each sensor node in The relevant information matrix at time, Indicates the first Each sensor node in The relevant information matrix at time, Indicates the first Each sensor node in Information matrix at any given time. Indicates the first Each sensor node in Information status at any given moment Indicates the first Each sensor node in Information status at any given moment According to Calculations show that Indicates the first Each sensor node in The information state vector at any given time.

[0071] This application proposes a consensus-based distributed consensus filtering algorithm based on dual event triggering. In some embodiments, the CIF algorithm is improved by introducing a consensus mechanism and a dual event triggering mechanism.

[0072] For example, the process of a consensus distributed consensus filtering algorithm based on dual event triggering may include the following steps: Step S1: Time update.

[0073] In this step, the solution can be obtained using formulas (18) to (20). Each sensor node in Forecast state estimation at time Forecast covariance error and .

[0074] (18) (19) (20) in, Indicates the first Each sensor node in The first time generated One sigma point, The predicted sigma point is the result of propagation through the system state equation. Let be the system process noise covariance matrix.

[0075] Step S2: Measurement update.

[0076] In this step, the observed information can be calculated according to formulas (21) to (23). , and .

[0077] (twenty one) (twenty two) (twenty three) in, This represents the observation vector calculated from the state estimation before the measurement update. Indicates the first The cross-covariance matrix between the state and measurement of each sensor node Indicates the first Measurement noise covariance matrix of each sensor node.

[0078] Step S3: Execute consensus distributed consistency filtering based on dual event triggering.

[0079] In this step, during the data transmission stage, parameters can be used... Sure Dynamic triggering conditions for real-time data transmission, when parameters When it equals 0, it can be considered that the first... Each sensor node in If the information state at a given time does not meet the triggering conditions for data transmission, then the calculation can be performed according to formulas (24) and (25). and the Each sensor node in Time forecast status .

[0080] (twenty four) (25) When parameter When it equals 1, it can be considered that the first... Each sensor node in The information state at time t satisfies the triggering condition for data transmission. Each sensor node can send its own sensor data to all neighboring sensor nodes and receive sensor data from all neighboring sensor nodes. In one example, the first... Sensor data from a single sensor node can include and , and It can be calculated according to formulas (26) and (27).

[0081] (26) (27) in, Indicates the first Each sensor node in Information matrix at time, Indicates the first Each sensor node in The information state vector at any given time.

[0082] In calculation and After that, initialization can be performed using formulas (28) and (29).

[0083] (28) (29) Then, a consensus mechanism can be used to determine whether to continue iterative calculations of the sensor node's information state. For example, when... When the value equals 1, it can be assumed that the iterative calculation process of the sensor node's information state needs to continue. When the value equals 0, it can be considered that the iterative calculation process of the information state of the sensor node needs to be terminated. The iterative calculation process of the information state of the sensor node can be represented by formulas (13) and (14).

[0084] After the iterative calculation process for determining the information state of the sensor node is completed, the current iteration count of the sensor node's information state can be recorded as follows: Then, event-triggered consensus filtering can be performed using formulas (30) to (32).

[0085] (30) (31) (32) The beneficial effects of the embodiments of this application will be illustrated by simulation analysis below.

[0086] Considering cloud-based autonomous driving applications, we use car modeling and simulation training data. Initial parameters and errors can be defined later based on actual conditions. The system state can be represented as follows: ,in and They can be represented as direction and Position information in the direction, and They are respectively direction and velocity information in direction, Let be the turning rate. The state-space model of the system can be represented by formula (33).

[0087] (33) The initial position of the target is known to be , and The speed of movement in both directions is Scanning time In a wireless sensor network composed of sensor nodes, the initial values ​​of the 1st to 10th sensor nodes are respectively... , , , , , , , , and , Figure 2 This is a schematic diagram illustrating the connectivity of 10 sensor nodes provided in an embodiment of this application.

[0088] For each sensor node, the data is observed, and the system's observation equation can be formula (34).

[0089] (34) in,( , ) indicates the first Each sensor node in direction and Initial value in direction, The root mean square error (RMSE) of the position, which represents the state estimation performance at a given time, can be expressed by formula (35).

[0090] (35) in, and They are Time of the first The true and estimated states of the system in the Monte Carlo simulation were obtained. In the simulation, the sampling time was 600 steps, and a total of 100 Monte Carlo experiments were performed. The initial state of the system was also determined. Initial covariance The variance of the system noise is Observation noise The variance is ,in .

[0091] A consensus-based distributed filtering algorithm triggered by dual events is used to estimate the predicted state of each sensor node, yielding state estimation curves for each node. These curves can be selected for subsequent comparison, with other sensor nodes exhibiting similar trends. Alternatively, a tracking algorithm for a single node, provided by relevant technologies, can be used to determine the state estimation curves for a subset of sensor nodes.

[0092] Figure 3 A comparative schematic diagram of various state estimation curves provided for embodiments of this application and related technologies, such as... Figure 3 As shown, DECIF-1 and DECIF-7 represent the state estimation curves of the first and seventh sensor nodes obtained using the scheme of this application embodiment, respectively, and n2 and n5 represent the state estimation curves of the second and fifth sensor nodes determined using the single-node tracking algorithm provided by related technologies, respectively. Figure 3 It can be seen that compared with the single-node tracking algorithms provided by related technologies, the DECIF algorithm has higher tracking accuracy, and in the triggered state, the DECIF algorithm can ensure consistent transmission among all sensor nodes. It can capture data at the 200th step of the sampling time. Here, the AMSE value is the square of the RMSE value mentioned above. For node 2, the triggering mechanism of dynamic events is analyzed. A comparison is made between the scheme using fixed threshold parameters and the scheme using dynamic threshold parameters, resulting in the communication frequency and... (Table 1 shows the values).

[0093] Table 1

[0094] As can be seen from Table 1, when using a fixed threshold parameter At that time, the communication frequency was low and the amount of data transmitted was small. When the value is large, the estimation accuracy decreases significantly, affecting system performance; when At that time, while increased communication frequency led to improved estimation accuracy, it did not significantly reduce resource consumption; At that time, a good balance was achieved between communication frequency and estimation accuracy. Compared to When the threshold value is 1.02, the scheme using a fixed threshold parameter and the scheme using a dynamic threshold parameter have similar event trigger probabilities, but the former has higher estimation accuracy. The dynamic threshold design focuses on reducing communication resource waste while ensuring performance, thus exhibiting superior performance.

[0095] For the analysis of the consistency iteration process, the threshold parameter of the event triggering mechanism. By setting the values ​​to 0.005, 0.01, 0.05, and 0.1 respectively, we can obtain the average number of iterations shown in Table 2.

[0096] Table 2

[0097] As shown in Table 2, the number of iterations required decreases as the threshold parameter increases. (The last sentence appears to be incomplete and possibly refers to a specific threshold parameter.) When the value is 0.01, comparing the number of iterations before the event-triggered mechanism was introduced into the consistency filtering algorithm with the number of iterations after the event-triggered mechanism was introduced into the consistency filtering algorithm, we can obtain... Figure 4 The diagram showing the comparison of iteration counts is as follows. Figure 4 In the diagram, CIF represents the number of iterations before the event-triggered mechanism is introduced into the consistency filtering algorithm, and DECIF represents the number of iterations after the event-triggered mechanism is introduced into the consistency filtering algorithm. Figure 4 It can be seen that without the introduction of an event-triggered mechanism, the consistency filtering algorithm typically requires a large number of iterations to ensure consistency across nodes. After the event-triggered mechanism is introduced, the number of iterations is significantly reduced, and the settings become more flexible, greatly alleviating the computational burden of consistency filtering.

[0098] The DECIF algorithm of this application embodiment is compared with the DICIF algorithm, DCIF-WAC algorithm and ET-DCIF-WAC algorithm proposed in related technologies in Table 3 below.

[0099] Table 3

[0100] The DCIF-WAC algorithm is a traditional consensus-based distributed CIF algorithm that achieves the highest accuracy under full trigger conditions, but it involves significant redundant data transmission. The ET-DCIF-WAC algorithm is an improvement on DCIF-WAC, and its performance has been proven to outperform the DUKF-WAC algorithm. Comparing the DECIF algorithm with these algorithms reveals that the DECIF algorithm has a lower trigger rate than the ET-DCIF-WAC algorithm, while its average RMSE value is similar to that of the ET-DCIF-WAC algorithm. This indicates that the DECIF algorithm can provide relatively excellent estimation performance while maintaining a low trigger rate.

[0101] This application proposes a cloud-based nonlinear distributed estimation algorithm for autonomous driving based on a dual-event triggering mechanism. By introducing innovative triggering conditions, this method effectively reduces redundant data transmission and network bandwidth consumption. Furthermore, by combining a consensus algorithm with the CIF algorithm, it improves the system's estimation efficiency and real-time processing capabilities. This invention optimizes related technologies to ensure excellent performance in autonomous driving scenarios with limited bandwidth and the need for real-time response.

[0102] This application proposes a method for reducing redundant data transmission without sacrificing estimation accuracy by setting two levels of triggering conditions. An adaptive triggering condition is designed for data transmission, adjusted based on the information vector error function, and redundant transmission is removed using the information state function. A static event trigger is set in the consensus iteration phase, utilizing the time-instance information vector error to reduce computational load. Therefore, this application proposes a method for reducing redundant data transmission based on a dual event triggering mechanism, combining adaptive triggering conditions with event triggers to reduce data transmission and computational load.

[0103] In this embodiment, the weighted average consensus algorithm is combined with the CIF algorithm to ensure data consistency and efficient estimation among sensor nodes. The consistency iteration stage uses the trace of the information matrix instead of the information matrix itself for consistency iteration, reducing data transmission while maintaining high estimation accuracy. This ensures the system's performance in real-time processing and efficient computation, and optimizes the system's real-time processing and computational performance.

[0104] This application proposes an overall scheme that can be applied to tracking estimation scenarios. The overall design of this scheme will not get stuck in local optima, and the computational complexity is lower than that of related model algorithms. The system has good overall adaptability and flexibility.

[0105] Compared with related technologies, the improved dual-event triggering mechanism proposed in this application effectively alleviates the problem of high data transmission burden, reduces the original computational burden, eliminates more redundant transmissions, and is more compatible with cloud-based autonomous driving systems. The nonlinear distributed estimation algorithm proposed in this application, which combines a weighted average consensus algorithm with the CIF algorithm, is simple to implement and has stable performance; compared with related technologies, it has better adaptability to dynamics and is more stable.

[0106] Those skilled in the art will understand that, in the above-described method of the specific implementation, the order in which each step is written does not imply a strict execution order and does not constitute any limitation on the implementation process. The specific execution order of each step should be determined by its function and possible internal logic.

[0107] Based on the data processing method proposed in the foregoing embodiments, this application also proposes a data processing apparatus, such as... Figure 5 As shown, the device includes: The first processing module 501 is used to transmit sensor data between the first sensor node and its neighboring sensor nodes when the information status of the first sensor node meets the triggering condition for data transmission. The second processing module 502 is used to perform iterative calculations on the information state of the first sensor node based on sensor data of the first sensor node and neighboring sensor nodes; when the difference between the information states obtained from two adjacent iterative calculations is less than or equal to a specified threshold, the iterative calculation process of the information state of the first sensor node ends; and based on the latest iterative calculation result of the information state of the first sensor node, the predicted state of the first sensor node at a future time is determined.

[0108] In some embodiments, the first processing module 501 is further configured to determine the triggering condition for the data transmission based on the information state error of the first sensor node and the sum of the information state errors of its neighboring nodes, wherein the information state error represents the difference between the information state of the first sensor node at the current moment and the information state of the first sensor node during the last data transmission; the sum of the information state errors represents the sum of the differences between the information state of each of the first sensor node's neighboring sensor nodes and the information state of the first sensor node during the last data transmission.

[0109] In some embodiments, the first processing module 501 is further configured to adjust the triggering conditions for the data transmission based on the difference between the information state of the first sensor node at the current moment and the information state of the first sensor node during the last data transmission.

[0110] In some embodiments, the second processing module 502 is specifically used to perform iterative calculations on the information state of the first sensor node based on sensor data of the first sensor node and neighboring sensor nodes when the information state of the first sensor node is the trace of a matrix.

[0111] In some embodiments, the second processing module 502 is further configured to, before performing iterative calculations on the information state of the first sensor node, determine the prediction error covariance of the first sensor node based on the latest determined prediction state of the first sensor node; and determine the information state of the first sensor node based on the prediction error covariance of the first sensor node.

[0112] In some embodiments, the second processing module 502 is specifically used to perform iterative calculations on the information state of the first sensor node based on the sensor data of the first sensor node and neighboring sensor nodes, and using a volumetric information weighted consistency filtering method.

[0113] In some embodiments, the second processing module 502 is further configured to repeatedly execute the step of iteratively calculating the information state of the first sensor node based on sensor data of the first sensor node and neighboring sensor nodes when the difference between the information states obtained from two adjacent iterative calculations is greater than the specified threshold, until the difference between the information states obtained from two adjacent iterative calculations is less than or equal to the specified threshold.

[0114] In practical applications, the first processing module 501 and the second processing module 502 can be implemented based on a processor and a communication device.

[0115] It should be noted that the description of the above device embodiments is similar to the description of the above method embodiments, and has similar beneficial effects. For technical details not disclosed in the device embodiments of this application, please refer to the description of the method embodiments of this application for understanding.

[0116] It should be noted that, in the embodiments of this application, if the above-described methods are implemented as software functional modules and sold or used as independent products, they can also be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the embodiments of this application, or the part that contributes to the prior art, 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 terminal, server, etc.) to execute all or part of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), magnetic disks, or optical disks. Thus, the embodiments of this application are not limited to any specific hardware and software combination.

[0117] Correspondingly, this application embodiment further provides a computer program product, the computer program product including computer executable instructions, which are used to implement any of the data processing methods provided in this application embodiment.

[0118] Accordingly, this application embodiment further provides a computer storage medium storing computer-executable instructions, which are used to implement any of the data processing methods provided in the above embodiments.

[0119] This application also provides an electronic device. Figure 6This is a schematic diagram of the composition structure of an electronic device provided in an embodiment of this application, as shown below. Figure 6 As shown, the electronic device 60 may include: Memory 601 is used to store executable instructions; The processor 602 is used to implement any of the above-described data processing methods when executing executable instructions stored in the memory 601.

[0120] The processor 602 mentioned above can be at least one of ASIC, DSP, DSPD, PLD, FPGA, CPU, controller, microcontroller, and microprocessor.

[0121] The aforementioned computer-readable storage medium and memory 601 may be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), a magnetic random access memory (FRAM), a flash memory, a magnetic surface memory, an optical disc, or a compact disc read-only memory (CD-ROM), etc.; it may also be various terminals that include one or any combination of the above-mentioned memories, such as mobile phones, computers, tablet devices, personal digital assistants, etc.

[0122] In some embodiments, the functions or modules of the apparatus provided in this application can be used to perform the methods described in the above method embodiments. The specific implementation can be referred to the description of the above method embodiments, and for the sake of brevity, it will not be repeated here.

[0123] The description of the various embodiments above tends to emphasize the differences between the various embodiments. The similarities or similarities between them can be referred to, and for the sake of brevity, they will not be repeated here.

[0124] The methods disclosed in the various method embodiments provided in this application can be arbitrarily combined to obtain new method embodiments without conflict.

[0125] The features disclosed in the various product embodiments provided in this application can be arbitrarily combined without conflict to obtain new product embodiments.

[0126] The features disclosed in the various method or device embodiments provided in this application can be arbitrarily combined without conflict to obtain new method or device embodiments.

[0127] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal (which may be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in the various embodiments of this application.

[0128] The embodiments of this application have been described above with reference to the accompanying drawings. However, this application is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of this application without departing from the spirit and scope of the claims. All of these forms are within the protection scope of this application.

Claims

1. A data processing method, characterized in that, The method includes: When the information state of the first sensor node meets the triggering condition for data transmission, the first sensor node and its neighboring sensor nodes transmit their respective sensor data to each other. Based on the sensor data of the first sensor node and its neighboring sensor nodes, the information state of the first sensor node is iteratively calculated. When the difference between the information states obtained from two adjacent iterations is less than or equal to a specified threshold, the iterative calculation process for determining the information state of the first sensor node ends. Based on the latest iterative calculation result of the information state of the first sensor node, the predicted state of the first sensor node at a future time is determined.

2. The method according to claim 1, characterized in that, The method further includes: Based on the information state error of the first sensor node and the sum of the information state errors of its neighboring sensor nodes, the triggering condition for the data transmission is determined. The information state error represents the difference between the information state of the first sensor node at the current moment and the information state of the first sensor node during the last data transmission. The sum of the information state errors represents the sum of the differences between the information state of each of the first sensor node's neighboring sensor nodes and the information state of the first sensor node during the last data transmission.

3. The method according to claim 1 or 2, characterized in that, The method further includes: The triggering conditions for the data transmission are adjusted based on the difference between the information state of the first sensor node at the current moment and the information state of the first sensor node during the last data transmission.

4. The method according to claim 1, characterized in that, The iterative calculation of the information state of the first sensor node based on sensor data from the first sensor node and its neighboring sensor nodes includes: When the information state of the first sensor node is the trace of a matrix, the information state of the first sensor node is iteratively calculated based on the sensor data of the first sensor node and its neighboring sensor nodes.

5. The method according to claim 1 or 4, characterized in that, Before performing iterative calculations on the information state of the first sensor node, the method further includes: Based on the latest determined forecast state of the first sensor node, the forecast error covariance of the first sensor node is determined; The information status of the first sensor node is determined based on the prediction error covariance of the first sensor node.

6. The method according to claim 1, characterized in that, Based on the sensor data of the first sensor node and its neighboring sensor nodes, iterative calculations are performed on the information state of the first sensor node, including: Based on the sensor data of the first sensor node and its neighboring sensor nodes, and using the volumetric information weighted consistency filtering method, the information state of the first sensor node is iteratively calculated.

7. The method according to claim 1, characterized in that, The method further includes: When the difference between the information states obtained from two adjacent iterations is greater than the specified threshold, the step of iteratively calculating the information state of the first sensor node based on the sensor data of the first sensor node and its neighboring sensor nodes is repeated until the difference between the information states obtained from two adjacent iterations is less than or equal to the specified threshold.

8. An electronic device, characterized in that, The electronic device includes a processor and a memory for storing computer programs capable of running on the processor; wherein, The processor is used to run the computer program to perform the method according to any one of claims 1 to 7.

9. A computer storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the computer program implements the method described in any one of claims 1 to 7.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the method described in any one of claims 1 to 7.

Images