Distributed data processing method and apparatus

Abstract

The application discloses a kind of distributed data processing method and device.Therein, the distributed data processing method includes: obtaining the information to be identified in the target area, wherein the information to be identified includes passenger trip information and / or driver driving information;In the case where the information to be identified includes the passenger trip information and the driver driving information, the input information to be identified is processed by a local decision model to obtain a target matching result;In the case where the information to be identified includes the driver driving information, the input information to be identified is processed by the local decision model to obtain cruise guidance information.Based on the technical scheme of the embodiment of the application, the accuracy of the determined target matching result and / or cruise guidance information can be improved.

Description

Technical Field

[0001] This invention relates to the field of intelligent cruise guidance technology in the context of vehicle networking, and particularly to a distributed data processing method and apparatus. Background Technology

[0002] While ride-sharing has become a mainstream mode of transportation, problems such as supply-demand imbalance and low vehicle utilization still exist. This is because drivers rely solely on their experience to cruise when they haven't received any orders. However, due to the varying distribution of vehicles and orders over time, drivers may not receive orders promptly, and passengers may wait for a long time without a successful match.

[0003] In existing technologies, route optimization algorithms and traffic flow prediction techniques from artificial intelligence are used to train decision-making models based on historical and real-time monitoring data. These models then provide drivers with matching information and / or cruise guidance information. However, areas vary in terms of activity levels, the number and distribution of drivers and passengers differ, and the customer base of each ride-sharing operator also varies. Therefore, the accuracy of matching information and / or cruise guidance information determined by the decision-making models is often poor. Summary of the Invention

[0004] This invention provides a distributed data processing method and apparatus to solve the technical problem of poor accuracy of matching information and / or cruise guidance information determined based on decision models.

[0005] According to one aspect of the present invention, a distributed data processing method is provided, wherein the method includes:

[0006] Acquire the information to be identified within the target area, wherein the information to be identified includes passenger travel information and / or driver driving information;

[0007] When the information to be identified includes the passenger travel information and the driver driving information, the input information to be identified is processed by a local decision model to obtain a target matching result. The local decision model is obtained by federated training of an initial decision model based on sample decision data from each edge. The sample decision data includes sample matching data and / or sample cruise guidance data.

[0008] When the information to be identified includes the driver's driving information, the input information to be identified is processed by the local decision model to obtain cruise guidance information, wherein the cruise guidance information includes cruise direction and / or cruise time.

[0009] According to another aspect of the present invention, a distributed data processing apparatus is provided, wherein the apparatus comprises:

[0010] The information acquisition module is used to acquire information to be identified within a target area, wherein the information to be identified includes passenger travel information and / or driver driving information;

[0011] The first information processing module is used to process the input information to be identified through a local decision model to obtain a target matching result when the information to be identified includes the passenger travel information and the driver driving information. The local decision model is obtained by federated training of an initial decision model based on sample decision data from each edge end. The sample decision data includes sample matching data and / or sample cruise guidance data.

[0012] The second information processing module is used to process the input information to be identified through the local decision model when the information to be identified includes the driver's driving information, to obtain cruise guidance information, wherein the cruise guidance information includes cruise direction and / or cruise time.

[0013] According to another aspect of the present invention, an electronic device is provided, the electronic device comprising:

[0014] At least one processor; and

[0015] A memory communicatively connected to the at least one processor; wherein,

[0016] The memory stores a computer program that can be executed by the at least one processor, which enables the at least one processor to perform the distributed data processing method according to any embodiment of the present invention.

[0017] According to another aspect of the present invention, a computer-readable storage medium is provided, the computer-readable storage medium storing computer instructions for causing a processor to execute and implement the distributed data processing method according to any embodiment of the present invention.

[0018] The technical solution of this invention involves acquiring information to be identified within the target area, wherein the information to be identified includes passenger travel information and / or driver driving information; when the information to be identified includes both passenger travel information and driver driving information, a local decision model processes the input information to obtain a target matching result, wherein the local decision model is obtained by federated training of an initial decision model based on sample decision data from each edge, wherein the sample decision data includes sample matching data and / or sample cruise guidance data; when the information to be identified includes the driver driving information, the local decision model processes the input information to obtain cruise guidance information, wherein the cruise guidance information includes cruise direction and / or cruise time. This improves the accuracy of the determined target matching result and / or cruise guidance information.

[0019] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of the present invention, nor is it intended to limit the scope of the invention. Other features of the invention will become readily apparent from the following description. Attached Figure Description

[0020] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0021] Figure 1 This is a flowchart of a distributed data processing method provided in Embodiment 1 of the present invention;

[0022] Figure 2 This is a flowchart of a distributed data processing method provided in Embodiment 2 of the present invention;

[0023] Figure 3 This is an architecture diagram of a distributed data processing system provided according to an embodiment of the present invention;

[0024] Figure 4 This is an architecture diagram for training a local decision-making model according to an embodiment of the present invention;

[0025] Figure 5 This is an overall signaling diagram of a distributed data processing method provided according to an embodiment of the present invention;

[0026] Figure 6 This is a flowchart of a distributed data processing method in a first preset scenario provided by an embodiment of the present invention;

[0027] Figure 7 This is a signaling diagram of a distributed data processing method in a first preset scenario provided by an embodiment of the present invention;

[0028] Figure 8 This is a flowchart of a distributed data processing method under a second preset scenario provided by an embodiment of the present invention;

[0029] Figure 9 This is a signaling diagram of a distributed data processing method in a second preset scenario provided by an embodiment of the present invention;

[0030] Figure 10 This is a flowchart of a distributed data processing method under a third preset scenario provided by an embodiment of the present invention;

[0031] Figure 11 This is a signaling diagram of a distributed data processing method under a third preset scenario provided by an embodiment of the present invention;

[0032] Figure 12 This is a schematic diagram of the structure of a distributed data processing device according to Embodiment 3 of the present invention;

[0033] Figure 13 This is a schematic diagram of the structure of an electronic device that implements the distributed data processing method of the present invention. Detailed Implementation

[0034] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0035] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0036] Example 1

[0037] Figure 1 The flowchart illustrates a distributed data processing method provided in Embodiment 1 of the present invention. This embodiment is applicable to distributed vehicle cruise guidance. The method can be executed by a distributed data processing device, which can be implemented in hardware and / or software and can be configured within computer software. Figure 1 As shown, the method includes:

[0038] S110. Obtain the information to be identified within the target area, wherein the information to be identified includes passenger travel information and / or driver driving information.

[0039] The target area can be understood as the coverage area of ​​the current edge server. In this embodiment of the invention, the target area can be preset according to scenario requirements, and is not specifically limited here.

[0040] The information to be identified can be understood as data to be identified. In this embodiment of the invention, the information to be identified can be preset according to scenario requirements, and is not specifically limited here. Optionally, the information to be identified includes passenger travel information and / or driver driving information.

[0041] The passenger travel information can be understood as the passenger's travel details. In this embodiment of the invention, the passenger travel information can be preset according to scenario requirements, and is not specifically limited here. Optionally, the passenger travel information may include at least one of the passenger's departure point, destination, and passenger location.

[0042] Specifically, passenger travel information can be obtained through the passenger's client. The passenger's client can be understood as a passenger application.

[0043] The driver driving information can be understood as the driver's driving information. In this embodiment of the invention, the driver driving information can be preset according to scenario requirements, and is not specifically limited here. Optionally, the driver driving information may include at least one of the driver's position and driving direction.

[0044] Specifically, the driver's driving information can be obtained through the driver's terminal, which is a driver client.

[0045] S120. If the information to be identified includes the passenger travel information and the driver driving information, the input information to be identified is processed by a local decision model to obtain a target matching result.

[0046] The local decision model can be understood as a model that determines the target matching result based on the information to be identified.

[0047] The target matching result can be understood as a model obtained by processing the input information to be identified through a local decision model. Optionally, the target matching result can be a matching result between the passenger travel information and the driver driving information.

[0048] The local decision model is obtained by federated training of the initial decision model based on the sample decision data of each edge terminal, and the sample decision data includes sample matching data and / or sample cruise guidance data.

[0049] The sample decision data can be understood as the data used to train the initial decision model. Optionally, the sample decision data may include sample matching data and / or sample cruise guidance data.

[0050] S130. If the information to be identified includes the driver's driving information, the input information to be identified is processed by the local decision model to obtain cruise guidance information.

[0051] The cruise guidance information can be understood as cruise information guiding the driver. In this embodiment of the invention, the cruise guidance information can be preset according to scenario requirements, and is not specifically limited here. Optionally, the cruise guidance information may include cruise direction and / or cruise time.

[0052] Optionally, the distributed data processing method further includes:

[0053] The target matching result is sent to the passenger terminal and the driver terminal, and / or the cruise guidance information is sent to the driver terminal.

[0054] The technical solution of this invention involves acquiring information to be identified within the target area, wherein the information to be identified includes passenger travel information and / or driver driving information; when the information to be identified includes both passenger travel information and driver driving information, a local decision model processes the input information to obtain a target matching result, wherein the local decision model is obtained by federated training of an initial decision model based on sample decision data from each edge, wherein the sample decision data includes sample matching data and / or sample cruise guidance data; when the information to be identified includes the driver driving information, the local decision model processes the input information to obtain cruise guidance information, wherein the cruise guidance information includes cruise direction and / or cruise time. This improves the accuracy of the determined target matching result and / or cruise guidance information.

[0055] Example 2

[0056] Figure 2This is a flowchart of a distributed data processing method provided in Embodiment 2 of the present invention. This embodiment focuses on processing the input information to be identified using a local decision model stored at the edge, as described in the above embodiments, and appending the target matching result. Figure 2 As shown, the method includes:

[0057] S210. For each edge, the initial decision model and the initial global model are trained based on the sample decision data of the current edge to obtain an intermediate decision model and an intermediate global model, wherein the initial global model is a student model distributed from the cloud.

[0058] The initial decision model can be understood as the initial decision model.

[0059] The initial global model can be understood as an initial global model. In this embodiment of the invention, the initial global model can be a student model distributed from the cloud.

[0060] The intermediate decision model can be understood as a model obtained by training the initial decision model based on the sample decision data. The intermediate global model can be understood as a model obtained by training the initial global model based on the sample decision data.

[0061] The student model can be understood as a lightweight deep learning model.

[0062] S220. Based on the intermediate decision model and the intermediate global model, determine the local decision model.

[0063] Optionally, determining the local decision model based on the intermediate decision model and the intermediate global model includes:

[0064] For each edge device, the first model parameters corresponding to the intermediate global model are sent to the cloud, so that the cloud performs weighted aggregation of the first model parameters sent by multiple edge devices to obtain the aggregated global parameters (see reference). Figure 3 );

[0065] The first model parameters are updated based on the aggregated global parameters to obtain the updated intermediate global model;

[0066] The local decision model is determined based on the intermediate decision model and the updated intermediate global model.

[0067] The first model parameter can be understood as the model parameter corresponding to the intermediate global model.

[0068] The aggregated global parameters can be understood as model parameters obtained by weighted aggregation of the first model parameters sent by multiple edge terminals through the cloud.

[0069] Optionally, the first model parameters corresponding to the updated intermediate global model include first local parameters and first global parameters, and the second model parameters corresponding to the intermediate decision model include second local parameters and second global parameters.

[0070] Here, the first local parameter can be understood as the parameter in the first model parameter that corresponds to the local data feature. The first global parameter can be understood as the parameter in the first model parameter that corresponds to the global data feature.

[0071] The second local parameter can be understood as the parameter in the second model parameter that corresponds to the local data feature. The second global parameter can be understood as the parameter in the second model parameter that corresponds to the global data feature.

[0072] Optionally, determining the local decision model based on the intermediate decision model and the updated intermediate global model includes:

[0073] For the intermediate decision model and the updated intermediate global model, the first local parameter is updated based on the second local parameter, and the target global model is obtained based on the updated first local parameter and the first global parameter.

[0074] The second global parameter is updated based on the first global parameter, and the local decision model is obtained based on the updated second global parameter and the second local parameter.

[0075] The step of updating the first local parameter based on the second local parameter can be understood as distilling the local data features learned by the intermediate decision model into the intermediate global model, so that the intermediate global model learns the local data features (see reference). Figure 4 Furthermore, the first global parameter is the model parameter corresponding to the global data features learned by the intermediate global model based on other edge terminals.

[0076] The step of updating the second global parameters based on the first global parameters can be understood as distilling the global data features learned by the intermediate global model into the intermediate decision model, so that the intermediate decision model learns the global data features (see reference). Figure 4Furthermore, the second local parameter is the model parameter corresponding to the local data features learned by the intermediate decision model based on the sample decision data at the current edge. Therefore, by updating the second global parameter based on the first global parameter, and obtaining the local decision model based on the updated second global parameter and the second local parameter, the determined local decision model can simultaneously possess the ability to identify global data features and local data features, thereby enabling the determined local decision model to achieve local personalized data processing based on accurate data processing.

[0077] Optionally, the distributed data processing method further includes:

[0078] The target matching results and the cruise guidance information are used as new sample decision data. The local decision model is then updated based on this new sample matching data and the global target model. This progressively improves the accuracy of data processing in the local decision model, ensuring the accuracy of the target matching results and / or cruise guidance information determined based on the local decision model.

[0079] S230. Obtain the information to be identified within the target area, wherein the information to be identified includes passenger travel information and / or driver driving information.

[0080] S240. If the information to be identified includes the passenger travel information and the driver driving information, the input information to be identified is processed by a local decision model to obtain a target matching result.

[0081] S250, If the information to be identified includes the driver's driving information, the input information to be identified is processed by the local decision model to obtain cruise guidance information.

[0082] The technical solution of this invention involves training an initial decision model and an initial global model on each edge device based on the sample decision data of that edge device to obtain an intermediate decision model and an intermediate global model. The initial global model is a student model distributed from the cloud. Based on the intermediate decision model and the intermediate global model, a local decision model is determined. This allows the determined local decision model to simultaneously identify both global and local data features, enabling it to achieve local personalized data processing based on accurate data processing.

[0083] Optional, see reference Figure 5 , Figure 6 and Figure 7Considering the first presupposed scenario, where the vehicle's equipment cannot be equipped with an intelligent model and can only collect data to make cruise guidance decisions based on the local decision-making model of the edge server, the following steps are taken: First, the vehicle's equipment sends a guidance request to the edge server based on the collected status information. Then, the edge server's local decision-making model makes a guidance decision based on the vehicle's status and sends it to the vehicle's equipment. Next, the edge server trains and updates its local decision-making model and global model based on the guidance records and sends the global model to the cloud server. Finally, the cloud server aggregates the global model from the edge and then sends the updated global model back to the edge server.

[0084] In this embodiment, it is assumed that the vehicle equipment cannot be equipped with an intelligent model and can only collect data to make cruise guidance decisions based on the local decision-making model of the edge server. First, the vehicle equipment collects its own position and orientation status information and sends a guidance request to the local decision-making model of the edge server.

[0085] Subsequently, the local decision-making model on the edge server makes a cruise guidance decision and sends it to the vehicle's equipment, storing this guidance record as a model training dataset. During the model training phase, the global model and the local decision-making model share sampled data and perform policy inter-distillation based on the calculated loss. The global model's distillation is passed to the local decision-making model as a global policy, while the local decision-making model's distillation is passed to the global model as an updated local policy. After a period of training and updates, the global model is uploaded to the cloud server.

[0086] Finally, the cloud server performs a weighted average aggregation on the global model based on the size of the training dataset on the edge server, and then distributes the aggregated global model to the edge server for global model updates.

[0087] Optional, see reference Figure 5 , Figure 8 and Figure 9 Consider the second presupposed scenario: the vehicle equipment can carry a local decision-making model to make guiding decisions based on collected data, but it lacks sufficient computing and storage capabilities to support model training and updates. Therefore, in this scenario, the vehicle equipment first makes a guiding decision based on the collected state information and sends a model update request to the edge server. Then, the edge server trains and updates the local decision-making model and the global model based on the data collected by the vehicle equipment, and sends the global model to the cloud server. Next, the cloud server aggregates the global model from the edge and then distributes the updated global model back to the edge server. Finally, the edge server distributes the local decision-making model back to the vehicle equipment.

[0088] In this embodiment, it is assumed that the vehicle equipment can carry a local decision-making model and make guidance decisions, but it does not have sufficient computing and storage capacity to support the training and updating of the model. First, the local decision-making model carried on the vehicle equipment makes a cruise guidance decision based on its own position and orientation status information, stores the guidance record, and sends a model update request to the edge server.

[0089] Subsequently, the edge server collects guidance records from the vehicle's equipment and constructs model training data. During the model training phase, the global model and the local decision model share sampled data and perform policy inter-distillation based on the calculated loss. The global model's distillation is passed to the local decision model as the global policy, while the local decision model's distillation is passed to the global model to update its local policy. After a period of training and updates, the global model is uploaded to the cloud server.

[0090] Finally, the global model generator in the cloud server performs a weighted average aggregation of the global models of each edge server based on the size of the training dataset of the edge server, and distributes the aggregated global model to the edge server for global model updates, and distributes the local decision model to the vehicle equipment for local decision model updates.

[0091] Optional, see reference Figure 5 , Figure 10 and Figure 11 Consider the third presupposed scenario, where there is no trusted central cloud node among the edge servers, and the edge servers have sufficient storage and computing power. Therefore, in this scenario, the vehicle device first initiates a guidance request to the edge server based on the collected status information. Then, the edge server's local decision model makes a guidance decision based on the vehicle device's status and sends it to the vehicle device. Next, the edge server trains and updates its local decision model and global model based on the guidance records, and broadcasts the global model to other edge servers. Finally, the edge server aggregates the global models from other nodes and updates them.

[0092] In this embodiment, it is assumed that there is no trusted central cloud node among the edge servers. First, the vehicle device collects its own position and orientation status information and sends a guidance request to the local decision-making model of the edge server.

[0093] Subsequently, the local decision-making model on the edge server makes a cruise guidance decision and sends it to the vehicle's equipment, storing this guidance record as a model training dataset. During the model training phase, the global model and the local decision-making model share sampled data and perform policy inter-distillation based on the calculated loss. The global model's distillation is passed to the local decision-making model as a global policy, while the local decision-making model's distillation is passed to the global model as an updated local policy. After a period of training and updates, the global model is broadcast to other edge servers.

[0094] Finally, the edge server performs a weighted average aggregation on the global model based on the training dataset size of other edge servers and updates the global model.

[0095] This invention addresses the intelligent cruise guidance service in the Internet of Vehicles (IoV) scenario, proposing a secure and efficient distributed solution based on federated policy distillation. Specifically, to tackle the data sharing challenge in scenarios with multiple coexisting operators, a secure and efficient federated framework is designed, enabling joint modeling of multiple operators while protecting user privacy. Furthermore, to address the heterogeneity of data characteristics among different operators, global knowledge is transferred through policy inter-distillation, constructing personalized local decision models for each edge operator and improving the robustness of these local decision models.

[0096] Example 3

[0097] Figure 12 This is a schematic diagram of the structure of a distributed data processing device provided in Embodiment 3 of the present invention. Figure 12 As shown, the device includes: an information acquisition module 310, a first information processing module 320, and a second information processing module 330; wherein,

[0098] Information acquisition module 310 is used to acquire information to be identified within a target area, wherein the information to be identified includes passenger travel information and / or driver driving information; first information processing module 320 is used to process the input information to be identified through a local decision model to obtain a target matching result when the information to be identified includes the passenger travel information and the driver driving information, wherein the local decision model is obtained by federated training of an initial decision model based on sample decision data from each edge, and the sample decision data includes sample matching data and / or sample cruise guidance data; second information processing module 330 is used to process the input information to be identified through the local decision model to obtain cruise guidance information when the information to be identified includes the driver driving information, wherein the cruise guidance information includes cruise direction and / or cruise time.

[0099] The technical solution of this invention involves acquiring information to be identified within the target area, wherein the information to be identified includes passenger travel information and / or driver driving information; when the information to be identified includes both passenger travel information and driver driving information, a local decision model processes the input information to obtain a target matching result, wherein the local decision model is obtained by federated training of an initial decision model based on sample decision data from each edge, wherein the sample decision data includes sample matching data and / or sample cruise guidance data; when the information to be identified includes the driver driving information, the local decision model processes the input information to obtain cruise guidance information, wherein the cruise guidance information includes cruise direction and / or cruise time. This improves the accuracy of the determined target matching result and / or cruise guidance information.

[0100] Optionally, the information acquisition module 310 is used for:

[0101] The passenger travel information is obtained through the passenger terminal, and the driver driving information is obtained through the driver terminal.

[0102] Optionally, the distributed data processing device further includes an information transmission module, used for:

[0103] The target matching result is sent to the passenger terminal and the driver terminal, and / or the cruise guidance information is sent to the driver terminal.

[0104] Optionally, the distributed data processing device further includes an intermediate model training module and a decision model determination module; wherein,

[0105] The intermediate model training module is used to train the initial decision model and the initial global model for each edge terminal before processing the input information to be identified through the local decision model stored at the edge terminal to obtain the target matching result. The initial global model is a student model distributed from the cloud.

[0106] The decision model determination module is used to determine the local decision model based on the intermediate decision model and the intermediate global model.

[0107] Optionally, the decision model determination module includes: a weighted aggregation unit, a global model update unit, and a decision model determination unit; wherein,

[0108] The weighted aggregation unit is used to send the first model parameters corresponding to the intermediate global model to the cloud for each edge end, so that the cloud performs weighted aggregation on the first model parameters sent by multiple edge ends to obtain the aggregated global parameters;

[0109] The global model update unit is used to update the first model parameters based on the aggregated global parameters to obtain the updated intermediate global model.

[0110] The decision model determination unit is used to determine the local decision model based on the intermediate decision model and the updated intermediate global model.

[0111] Optionally, the first model parameters corresponding to the updated intermediate global model include first local parameters and first global parameters, and the second model parameters corresponding to the intermediate decision model include second local parameters and second global parameters.

[0112] Optionally, the decision model determination unit is used for:

[0113] For the intermediate decision model and the updated intermediate global model, the first local parameter is updated based on the second local parameter, and the target global model is obtained based on the updated first local parameter and the first global parameter.

[0114] The second global parameter is updated based on the first global parameter, and the local decision model is obtained based on the updated second global parameter and the second local parameter.

[0115] Optionally, the distributed data processing device further includes a model update module, used for:

[0116] The target matching result and the cruise guidance information are used as the newly added sample decision data, and the local decision model is updated based on the newly added sample matching data and the target global model.

[0117] Optionally, the passenger travel information includes at least one of the passenger's departure point, destination, and passenger location, and the driver driving information includes at least one of the driver's position and driving direction.

[0118] The distributed data processing device provided in the embodiments of the present invention can execute the distributed data processing method provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of executing the method.

[0119] Example 4

[0120] Figure 13A schematic diagram of an electronic device 10 that can be used to implement embodiments of the present invention is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices (e.g., helmets, glasses, watches, etc.), and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the invention described and / or claimed herein.

[0121] like Figure 13 As shown, the electronic device 10 includes at least one processor 11 and a memory, such as a read-only memory (ROM) 12 or a random access memory (RAM) 13, communicatively connected to the at least one processor 11. The memory stores computer programs executable by the at least one processor. The processor 11 can perform various appropriate actions and processes based on the computer program stored in the ROM 12 or loaded from storage unit 18 into the RAM 13. The RAM 13 may also store various programs and data required for the operation of the electronic device 10. The processor 11, ROM 12, and RAM 13 are interconnected via a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0122] Multiple components in electronic device 10 are connected to I / O interface 15, including: input unit 16, such as keyboard, mouse, etc.; output unit 17, such as various types of displays, speakers, etc.; storage unit 18, such as disk, optical disk, etc.; and communication unit 19, such as network card, modem, wireless transceiver, etc. Communication unit 19 allows electronic device 10 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0123] Processor 11 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. Processor 11 performs the various methods and processes described above, such as distributed data processing methods.

[0124] In some embodiments, the distributed data processing method may be implemented as a computer program tangibly contained in a computer-readable storage medium, such as storage unit 18. In some embodiments, part or all of the computer program may be loaded and / or mounted on electronic device 10 via ROM 12 and / or communication unit 19. When the computer program is loaded into RAM 13 and executed by processor 11, one or more steps of the distributed data processing method described above may be performed. Alternatively, in other embodiments, processor 11 may be configured to perform the distributed data processing method by any other suitable means (e.g., by means of firmware).

[0125] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), payload-programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0126] Computer programs used to implement the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when executed by the processor, the computer programs cause the functions / operations specified in the flowcharts and / or block diagrams to be performed. The computer programs may be executed entirely on a machine, partially on a machine, or as a standalone software package, partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0127] In the context of this invention, a computer-readable storage medium can be a tangible medium that may contain or store a computer program for use by or in conjunction with an instruction execution system, apparatus, or device. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination thereof. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0128] To provide interaction with a user, the systems and techniques described herein can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the electronic device. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0129] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as data servers), or computing systems that include middleware components (e.g., application servers), or computing systems that include frontend components (e.g., user computers with graphical user interfaces or web browsers through which users can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., communication networks). Examples of communication networks include local area networks (LANs), wide area networks (WANs), blockchain networks, and the Internet.

[0130] A computing system can include clients and servers. Clients and servers are generally located far apart and typically interact through communication networks. The client-server relationship is created by computer programs running on the respective computers and having a client-server relationship with each other. The server can be a cloud server, also known as a cloud computing server or cloud host, which is a hosting product within the cloud computing service system to address the shortcomings of traditional physical hosts and VPS services, such as high management difficulty and weak business scalability.

[0131] It should be understood that the various forms of processes shown above can be used, with steps reordered, added, or deleted. For example, the steps described in this invention can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution of this invention can be achieved, and this is not limited herein.

[0132] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. A distributed data processing method, characterized in that, include: Acquire the information to be identified within the target area, wherein the information to be identified includes passenger travel information and / or driver driving information; When the information to be identified includes the passenger travel information and the driver driving information, the input information to be identified is processed by a local decision model to obtain a target matching result. The local decision model is obtained by federated training of an initial decision model based on sample decision data from each edge. The sample decision data includes sample matching data and / or sample cruise guidance data. When the information to be identified includes the driver's driving information, the input information to be identified is processed by the local decision model to obtain cruise guidance information, wherein the cruise guidance information includes cruise direction and / or cruise time; Before processing the input information to be identified using a local decision model stored at the edge to obtain the target matching result, the process further includes: For each edge, the initial decision model and the initial global model are trained based on the sample decision data of the current edge to obtain an intermediate decision model and an intermediate global model, wherein the initial global model is a student model distributed from the cloud. For each edge device, the first model parameters corresponding to the intermediate global model are sent to the cloud, so that the cloud performs weighted aggregation on the first model parameters sent by multiple edge devices to obtain aggregated global parameters; The first model parameters are updated based on the aggregated global parameters to obtain the updated intermediate global model; For the intermediate decision model and the updated intermediate global model, the first local parameters are updated based on the second local parameters, and the target global model is obtained based on the updated first local parameters and the first global parameters; the first local parameters are the parameters in the first model parameters that correspond to the local data features; the first global parameters are the parameters in the first model parameters that correspond to the global data features; the second local parameters are the parameters in the second model parameters that correspond to the local data features. The second global parameter is updated based on the first global parameter, and the local decision model is obtained based on the updated second global parameter and the second local parameter; the second global parameter is the parameter in the second model parameter that corresponds to the global data feature.

2. The method according to claim 1, characterized in that, The step of obtaining the information to be identified within the target area includes: The passenger travel information is obtained through the passenger terminal, and the driver driving information is obtained through the driver terminal.

3. The method according to claim 2, characterized in that, Also includes: The target matching result is sent to the passenger terminal and the driver terminal, and / or the cruise guidance information is sent to the driver terminal.

4. The method according to claim 1, characterized in that, The updated intermediate global model corresponds to the first model parameters including the first local parameters and the first global parameters, and the intermediate decision model corresponds to the second model parameters including the second local parameters and the second global parameters.

5. The method according to any one of claims 1-4, characterized in that, Also includes: The target matching result and the cruise guidance information are used as the newly added sample decision data, and the local decision model is updated based on the newly added sample matching data and the target global model.

6. The method according to claim 1, characterized in that, The passenger travel information includes at least one of the passenger's departure point, destination, and passenger location, and the driver driving information includes at least one of the driver's position and driving direction.

7. A distributed data processing device, characterized in that, include: The information acquisition module is used to acquire information to be identified within a target area, wherein the information to be identified includes passenger travel information and / or driver driving information; The first information processing module is used to process the input information to be identified through a local decision model to obtain a target matching result when the information to be identified includes the passenger travel information and the driver driving information. The local decision model is obtained by federated training of an initial decision model based on sample decision data from each edge end. The sample decision data includes sample matching data and / or sample cruise guidance data. The second information processing module is used to process the input information to be identified through the local decision model when the information to be identified includes the driver's driving information, to obtain cruise guidance information, wherein the cruise guidance information includes cruise direction and / or cruise time. The intermediate model training module is used to train the initial decision model and the initial global model for each edge terminal before processing the input information to be identified through the local decision model stored at the edge terminal to obtain the target matching result. The initial global model is a student model distributed from the cloud. The decision model determination module includes: a weighted aggregation unit, a global model update unit, and a decision model determination unit; The weighted aggregation unit is used to send the first model parameters corresponding to the intermediate global model to the cloud for each edge end, so that the cloud performs weighted aggregation on the first model parameters sent by multiple edge ends to obtain aggregated global parameters; The global model update unit is used to update the first model parameters based on the aggregated global parameters to obtain the updated intermediate global model. The decision model determination unit is configured to, for the intermediate decision model and the updated intermediate global model, update the first local parameters based on the second local parameters, and obtain the target global model based on the updated first local parameters and the first global parameters; update the second global parameters based on the first global parameters, and obtain the local decision model based on the updated second global parameters and the second local parameters; wherein the first local parameters are the parameters in the first model parameters corresponding to the local data features; the first global parameters are the parameters in the first model parameters corresponding to the global data features; the second local parameters are the parameters in the second model parameters corresponding to the local data features; and the second global parameters are the parameters in the second model parameters corresponding to the global data features.

Images