Driving situation awareness energy management method and device, equipment and storage medium

Abstract

The invention relates to a driving situation perception energy management method, device and equipment and a storage medium, and belongs to the technical field of data processing.The method comprises the steps that multi-modal data are collected and subjected to integration processing, and integrated data are obtained, the multi-modal data comprises real-time traffic jam information, vehicle-mounted sensor temperature information and driver driving style information; inputting the integrated data into a pre-established multi-modal deep security clustering model for security clustering analysis to obtain a multi-modal clustering analysis result; and personalized energy management rules are generated based on the multi-modal clustering analysis result, and the personalized energy management rules comprise a real-time traffic management rule, a real-time temperature management rule and a driving style management rule. The energy utilization rate of the new energy automobile in various driving scenes is improved, unnecessary energy consumption is reduced, and it is guaranteed that the automobile can be rapidly adjusted to the optimal working state under different working conditions.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to a driving situation awareness energy management method, device, equipment and storage medium. Background Art

[0002] The energy management system of new energy vehicles is the core technology of new energy vehicles, and plays a vital role in optimizing energy use, extending driving range, ensuring battery safety and improving driving performance. For hybrid vehicles, the main purpose of the new energy management system for hybrid vehicles is to meet the driver's demand for traction power, maintain the battery charging state and optimize the transmission system efficiency, fuel consumption and exhaust emissions.

[0003] With the rapid development and widespread application of artificial intelligence and machine learning technologies, the energy management system of new energy vehicles has achieved intelligent upgrades and personalized optimization. These advanced technologies enable the energy management system to more accurately monitor and intelligently manage vehicle energy consumption. Through in-depth analysis and prediction of massive real-time data, the system can dynamically and flexibly adjust the energy allocation strategy, thereby maximizing energy efficiency in various driving scenarios and significantly reducing overall energy consumption.

[0004] Traditional and classic energy management strategies can be divided into two categories: one is rule-based strategies, and the other is optimization algorithm-based strategies. Since rule-based strategies are based on fixed rules, they may not be able to adapt to all possible working conditions and conditions, especially when facing complex or changeable tasks, it is difficult to globally optimize them to improve performance or reduce energy consumption. For strategies based on optimization algorithms, the randomness and uncertainty of the actual driving environment make it difficult to obtain global information, and thus real-time optimization cannot be achieved. In addition, the optimization algorithm places high computing power requirements on the computing unit and requires accurate subsystem models, which makes it difficult to directly apply to real-time control of energy management. Therefore, how to effectively manage the energy of new energy vehicles is a technical problem that needs to be solved urgently. Summary of the invention

[0005] In view of this, it is necessary to provide a driving situation awareness energy management method, device, equipment and storage medium to improve the energy utilization rate of new energy vehicles in various driving scenarios and reduce unnecessary energy consumption.

[0006] In order to solve the above problems, the present invention provides a driving situation awareness energy management method, comprising: Collecting multimodal data and integrating the multimodal data to obtain integrated data, wherein the multimodal data includes real-time traffic congestion information, vehicle sensor temperature information, and driver driving style information; Inputting the integrated data into a pre-established multimodal deep security clustering model to perform security clustering analysis to obtain a multimodal clustering analysis result; Generate personalized energy management rules based on the multimodal cluster analysis results.

[0007] In a possible implementation, when the multimodal data is integrated, the method further includes: Processing the multimodal data based on a federated learning mechanism; The processing of the multimodal data based on the federated learning mechanism specifically includes: Constructing a federated learning model including a central server and a set of client devices, wherein each client stores an initial global model distributed by the central server; Each client is used as a node of federated learning. Multimodal data corresponding to each vehicle is used for iterative training on each client to obtain multiple local training models and upload them to the central server. After the central server receives the local training model uploaded by each client, it uses a preset aggregation algorithm to aggregate the parameters in each local training model, generates a first global model, and distributes it to each client; After the client receives the first global model, it uses the multimodal data corresponding to each vehicle to perform iterative training again until the preset training conditions are met and the multimodal data processing is completed.

[0008] In a possible implementation, the pre-established multimodal deep security clustering model includes a security module; The security module is: ,in, is the feature vector of the input sample, For the local functions of the device, , and are learnable parameters.

[0009] In a possible implementation, inputting the integrated data into a pre-established multimodal deep security clustering model for security clustering analysis includes: Determine an anomaly detection safety mechanism in the multimodal deep security clustering model, and monitor abnormal data in the integrated data based on the anomaly detection safety mechanism.

[0010] In a possible implementation, the method further includes: Clustering risk is used as a performance evaluation indicator of the multimodal deep security clustering model to optimize the multimodal deep security clustering model.

[0011] In a possible implementation, the method further includes: Collect rule feedback information, and optimize the multimodal deep security clustering model and adjust the personalized energy management rules based on the rule feedback information.

[0012] The present invention also provides a driving situation awareness energy management device, comprising: A data integration module, used to collect multimodal data and integrate the multimodal data to obtain pre-integrated data, wherein the multimodal data includes real-time traffic congestion information, vehicle sensor temperature information and driver driving style information; A multimodal clustering analysis module, used for inputting the integrated data into a pre-established multimodal deep security clustering model for clustering analysis to obtain a multimodal clustering analysis result; A rule generation module is used to generate personalized energy management rules based on the multimodal clustering analysis results.

[0013] In a third aspect, the present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps in the above-mentioned driving situation awareness energy management method when executing the computer program.

[0014] In a fourth aspect, the present invention further provides a computer storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps in the above-mentioned driving situation awareness energy management method are implemented.

[0015] The present invention combines multimodal data such as real-time traffic congestion information, vehicle-mounted sensor temperature information, and driver driving style information, and integrates the multimodal data to improve the processing speed of data in the multimodal deep safety clustering model; then uses the pre-established multimodal deep safety clustering model to perform safety clustering analysis on the integrated data, which can effectively reduce the impact of data non-independent and identically distributed on model accuracy, improve the accuracy and safety of model predictions, and require less calculations, improve the accuracy and stability of clustering when the data set is insufficient, and finally generate personalized energy management rules based on the multimodal clustering analysis results, thereby improving the energy utilization rate of new energy vehicles in various driving scenarios, reducing unnecessary energy consumption, and ensuring that the vehicle can be quickly adjusted to the optimal working state when facing different working conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the 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 work.

[0017] Figure 1 A method flow chart of an embodiment of a driving situation awareness energy management method provided by the present invention; Figure 2 A schematic diagram of the structure of an embodiment of a driving situation awareness energy management device provided by the present invention; Figure 3 A schematic diagram of the structure of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0018] The preferred embodiments of the present invention will be described in detail below in conjunction with the accompanying drawings, wherein the accompanying drawings constitute a part of the present invention and are used together with the embodiments of the present invention to illustrate the principles of the present invention, but are not used to limit the scope of the present invention.

[0019] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present invention. The appearance of the phrase in various places in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0020] A specific embodiment of the present invention discloses a driving situation awareness energy management method, comprising: Step 101: Collect multimodal data and integrate the multimodal data to obtain integrated data, wherein the multimodal data includes real-time traffic congestion information, vehicle sensor temperature information, and driver driving style information; Step 102: Input the integrated data into a pre-established multimodal deep security clustering model to perform security clustering analysis and obtain a multimodal clustering analysis result; Step 103: Generate personalized energy management rules based on the multimodal cluster analysis results.

[0021] It is understandable that multimodal data is collected through various sensors on the vehicle, such as cameras, radars, temperature sensors, etc. The multimodal data includes but is not limited to real-time traffic congestion information, vehicle-mounted sensor temperature information, driver driving style information, etc. The collected data is then transmitted to the central control unit through a high-bandwidth, low-latency bus. The central control unit performs preliminary processing on the data, including data compression, slicing, timestamp marking, etc., to ensure the time synchronization of the data and give priority to emergency information. In addition, for large amounts of data transmission, such as real-time traffic video and temperature sensor data, the system uses specific compression algorithms and data slicing strategies to ensure efficient and stable transmission in low-bandwidth environments. Then, through multimodal data fusion technology, the multimodal data is integrated into a unified representation. Among them, personalized energy management rules include real-time traffic management rules, real-time temperature management rules, and driving style management rules.

[0022] The present invention combines multimodal data such as real-time traffic congestion information, vehicle-mounted sensor temperature information, and driver driving style information, and integrates the multimodal data to improve the processing speed of data in the multimodal deep safety clustering model; then uses the pre-established multimodal deep safety clustering model to perform safety clustering analysis on the integrated data, which can effectively reduce the impact of data non-independent and identically distributed on model accuracy, improve the accuracy and safety of model predictions, and require less calculations, improve the accuracy and stability of clustering when the data set is insufficient, and finally generate personalized energy management rules based on the multimodal clustering analysis results, thereby improving the energy utilization rate of new energy vehicles in various driving scenarios, reducing unnecessary energy consumption, and ensuring that the vehicle can be quickly adjusted to the optimal working state when facing different working conditions.

[0023] It is understandable that for ease of processing, after collecting data from multiple sensors, including vision, radar, LiDAR and GPS data, multimodal data fusion technology is used to integrate data from different sources into a unified representation.

[0024] When integrating and processing multimodal data, the above method further includes: Process multimodal data based on federated learning mechanism; Processing multimodal data based on the federated learning mechanism, including: Constructing a federated learning model including a central server and a set of client devices, wherein each client stores an initial global model distributed by the central server; Each client is used as a node of federated learning. Multimodal data corresponding to each vehicle is used for iterative training on each client to obtain multiple local training models and upload them to the central server. After the central server receives the local training model uploaded by each client, it uses a preset aggregation algorithm to aggregate the parameters in each local training model, generates a first global model, and distributes it to each client; After the client receives the first global model, it uses the multimodal data corresponding to each vehicle to perform iterative training again until the preset training conditions are met and the multimodal data processing is completed.

[0025] It should be noted that in order to ensure the security of data transmission, AES encryption and TLS authentication mechanisms can be used to prevent data from being tampered with or illegally accessed during transmission. At the same time, multimodal data can also be processed based on the federated learning mechanism to achieve encryption purposes.

[0026] Specifically, through the federated learning mechanism, a federated learning model including a central server and a set of client devices is constructed. The central server distributes an initial global model to all clients, which the clients receive and store. Each client is used as a node of federated learning. The multimodal data corresponding to each vehicle is used for local training at each client. The global model is trained multiple times through local data to obtain multiple local training models and upload them to the central server to generate locally updated model parameters and weights. After receiving the local model parameters uploaded by all clients, the central server uses the preset FedAvg aggregation algorithm to aggregate these parameters to form a new global model. The central server adjusts the weights according to the data volume or data quality, and distributes the newly generated global model to all clients again. The client receives the new global model, adjusts the local model parameters, and conducts the next round of local training. The above process is repeated for multiple rounds. Based on the preset aggregation mechanism, the differential privacy strategy is used in each iteration to dynamically adjust the privacy parameters for training. The processing of multimodal data is completed. By using each vehicle as a node of federated learning and training the model locally with its own data, the risk of data leakage is reduced. Through a secure aggregation mechanism, the model updates of each vehicle's local training are aggregated to the central server. This approach ensures continuous improvement of the model while protecting the privacy of users. At the same time, differential privacy strategies are applied during the training process, and random noise is added during the model update process to protect the privacy of individual contributors and prevent the original data from being inferred through the model.

[0027] The present invention aims at two requirements of existing intelligent energy management technology for new energy vehicles: to improve the safety and stability of the vehicle multimodal data clustering process and to improve the flexibility of the vehicle to switch working modes in real time according to the driving situation while ensuring the clustering accuracy. By using sensors to collect diversified driving situation information, the problems of high battery energy consumption and insufficient intelligent motor power control in different driving scenarios are analyzed, and an intelligent system integrating multiple technologies is proposed, which covers a comprehensive solution for driving situation perception energy management including traffic congestion identification, temperature data processing and energy consumption optimization.

[0028] In the driving situation awareness module, the main challenge faced by the system is how to achieve reliable clustering analysis under multimodal data, especially in the absence of ground truth labels. When the true labels are available, the security of the model can be measured by comparing the clustering results with the true labels. The present invention proposes to efficiently cluster and analyze multimodal data while ensuring data privacy by combining a federated learning mechanism with a multimodal deep security clustering model. Once the expected secure multi-view clustering is achieved, its generalization and security capabilities will be more suitable for actual scenarios.

[0029] The robust federated algorithm aims to minimize the following objective function: , where m is the vehicle information collection device, is the local objective function of the kth device. That is, empirical risk, where is the number of local samples available. User defined term To specify the degree of influence of each device, two natural settings are or ,in is the total number of samples. Federated learning provides a core optimization problem, which makes the uploading of vehicle data to the training of multimodal deep security clustering model more private. That is, the minimization problem of global model parameters: .here, Represents the global loss function, which is the local loss function on all devices participating in federated learning The contribution of each device to the global model is represented by a weight, which is usually proportional to the amount or importance of the data on the device.

[0030] In one embodiment of the present invention, the pre-established multimodal deep security clustering model includes a security module; The security modules are: ,in, is the feature vector of the input sample, For the local functions of the device, , and are learnable parameters.

[0031] It can be understood that the multimodal deep security clustering model consists of a feature extraction module F and a clustering assignment module Composition, of which and Can be composed of deep neural networks.

[0032] In the feature extraction stage, the feature extraction module F adds noise to the gradient update through differential privacy technology to prevent the leakage of individual data information. When transmitting data between the client and the central server, TSS encryption technology is used to protect the security of the data. In the cluster allocation stage, the cluster allocation module C uses technologies such as secure multi-party computing (SMC) or homomorphic encryption to aggregate model parameters without exposing specific data. In order to achieve the security of multiple views, each information collection device may collect different amounts of data during the information collection stage. Therefore, the model needs to introduce a security module S to ensure the privacy and security of the data. Security module Get the feature detector output combination: ,in, is the feature vector of the input sample, For the local functions of the device, , and is a learnable parameter. It should be noted that deep secure clustering proposes a new framework to reduce the degradation of clustering performance caused by the increase in data transmission views uploaded from the vehicle to the cloud. The objective of the proposed deep secure multi-view clustering framework is expressed as: , where λ is an adjustable learning parameter and Represent the clustering assignment module and clustering loss respectively, and The security module and feature extractor functions, The feature vector extracted from the sample.

[0033] Introducing a multimodal deep security clustering model to learn the clustering risks of newly added datasets As an indicator for evaluating the model: Before the amount of collected data increases, the supervised clustering loss that the model learns from the data is: In this architecture, the security module output is input into a fully connected layer to obtain the hidden features h(p). Clustering loss function Combined with hidden features to obtain the clustering assignment matrix With the Gaussian kernel matrix In this way, the basis equation of deep security clustering can be obtained, and the proposed goal of deep security clustering can be expressed as: , , are the parameters of the security module, θ are the parameters of the feature extractor and cluster assignment module, and K is the number of clusters. is the hidden feature matrix, which represents the hidden feature vector of the i-th sample in the p-th view.

[0034] In one embodiment of the present invention, the integrated data is input into a pre-established multi-modal deep security clustering model for security clustering analysis, including: Determine the anomaly detection safety mechanism in the multimodal deep security clustering model, and monitor abnormal data in the integrated data based on the anomaly detection safety mechanism.

[0035] It is understandable that the deep security clustering algorithm is used in the cloud to perform efficient security clustering analysis on multimodal data from multiple vehicles. The algorithm introduces a security mechanism based on anomaly detection, identifies potential attacks or failures by monitoring abnormal patterns in data streams, automatically extracts important information from multimodal data streams and excludes meaningless noise, ensures that the system clustering performance does not decrease when adding new modal data, and maintains high security and stability of the system.

[0036] In one embodiment of the present invention, the method further comprises: Clustering risk is used as the performance evaluation indicator of the multimodal deep security clustering model to optimize the multimodal deep security clustering model.

[0037] It is understandable that in order to make the energy management more efficient, it is necessary to evaluate the performance of the model to judge the quality of the energy rules. Under the given safety factor λ of the vehicle system, the optimization can be regarded as a conventional multi-view deep clustering process, in which the model is trained to extract complementary information from the cloud data. The safety factor is then optimized so that the model automatically selects appropriate features from the feature extractor. Therefore, compared with the traditional clustering method, the whole process greatly improves the adaptability and robustness of the system in complex driving scenarios when processing multi-view data, and ensures the clustering accuracy of multimodal data. And by optimizing the clustering loss function, the adaptability and robustness of the model when processing multi-view data are enhanced. And when using the clustering risk of the newly added data set as the model performance evaluation indicator, it is ensured that the model can self-optimize according to the actual situation, reduce the clustering risk of the system when facing multi-modal data, and improve the interpretability and robustness of the system model.

[0038] In one embodiment of the present invention, the method further comprises: Collect rule feedback information, and optimize the multimodal deep security clustering model and adjust the personalized energy management rules based on the rule feedback information.

[0039] It is understandable that based on the results of cluster analysis, personalized energy management rules are generated, and these rules dynamically adjust the energy allocation strategy according to different driving modes, road conditions and environmental factors. The optimized energy management strategy is sent to the energy management system of the car, and decisions can be quickly generated according to the predefined rule base, which is suitable for scenarios with high real-time requirements. Under the safety factor λ given by the vehicle-mounted system, the optimization can be regarded as a conventional multi-view deep clustering process, in which the model is trained to extract information from cloud data and optimize the safety factor so that the model automatically selects appropriate features from the feature extractor, which greatly improves the adaptability and robustness of the system in complex driving scenarios and ensures the clustering speed and accuracy of multimodal data. In addition, the present invention integrates IoT technology and uses vehicle operation data including sensor data and driving behavior data collected by IoT technology to continuously monitor the vehicle's operating status and energy usage. At the same time, a real-time operating system is used to ensure the timeliness of data processing and decision-making, meet the low latency requirements of the vehicle-mounted system, collect feedback information for subsequent model optimization and rule adjustment, and form a closed-loop optimization process.

[0040] The present invention proposes to efficiently cluster and analyze multimodal data while ensuring data privacy by combining a federated learning mechanism with a multimodal deep security clustering model. Federated learning allows vehicles to train models locally and transmit model updates to the cloud for global aggregation, avoiding the transmission of original data and effectively protecting the privacy of vehicle data. The core of the DSMVC algorithm is its ability to maintain the security of the system when adding new perspectives, which is crucial for processing high-dimensional and complex structured data.

[0041] Compared with the prior art, the present invention shows significant advantages in the following aspects: First, it can greatly improve the robustness and adaptability of the system in complex driving scenarios, and ensure the clustering accuracy of multimodal data. Secondly, the intelligent energy management system of new energy vehicles can maintain a high level of safety and reliability in the face of ever-changing environments and increased new perspectives. Finally, the system can generate personalized energy management rules and send them to the energy management system of the vehicle. These rules dynamically adjust the energy allocation strategy according to different driving modes, road conditions and environmental factors. By optimizing the energy management strategy, the present invention can significantly save energy and maintenance costs by reducing energy consumption and extending the service life of equipment while reducing environmental pollution. This environmentally friendly design not only meets the requirements of modern society for green technology, but also brings greater economic returns to users.

[0042] The present invention takes into account the ease of user operation and the scalability of the system when designing the system. Through modular design, the system can flexibly add or replace components and support users to adjust system settings according to their own needs. This design not only simplifies the operation process of the system, but also ensures that the system is easy to maintain and upgrade during design, reducing long-term operating costs. The modular design enables the system to have good scalability and maintainability. Users can adjust system functions according to actual needs, further improving the applicability of the system and user experience.

[0043] In order to better implement the driving situation awareness energy management method in the embodiment of the present invention, based on the driving situation awareness energy management method, correspondingly, Figure 2 As shown, the embodiment of the present invention further provides a driving situation awareness energy management device 200, and the driving situation awareness energy management device 200 includes: The data integration module 201 is used to collect multimodal data and integrate the multimodal data to obtain pre-integrated data, wherein the multimodal data includes real-time traffic congestion information, vehicle sensor temperature information and driver driving style information; The multimodal cluster analysis module 202 is used to input the integrated data into a pre-established multimodal deep security clustering model for cluster analysis to obtain a multimodal cluster analysis result; The rule generation module 203 is used to generate personalized energy management rules based on the multimodal cluster analysis results.

[0044] The driving situation awareness energy management device 200 provided in the above embodiment can implement the technical solution described in the above driving situation awareness energy management method embodiment. The specific implementation principles of the above modules or units can refer to the corresponding contents in the above driving situation awareness energy management method embodiment, which will not be repeated here.

[0045] like Figure 3 As shown, the present invention also provides an electronic device 300. The electronic device 300 includes a processor 301, a memory 302 and a display 303. Figure 3 Only some components of the electronic device 300 are shown, but it should be understood that it is not required to implement all of the components shown, and more or fewer components may be implemented instead.

[0046] In some embodiments, the processor 301 may be a central processing unit (CPU), a microprocessor or other data processing chip, used to run program codes or process data stored in the memory 302, such as the driving situation awareness energy management method of the present invention.

[0047] In some embodiments, the processor 301 may be a single server or a server group. The server group may be centralized or distributed. In some embodiments, the processor 301 may be local or remote. In some embodiments, the processor 301 may be implemented in a cloud platform. In one embodiment, the cloud platform may include a private cloud, a public cloud, a hybrid cloud, a community cloud, a distributed cloud, an internal cloud, a multi-cloud, etc., or any combination thereof.

[0048] In some embodiments, the memory 302 may be an internal storage unit of the electronic device 300, such as a hard disk or memory of the electronic device 300. In other embodiments, the memory 302 may also be an external storage device of the electronic device 300, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc., equipped on the electronic device 300.

[0049] Furthermore, the memory 302 may include both an internal storage unit of the electronic device 300 and an external storage device. The memory 302 is used to store application software installed in the electronic device 300 and various data.

[0050] In some embodiments, the display 303 may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, an OLED (Organic Light-Emitting Diode) touch device, etc. The display 303 is used to display information of the electronic device 300 and to display a visual user interface. The components 301-303 of the electronic device 300 communicate with each other through a system bus.

[0051] In some embodiments, when the processor 301 executes the driving situation awareness energy management program in the memory 302, the following steps may be implemented: Collecting multimodal data and integrating and processing the multimodal data to obtain pre-integrated data, wherein the multimodal data includes real-time traffic congestion information, vehicle sensor temperature information, and driver driving style information; Input the integrated data into a pre-established multimodal deep security clustering model for cluster analysis to obtain multimodal clustering analysis results; Generate personalized energy management rules based on multimodal clustering analysis results.

[0052] It should be understood that: when the processor 301 executes the driving situation awareness energy management program in the memory 302, in addition to the above functions, other functions can also be implemented. For details, please refer to the description of the corresponding method embodiment above.

[0053] Furthermore, the embodiment of the present invention does not specifically limit the type of the electronic device 300 mentioned, and the electronic device 300 may be a portable electronic device such as a mobile phone, a tablet computer, a personal digital assistant (PDA), a wearable device, a laptop computer, etc. Exemplary embodiments of portable electronic devices include but are not limited to portable electronic devices equipped with IOS, Android, Microsoft or other operating systems. The above-mentioned portable electronic devices may also be other portable electronic devices, such as a laptop computer with a touch-sensitive surface (e.g., a touch panel). It should also be understood that in some other embodiments of the present invention, the electronic device 300 may not be a portable electronic device, but a desktop computer with a touch-sensitive surface (e.g., a touch panel).

[0054] Correspondingly, an embodiment of the present invention also provides a computer-readable storage medium, which is used to store computer-readable programs or instructions. When the program or instructions are executed by a processor, the steps or functions in the driving situation awareness energy management method provided by the above-mentioned method embodiments can be implemented.

[0055] Those skilled in the art will appreciate that all or part of the processes of the above-mentioned embodiments can be implemented by instructing related hardware through a computer program, and the program can be stored in a computer-readable storage medium, wherein the computer-readable storage medium is a disk, an optical disk, a read-only storage memory, or a random access memory, etc.

[0056] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by any technician familiar with the technical field within the technical scope disclosed by the present invention should be covered within the protection scope of the present invention.

Claims

1. A driving situation awareness energy management method, characterized in that: include: Collecting multimodal data and integrating the multimodal data to obtain integrated data, wherein the multimodal data includes real-time traffic congestion information, vehicle sensor temperature information, and driver driving style information; Inputting the integrated data into a pre-established multimodal deep security clustering model to perform security clustering analysis to obtain a multimodal clustering analysis result; Generate personalized energy management rules based on the multimodal cluster analysis results.

2. The driving situation awareness energy management method according to claim 1, characterized in that: When the multimodal data is integrated and processed, the method further includes: Processing the multimodal data based on a federated learning mechanism; The processing of the multimodal data based on the federated learning mechanism specifically includes: Constructing a federated learning model including a central server and a set of client devices, wherein each client stores an initial global model distributed by the central server; Each client is used as a node of federated learning. Multimodal data corresponding to each vehicle is used for iterative training on each client to obtain multiple local training models and upload them to the central server. After the central server receives the local training model uploaded by each client, it uses a preset aggregation algorithm to aggregate the parameters in each local training model, generates a first global model, and distributes it to each client; After the client receives the first global model, it uses the multimodal data corresponding to each vehicle to perform iterative training again until the preset training conditions are met and the multimodal data processing is completed.

3. The driving situation awareness energy management method according to claim 2, characterized in that: The pre-established multimodal deep security clustering model includes a security module; The security module is: ， in, is the feature vector of the input sample, For the local functions of the device, , and are learnable parameters.

4. The driving situation awareness energy management method according to claim 3, characterized in that: The step of inputting the integrated data into a pre-established multi-modal deep security clustering model for security clustering analysis includes: Determine an anomaly detection safety mechanism in the multimodal deep security clustering model, and monitor abnormal data in the integrated data based on the anomaly detection safety mechanism.

5. The driving situation awareness energy management method according to claim 1, characterized in that: The method further comprises: Clustering risk is used as a performance evaluation indicator of the multimodal deep security clustering model to optimize the multimodal deep security clustering model.

6. The driving situation awareness energy management method according to claim 1, characterized in that: The method further comprises: Collect rule feedback information, and optimize the multimodal deep security clustering model and adjust the personalized energy management rules based on the rule feedback information.

7. A driving situation awareness energy management device, characterized in that: include: A data integration module, used to collect multimodal data and integrate the multimodal data to obtain pre-integrated data, wherein the multimodal data includes real-time traffic congestion information, vehicle sensor temperature information and driver driving style information; A multimodal clustering analysis module, used for inputting the integrated data into a pre-established multimodal deep security clustering model for clustering analysis to obtain a multimodal clustering analysis result; A rule generation module is used to generate personalized energy management rules based on the multimodal clustering analysis results.

8. An electronic device, characterized in that: comprising a memory and a processor, wherein: The memory is used to store programs; The processor is coupled to the memory and is used to execute the program stored in the memory to implement the steps in the driving situation awareness energy management method described in any one of claims 1 to 6 above.

9. A computer-readable storage medium, characterized in that: Used to store computer-readable programs or instructions, which, when executed by a processor, can implement the steps of the driving situation awareness energy management method described in any one of claims 1 to 6 above.

Images