Vehicle-cloud integrated multi-scene prediction method based on big data driving, electronic equipment and medium

Abstract

The invention discloses a vehicle-cloud integrated multi-scene prediction method based on big data driving, electronic equipment and a medium. The method can comprise the steps that a vehicle end integrates various controllers, collects corresponding controller data, and sends the data to a cloud end through a data interaction layer; the cloud receives the controller data, and processes and predicts the controller data; and the cloud end pushes the prediction result to the user end and / or the research and development end. According to the vehicle-cloud collaborative multi-scene prediction and control system, the vehicle-cloud collaborative multi-scene prediction and control system is constructed based on the vehicle end data and the cloud computing capability and in combination with the deep reinforcement learning model for the intelligent demand of the vehicle multi-scene, and the economy, the safety and the driving performance of the vehicle are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of intelligent transportation, more particularly, to a vehicle-cloud integrated multi-scene prediction method based on big data driving, an electronic device and a medium. BACKGROUND

[0002] At present, the vehicle end is in the stage of using rule control algorithm and traditional simulation test real vehicle test, each professional adopts its own special model development, and the fixed vehicle function parameters are adjusted and optimized, and the verification is performed through the virtual environment simulation algorithm. The whole optimization process needs a large amount of test and data record, repeated iteration, and long time consumption. The traditional vehicle end data acquisition and optimization need to be matched and developed according to different scenes and user experience, the generalization of complex scenes is poor, the calibration period is long, and the development efficiency is low; the vehicle end algorithm power is limited, the model response is delayed when multi-task scheduling; the vehicle end domain data processing and storage are redundant, the cross-domain model management is scattered and the access efficiency is low, which causes resource waste and increase of development cost.

[0003] The cloud end multi-scene prediction utilizes the development of special models, and each business scene is customized and developed, and a special model is developed for each newly added scene, the common technology is repeatedly developed, the resource utilization rate is low, and the development period is long.

[0004] At present, a vehicle-cloud integrated multi-scene prediction method based on big data driving needs to be developed.

[0005] The information disclosed in the background section of this application is only intended to deepen the understanding of the general background of the application, and should not be regarded as acknowledging or implying in any form that the information constitutes prior art known to those skilled in the art. SUMMARY

[0006] The present application provides a vehicle-cloud integrated multi-scene prediction method based on big data driving, an electronic device and a medium, which can meet the intelligent demand of vehicle multi-scene (energy management, state estimation, control decision, etc.), based on vehicle end data and cloud computing ability, combined with deep reinforcement learning model, construct vehicle-cloud collaborative multi-scene prediction and control system, improve vehicle economy, safety and driving performance.

[0007] In a first aspect, the present application provides a vehicle-cloud integrated multi-scene prediction method based on big data driving, comprising:

[0008] The vehicle end integrates multiple controllers, collects corresponding controller data, and sends the controller data to the cloud end through the data interaction layer;

[0009] The cloud end receives the controller data, processes and predicts the controller data;

[0010] The cloud pushes the prediction result to the user end and / or the research and development end.

[0011] Preferably, the controller includes a VCU vehicle control unit, a BMS battery management system, a MCU motor controller, and an ECU electronic control unit.

[0012] Preferably, processing the controller data includes:

[0013] analyzing the controller data and predicting the aging trend of components;

[0014]

[0015] establishing a data label system and storing the data in categories.

[0016] Preferably, the processed controller data is pushed to a smart general prediction model platform of the cloud to predict the controller data.

[0017] Preferably, predicting the controller data includes:

[0018] based on the input controller data, adjusting the pre-trained time series prediction model according to the scene;

[0019] using the model to perform multi-scene prediction, including vehicle health prediction, battery health state prediction, and energy consumption prediction.

[0020] Preferably, the vehicle health prediction includes:

[0021] analyzing the controller data and predicting the aging trend of components;

[0022] The battery health state prediction includes:

[0023] estimating the remaining life of the battery through charging behavior and discharging curve;

[0024] The energy consumption prediction includes:

[0025] predicting the vehicle energy consumption trend according to driving style and road condition data.

[0026] Preferably, it further includes:

[0027] comparing and verifying the prediction result through historical data, iteratively optimizing the model parameters, and improving the prediction accuracy.

[0028] Preferably, it further includes:

[0029] constructing a database and storing the prediction result data of different scenes in categories in the database.

[0030] In a second aspect, the embodiments of the present disclosure further provide an electronic device, which comprises:​

[0031] a memory storing executable instructions;

[0032] a processor running the executable instructions in the memory to implement the big data driven vehicle-cloud integrated multi-scenario prediction method.

[0033] In a third aspect, the embodiments of the present disclosure further provide a computer readable storage medium storing a computer program, which, when executed by a processor, implements the big data driven vehicle-cloud integrated multi-scenario prediction method.

[0034] The beneficial effects are as follows:

[0035] 1) Vehicle-cloud collaborative architecture innovation: build an integrated architecture of "vehicle-end data acquisition-cloud-end intelligent processing", break through the limitations of traditional single vehicle-end or cloud-end processing, and realize real-time data uploading, model cloud computing and result bidirectional pushing.

[0036] 2) General prediction model method strategy design: develop a cloud intelligent general prediction model platform based on a general time series prediction framework, and replace the traditional special model development mode by adapting to multi-scenario prediction needs such as vehicle health, battery life and energy consumption through "model fine-tuning" to realize cross-scenario reuse of methodology.

[0037] 3) Multi-dimensional data fusion analysis: integrate multi-source heterogeneous data such as driving behavior, vehicle maintenance cycle, vehicle health status and charging behavior to form a three-dimensional data label system and provide more comprehensive and accurate data input basis for multi-scenario prediction.

[0038] 4) Optimize user experience and service value: push personalized services to users based on accurate prediction results (such as battery maintenance reminders and energy optimization suggestions) to help users use vehicles scientifically and improve vehicle use economy and safety.

[0039] 5) Strengthen data asset accumulation and iteration: build a prediction knowledge base to deposit historical prediction data and analysis results, provide data support for vehicle development optimization and new function design, and form a virtuous cycle of "data-prediction-application-iteration".

[0040] 6) Strong scene adaptability and expansibility: the model platform supports rapid adaptation to new scenes and can be applied to vehicle fault prediction and user behavior analysis by fine-tuning, meeting the diversified intelligent needs in the future.

[0041] The method and apparatus of the present application have other characteristics and advantages that will be apparent from, or elaborated on in, the accompanying drawings and the following detailed description, which, taken together, serve to explain certain principles of the present application. BRIEF DESCRIPTION OF DRAWINGS

[0042] The above and other objects, features and advantages of the present application will become more apparent from the following detailed description when taken in conjunction with the accompanying drawings in which like reference characters refer to like parts throughout the several views.

[0043] Figure 1 A flow chart showing steps of a big data driven vehicle-cloud integrated multi-scenario prediction method according to one embodiment of the present application is shown.

[0044] Figure 2 Vehicle-cloud integrated prediction architecture diagrams according to one embodiment of the present application are shown. DETAILED DESCRIPTION

[0045] Preferred embodiments of the present application will be described in more detail below. Although the following describes preferred embodiments of the present application, it should be understood that the present application can be implemented in various forms and should not be limited by the embodiments set forth herein.

[0046] To facilitate understanding of the scheme and effects of the embodiments of the present application, three specific application examples are given below. Those skilled in the art should understand that the examples are only for the convenience of understanding the present application, and any specific details thereof are not intended to limit the present application in any way.

[0047] Example 1

[0048] Figure 1 A flow chart showing steps of a big data driven vehicle-cloud integrated multi-scenario prediction method according to one embodiment of the present application is shown.

[0049] As shown in Figure 1 the big data driven vehicle-cloud integrated multi-scenario prediction method includes:

[0050] Step 101, a plurality of controllers are integrated at the vehicle end, corresponding controller data is collected, and the controller data is sent to the cloud end through a data interaction layer;

[0051] Step 102, the cloud end receives the controller data, processes and predicts the controller data;

[0052] Step 103, the cloud end pushes the prediction result to the user end and / or the research and development end.

[0053] In one example, the controller includes a VCU vehicle control unit, a BMS battery management system, a MCU motor controller, and an ECU electronic control unit.

[0054] In one example, processing the controller data includes:

[0055] Parsing the controller data into a unified format;

[0056] Data cleaning and normalization to improve data quality;

[0057] Establishing a data tag system for classified storage.

[0058] In one example, the processed controller data is pushed to a cloud-based intelligent general prediction model platform for prediction.

[0059] In one example, predicting the controller data includes:

[0060] Based on the input controller data, adjusting the pre-trained time series prediction model for scene adaptation;

[0061] Using the model to perform multi-scene prediction, including vehicle health prediction, battery health state prediction, and energy consumption prediction.

[0062] In one example, the vehicle health prediction includes:

[0063] Analyzing the controller data to predict component aging trends;

[0064] The battery health state prediction includes:

[0065] Estimating the remaining life of the battery through charging behavior and discharging curve;

[0066] The energy consumption prediction includes:

[0067] Predicting the vehicle energy consumption trend based on driving style and road condition data.

[0068] In one example, it also includes:

[0069] Comparing and verifying the prediction results through historical data, iteratively optimizing model parameters, and improving prediction accuracy.

[0070] In one example, it also includes:

[0071] Building a database to store the prediction result data of different scenes in the database.

[0072] Figure 2 The vehicle-cloud integrated prediction architecture according to one embodiment of the present application is shown.

[0073] Specifically, asFigure 2 As shown, the vehicle end is based on a vehicle platform, integrates various controllers (VCU vehicle controller, BMS battery management system, MCU motor controller, ECU electronic control unit, etc.), and realizes the control functions of core modules such as vehicle power, battery, and electronic system. Through vehicle-mounted sensors, BMS battery management system, vehicle ECU and other devices, real-time collection of user vehicle data is realized, including but not limited to:

[0074] Driving behavior data: driving style (aggressive / flat), on / off state, gear shifting state, charging behavior;

[0075] Vehicle state data: overall vehicle health status, battery health status, vehicle maintenance cycle;

[0076] Energy consumption data: overall vehicle energy consumption, battery charging and discharging parameters.

[0077] Data integration: preliminary structured integration of scattered sensor data and system state data to form standardized data units, preparing for uploading to the cloud.

[0078] The communication link between the vehicle end and the cloud established by the TBOX (vehicle-mounted intelligent terminal) is the data interaction layer. On the one hand, controller data (such as vehicle status, sensor information) is uploaded to the cloud, and on the other hand, instructions (such as control strategies, diagnosis instructions) are received from the cloud to drive the vehicle to perform corresponding operations.

[0079] The cloud architecture includes:

[0080] Data conversion platform: responsible for data access and distribution functions, collects, cleans and formats vehicle end data through data gateway, and supports the issuance of cloud algorithm results or instructions to the vehicle end;

[0081] Inference platform: integrates fault reasoning, prediction, recommendation and other algorithms, and realizes real-time analysis functions such as vehicle fault diagnosis and state prediction (such as health degree, aging trend) based on vehicle end data;

[0082] Training platform: provides algorithm development environment, covering data labeling, model training (such as machine learning, deep learning algorithms), simulation verification and other processes, and supports iterative optimization of inference platform algorithms;

[0083] Data platform: includes data center (responsible for data storage, management, labeling) and visualization BI module, supports data dashboard display, algorithm simulation analysis and visual decision-making, and provides data support for algorithm integration.

[0084] Data processing flow:

[0085] Data analysis: analyze uploaded heterogeneous data (such as CAN bus data, protocol data) and convert it into a unified format;

[0086] Data preprocessing: Perform data cleaning (remove outliers, null values), normalization processing to improve data quality;

[0087] Data management and query: Establish data tag system, classified storage (such as driving behavior class, vehicle health class), support researchers to query and call as needed.

[0088] Push the processed data to the "general time series prediction platform" of the cloud intelligent general prediction model platform.

[0089] Model execution prediction:

[0090] Model fine-tuning: Based on the input data features, fine-tune the pre-trained time series prediction model (such as LSTM, Transformer) for scene adaptation, for example, optimize the model parameters for battery health prediction;

[0091] Prediction service: Use the model to perform multi-scenario prediction, including:

[0092] Whole vehicle health prediction: Analyze sensor data to predict component aging trends;

[0093] Battery health state prediction: Estimate battery remaining life through charging behavior and discharge curve;

[0094] Energy consumption prediction: Combine driving style and road condition data to predict whole vehicle energy consumption trends.

[0095] Verification and optimization: Compare and verify the prediction results with historical data, iteratively optimize model parameters, and improve prediction accuracy.

[0096] Build a database to store prediction results (such as whole vehicle health prediction value, battery life prediction value, energy consumption prediction value) into the database, establish a "prediction knowledge base", and store prediction results data of different scenarios, forming reusable knowledge assets.

[0097] Through data interaction and functional cooperation of cloud platforms, form a "data input-algorithm training-reasoning application-visualization output" closed loop link, realize the deep utilization and value mining of vehicle data.

[0098] Push the prediction results to the user end and / or research and development end, including:

[0099] User empowerment: Push the prediction results (such as battery maintenance reminders, energy consumption optimization suggestions) to users through car machine systems or mobile phone APPs to assist users in optimizing vehicle behavior.

[0100] R&D end support: push deep analysis results (such as vehicle failure prediction trend, user behavior statistics) to R&D personnel, provide data support for vehicle iteration optimization and new function development, form a closed-loop ecology of "data collection-prediction analysis-application optimization".

[0101] The vehicle end of the application is responsible for real-time data collection and instruction execution. The cloud end realizes vehicle state analysis, strategy optimization and function iteration through data conversion, intelligent algorithm and data management. Finally, through vehicle-cloud two-way communication, an integrated intelligent service system of "data-driven decision-making and decision-making feeding control" is achieved, and the intelligent level of the vehicle and the user experience are improved.

[0102] The application proposes a multi-scene intelligent prediction and control method based on a big data driven vehicle-cloud collaborative architecture. The vehicle end dynamically adjusts the control strategy (such as energy distribution priority, torque vector control parameter) based on the cloud end pre-trained model output (such as battery health prediction value, user behavior pattern) combined with real-time sensor data. The cloud end builds a multi-modal time series data fusion framework, integrates vehicle sensor data (such as battery voltage, motor temperature), user behavior logs (such as driving habits, charging frequency) and external environment data (such as weather, road conditions). A migratable feature engineering module is designed to extract cross-scene common features (such as energy consumption pattern, component aging trend) through a dynamic feature selection algorithm, and multi-task learning is achieved combined with an attention mechanism. A lightweight time series prediction model (such as an improved Transformer architecture) is developed to support fast training and online incremental updating of the cloud end, and a vehicle-end adapted version is generated through model distillation technology to realize vehicle-cloud collaborative optimization closed loop and build a two-way feedback iteration system. The vehicle end feeds back control effect data (such as energy consumption optimization rate, fault warning accuracy) to the cloud end model training, and the cloud end updates model parameters and control strategies through OTA, develops scene adaptive transfer learning algorithm, automatically adjusts model weights according to different vehicle models and use environments, realizes "one cloud multiple ends" fast adaptation, and reduces the development cost of special models.

[0103] Example 2

[0104] The present disclosure provides an electronic device, comprising: a memory storing executable instructions; a processor running the executable instructions in the memory to implement the above-mentioned big data driven vehicle-cloud integrated multi-scene prediction method.

[0105] The electronic device according to the embodiments of the present disclosure comprises a memory and a processor.

[0106] The memory is configured to store non-transitory computer readable instructions. Specifically, the memory can include one or more computer program products that can include various forms of computer readable storage media, such as volatile memory and / or non-volatile memory. The volatile memory, for example, can include random access memory (RAM), cache memory, and / or the like. The non-volatile memory, for example, can include read only memory (ROM), hard disk, flash memory, and / or the like.

[0107] The processor can be a central processing unit (CPU) or other form of processing unit that has data processing and / or instruction execution capabilities, and can control other components in the electronic device to perform desired functions. In one embodiment of the present disclosure, the processor is configured to execute the computer readable instructions stored in the memory.

[0108] Those skilled in the art will understand that, in order to solve the technical problem of how to obtain a good user experience effect, the embodiment can also include well-known structures such as a communication bus, an interface, and the like, which should also be included in the protection scope of the present disclosure.

[0109] Detailed descriptions of the embodiments can refer to the corresponding descriptions in the foregoing embodiments, which will not be repeated here.

[0110] Example 3

[0111] The embodiment of the present disclosure provides a computer readable storage medium, which stores a computer program. The computer program is executed by a processor to implement the vehicle-cloud integrated multi-scenario prediction method based on big data driving.

[0112] The computer readable storage medium according to the embodiment of the present disclosure has non-transitory computer readable instructions stored thereon. When the non-transitory computer readable instructions are executed by a processor, all or part of the steps of the method of the embodiments of the present disclosure are executed.

[0113] The computer readable storage medium described above includes, but is not limited to, optical storage media (such as CD-ROM and DVD), magneto-optical storage media (such as MO), magnetic storage media (such as magnetic tape or mobile hard disk), media with built-in rewritable non-volatile memory (such as memory card), and media with built-in ROM (such as ROM cartridge).

[0114] Those skilled in the art will understand that the purpose of the above description of the embodiments of the present disclosure is only to exemplarily illustrate the beneficial effects of the embodiments of the present disclosure, and is not intended to limit the embodiments of the present disclosure to any examples given.

[0115] Having described various embodiments of the application, it is to be understood that the above description is meant to be illustrative only, and that many modifications and variations of the embodiments are possible without departing from the scope and spirit of the described embodiments. Many modifications and variations of the described embodiments are possible in light of the above teachings. It is, therefore, to be understood that within the scope of the appended claims and their equivalents, the described embodiments can be practiced otherwise than as specifically described.

Claims

1. A big data driving-based vehicle-cloud integrated multi-scenario prediction method, characterized in that, The method comprises the following steps: The vehicle end integrates various controllers, collects corresponding controller data, and sends the data to the cloud through a data interaction layer; The cloud receives the controller data, processes and predicts the data; The cloud pushes the prediction results to the user end and / or the research and development end.

2. The big data driven vehicle cloud integrated multi-scenario prediction method according to claim 1, wherein, The controllers include a VCU whole vehicle controller, a BMS battery management system, a MCU motor controller, and an ECU electronic control unit.

3. The big data driven vehicle cloud integrated multi-scenario prediction method according to claim 1, wherein, Processing the controller data includes the following steps: Analyzing the controller data and converting it into a unified format; Performing data cleaning and normalization to improve data quality; Establishing a data label system for classified storage.

4. The big data driven vehicle cloud integrated multi-scenario prediction method of claim 1, wherein, The processed controller data is pushed to the intelligent general prediction model platform of the cloud to predict the controller data.

5. The big data driven vehicle cloud integrated multi-scenario prediction method according to claim 4, wherein, Predicting the controller data includes the following steps: Based on the input controller data, the pre-trained time series prediction model is adjusted for scene adaptation; Using the model to perform multi-scenario prediction, including whole vehicle health prediction, battery health state prediction, and energy consumption prediction.

6. The big data driven car-cloud integrated multi-scenario prediction method according to claim 1, wherein, The whole vehicle health prediction includes the following steps: Analyzing the controller data to predict the aging trend of components; The battery health state prediction includes the following steps: Estimating the remaining life of the battery through charging behavior and discharge curve; The energy consumption prediction includes the following steps: According to the driving style and road condition data, the whole vehicle energy consumption trend is predicted.

7. The big data driven car-cloud integrated multi-scenario prediction method according to claim 1, wherein, The method further includes the following steps: Comparing and verifying the prediction results with historical data to iteratively optimize model parameters and improve prediction accuracy.

8. The big data driven car-cloud integrated multi-scenario prediction method according to claim 1, wherein, The method further includes the following steps: Building a database to store the prediction result data of different scenarios in the database.

9. An electronic device, comprising: The electronic device includes: A memory storing executable instructions; A processor running the executable instructions in the memory to implement the multi-scenario prediction method based on big data driving of vehicle-cloud integration according to any one of claims 1-8.

10. A computer-readable storage medium, characterized in that, The computer readable storage medium stores a computer program which is executed by the processor to implement the multi-scenario prediction method based on big data driving of vehicle-cloud integration according to any one of claims 1-8.

Images