Control method and device of vehicle, electronic equipment and vehicle

Abstract

Embodiments of the present application provide a vehicle control method, device, electronic equipment and vehicle. The method comprises: obtaining road condition information of a to-be-traveled road of the vehicle; determining, based on the road condition information, driving style data of the vehicle on the to-be-traveled road, wherein the driving style data is used to represent the driving style of the vehicle; predicting, based on the road condition information and the driving style data, a predicted vehicle speed sequence of the vehicle, wherein the predicted vehicle speed sequence is used to represent the vehicle speed of the vehicle when driving at different positions on the to-be-traveled road; determining, based on the predicted vehicle speed sequence, control data of the vehicle, wherein the control data is used to represent the start-stop strategy and power distribution strategy of a range extender in the vehicle; controlling the start-stop time of the range extender according to the start-stop strategy, and controlling the power generation of the range extender according to the power distribution strategy. The present application solves the technical problem of low fuel economy of the vehicle.

Description

Technical Field

[0001] This application relates to the field of vehicle intelligent control and energy management technology, and in particular to a vehicle control method, device, electronic device, and vehicle. Background Technology

[0002] Currently, range-extended electric vehicles (REEVs) have become a key technological route for new energy vehicles, experiencing rapid market share growth in recent years. This type of vehicle effectively combines the smoothness of pure electric drive, low-cost charging advantages, and the absence of range anxiety. Its cost is also more advantageous compared to comparable pure electric models, making it uniquely adaptable to areas with limited charging options, long-distance driving scenarios, and cold northern regions. However, despite advancements in range-extending technology, current energy management strategies still have some significant shortcomings, resulting in persistent technical issues related to low fuel economy. Summary of the Invention

[0003] This application provides a vehicle control method, device, electronic device, and vehicle, aiming to solve the technical problem of low fuel economy in vehicles.

[0004] According to one embodiment of this application, a vehicle control method is provided. The method may include: acquiring road condition information of the road to be driven by the vehicle; determining driving style data of the vehicle on the road to be driven based on the road condition information, wherein the driving style data is used to characterize the driving style of the vehicle; predicting a predicted vehicle speed sequence based on the road condition information and the driving style data, wherein the predicted vehicle speed sequence is used to characterize the vehicle speed when driving at different positions on the road to be driven; determining control data of the vehicle based on the predicted vehicle speed sequence, wherein the control data is used to characterize the start-stop strategy and power distribution strategy of the range extender in the vehicle; controlling the start-stop timing of the range extender according to the start-stop strategy, and controlling the power generation of the range extender according to the power distribution strategy.

[0005] The above-mentioned optional embodiments of this application can achieve the following beneficial effects: by fusing road condition information and driving style data for deep learning prediction, the speed change trend of the vehicle on the road to be driven can be accurately predicted. This allows for advance planning of the range extender's start-up and load management strategies, avoiding temporary starts in inefficient areas or excessive reliance on battery power, thus optimizing energy use, extending battery life, and improving the driving experience. Specifically, by acquiring road condition information and using the road condition information and quantified driving style data as input, a predicted vehicle speed sequence is obtained. Based on the predicted vehicle speed sequence, the control data of the range extender (i.e., the engine) in the vehicle is determined. Through precise control of the engine and battery, dynamic balance energy distribution is achieved, significantly improving the vehicle's fuel economy, extending the range extender's life, and reducing start-stop smoothness (NVH). Simultaneously, it enables the vehicle to achieve optimal energy management under different operating conditions and driving habits.

[0006] Optionally, based on road condition information, the driving style data of the vehicle on the road to be driven is determined, including: determining the road type of the road to be driven based on road condition information; filtering at least one historical driving information that matches the road type from the vehicle's historical driving information set, wherein the historical driving information set is used to characterize the vehicle's historical driving state; and determining the driving style data based on at least one historical driving information.

[0007] The above-mentioned optional embodiments of this application can achieve the following beneficial effects: by integrating real-time road condition information and historical driving data, the driving style of the vehicle in a specific road type can be accurately identified, thereby allowing for personalized adjustment of the range extender's operating parameters. This avoids the neglect of driver habits and road conditions in traditional control strategies, achieving the goal of improving energy utilization efficiency and driving experience.

[0008] Optionally, based on at least one historical driving information, driving style data is determined, including: performing clustering processing on at least one historical driving information to obtain clustering results, wherein the clustering results are used to characterize the degree of correlation between the historical driving information and different types of driving style data; and transforming the clustering results to obtain the vehicle's driving style data.

[0009] The above-mentioned optional embodiments of this application can achieve the following beneficial effects: by clustering and analyzing historical driving information and converting it into driving style data, the driving habits and styles of vehicles can be accurately identified and quantified, thereby enabling the formulation of personalized energy management strategies that match driving styles. This avoids energy waste or poor driving experience caused by using fixed or single energy allocation schemes, and achieves the goal of optimizing the range extender operating point and battery state of charge strategy under different driving scenarios, thereby improving the overall energy utilization efficiency and driving comfort of the vehicle.

[0010] Optionally, the road condition information includes short field of view (SFL) information, which is used to characterize the road condition of the vehicle within the target range. Based on the road condition information and driving style data, a predicted vehicle speed sequence is predicted, including: inputting the SFL information and driving style data into a prediction model to predict the predicted vehicle speed sequence. The prediction model is trained based on driving style data samples, SFL information samples, and vehicle speed sequence samples corresponding to the driving style data samples and SFL information samples. The driving style data samples are used to characterize the vehicle's historical driving style, and the vehicle speed sequence samples are used to characterize the vehicle's historical speed sequence.

[0011] The above-mentioned optional embodiments of this application can achieve the following beneficial effects: by fusing short field of vision information and driving style data into a deep learning prediction model, it is possible to accurately predict the speed change of the road segment to be driven, thereby dynamically adjusting the start-stop strategy and power distribution of the range extender, avoiding uneven energy distribution or ineffective operation of the range extender due to information lag or single processing in traditional control, thus achieving the goal of improving fuel economy and ensuring driving smoothness.

[0012] Optionally, the road condition information also includes long field of view information, which is used to characterize the road condition status of the entire road to be driven. The method further includes: inputting the long field of view information into the prediction model to predict the vehicle speed of the entire road to be driven.

[0013] The above-mentioned optional embodiments of this application can achieve the following beneficial effects: by integrating long-distance predicted road condition information into the prediction model, the speed change trend of the entire road to be driven can be comprehensively predicted, thereby globally optimizing the energy distribution strategy, avoiding the limitations of local optimization caused by relying only on short-term information or real-time feedback, and achieving the goal of improving the energy utilization efficiency and driving comfort of the vehicle throughout the entire journey.

[0014] Optionally, based on the predicted vehicle speed sequence, the vehicle control data is determined, including: determining the state of charge (SOC) of the battery in the vehicle; determining the SOC trajectory of the battery on the road to be traveled based on the predicted vehicle speed and SOC throughout the journey, wherein the SOC trajectory is used to characterize the SOC of the battery at different times; determining the energy consumption data of the range extender under multiple equivalent factors based on the SOC trajectory and the predicted vehicle speed sequence, wherein the equivalent factors are used to balance the energy distribution between the range extender and the battery; determining a target equivalent factor from the multiple equivalent factors based on the energy consumption data; determining the power contribution ratio of the battery and the range extender using the target equivalent factor; and converting the power contribution ratio to obtain control data.

[0015] The above-mentioned optional embodiments of this application can achieve the following beneficial effects: by comprehensively analyzing and predicting the vehicle speed sequence and the battery state of charge, the state of charge trajectory can be scientifically planned, and the energy consumption distribution under different equivalent factors can be accurately calculated. Thus, the optimal equivalent factor (i.e., the target equivalent factor) can be reasonably selected, and the power contribution of the battery and the range extender can be allocated in the optimal ratio. This avoids the blindness or singularity of energy allocation in traditional control, and achieves the goal of achieving the lowest energy consumption operation and maximizing fuel economy throughout the entire range.

[0016] Optionally, based on road condition information, determine the vehicle's driving style data on the road to be driven, including: converting the road condition information in the spatial domain according to a preset step size to obtain equally spaced information in the time domain; and determining the driving style data based on the equally spaced information.

[0017] The above-mentioned optional embodiments of this application can achieve the following beneficial effects: by converting the road condition information in the spatial domain into information in the time domain with equal intervals according to a preset step size, the information can be standardized and processed continuously, thereby more accurately identifying and analyzing driving styles, avoiding data analysis errors caused by discontinuous information or inconsistent formats, and achieving the goal of improving the accuracy of driving style recognition and the performance of prediction models.

[0018] According to one embodiment of this application, a vehicle control device is also provided, comprising: an acquisition unit for acquiring road condition information of a road to be driven by the vehicle; a first determination unit for determining driving style data of the vehicle on the road to be driven based on the road condition information, wherein the driving style data is used to characterize the driving style of the vehicle; a prediction unit for predicting a predicted vehicle speed sequence based on the road condition information and the driving style data, wherein the predicted vehicle speed sequence is used to characterize the vehicle speed when driving at different positions on the road to be driven; a second determination unit for determining control data of the vehicle based on the predicted vehicle speed sequence, wherein the control data is used to characterize the start-stop strategy and power distribution strategy of the range extender in the vehicle; and a control unit for controlling the start-stop timing of the range extender according to the start-stop strategy and controlling the power generation of the range extender according to the power distribution strategy.

[0019] According to another aspect of the embodiments of this application, an electronic device is provided, including a processor; a memory for storing processor-executable instructions; wherein the processor is configured to execute instructions to implement the method described above.

[0020] According to another aspect of the embodiments of this application, a computer-readable storage medium is provided, wherein a computer program is stored in the computer-readable storage medium, and the computer program is configured to perform the above-described method when run by a processor.

[0021] According to another aspect of the embodiments of this application, a computer program product is provided, including a computer program that, when executed by a processor, implements the above-described method.

[0022] According to another aspect of the embodiments of this application, a vehicle is provided, including an on-board processor and an on-board memory, wherein the on-board memory is used to store a computer program; and the on-board processor is used to execute the computer program stored in the memory to implement the above method.

[0023] It should be noted that the general descriptions above and the detailed descriptions that follow are merely examples and explanations of this application and do not constitute a limitation on this application. Attached Figure Description

[0024] Figure 1 This is a flowchart of a vehicle control method provided in an embodiment of this application;

[0025] Figure 2 This is a flowchart of a scene- and style-based vehicle speed sequence prediction method provided in an embodiment of this application;

[0026] Figure 3 This is a flowchart of a three-layer optimization control method for a range extender provided in an embodiment of this application;

[0027] Figure 4 This is a flowchart of a traffic information processing method provided in an embodiment of this application;

[0028] Figure 5 This is a flowchart of a scenario prediction-based range-extended vehicle control method provided in an embodiment of this application;

[0029] Figure 6 This is a flowchart of control command execution and status feedback provided in one embodiment of this application;

[0030] Figure 7 This is a structural diagram of a vehicle control device provided in one embodiment of this application;

[0031] Figure 8 This is a structural diagram of an electronic device provided in one embodiment of this application. Detailed Implementation

[0032] To make the technical problems, technical solutions, and beneficial effects solved by this application clearer, the following detailed description is provided in conjunction with embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0033] Currently, range-extended electric vehicles (REEVs) are one of the important technological routes for new energy vehicles, and their market share has increased rapidly in recent years. This type of vehicle can effectively combine the smoothness of pure electric drive, the advantages of low-cost charging, and the absence of range anxiety. Compared with pure electric models of the same class, it has a cost advantage, thus demonstrating unique market adaptability in areas with inconvenient charging, long-distance self-driving scenarios, and cold northern regions.

[0034] However, the field of range extender technology is developing towards miniaturization, high efficiency, high quietness, and intelligent management. On the power side, this typically involves reducing the size and weight of the range extender by employing technologies such as flux motors, continuously improving engine thermal efficiency and optimizing it, and using technologies like harmonic injection to improve the NVH (noise, vibration, and harshness) experience. Predictive energy management and continuously variable transmission (CVT) management, as important components of intelligent management, are considered key technologies for optimizing electromechanical operation and energy consumption, and balancing fuel consumption and driving experience. Despite the progress made in range extender technology, current energy management strategies still have some significant shortcomings.

[0035] First, the aforementioned methods suffer from a significant "battery depletion" problem. Traditional control strategies are largely rule-based, such as triggering startup or instantaneous optimization based on a fixed State of Charge (SOC) threshold. This lack of anticipation of future driving conditions causes the range extender to frequently operate in inefficient zones. Consequently, when the battery level is low, many range-extended vehicles experience three major pain points: soaring fuel consumption, reduced power, and increased vibration and noise. Second, these methods typically only passively respond to real-time vehicle conditions (such as current speed, power demand, and battery SOC), failing to effectively utilize prior information like navigation, real-time traffic information (V2X), and road gradient for forward planning. Therefore, energy management lacks foresight. For example, before entering a long uphill climb or congested area, the energy distribution strategy cannot be intelligently adjusted in advance, potentially leading to forced startup of the range extender under high load on an incline or rapid battery depletion in congestion, impacting economy and comfort. Third, traditional rule-based control strategies have poor adaptability, making it difficult to achieve an optimal balance of multiple objectives (economy, comfort, and battery life) across the entire driving range. While real-time optimization strategies (such as ECMS) offer some improvement, their optimization scope is limited, making it difficult to cope with complex and ever-changing real-world traffic scenarios and individual driver habits. Furthermore, they place high demands on computational resources, resulting in insufficient global optimization capabilities. Existing systems typically employ fixed control logic, making it difficult to adaptively adjust to different driver styles (aggressive, mild), travel habits (commuting, long-distance), and real-time changing traffic environments, thus failing to fully realize the potential of predictive control.

[0036] In related technologies, a route prediction-based energy management method for range-extended electric vehicles (REEVs) is proposed. This method first determines the current travel route based on the vehicle's current starting and ending points. Then, based on the predicted speed and vehicle weight at each segment of the predicted route, the predicted total energy consumption required to complete the entire trip is calculated. The core decision-making point of this method is to determine whether the vehicle's remaining battery power can meet the predicted energy consumption. If the remaining battery power does not meet the predicted energy consumption, the system determines the preset operating stages of the range extender and the control strategies for each stage, and controls the power generation of the range extender and the charging and discharging state of the power battery according to these strategies. If the remaining battery power meets the predicted energy consumption, the power battery is controlled to discharge to provide power to the vehicle. In other words, this method simply improves fuel economy by combining route planning and energy management to optimize the operating range of the range extender. However, this method only makes decisions based on the total energy consumption prediction for the entire trip. This is a very macroscopic and static prediction, resulting in coarse prediction granularity and an inability to cope with dynamic changes. For example, it may assume that the battery is sufficient for the entire trip and choose to drive on pure electric power. However, if the vehicle suddenly needs to climb a long, steep hill, it will face a sudden high power demand, which may lead to insufficient power response or force the range extender to start and operate at high load in an inefficient zone, resulting in higher fuel consumption and a poorer experience. At the same time, this method lacks real-time rolling optimization, and the strategy is rigid. Once the actual road conditions deviate from the prediction (such as encountering sudden congestion or changing the route temporarily), the entire plan may become invalid, and the system will revert to the traditional passive control mode, unable to make online real-time corrections, thus resulting in technical problems with poor fault tolerance.

[0037] It's important to note that the effectiveness of the above methods depends entirely on the accuracy of the predicted route and speed. If the navigation map's data on road gradients, speed limits, etc., is inaccurate, the predicted energy consumption will have significant deviations. Furthermore, unexpected accidents and weather changes can lead to completely incorrect predictions, resulting in a technical problem of low control accuracy. Additionally, while the methods rely on "historical travel routes" and "predicted speeds," they don't address how to adapt to different driving styles, potentially leading to lower fuel economy.

[0038] To address the aforementioned problems, embodiments of this application provide a vehicle control method. This method may include: acquiring road condition information of the road to be traveled; determining driving style data of the vehicle on the road to be traveled based on the road condition information, wherein the driving style data characterizes the vehicle's driving style; predicting a predicted vehicle speed sequence based on the road condition information and the driving style data, wherein the predicted vehicle speed sequence characterizes the vehicle speed at different locations on the road to be traveled; determining control data of the vehicle based on the predicted vehicle speed sequence, wherein the control data characterizes the start-stop strategy and power distribution strategy of the range extender in the vehicle; controlling the start-stop timing of the range extender according to the start-stop strategy, and controlling the power generation of the range extender according to the power distribution strategy.

[0039] The vehicle control method provided in this application achieves the following technical effects: acquiring road condition information, using the road condition information and quantified driving style data as input to obtain a predicted vehicle speed sequence, determining the control data of the range extender (i.e., the engine) in the vehicle based on the predicted vehicle speed sequence, and achieving precise control of the engine and battery through the control data, thereby solving the technical problem of low vehicle fuel economy and achieving the technical effect of improving vehicle fuel economy.

[0040] Example 1

[0041] This application provides a vehicle control method, with reference to... Figure 1 , Figure 1 This is a flowchart of a vehicle control method according to an embodiment of this application, which may include the following steps:

[0042] S102: Obtain road condition information for the road the vehicle is about to travel on.

[0043] In step S102, the aforementioned road to be traveled can be the vehicle's intended route, used to determine the vehicle's travel path, and can be the road the vehicle will travel on. The aforementioned road condition information can be scene information, including but not limited to: long-field-of-view scene information and short-field-of-view scene information, which can be used to characterize the road condition status of a portion of the road to be traveled, or to characterize the road condition status of the entire road to be traveled, for example, it can be used to characterize traffic flow, vehicle speed, and other data in the road to be traveled. The aforementioned vehicle can be a range-extended vehicle.

[0044] Optionally, road condition information about the road the vehicle is about to travel on can be collected through the vehicle's intelligent connected system and Advanced Driver Assistance Systems (ADAS) map data. This road condition information may include, but is not limited to, real-time traffic flow, road gradient, speed limit signs, road type (e.g., highway, city street) and other external conditions that may affect vehicle speed.

[0045] For example, the intelligent connectivity module equipped in a range-extended electric vehicle (REEV) can acquire the necessary scenario information for driving. This scenario information can include road condition information, encompassing both long-range and short-range field-of-view data. Suppose a REEV is preparing to travel from city A to city B, and its navigation system has already planned the entire route. Based on this planned route, information such as traffic density, average speed, and gradient can be obtained from the ADAS map. Simultaneously, real-time traffic data can be integrated to determine if there are traffic congestion or construction zones ahead, thus obtaining the road condition information for the route.

[0046] S104: Based on road condition information, determine the vehicle's driving style data on the road to be traveled.

[0047] In step S104, the above driving style data can be used to characterize the driving style of the vehicle, which can be the driving style of the vehicle, and can be used to quantify the aggressiveness of the vehicle during driving. It can also be called driving style data.

[0048] Optionally, driving style data for different types of road sections can be determined in advance. After obtaining road condition information, the road conditions of the road to be driven can be determined based on the road condition information, such as whether it is an uphill or downhill section, an urban section or a suburban section, etc. Based on this information, the driving style data of the vehicle on that type of road can be determined.

[0049] Optionally, after collecting road condition information, the collected information, combined with vehicle operation data captured by the vehicle's built-in sensors (such as accelerator pedal depth, steering angle, and braking force), can be used to identify and quantify the vehicle's driving style using a pre-trained model, such as a Gaussian Mixture Model (GMM), to obtain the vehicle's driving style data on the road to be driven. Based on this driving style data, the vehicle's driving style data for that type of road can be determined.

[0050] For example, by acquiring road condition information, the system can analyze the driver's past driving records on similar roads and discover a behavior pattern where the vehicle tends to accelerate rapidly and brake suddenly at traffic lights, indicating a relatively aggressive driving style. Therefore, the system's GMM model outputs a high probability value for an aggressive driving style, which is then quantified into a specific numerical value, such as 0.8, indicating a current driving style that leans towards aggressiveness.

[0051] In this embodiment, after acquiring road condition information, driving style can be identified in real time based on the road condition information. After adding driving style parameters, the root mean square error (RMSE) of short-term vehicle speed prediction is significantly reduced, thereby effectively improving prediction accuracy. At the same time, it can adapt to the differences in energy demand of drivers with different styles. For example, aggressive drivers consume more energy than conservative drivers, and the range extender start-up threshold is dynamically adjusted to achieve personalized control.

[0052] S106: Based on road condition information and driving style data, the predicted vehicle speed sequence is obtained.

[0053] In step S106, the aforementioned predicted vehicle speed sequence may refer to data generated by predicting the expected vehicle speed at various locations (or time points) on the road to be driven, and may be determined based on the vehicle's current driving status, road conditions, and driving style data.

[0054] Optionally, road condition information, including but not limited to traffic flow, historical average vehicle speed, and road gradient, is obtained from the ADAS map, encompassing both long-field-of-view (covering the entire path) and short-field-of-view (the local road ahead of the vehicle). Subsequently, real-time identification of the vehicle's driving style data quantifies the vehicle's aggressiveness into a numerical value, which is then used as input to the prediction model. Next, a deep learning model, combining scene information and driving style data, predicts a short-term vehicle speed sequence to obtain a predicted speed sequence. This prediction result can guide the start-stop and power distribution decisions of the range extender, achieving optimal control of the vehicle's driving state.

[0055] Optionally, road condition information and driving style data are used as input to predict the expected vehicle speed at different driving positions using a deep learning model (e.g., LSTM+Bi-LSTM+LSTM architecture) to obtain a predicted speed sequence. This deep learning model not only considers macroscopic road condition trends but also refines to local speed limits and stopping points, as well as vehicle driving style preferences, generating speed sequences covering both full-range and short-term predictions.

[0056] For example, it is predicted that vehicles will travel at approximately 110 km / h on the highway section before entering city B, while upon entering the urban area, the speed is expected to decrease to around 50 km / h due to traffic lights and speed limits. Furthermore, considering aggressive driving habits, it is predicted that vehicle speeds within the urban area will be slightly above average, assuming traffic conditions permit.

[0057] S108: Determine vehicle control data based on predicted vehicle speed sequence.

[0058] In step S108, the aforementioned control data can refer to specific instruction parameters used to directly guide the operation of the vehicle (especially the range extender), including but not limited to the start-stop timing of the range extender, power generation, and battery charging / discharging status. It can be the actual operating instructions output by the energy management system, used to adjust the operating status of the vehicle's power system, and may include, but is not limited to, the power demand of the range extender, battery discharge or charging instructions, and engine start-stop signals. It can ensure that the vehicle operates in the most efficient manner under different driving conditions. The aforementioned start-stop strategy can be used to determine when the range extender starts and when it stops. This strategy reduces unnecessary engine starts, thereby reducing fuel consumption, improving comfort, and extending engine life. The aforementioned power distribution strategy can be used to distribute the total power of the power system between the range extender and the battery to meet the vehicle's power needs while optimizing energy efficiency. This can include adjustments to parameters such as the range extender's power generation, battery charging / discharging power, and the motor's auxiliary drive power.

[0059] For example, the aforementioned start-stop strategy can be generated based on predicted vehicle speed sequences and the battery's state of charge (SOC) trajectory. A dynamic programming (DP) algorithm is used to plan the global SOC trajectory to ensure the battery is in an optimal state of charge and discharge throughout the journey, avoiding deep discharge or overcharging. Subsequently, the start-stop strategy of the range extender is adjusted in real time by combining the calculation results of the Pondrigen Minimum Principle (PMP) and the Equivalent Minimum Fuel Consumption Strategy (ECMS) algorithm. For instance, it can predict when approaching a long uphill section or when the load is heavy, and activate the range extender in advance to reserve power; while in low-demand phases, such as low-speed urban driving, the range extender can be kept off, driving in pure electric mode, reducing the frequency of engine start-stop, and improving fuel economy and driving experience.

[0060] For another example, the aforementioned power allocation strategy can be implemented using the ECMS algorithm. Based on the predicted vehicle speed sequence, driving style recognition results, and the optimized equivalent factor (λ), the optimal power allocation between the range extender and the battery can be calculated. For instance, for routes primarily involving high-speed driving, a higher engine power can be set to meet the high power demands of high-speed driving; while in urban congestion or low-speed driving conditions, the ECMS algorithm reduces the power output of the range extender to allow the vehicle to rely on battery power as much as possible, reducing fuel consumption. Furthermore, the power allocation strategy can be dynamically adjusted based on driving style. For aggressive drivers, the range extender will reserve more power to handle sudden acceleration needs, while for conservative drivers, the system tends to reduce the total power of the range extender for more economical operation.

[0061] Optionally, based on the predicted vehicle speed sequence, a three-level optimization strategy (dynamic programming (DP), Pondrikin minimum principle (PMP), and equivalent fuel consumption minimum strategy (ECMS)) is used to calculate the optimal start-stop timing and power distribution strategy of the range extender to obtain the corresponding control data. The vehicle is then controlled using the control data to optimize energy utilization efficiency and reduce energy consumption.

[0062] For example, based on the predicted vehicle speed sequence, the DP algorithm can be used to plan the global SOC trajectory to ensure that the battery maintains an ideal charge range of 20%-90% throughout the journey. Then, the PMP algorithm dynamically optimizes the predicted information every kilometer using an equivalent factor λ to adapt to changes in traffic conditions in the short to medium term. Finally, the ECMS algorithm determines the power output that the range extender should provide at the current moment based on the real-time vehicle speed and the optimized λ value, ensuring that the vehicle operates efficiently under all conditions. Especially under aggressive driving styles, the system will activate the range extender in advance to reserve sufficient power and avoid power delays or insufficient power during temporary acceleration.

[0063] S110: Control the start-up and shutdown timing of the range extender according to the start-up and shutdown strategy, and control the power generation of the range extender according to the power distribution strategy.

[0064] In step S110, after obtaining the control data, the vehicle's control system can start or stop the range extender in a timely manner according to the start-stop strategy and power distribution strategy calculated above, and adjust the power generation of the range extender to meet the needs of vehicle operation, while achieving a balance between the lowest energy consumption and the best comfort.

[0065] Optionally, the start-stop timing of the range extender can be controlled according to the start-stop strategy, and the power generation of the range extender can be controlled according to the power distribution strategy. Furthermore, the range extender control commands can be deployed to a multi-core heterogeneous hardware platform, and the control actions can be executed through the hardware platform; vehicle status parameters can be collected in real time and fed back to the control module to dynamically adjust the range extender control parameters and form a control closed loop.

[0066] For example, before a vehicle enters an uphill section, if the speed is predicted to decrease, and considering the high power demand on the slope, control data can be used to instruct the range extender to increase power generation in advance to assist the electric motor in overcoming the slope resistance. When the vehicle enters the low-speed driving phase in the city, if the speed is predicted to remain at a low level for an extended period, control data can be used to reduce the workload of the range extender, allowing the vehicle to operate in pure electric mode as much as possible, reducing fuel consumption, and also reducing vibration and noise from engine startup.

[0067] Based on steps S102 to S110 above, road condition information of the road to be driven by the vehicle is obtained; based on the road condition information, driving style data of the vehicle on the road to be driven is determined, wherein the driving style data is used to characterize the driving style of the vehicle; based on the road condition information and driving style data, a predicted vehicle speed sequence is predicted, wherein the predicted vehicle speed sequence is used to characterize the vehicle speed when driving at different positions on the road to be driven; based on the predicted vehicle speed sequence, control data of the vehicle is determined, wherein the control data is used to characterize the start-stop strategy and power distribution strategy of the range extender in the vehicle; according to the start-stop strategy, the start-stop timing of the range extender is controlled, and according to the power distribution strategy, the power generation of the range extender is controlled. That is, this embodiment obtains road condition information, uses the road condition information and quantified driving style data as input to obtain a predicted vehicle speed sequence, and determines the control data of the range extender (i.e., the engine) in the vehicle based on the predicted vehicle speed sequence. Through the control data, the purpose of precise control of the engine and battery is achieved, thereby solving the technical problem of low fuel economy of the vehicle and realizing the technical effect of improving the fuel economy of the vehicle.

[0068] The method described in this embodiment will now be further explained.

[0069] As an optional embodiment, step S104, based on road condition information, determines the driving style data of the vehicle on the road to be driven, including: determining the road type of the road to be driven based on road condition information; filtering at least one historical driving information that matches the road type from the vehicle's historical driving information set, wherein the historical driving information set is used to characterize the vehicle's historical driving state; and determining the driving style data based on at least one historical driving information.

[0070] In this embodiment, the road type mentioned above may include, but is not limited to, highways, arterial roads, and secondary roads. The historical driving information set may include historical driving information of the vehicle at multiple historical moments, including but not limited to: vehicle speed, accelerator pedal opening, longitudinal acceleration, lateral acceleration, steering wheel angle, etc., which can be used to characterize historical driving states and may include multiple raw signals. The historical driving information may be core features selected from the historical driving information set, or it may be data calculated based on the historical driving dataset, including but not limited to: average accelerator pedal position, standard deviation of accelerator pedal rate of change, maximum accelerator pedal rate of change, maximum longitudinal acceleration, maximum vehicle speed, mean accelerator pedal rate of change, maximum lateral acceleration, etc.

[0071] Optionally, after obtaining road condition information, driving scenarios can be divided based on preprocessed road condition information (i.e., scenario information). The driving scenario can be determined by "road type + driving behavior". The driver's operation signal and vehicle status signal features (i.e., at least one historical driving information) under the divided road type are extracted. Clustering algorithm is used to analyze at least one historical driving information to identify the driving style of the vehicle in the current road type and quantify the aggressiveness of the driving style.

[0072] Optionally, road condition information is acquired, which may include long-field-of-view (LVR) and short-field-of-view (SVR) information. The LVR information can be used for long-term vehicle speed prediction and may include, but is not limited to: traffic flow data for the entire navigation route (less than or equal to a long distance, such as 160 km), historical average vehicle speed data, and gradient data. The SVR information can be used for short-term vehicle speed prediction and may include, but is not limited to: speed limit data within a short distance (such as 3 km) ahead of the vehicle, historical average vehicle speed data, and potential stopping point information. Based on the road condition information, the road type of the road to be driven can be determined. Driving behavior may include acceleration, braking, starting, and turning. The ADAS map signal is analyzed by the Environmental-Historical Route Information Processor (E-HR) module, and the positional relationship between the vehicle and scene feature points is calculated by accumulating offset values, with equal-interval data steps set to 1m (short-term prediction) or 20m (long-term prediction).

[0073] In this embodiment, personalized control is achieved through "scenario-based style recognition + dynamic adaptation". Exclusive scenarios are divided according to "road type + driving behavior", and scenario features such as accelerator pedal change rate and longitudinal acceleration are extracted. GMM soft clustering is used to quantify the aggressiveness of driving style (0-1 index) in real time. The driving style data is integrated into the vehicle speed prediction and power distribution process to increase the range extender start-up threshold for vehicles with aggressive driving style to reserve power, and optimize the power generation strategy for vehicles with conservative driving style to reduce energy consumption. This allows the control logic to actively adapt to individual habits and scenario differences.

[0074] Optionally, real-time traffic information provided by ADAS maps can be combined with historical driving information recorded by the vehicle. By filtering historical driving information that matches the current road type (such as highways, main roads, or urban roads), the vehicle's operating habits under similar road conditions, such as acceleration mode, braking frequency, and driving aggressiveness, can be analyzed to determine the vehicle's driving style data for the current road. This process allows the range extender's power distribution and start-stop strategy to proactively adapt to the driver's individual needs and road condition changes, improving fuel economy, reducing unnecessary battery charging and discharging, and significantly reducing NVH (noise, vibration, and harshness) issues caused by the range extender's start-stop function, thus enhancing driving comfort and quietness. Simultaneously, by integrating driving style data into the vehicle control logic, the vehicle can more intelligently predict energy demand and pre-store or release energy, avoiding the disjointed driving experience caused by powertrain response lag, achieving a smoother, more pure electric driving experience for range-extended vehicles.

[0075] Optionally, traffic information provided by the advanced driver assistance system (ADAS) map can be used to identify the type of road the vehicle is about to travel on. This could include highways, urban arterial roads, rural roads, or steep mountain roads. Identifying the road type is crucial for subsequent analysis and prediction, as different road types often correspond to different average vehicle speeds, traffic conditions, and potential driving behavior patterns. Historical driving information matching the upcoming road type can be selected from the vehicle's historical driving data set. This set includes past driving conditions, such as vehicle speed, engine speed, battery state of charge (SOC), and records of driver operating habits. By filtering, historical driving information from previous trips on similar road types can be found, providing a reference for subsequent analysis.

[0076] Optionally, road condition information, including traffic flow, gradient, speed limits, and potential parking spots, can be obtained from ADAS maps, encompassing both long-field-of-view (the entire navigation route) and short-field-of-view (within a preset distance ahead of the vehicle). Based on this road condition information, road types can be identified; for example, analyzing gradient and speed limit data can determine whether the current road is a highway, arterial road, or secondary road. At least one historical driving record matching the current road type (highway, arterial road, secondary road, etc.) is selected from the historical driving information set. This record can be used to identify and quantify the vehicle's driving style. For example, acceleration behavior requires a speed > 10 km / h, longitudinal acceleration > 0.1 m / s², and accelerator pedal opening > 10% for at least 1 second. These criteria can be used to determine instances of driving behaviors such as acceleration, braking, starting, and turning. Finally, based on the selected historical driving information, statistical analysis or machine learning models can be used to determine the vehicle's driving style data. This driving style data can be used to characterize a driver's driving habits, such as whether they tend to drive aggressively (e.g., frequent acceleration and braking), drive gently (stable driving with less sudden acceleration or braking), or something in between.

[0077] For example, Gaussian mixture models can be used to analyze the characteristics of driver operation signals and vehicle state signals to identify the current driving style and quantify the degree of aggression. This process involves filtering key features (i.e., at least one piece of historical driving information) from multiple raw signals (i.e., the set of historical driving information), such as accelerator pedal opening and longitudinal acceleration, and inputting at least one piece of historical driving information into the Gaussian mixture model for soft clustering to calculate the posterior probability of different styles.

[0078] Alternatively, a sliding window can be used to update the style index in real time to adapt to short-term changes in driving style. This means that this embodiment can instantly reflect changes in vehicle style under different scenarios, thereby providing more personalized guidance for the control of the range extender.

[0079] Through the steps described above, based on road condition information obtained from ADAS maps and matching data filtered from historical vehicle driving information sets, a vehicle's driving style can be effectively identified and quantified. This process not only considers the characteristics of the current road but also incorporates the vehicle's past behavior under similar road conditions, resulting in more accurate and personalized driving style data. This data is then used to predict vehicle speed sequences and formulate start-stop and power distribution strategies for the range extender to achieve optimized control of vehicle energy management, ensuring driving safety, energy conservation and emission reduction, and improving the driving experience.

[0080] Optionally, the aforementioned historical driving information set can be data collected over a period of time, or data collected and updated in real time according to a preset collection frequency.

[0081] As an optional implementation method, determining driving style data based on at least one historical driving information includes: performing clustering processing on at least one historical driving information to obtain clustering results, wherein the clustering results are used to characterize the degree of correlation between the historical driving information and different types of driving style data; and transforming the clustering results to obtain the vehicle's driving style data.

[0082] In this embodiment, the clustering results can be used to characterize the degree of correlation between historical driving information and different types of driving style data, based on the posterior probability of the vehicle.

[0083] Optionally, at least one historical driving information point matching the current road type can be selected from the historical driving information set. This historical driving information may include, but is not limited to, data such as vehicle speed changes, accelerator pedal usage, and longitudinal acceleration. At least one historical driving information point can be input into a Gaussian Mixture Model (GMM) for cluster analysis. This model can divide the input data into multiple Gaussian distribution clusters, each representing a specific driving style. Through clustering, the degree of correlation between different driving styles (e.g., aggressive, moderate, and conservative) and historical driving information can be determined, i.e., the clustering result. The posterior probability output by the GMM can be used to quantify the vehicle's driving style. For each driving style, the GMM can calculate a probability value, which can be the driving style data, representing the probability that given historical data belongs to that style, or an aggression index (AgrScore) ranging from 0 to 1, where 0 represents the most conservative driving style and 1 represents the most aggressive.

[0084] Optionally, a sliding window technique can be used to achieve real-time recognition and adaptation of driving style. That is, the latest driving information can be continuously collected and the data within the window updated. The driving information within the sliding window can be used to recalculate the posterior probability of the GMM, thereby updating the quantification index of the driving style. This process can be executed every 100 milliseconds, ensuring the real-time nature and accuracy of the driving style data. Finally, the driving style data obtained through GMM clustering and sliding window techniques can be converted into practically usable control parameters. These parameters will then be used to adjust the vehicle speed prediction model and the range extender control strategy. For example, if the vehicle's style is identified as more aggressive, the range extender's activation threshold can be increased to ensure sufficient power reserve when needed and avoid power delay; if the style is more conservative, the power generation strategy can be optimized to reduce unnecessary energy consumption.

[0085] As another optional example, real-time driving style recognition can include: acquiring a historical driving information set; based on this set, multiple raw signals can be calculated, such as vehicle speed, accelerator pedal opening, longitudinal acceleration, lateral acceleration, and steering wheel angle. The data type to be filtered can be determined based on road type, allowing for the selection of eight core features from these signals: average accelerator pedal position, standard deviation of accelerator pedal rate of change, maximum accelerator pedal rate of change, maximum longitudinal acceleration, maximum vehicle speed, average accelerator pedal rate of change, and maximum lateral acceleration, to obtain at least one piece of historical driving information.

[0086] Furthermore, after obtaining historical driving information, a Gaussian mixture model can be used. The number of clusters in this model can be set to 3 (corresponding to aggressive, moderate, and conservative types), the number of iterations can be set to 1000, and the termination tolerance can be set to 1×10⁻⁶. -7 The radical exponent, calculated based on the posterior probability output by the GMM, can be calculated using the following formula: ,in, For the posterior probability of the calm class, For the posterior probability of the ordinary class, The posterior probability of the aggressive class is used, with an exponent ranging from 0 to 1 (0 being the most conservative and 1 being the most aggressive), to obtain the aggressiveness index (i.e., the quantification parameter of the degree of aggressiveness). This quantification parameter of the degree of aggressiveness can be used as driving style data to determine the driving style of the vehicle.

[0087] Optionally, the sliding window size can be determined through real vehicle calibration (for example, by taking the 10 most recent samples), and the average of the aggressive index within the window can be used as the current driving style index. The update period is set to 100ms to complete the real-time update of the collected data.

[0088] As an optional implementation, the road condition information includes short field of view (SFL) information, which is used to characterize the road condition status of the vehicle within the target range. Based on the road condition information and driving style data, a predicted vehicle speed sequence is predicted, including: inputting the SFL information and driving style data into a prediction model to predict the predicted vehicle speed sequence. The prediction model is trained based on driving style data samples, SFL information samples, and vehicle speed sequence samples corresponding to the driving style data samples and SFL information samples. The driving style data samples are used to characterize the vehicle's historical driving style, and the vehicle speed sequence samples are used to characterize the vehicle's historical speed sequence.

[0089] In this embodiment, the aforementioned road condition information may include short-field-of-view information, which can characterize the road condition status of the vehicle on the road to be driven within the target range. This short-field-of-view information may include, but is not limited to, long-field-of-view traffic flow, historical average vehicle speed in the short-field-of-view, and driving style sequences.

[0090] Optionally, short-field-of-view information and driving style index (i.e., driving style data) are used as input data and fed into the prediction model to predict the vehicle speed sequence. The prediction model is trained based on driving style data samples, short-field-of-view information samples, and corresponding vehicle speed sequence samples. It can be a deep learning model, and the specific type of prediction model is not limited here.

[0091] Optionally, the preprocessed road condition information (which may include short field of view information and long field of view information) and the quantified driving style results (i.e., driving style data) are used as input data and fed into a deep learning model. The deep learning model then predicts the future long-term vehicle speed sequence and short-term vehicle speed sequence, respectively. The long-term vehicle speed sequence corresponds to the entire navigation path, and the short-term vehicle speed sequence corresponds to a preset distance in front of the vehicle. This can be used to predict the vehicle speed sequence.

[0092] In this embodiment, the range extender needs to predict energy demand in advance (e.g., generate electricity before climbing a hill). The vehicle speed prediction sequence provided in this step provides prior information for the range extender's power planning and is the core of achieving predictive control.

[0093] Optionally, this embodiment leverages the advantages of predictive models, combining real-time vehicle operating habits with immediate road conditions ahead, to make the control logic of the range-extended vehicle more intelligent and personalized. This optimizes vehicle energy management in complex traffic environments, significantly reduces engine start-stop frequency, lowers fuel consumption, and ensures timely and comfortable vehicle power response, thereby enhancing the user's driving experience.

[0094] As an optional implementation, the road condition information also includes long field of view information, which is used to characterize the road condition status of the entire road to be driven. The method further includes: inputting the long field of view information into the prediction model to predict the vehicle speed of the entire road to be driven.

[0095] In this embodiment, long-field-of-view information can also be input into the prediction model to predict the vehicle's speed throughout the entire journey on the road to be traveled. By utilizing long-field-of-view information, a global prediction of the road to be traveled is achieved.

[0096] Optionally, the input parameters only include long-field-of-view information, which can be traffic flow data, historical average vehicle speed data, and gradient data. In this case, the input data does not include driving style parameters. The long-field-of-view information can be input into the prediction model to predict the vehicle's speed throughout the entire journey on the road to be traveled.

[0097] Optionally, the above predicted vehicle speed sequence can be the vehicle speed of a small segment of the road to be traveled, and the full-range predicted vehicle speed can be the full-range predicted vehicle speed sequence of the vehicle on the road to be traveled.

[0098] Optionally, the above prediction model can adopt a network architecture of "Long Short-Term Memory (LSTM) + Bidirectional Long Short-Term Memory (Bi-LSTM) + LSTM". The probability of the dropout layer in the network can be set to 0.5 to prevent overfitting. The sample set can be divided into training set, test set and validation set in a ratio of 6:2:2. The short-term vehicle speed prediction RMSE target is ≤13.21 km / h, and the long-term vehicle speed prediction RMSE target is ≤14.76 km / h.

[0099] In this embodiment, not only is short-distance road condition information ahead of the vehicle considered, but long-range visual information covering the entire driving path is also incorporated, including macroscopic data such as road type, average traffic flow, and gradient changes. Through deep learning model processing and analysis, combined with driving style data, the system can predict the entire speed change sequence, providing global guidance for the start-stop and power distribution of the range extender. This predictive capability allows the vehicle to plan energy reserves and release in advance during long-distance driving, such as cross-city trips or mountain driving, avoiding insufficient power or excessive battery discharge in high-energy-consuming road sections (such as uphill or high-speed driving). At the same time, the predictive model helps reduce unnecessary engine starts in low-speed congested areas, reducing fuel consumption and mechanical wear, improving driving economy and NVH performance, and providing users with an environmentally friendly and comfortable driving environment.

[0100] Figure 2 This is a flowchart of a scene- and style-based vehicle speed sequence prediction method provided in an embodiment of this application, as shown below. Figure 2 As shown, the method may include the following steps.

[0101] Step S202: Input road condition information and driving style data into the prediction model.

[0102] In this embodiment, road condition information and driving style data can be input into the prediction model. The road condition information can be scene information, including both long-field-of-view and short-field-of-view information. The driving style data can be a driving style index.

[0103] Step S204: Use the prediction model to classify and predict road condition information.

[0104] In this embodiment, a prediction model can be used to predict short-field-of-view information and driving style data in road condition information to obtain short-term speed prediction results (i.e., predicted vehicle speed sequences). Furthermore, long-field-of-view information can be predicted to obtain long-term speed prediction results (i.e., full-range predicted vehicle speed).

[0105] Step S206: The prediction model outputs the prediction results.

[0106] In this embodiment, the prediction model can adopt a network architecture of "LSTM (front view) + Bi-LSTM (fusion) + LSTM (output)" to process short-view and long-view information respectively to obtain the predicted vehicle speed sequence and the full-range predicted vehicle speed. The predicted results can be output.

[0107] In this embodiment, the problem of coarse prediction granularity and inability to cope with dynamic changes is solved by using a method of "layered acquisition of long / short field of view scenes + variable step size data reconstruction". Long field of view information can cover macroscopic information such as slope and traffic flow throughout the navigation process, while short field of view information can focus on microscopic dynamics such as speed limits and potential parking spots within the target range ahead of the vehicle (e.g., 3km). Combined with differentiated step size processing for short and long time periods, accurate prediction of "macroscopic trends + microscopic details" is achieved. Simultaneously, dynamic correction of driving style is incorporated, enabling the prediction to respond in real time to changes in road conditions and operating habits.

[0108] As an optional implementation, the vehicle control data is determined based on the predicted vehicle speed sequence, including: determining the state of charge (SOC) of the battery in the vehicle; determining the SOC trajectory of the battery on the road to be traveled based on the predicted vehicle speed and SOC throughout the journey, wherein the SOC trajectory is used to characterize the SOC of the battery at different times; determining the energy consumption data of the range extender under multiple equivalent factors based on the SOC trajectory and the predicted vehicle speed sequence, wherein the equivalent factors are used to balance the energy distribution between the range extender and the battery; determining a target equivalent factor from the multiple equivalent factors based on the energy consumption data; determining the power contribution ratio of the battery and the range extender using the target equivalent factor; and converting the power contribution ratio to obtain control data.

[0109] In this embodiment, the state of charge trajectory can refer to the trend or path of the battery's state of charge (SOC) changing over time during vehicle operation. It can be used to describe the expected path of how the battery's SOC changes throughout the entire journey from the vehicle's departure to its destination. It can be determined based on the predicted vehicle speed throughout the journey and may include, but is not limited to, the initial value, rate of change, peak value, and endpoint value of SOC.

[0110] Optionally, by using dynamic programming (DP) algorithms to plan the SOC trajectory, globally optimal energy management can be achieved, ensuring that the battery SOC remains within an ideal range during driving. This avoids power degradation caused by excessively low SOC or charging risks associated with excessively high SOC. The planning objective for the SOC trajectory can be to maintain the battery SOC between 20% and 90%, thereby extending battery lifespan.

[0111] In this embodiment, the aforementioned equivalence factor (λ) can be a parameter used to balance the contributions of different energy sources (e.g., batteries and range extenders). It can be used to convert different energy consumptions into comparable units and to equate the fuel consumed by the range extender with the electricity consumed by the battery. The value of λ can be used to reflect the proportion of energy contribution from the range extender and the battery. A high λ value means a greater tendency to use the range extender, while a low λ value means priority is given to battery energy.

[0112] Optionally, by dynamically adjusting the equivalent factor λ using the Pontryagin's Minimum Principle (PMP) algorithm, the energy management strategy can be optimized under medium-term vehicle speed prediction sequences and gradient changes. The equivalent factor λ can be pre-set to a range of 2.0-6.0, and an optimal λ value (i.e., the target equivalent factor) can be found through iterative optimization to ensure that the overall energy consumption cost, including fuel consumption and battery degradation costs, is minimized while meeting the SOC target trajectory.

[0113] In this embodiment, the aforementioned energy consumption data can refer to the energy consumed by a vehicle (especially a range-extended vehicle) during operation under different energy management strategies. This can include the electrical energy consumption of the battery and the fuel consumption of the range extender (such as an engine). The data can be determined based on the actual driving conditions of the vehicle, such as vehicle speed, slope, driving style, and the current state of energy (such as battery SOC) and equivalent factor λ.

[0114] Optionally, energy consumption data can serve as a key indicator for evaluating and optimizing energy management strategies. The energy consumption data of the range extender can be calculated under multiple equivalent factors λ, the total energy consumption under different strategies can be compared, and a target equivalent factor that achieves the lowest energy consumption can be selected. This process may involve simulating the operating state of the range extender, considering engine thermal efficiency, start-stop costs, and battery charging and discharging efficiency, thereby achieving efficient energy utilization, reducing overall energy consumption, and improving vehicle fuel economy and driving range.

[0115] In this embodiment, the current state of charge (SOC) of the vehicle's battery is monitored and determined. SOC is an indicator of the battery's charge / discharge level, typically ranging from 0% to 100%. It is a key parameter in the energy management system's decision-making process because the battery's SOC directly affects the distance the vehicle can drive using the battery. It is also an important basis for deciding when and how to activate the range extender (such as the engine). The current SOC of the vehicle's battery can be obtained through real-time sensor data. Based on the predicted vehicle speed sequence and the current battery SOC, the theoretical trajectory of the battery's SOC over time on the planned route is calculated—that is, the SOC trajectory. This SOC trajectory is the core of the energy management strategy, guiding the allocation of energy between the battery and the range extender throughout the entire journey on the planned route. This maintains the SOC within a predetermined ideal range, ensuring the battery does not over-discharge while also considering the need for the range extender to activate in a timely manner to replenish energy.

[0116] Optionally, a dynamic programming (DP) algorithm can be used to plan the global State of Charge (SOC) trajectory. This process involves using the SOC state and engine power as control variables and finding an optimal trajectory for SOC changes by minimizing a cost function that includes engine fuel consumption, battery energy consumption, and engine start-stop penalties. Furthermore, the battery SOC trajectory planned by the DP algorithm can be combined with the predicted vehicle speed sequence to calculate the range extender's energy consumption data under different equivalent factors λ. A second-layer power management algorithm (PMP) can then be used to optimize the equivalent factor λ within each fixed short distance segment. This allows for finding a λ value that minimizes total energy consumption based on the current SOC trajectory and predicted vehicle speed. Based on the energy consumption data calculated under multiple equivalent factors λ, a target equivalent factor λ can be selected. This λ value maximizes energy savings while also considering factors such as driving experience and equipment wear.

[0117] Optionally, once the target equivalence factor is available, the power contribution ratio of the battery and the range extender at different driving stages can be calculated based on the target equivalence factor. Based on this power contribution ratio, it can be determined whether to use the energy stored in the battery for driving (i.e., pure electric mode) or to activate the range extender to provide additional power at a certain moment.

[0118] In this embodiment, a third-layer ECMS algorithm, combined with the target equivalence factor optimized by PMP, can be used to dynamically adjust the power distribution between the battery and the range extender, thereby achieving instantaneous range extender power management. Finally, the power contribution ratio of the battery and the range extender can be converted into specific control commands to obtain control data. This control data can be used to determine the range extender's start-up timing, the engine's power output, and the battery's charging and discharging strategy.

[0119] Alternatively, control data can be executed via hardware platforms (such as MCUs and MPUs) to adjust the operating status of the range extender and battery in real time to ensure that the predicted energy distribution strategy is implemented accurately.

[0120] In this embodiment, a three-layer optimization control method for the range extender is proposed. This method uses the predicted vehicle speed sequence as the core input and determines the start-stop timing and power generation of the range extender through a three-layer optimization strategy. The first layer employs a dynamic programming (DP) algorithm to optimize the energy demand of a relatively long road segment in the future, obtaining the optimal State of Charge (SOC) trajectory. The second layer uses the Pondriagin minimum principle based on the equivalent fuel consumption minimum strategy to optimize the equivalent factors of the mid-term speed prediction sequence and gradient. The third layer uses the equivalent fuel consumption minimum strategy, combined with the equivalent fuel factor optimized by PMP, to achieve instantaneous power allocation of the range extender.

[0121] Optionally, the three-layer optimization control method for the range extender may include: The first layer utilizes the DP algorithm for global SOC planning. This layer can set control conditions such as an initial SOC (e.g., 35%-36%) and a final SOC (e.g., 23%-25%). Input data may also include predicted vehicle speed sequences and gradient information. The state variable is battery SOC, the control variable is engine power (range 0-maximum engine power), and the cost function is the sum of engine fuel consumption, battery energy consumption, and engine start-up penalty. The second layer utilizes the PMP algorithm for mid-term equivalent factor optimization. Control conditions for this layer may include: inputs including a 1km forward vehicle speed sequence (divided into 5m segments, totaling 200 data points) and the target SOC trajectory planned by DP. The parameters for this layer can be: an equivalent factor n with a search range of 2.0-6.0, using a bisection iterative optimization method with an iteration step size of 0.01, and a termination condition of SOC deviation ≤1%. The third layer may utilize the ECMS algorithm for instantaneous range extender control. The control conditions of the third layer can be to divide the engine operating point set into different sets according to different speed levels. Each speed level corresponds to a set of candidate engine operating points, and the ECMS calculates the optimal operating point in this set. The parameters of this layer can be to divide 0-180km / h into several speed levels. For example, when the vehicle speed is 85~100km / h, the engine operating point set is selected in the range of 2200 revolutions per minute (rpm) to 2600rpm to ensure that the driving noise of the vehicle can mask the noise of the engine itself.

[0122] Optionally, the third layer of the three-layer optimized control of the range extender can directly realize the instantaneous distribution of the range extender's power to respond to the driver's real-time needs.

[0123] Figure 3 This is a flowchart of a three-layer optimization control method for a range extender provided in an embodiment of this application, as shown below. Figure 3 As shown, the method may include the following steps.

[0124] Step S302: Input the state of charge and the predicted speed for the entire journey.

[0125] In this embodiment, the state of charge (SOC) of the vehicle's battery at the current moment is obtained. The SOC and the predicted vehicle speed over the entire journey can be input.

[0126] In step S304, the DP layer processes the state of charge and the predicted speed throughout the journey to obtain the globally optimal SOC trajectory.

[0127] In this embodiment, the first layer is the DP layer, which is used to determine the optimal trajectory of the battery's State of Charge (SOC) throughout the entire journey, ensuring that the battery remains within its ideal operating range during driving. The DP algorithm calculates the energy consumption under different states to obtain a globally optimal change in battery SOC, thus obtaining the optimal SOC trajectory. This state of charge trajectory can be used as the state of charge trajectory.

[0128] In step S306, the PMP layer determines the target equivalence factor based on the charged state trajectory and the predicted vehicle speed sequence.

[0129] In this embodiment, within each fixed short distance of the trip, the PMP algorithm is used to adjust the equivalence factor to adapt to real-time changing traffic conditions (such as predicted vehicle speed and road gradient). The equivalence factor is an important parameter that can be used to balance the proportion of energy consumption between the battery and the range extender (such as the engine). By dynamically optimizing the input value in the second layer, it can more accurately match the current driving conditions and avoid rigidity in energy allocation.

[0130] Optionally, the charge state trajectory and predicted vehicle speed sequence obtained from the first-layer DP algorithm can be used to determine the energy consumption data of the range extender under multiple equivalent factors. Based on the energy consumption data, the target equivalent factor can be determined from the multiple equivalent factors.

[0131] In step S308, the ECMS layer determines the power contribution ratio of the battery and the range extender based on the target equivalence factor.

[0132] In this embodiment, the ECMS layer can adjust the power output of the range extender in real time based on the target equivalence factor to match the instantaneous power demand.

[0133] Optionally, the ECMS algorithm can quickly calculate a suitable energy allocation strategy based on the target equivalence factor optimized by the PMP algorithm and combined with the current vehicle state. This energy allocation strategy can be used to determine the power contribution ratio of the battery and the range extender. The ECMS algorithm can respond instantly to the driver's operations and the vehicle's real-time needs, ensuring that the range extender operates in the high-efficiency range while keeping the battery SOC within a safe range.

[0134] Step S310: Start and stop the range extender.

[0135] In this embodiment, based on the energy distribution strategy, a power control command is output, which is used to control the start and stop of the range extender.

[0136] In this embodiment, the initial state of charge (SOC) of the battery is first determined. Then, based on the predicted vehicle speed sequence for the entire journey, methods such as dynamic programming (DP) or equivalent fuel consumption minimization strategy (ECMS) are used to calculate the optimal SOC trajectory that meets the requirements throughout the journey, keeping the SOC within a stable range that is beneficial to battery life. Subsequently, based on the determined SOC trajectory and the predicted vehicle speed sequence, the energy consumption of the range extender and battery is simulated under various equivalent factor (λ) settings. By comparing the energy consumption data under different λ values, a target equivalent factor that maintains the rationality of the SOC trajectory while minimizing overall energy consumption is selected. Based on the selected target equivalent factor, the power output ratio of the battery and range extender in different driving stages (such as urban congestion, highway cruising, or mountain uphill driving) is further determined to ensure efficient energy distribution under any road conditions. Finally, by converting the power contribution ratio into corresponding control parameters, control data such as battery charging and discharging current, range extender target speed and torque are obtained, thereby achieving precise control of the vehicle energy management system. This improves the vehicle's energy utilization and economic performance, optimizes the driving experience, and reduces the noise and vibration caused by the range extender's operation, achieving the goal of efficient, quiet and smooth operation of range-extended electric vehicles under complex road conditions.

[0137] In this embodiment, to address the problem of "lack of real-time rolling optimization and rigid strategy", a "three-layer progressive rolling optimization" is used to break the rigidity of the strategy. The first layer, the DP algorithm, provides the global SOC trajectory basis. The second layer, the PMP algorithm, dynamically optimizes the equivalent factor based on the latest short-term prediction data for each fixed short distance. The third layer, the ECMS algorithm, achieves instantaneous power allocation with millisecond-level response. Combined with the real-time vehicle status feedback closed loop, a dynamic optimization chain of "global planning - mid-term adjustment - instantaneous response" is formed to ensure that the strategy always adapts to the current operating conditions.

[0138] As an optional embodiment, step S104, based on road condition information, determines the driving style data of the vehicle on the road to be driven, including: converting the road condition information in the spatial domain according to a preset step size to obtain equally spaced information in the time domain; and determining the driving style data based on the equally spaced information.

[0139] In this embodiment, after acquiring road condition information, the information can be preprocessed. The preprocessing process may include: transforming the spatial domain road condition information according to a preset step size to obtain equally spaced information in the time domain. After obtaining the equally spaced information, driving style data can be determined based on it.

[0140] Optionally, the intelligent connected module on the range-extended vehicle acquires road condition information required for driving, including long-field-of-view and short-field-of-view information. The acquired scene information is then analyzed and reconstructed, converting the spatial road condition information into standardized, equally spaced data that matches the vehicle speed, thus completing the road condition information preprocessing process. This step provides basic scene data for subsequent vehicle speed prediction and range-extended control, unlike the passive response mode of traditional range-extended control that relies solely on real-time vehicle speed.

[0141] Optionally, by reorganizing and formatting the non-uniformly distributed road condition information (such as speed limits, gradients, and traffic flow) obtained from ADAS maps and sensors according to preset time intervals (e.g., every second or adaptively adjusted based on vehicle speed), the data input to the prediction model is ensured to have a unified time dimension, forming an equally spaced time series. This processing method not only improves data availability and model training efficiency but also, by introducing driving style recognition algorithms, such as Gaussian Mixture Models (GMMs) or cluster analysis, combines the standardized information with driver operation signals, more accurately quantifying the vehicle's driving characteristics under different driving styles. Ultimately, by fusing this deeply analyzed driving style data with the prediction model, the system can intelligently adjust the range extender's operating mode according to changes in driver habits and road conditions, achieving dual optimization of energy saving and emission reduction while improving driving comfort.

[0142] Figure 4 This is a flowchart of a traffic information processing method provided in an embodiment of this application, such as... Figure 4 As shown, the method may include the following steps:

[0143] Step S402: Obtain map information.

[0144] In this embodiment, ADAS map information is obtained to obtain road condition information.

[0145] Step S404: Classify and extract road condition information.

[0146] In this embodiment, long field of view information and short field of view information are obtained by classifying and extracting road condition information that is not uniformly distributed in space from ADAS maps and sensors.

[0147] Step S406: Analyze and reconstruct the long field of view information and the short field of view information.

[0148] In this embodiment, long-field-of-view and short-field-of-view information are reorganized and format-converted according to preset time intervals (e.g., every second or adaptively adjusted based on vehicle speed) to ensure that the data input to the prediction model has a unified time dimension, forming an equally spaced time series. This processing method not only improves data availability and model training efficiency, but also, by introducing driving style recognition algorithms, such as Gaussian Mixture Models (GMMs) or cluster analysis, combines the standardized information with driver operation signals to more accurately quantify the vehicle's driving characteristics under different driving styles.

[0149] Step S408: Standardize the parsed and reconstructed data.

[0150] In this embodiment, on the one hand, multi-dimensional data such as ADAS maps, traffic flow, and driver operation are integrated, and outliers are eliminated through cross-validation; on the other hand, data dimensions are unified through preprocessing such as 0-1 standardization and One-Hot encoding.

[0151] Step S410, scene division.

[0152] In this embodiment, driving scenarios are divided based on preprocessed road condition information (i.e., scenario information). The driving scenario can be determined by "road type + driving behavior". The driver's operation signal and vehicle status signal features (i.e., at least one historical driving information) under the divided road type are extracted. A clustering algorithm is used to analyze at least one historical driving information to identify the driving style of the vehicle in the current road type and quantify the aggressiveness of the driving style.

[0153] Step S412: Output standardized scene data and corresponding scene labels.

[0154] In this embodiment, standardized scene data and corresponding scene labels can be output.

[0155] Optionally, by integrating this in-depth analysis of driving style data with predictive models, the system can intelligently adjust the range extender's operating mode according to changes in driver habits and road conditions, achieving a dual optimization of energy conservation, emission reduction, and driving comfort.

[0156] In this embodiment, to address the issue of "high dependence on the accuracy and reliability of input data", data fault tolerance is improved through "multi-source fusion + preprocessing enhancement": on the one hand, multi-dimensional data such as ADAS maps, traffic flow, and driver operation are fused, and outliers are eliminated through cross-validation; on the other hand, data dimensions are unified through preprocessing such as 0-1 standardization and One-Hot encoding, and a sliding window is used to smooth the driving style recognition results, reducing the impact of single data errors. Even if some input data has deviations, control accuracy can still be maintained through multi-source complementarity.

[0157] Optionally, to address the issue of "single optimization objective without considering multi-objective balance," a "multi-objective collaborative optimization system" is constructed to achieve balanced control: the cost function comprehensively incorporates three core indicators: fuel consumption, battery energy consumption, and range extender start-up penalty. The DP algorithm ensures that the battery SOC is always within the long-life range of 20%-90%. Through PMP and ECMS collaborative optimization, the range extender's operating point is concentrated in the high-efficiency and low-noise region. At the same time, combined with driving style prediction, power is reserved in advance to achieve a dynamic balance between fuel saving, battery protection, NVH optimization, and power response.

[0158] In this embodiment, through the whole-chain technological innovation of "scenario prediction - style adaptation - three-layer control - heterogeneous deployment", multi-dimensional performance upgrades are achieved to address the core pain points of traditional range-extended vehicles such as "poor fuel economy, short battery life, fragmented driving experience, and insufficient reliability".

[0159] Optionally, through the synergistic effect of dual-dimensional vehicle speed prediction and three-layer energy optimization, accurate prediction and dynamic adaptation of energy distribution in the range extender system can be achieved, thereby significantly improving fuel economy and reducing operating costs. On the one hand, the dual-dimensional input of "long / short field of vision scenario information + driving style" significantly reduces the RMSE of vehicle speed prediction, providing a reliable prior basis for energy planning; on the other hand, the three-layer strategy of DP global SOC planning, PMP equivalent factor optimization, and ECMS instantaneous control can avoid the problems of "excessive power consumption in the early stage and frequent inefficient operation of the range extender in the later stage" in traditional control, improve the overall fuel saving rate, and reduce the user's daily fuel expenditure.

[0160] Optionally, by planning the global SOC trajectory through the DP algorithm, the battery can always operate in the optimal range of 20%-90%, avoiding deep charging and discharging (deep discharging will accelerate the degradation of electrode material structure). Combined with the "shallow charging and shallow discharging" operation logic, the equivalent cycle number of lithium iron phosphate batteries can be increased, extending battery life and reducing the user's later battery replacement costs.

[0161] Optionally, the three-layer control strategy reduces the number of low-speed start-ups of the range extender by predicting energy demand in advance, shortens engine operating time, avoids wear and tear on mechanical components caused by frequent start-stop (frequent starts will aggravate internal friction and aging of the range extender), significantly extends the service life of the range extender, and reduces engine maintenance frequency.

[0162] In this embodiment, the scenario-based driving style recognition based on "road type + driving behavior" can accurately match different driver habits. For aggressive drivers, it can start the engine in advance to reserve power and avoid power lag during acceleration; for conservative drivers, it can optimize the power generation strategy to reduce energy consumption fluctuations. The recognition results are highly consistent with subjective feelings, solving the problem of the "one-size-fits-all" control separation of traditional systems. At the same time, the reduction of the number of start-stop cycles of the range extender, the division of engine operating points according to vehicle speed levels, and the optimization of operating points (for example, matching the high-efficiency power generation range in high-speed mode and avoiding high-frequency whistling in low-speed mode) significantly reduce cabin noise. Combined with the "seamless start-stop" control logic, users can hardly perceive the intervention of the range extender in scenarios such as urban congestion.

[0163] Optionally, by using short-term vehicle speed prediction and dynamic adjustment of PMP equivalent factors, the range extender can switch to a high-performance power generation point in a short time, eliminating the power step-like feeling in parallel mode, improving the linearity of torque output during full-throttle acceleration, and achieving a "pure electric sporty" driving experience.

[0164] In this embodiment, the deep fusion of long-field-of-view scene information (such as slope and traffic flow) with the navigation path enables advance planning of energy strategies for high-speed power conservation and low-speed pure electric operation, avoiding the phenomenon of "sudden power drop". The multi-core heterogeneous deployment method resolves the contradiction between the computational power requirements of complex algorithms and real-time control response. The MPU's 1GHz clock speed meets the high-energy-consuming computational needs of LSTM models and DP algorithms, while the MCU's 400MHz clock speed ensures rapid execution of control commands, with inter-core communication accuracy reaching 1ms. Compared to traditional single-chip solutions, this architecture allows for algorithm implementation in vehicles without additional hardware upgrades, reducing R&D and mass production costs for automakers and adapting to the platform transformation needs of current mainstream range-extended electric vehicles.

[0165] Figure 5 This is a flowchart of a scenario prediction-based range-extended vehicle control method provided in an embodiment of this application, as shown below. Figure 5 As shown, the method may include:

[0166] Step S502, scene information processing.

[0167] In this embodiment, long / short field-of-view scene information from the ADAS map is introduced, and the information is reconstructed through the E-HR module. At the same time, driving style sequences are added as short-term prediction inputs.

[0168] Step S504, vehicle speed prediction.

[0169] In this embodiment, the "LSTM+Bi-LSTM+LSTM" architecture is used to fuse front and rear field information, resulting in a significant decrease in both short-term and long-term prediction RMSE and a significant improvement in robustness.

[0170] Step S506, driving style recognition.

[0171] In this embodiment, to address the problem that the technology lacks scene segmentation and cannot adapt to the style differences of the same driver in different scenarios, a scenario segmentation based on "road type + driving behavior" is proposed. Scene-specific features are determined through multi-step feature filtering, and GMM soft clustering is used to output posterior probabilities. Combined with a sliding window, real-time recognition is achieved. The recognition results are completely consistent with subjective feelings, providing a personalized adaptation basis for range extender control, thereby realizing the recognition of vehicle driving style.

[0172] Optionally, driving style recognition may include three processes: feature selection, GMM clustering, and radical quantization.

[0173] Step S508, three-layer control of the range extender.

[0174] In this embodiment, a "DP+PMP+ECMS three-layer range extender optimization control strategy" is proposed. DP plans the global SOC trajectory to solve the global optimization problem, PMP optimizes the equivalent factor to adapt to mid-term traffic dynamics, and ECMS realizes instantaneous power distribution. The three-layer collaborative control reduces the overall fuel consumption of the vehicle, reduces engine operating time, and reduces the number of low-speed start-ups.

[0175] Step S510: Execute control data and obtain feedback data after executing control data.

[0176] In this embodiment, the range extender control commands are deployed to a multi-core heterogeneous hardware platform, and the control actions are executed through the hardware platform; vehicle status parameters are collected in real time and fed back to the control module, and the range extender control parameters are dynamically adjusted to form a control closed loop.

[0177] Optionally, the MCU sends control commands to the EMS (engine power 26kW) and the motor controller (torque 40kW) via the CANFD bus; the Vgate3 recorder collects the engine speed at 2500rpm, torque at 100Nm, and battery SOC at 34.8% in real time. After the data is transmitted back, if the SOC deviates from the target (33.5% < 35%), the PMP re-optimizes λ=4.5, and the ECMS adjusts the range extender power to 45kW to ensure that the SOC returns to the target.

[0178] Step S512: Determine whether the task is completed.

[0179] In this embodiment, it can be determined whether the task is completed. If the task is completed, step S514 can be executed; otherwise, step S502 can be executed.

[0180] Step S514: Determine that the vehicle has completed the driving process.

[0181] In this embodiment, a "multi-core heterogeneous deployment scheme" is constructed, in which the real-time control module (EHR, ECMS) is deployed on the MCU to ensure a 10ms response, and the complex computing module (style recognition, vehicle speed prediction) is deployed on the MPU to provide high computing power. Inter-core communication is realized through Mid-SOA to solve the contradiction between computing power and real-time performance in engineering implementation.

[0182] In this embodiment, road condition information is first acquired, including both long-field-of-view (traffic flow related to the entire navigation path, historical average vehicle speed, and gradient) and short-field-of-view (speed limit within a preset distance ahead of the vehicle and potential stopping points). Driving scenarios are then categorized by "road type + driving behavior," and the core features of driver operation signals and vehicle status signals within each scenario are extracted. A Gaussian mixture model combined with a sliding window is used to identify driving style data in real time, and the aggressiveness of the driving style data is quantified. Using the preprocessed road condition information and quantified driving style data as input, a "LSTM+Bi-LSTM+LSTM" architecture is used to predict future long-term and short-term vehicle speed sequences, respectively. A three-layer optimization strategy—"Dynamic Programming (DP) - Pondriagen Minimum Principle (PMP) - Equivalent Fuel Consumption Minimum Strategy (ECMS)"—is employed to sequentially achieve global SOC trajectory planning, mid-term equivalent factor optimization, and instantaneous range extender power allocation. This method significantly improves the fuel economy of range-extended vehicles, reduces the operating time and start-stop frequency of the range extender (engine), and optimizes the driving experience, making range extension control more aligned with actual driving scenarios and driver operating habits.

[0183] Optionally, the vehicle obtains long-field-of-view information (e.g., historical average speed of 60km / h and gradient of 0-3%) and short-field-of-view information (e.g., real-time traffic flow 3km ahead: 20km / h in congested sections and 80km / h in uncongested sections) through the ADAS map; the environment-historical route information processor module parses the information, converts it into equally spaced data in 1-meter (m) increments, and performs one-hot encoding on the road type (e.g., highway = 1000, main road = 0100). The microcontroller unit (MCU) in the vehicle can acquire signals such as accelerator pedal opening (30%) and longitudinal acceleration (1.5m / s²) at a frequency of 50 Hz, and filter samples according to the "main road acceleration" scenario; the microprocessor unit (MPU) extracts 8 core features and inputs them into the GMM model, calculates the aggressive class posterior probability of 0.6 and the normal class posterior probability of 0.4, and outputs an aggressive index of 0.6; the style index is updated to 0.58 through a sliding window (Q=10).

[0184] Optionally, the MPU inputs short-field information plus a style index of 0.58 into a deep learning model to predict the vehicle speed sequence over the next 1.5km (20-30km / h in congested areas, 50-60km / h in uncongested areas, RMSE=7.634km / h); at the same time, it predicts the global vehicle speed trend based on long-field information (80km / h on highways, 40-60km / h on main roads, RMSE=12.78km / h).

[0185] In this embodiment, the MPU plans the global SOC trajectory using the DP algorithm (gradually decreasing from 35% to 23%, maintaining SOC=30% in high-speed sections); the PMP algorithm is triggered every 1km, taking the short-term predicted vehicle speed as input, and iteratively obtaining the equivalent factor λ=1.2; the MCU calculates the range extender operating point through ECMS based on the current vehicle speed of 50km / h (high-speed mode) and λ=3.2: engine power 26kW (generator) and motor power 40kW (auxiliary drive).

[0186] Figure 6 This is a flowchart of control command execution and status feedback provided in one embodiment of this application, such as... Figure 6 As shown, the process may include the following steps.

[0187] Step S602: Obtain the control command for the range extender.

[0188] In this embodiment, control commands for the range extender can be obtained.

[0189] Step S604: Execute control commands.

[0190] In this embodiment, the range extender executes start / stop / adjustment of operating point according to control instructions.

[0191] Step S606: Collect engine speed or torque, and battery state of charge.

[0192] In this embodiment, engine speed or torque, as well as the state of charge of the battery, can be collected, and the state of charge can be further verified.

[0193] Step S608: Verify whether the SOC deviates from the planned trajectory.

[0194] In this embodiment, it can be verified whether the SOC deviates from the planned trajectory. If it does, step S610 can be executed; otherwise, step S612 can be executed.

[0195] Step S610: Recalculate the SOC trajectory and equivalent factor.

[0196] In this embodiment, if the SOC deviates from the planned trajectory to a certain extent, the SOC trajectory and equivalent factor can be recalculated.

[0197] Step S612: Output feedback results.

[0198] In this embodiment, if the SOC does not deviate from the planned trajectory to a certain extent, a feedback result can be output.

[0199] In this application embodiment, a vehicle control method is proposed. The method acquires road condition information, uses the road condition information and quantified driving style data as input to obtain a predicted vehicle speed sequence, and determines the control data of the range extender (i.e., engine) in the vehicle based on the predicted vehicle speed sequence. By precisely controlling the engine and battery, the technical effect of improving the vehicle's fuel economy is achieved.

[0200] Example 2

[0201] According to an embodiment of this application, a vehicle control device is also provided. It should be noted that this vehicle control device can be used to execute the vehicle control method in Embodiment 1.

[0202] This application also provides a vehicle control device 700, please refer to... Figure 7 , Figure 7 This is a structural diagram of a vehicle control device according to an embodiment of this application, as shown below. Figure 7 As shown, the vehicle control device 700 may include: an acquisition unit 702, a first determination unit 704, a prediction unit 706, a second determination unit 708, and a control unit 710.

[0203] The acquisition unit 702 is used to acquire road condition information of the road to which the vehicle is to travel.

[0204] The first determining unit 704 is used to determine the driving style data of the vehicle on the road to be driven based on road condition information, wherein the driving style data is used to characterize the driving style of the vehicle.

[0205] The prediction unit 706 is used to predict the vehicle speed sequence based on road condition information and driving style data, wherein the predicted speed sequence is used to characterize the vehicle speed when driving at different positions on the road to be driven.

[0206] The second determining unit 708 is used to determine the vehicle's control data based on the predicted vehicle speed sequence, wherein the control data is used to characterize the start-stop strategy and power distribution strategy of the range extender in the vehicle.

[0207] The control unit 710 is used to control the start-up and stop timing of the range extender according to the start-up and stop strategy, and to control the power generation of the range extender according to the power distribution strategy.

[0208] The vehicle control device provided in this application embodiment achieves the following technical effects: The acquisition unit acquires road condition information of the road to be driven; the first determination unit determines the driving style data of the vehicle on the road to be driven based on the road condition information, wherein the driving style data characterizes the vehicle's driving style; the prediction unit predicts the vehicle's predicted speed sequence based on the road condition information and the driving style data, wherein the predicted speed sequence characterizes the vehicle's speed at different positions on the road to be driven; the second determination unit determines the vehicle's control data based on the predicted speed sequence, wherein the control data characterizes the start-stop strategy and power distribution strategy of the range extender in the vehicle; the control unit controls the start-stop timing of the range extender according to the start-stop strategy and controls the power generation of the range extender according to the power distribution strategy, thereby solving the technical problem of low fuel economy in vehicles and achieving the technical effect of improving vehicle fuel economy.

[0209] It should be noted that the above-mentioned units can be implemented by software or hardware. For the latter, they can be implemented in the following ways, but are not limited to: all the above modules are located in the same processor; or, the above modules are located in different processors in any combination.

[0210] Example 3

[0211] This application also provides an electronic device 80, please refer to... Figure 8 , Figure 8 This is a structural diagram of an electronic device provided in one embodiment of the present application, including a memory 810 and a processor 820. The memory 810 is used to store computer programs; the processor 820 is used to execute the programs stored in the memory 810 to implement the vehicle control method described in any embodiment of the present application.

[0212] Optionally, in this embodiment, the processor can be configured to perform the following steps via a computer program:

[0213] Step S1: Obtain road condition information for the road the vehicle is to travel on;

[0214] Step S2: Based on road condition information, determine the driving style data of the vehicle on the road to be driven, wherein the driving style data is used to characterize the driving style of the vehicle.

[0215] Step S3: Based on road condition information and driving style data, predict the vehicle's predicted speed sequence, where the predicted speed sequence is used to characterize the vehicle's speed when traveling at different locations on the road to be traveled.

[0216] Step S4: Based on the predicted vehicle speed sequence, determine the vehicle's control data, wherein the control data is used to characterize the start-stop strategy and power distribution strategy of the range extender in the vehicle.

[0217] Step S5: Control the start-up and stop timing of the range extender according to the start-up and stop strategy, and control the power generation of the range extender according to the power distribution strategy.

[0218] The electronic device provided in this application embodiment achieves the following technical effects: acquiring road condition information, using the road condition information and quantified driving style data as input to obtain a predicted vehicle speed sequence, determining the control data of the range extender (i.e., the engine) in the vehicle based on the predicted vehicle speed sequence, and achieving the purpose of precise control of the engine and battery through the control data, thereby solving the technical problem of low fuel economy of the vehicle and achieving the technical effect of improving the fuel economy of the vehicle.

[0219] Those skilled in the art will understand that Figure 8 The structure shown is for illustrative purposes only. Electronic devices can also be smartphones (such as Android phones, iOS phones, etc.), tablets, PDAs, and mobile internet devices (MIDs) and other terminal devices. Figure 8 This does not limit the structure of the aforementioned electronic device. For example, electronic device 80 may also include components that are more... Figure 8 The more or fewer components shown (such as network interfaces, display devices, etc.), or having the same Figure 8 The different configurations shown.

[0220] Example 4

[0221] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the vehicle control method described in any embodiment of this application.

[0222] Optionally, in this embodiment, the storage medium may be configured to store a computer program for performing the following steps:

[0223] Step S1: Obtain road condition information for the road the vehicle is to travel on;

[0224] Step S2: Based on road condition information, determine the driving style data of the vehicle on the road to be driven, wherein the driving style data is used to characterize the driving style of the vehicle.

[0225] Step S3: Based on road condition information and driving style data, predict the vehicle's predicted speed sequence, where the predicted speed sequence is used to characterize the vehicle's speed when traveling at different locations on the road to be traveled.

[0226] Step S4: Based on the predicted vehicle speed sequence, determine the vehicle's control data, wherein the control data is used to characterize the start-stop strategy and power distribution strategy of the range extender in the vehicle.

[0227] Step S5: Control the start-up and stop timing of the range extender according to the start-up and stop strategy, and control the power generation of the range extender according to the power distribution strategy.

[0228] The electronic device provided in this application embodiment achieves the following technical effects: acquiring road condition information, using the road condition information and quantified driving style data as input to obtain a predicted vehicle speed sequence, determining the control data of the range extender (i.e., the engine) in the vehicle based on the predicted vehicle speed sequence, and achieving the purpose of precise control of the engine and battery through the control data, thereby solving the technical problem of low fuel economy of the vehicle and achieving the technical effect of improving the fuel economy of the vehicle.

[0229] Optionally, in this embodiment, the storage medium may include, but is not limited to, various media capable of storing computer programs, such as USB flash drives, read-only memory (ROM), random access memory (RAM), portable hard drives, magnetic disks, or optical disks.

[0230] The electronic device provided in this application embodiment achieves the following technical effects:

[0231] In this application, "multiple" refers to two or more.

[0232] In this application, unless otherwise expressly defined, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection between two components. Those skilled in the art can understand the specific meaning of the above terms in this application based on the specific circumstances.

[0233] The terms “first,” “second,” “third,” “fourth,” etc., in this application (if present) are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence.

[0234] In this application, the term "and / or" is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent three cases: A existing alone, A and B existing simultaneously, and B existing alone. Additionally, in this application, the character " / " generally indicates that the preceding and following related objects have an "or" relationship.

[0235] Unless otherwise specified, all steps in this application may be performed sequentially or randomly. For example, if a method includes steps A and B, it means that the method may include steps A and B performed sequentially, or it may include steps B and A performed sequentially. For example, if a method may also include step C, it means that step C may be added to the method in any order. For example, the method may include steps A, B, and C, or it may include steps A, C, and B, or it may include steps C, A, and B, etc.

[0236] The above are merely preferred embodiments of this application and are not intended to limit this application. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this application should be included within the protection scope of this application.

Claims

1. A method for controlling a vehicle, characterized in that, include: Obtain road condition information for the road the vehicle is to travel on; Based on the road condition information, the driving style data of the vehicle on the road to be driven is determined, wherein the driving style data is used to characterize the driving style of the vehicle; Based on the road condition information and the driving style data, a predicted vehicle speed sequence is obtained, wherein the predicted vehicle speed sequence is used to characterize the vehicle speed when driving at different positions on the road to be driven. Based on the predicted vehicle speed sequence, control data for the vehicle is determined, wherein the control data is used to characterize the start-stop strategy and power distribution strategy of the range extender in the vehicle. The start-stop timing of the range extender is controlled according to the start-stop strategy, and the power generation power of the range extender is controlled according to the power distribution strategy.

2. The method according to claim 1, characterized in that, The step of determining the vehicle's driving style data on the road to be driven based on the road condition information includes: Based on the traffic information, the road type of the road to be traveled is determined; From the vehicle's historical driving information set, at least one historical driving information that matches the road type is selected, wherein the historical driving information set is used to characterize the vehicle's historical driving status; The driving style data is determined based on at least one of the historical driving information.

3. The method according to claim 2, characterized in that, The determination of the driving style data based on at least one of the historical driving information includes: Clustering is performed on at least one of the historical driving information to obtain clustering results, wherein the clustering results are used to characterize the degree of correlation between the historical driving information and different types of driving style data; The clustering results are transformed to obtain the vehicle's driving style data.

4. The method according to claim 3, characterized in that, The road condition information includes short field of view (SFR) information, which characterizes the road condition status of the vehicle within the target range of the road to be driven. The step of predicting the vehicle's predicted speed sequence based on the road condition information and the driving style data includes: The short field of view information and the driving style data are input into the prediction model to predict the predicted vehicle speed sequence. The prediction model is trained based on driving style data samples, short field of view information samples, and vehicle speed sequence samples corresponding to the driving style data samples and the short field of view information samples. The driving style data samples are used to characterize the historical driving style of the vehicle, and the vehicle speed sequence samples are used to characterize the historical vehicle speed sequence of the vehicle.

5. The method according to claim 4, characterized in that, The road condition information also includes long-field-of-view information, which is used to characterize the overall road condition of the road to be traveled. The method further includes: The long field of view information is input into the prediction model to predict the vehicle's speed throughout the entire journey on the road to be driven.

6. The method according to claim 5, characterized in that, The process of determining the vehicle's control data based on the predicted vehicle speed sequence includes: Determine the state of charge of the battery in the vehicle; Based on the predicted vehicle speed and the state of charge, the state of charge trajectory of the battery in the road to be driven is determined, wherein the state of charge trajectory is used to characterize the state of charge of the battery at different times. Based on the state of charge trajectory and the predicted vehicle speed sequence, the energy consumption data of the range extender under multiple equivalent factors are determined, wherein the equivalent factors are used to balance the energy distribution state between the range extender and the battery. Based on the energy consumption data, a target equivalent factor is determined from among the multiple equivalent factors; Using the target equivalence factor, the power contribution ratio of the battery and the range extender is determined; The power contribution ratio is converted to obtain the control data.

7. The method according to claim 1, characterized in that, The step of determining the vehicle's driving style data on the road to be driven based on the road condition information includes: According to a preset step size, the road condition information in the spatial domain is transformed to obtain equally spaced information in the equally spaced time domain; Based on the equal spacing information, the driving style data is determined.

8. A vehicle control device, characterized in that, include: The acquisition unit is used to acquire road condition information of the road to be traveled by the vehicle; The first determining unit is configured to determine the driving style data of the vehicle on the road to be driven based on the road condition information, wherein the driving style data is used to characterize the driving style of the vehicle. The prediction unit is used to predict the vehicle's speed sequence based on the road condition information and the driving style data, wherein the predicted speed sequence is used to characterize the vehicle's speed when it is traveling at different positions on the road to be driven. The second determining unit is used to determine the control data of the vehicle based on the predicted vehicle speed sequence, wherein the control data is used to characterize the start-stop strategy and power distribution strategy of the range extender in the vehicle. The control unit is configured to control the start-up and stop timing of the range extender according to the start-up and stop strategy, and to control the power generation of the range extender according to the power distribution strategy.

9. An electronic device, characterized in that, Including processor and memory, among which, Memory, used to store computer programs; A processor for executing a program stored in memory to implement the method described in any one of claims 1-7.

10. A vehicle, characterized in that, The vehicle includes an on-board processor and an on-board memory, the on-board memory being used to store computer programs; the on-board processor being used to execute the computer programs stored in the memory, the computer programs being executed by the processor to implement the method described in any one of claims 1 to 7.

Images