Vehicle energy management method and system and vehicle

Abstract

The invention provides a vehicle energy management method and system and a vehicle, and belongs to the technical field of vehicles, and the method comprises the steps: transmitting vehicle identification information to a cloud after the vehicle is powered on, enabling the cloud to obtain the historical driving data of the vehicle based on the vehicle identification information to obtain driving route data, and training a preset first energy management model, target model parameters are obtained, and then a target energy management model is obtained; and determining a target driving route from the plurality of historical driving routes through the application core based on the route data of the plurality of historical driving routes and the current working condition data of the vehicle, and obtaining control parameters for controlling the vehicle to drive on the target driving route based on the target energy management model and the feature data corresponding to the target driving route, and the control parameters are sent to the execution core so that the execution assembly can control the vehicle based on the control parameters. Through the method provided by the invention, the inherent defects of a traditional fixed strategy and a single deployment architecture can be overcome.

Description

Technical Field

[0001] This invention belongs to the field of vehicle technology, specifically relating to a vehicle energy management method, system, and vehicle. Background Technology

[0002] Current vehicle energy management primarily relies on fixed rules and parameters pre-set on the vehicle side, making it difficult to achieve globally optimal energy consumption based on dynamic factors such as real-time road conditions and driving style. While vehicle-cloud collaborative control offers a possibility of overcoming this bottleneck by utilizing cloud computing power, the complexity of the model and frequent updates lead to insufficient stability. Furthermore, its over-reliance on real-time cloud decision-making means that functionality fails in the event of network anomalies, compromising reliability. Summary of the Invention

[0003] In view of the above problems, embodiments of this application provide a vehicle energy management method, system, and vehicle to overcome or at least partially solve the above problems.

[0004] In a first aspect, this application provides a vehicle energy management method applied to a multi-core heterogeneous controller on the vehicle side, the multi-core heterogeneous controller including an application core and an execution core, the method comprising: After the vehicle is powered on, vehicle identification information is sent to the cloud. The cloud then uses this information to obtain the vehicle's historical driving data, and based on this data, obtains driving route data and trains a preset first energy management model to obtain target model parameters. The driving route data includes route data for multiple historical driving routes and feature data corresponding to each route. The feature data includes a vehicle speed-acceleration transition probability matrix, the correspondence between acceleration and latitude / longitude, and the correspondence between vehicle speed and latitude / longitude. The system receives the target model parameters and driving route data from the cloud and configures a preset second energy management model based on the target model parameters. A target energy management model is used; an energy management planning step is executed through the application core, the energy management planning step including: determining a target driving route from the multiple historical driving routes based on route data and the vehicle's current operating condition data; obtaining control parameters for controlling the vehicle to drive on the target driving route based on the target energy management model and the feature data corresponding to the target driving route; sending the control parameters to the execution core; and sending the control parameters to the execution component through the execution core, so that the execution component controls the vehicle based on the control parameters; wherein, the execution component includes at least an engine controller, a motor controller, and a thermal management controller.

[0005] In some embodiments, the first energy model and the second energy model include at least: a driving energy consumption sub-model, a high-voltage accessory energy consumption model, a low-voltage accessory energy consumption model, a battery thermal management sub-model, and a driver style sub-model.

[0006] In some embodiments, the current operating condition data includes the current location of the vehicle, and the route data includes the start location, end location, and number of trips for each historical driving route. Determining a target driving route from the multiple historical driving routes based on the route data and the current operating condition data of the vehicle includes: determining the target location of the vehicle based on the current location and the start location; and determining the target driving route from the multiple historical driving routes based on the target location and the number of trips. The target driving route is the route with the highest driving frequency between the target location and the current location.

[0007] In some embodiments, the current operating condition data further includes the current time of the vehicle, and the route data further includes the time period of each historical driving route. Determining the target location of the vehicle based on the current location and the starting location includes: based on the current location, filtering out a first driving route from the plurality of historical driving routes in which the starting location matches the current location; based on the current time, selecting a second driving route from the first driving route in which the current time falls within the time period; obtaining at least one target location corresponding to the second driving route, and determining the target location with the highest vehicle dwell frequency among the at least one target location as the target location.

[0008] In some embodiments, obtaining the control parameters for controlling the vehicle to travel on the target driving route includes: determining the energy management control parameters of the vehicle on the target driving route based on the target energy management model, the route data of the target driving path, and the feature data, wherein the energy management control parameters include energy consumption curves, state of charge trajectory curves, power battery drive power limit curves, and power battery recovery power limit curves; and determining the control parameters based on the energy management control parameters, as well as preset vehicle hybrid control strategies and thermal management temperature strategies.

[0009] In some embodiments, determining the control parameters based on the energy management control parameters, and preset vehicle hybrid control strategies and thermal management temperature strategies, includes: determining the total energy consumption demand of the vehicle on the target driving route based on the energy consumption curve; determining the total change in battery charge of the vehicle on the target driving route based on the state of charge trajectory curve; determining the power constraint conditions of the power battery based on the power battery drive power limit curve and the power battery recovery power limit curve; and determining the control parameters for the vehicle to travel on the target driving route based on the total energy consumption demand, the total change in battery charge, the power constraint conditions of the power battery, and the vehicle hybrid control strategy and thermal management temperature strategy, wherein the control parameters include at least the generator's power output and speed, the start / stop time and power level of the power battery thermal management controller, and the air conditioning temperature.

[0010] In some embodiments, the application core sends the control parameters to the execution core, including: The application core sends the control parameters to the execution core via internal inter-process communication.

[0011] In some embodiments, after the control parameters are sent to the execution component through the execution core so that the execution component controls the vehicle based on the control parameters, the method further includes: during the process of the vehicle traveling along the target driving route, acquiring actual energy consumption data and predicted energy consumption data of the vehicle at each sampling time point, wherein the predicted energy consumption data is determined based on the driving mileage corresponding to the sampling time point and the energy consumption curve; when the mileage traveled by the vehicle along the target driving route reaches a first preset mileage, if the first energy consumption error between the first actual energy consumption data and the first predicted energy consumption data collected at each sampling time point within the first preset mileage exceeds a preset threshold, and the changing trend of the first energy consumption error at each sampling time point meets a preset trend, then offset compensation is performed on the energy consumption curve based on the offset between the actual energy consumption data and the predicted energy consumption data.

[0012] In some embodiments, after offset compensation of the energy consumption curve based on the offset between the actual energy consumption data and the predicted energy consumption data, the method further includes: when the mileage traveled by the vehicle along the target driving route reaches a second preset mileage, if the second energy consumption error between the second actual energy consumption data and the second predicted energy consumption data collected at each sampling time point within the second preset mileage exceeds the preset threshold, a self-learning flag is sent to the cloud so that the cloud retrains the first energy management model after the vehicle is powered on again to update the target energy management model.

[0013] In some embodiments, after the execution core sends the control parameters to the execution component so that the execution component controls the vehicle based on the control parameters, the method further includes: during the process of the vehicle traveling according to the target driving route, acquiring the actual energy consumption data and predicted energy consumption data of the vehicle at each sampling time point, wherein the predicted energy consumption data is determined based on the driving mileage corresponding to the sampling time point and the energy consumption curve; If the vehicle travels a distance of a first preset mileage along the target route, and the first energy consumption error between the first actual energy consumption data and the first predicted energy consumption data collected at each sampling time point within the first preset mileage exceeds a preset threshold, then a self-learning flag is sent to the cloud so that the cloud can retrain the first energy management model after the vehicle is powered on again, in order to update the target energy management model.

[0014] A second aspect of this application provides a vehicle energy management method applied in a cloud environment, the method comprising: The system receives vehicle identification information sent by the vehicle terminal; based on the vehicle identification information, it obtains the vehicle's historical driving data; based on the historical driving data, it obtains driving route data and trains a preset first energy management model to obtain target model parameters; wherein, the driving route data includes route data of multiple historical driving routes and feature data corresponding to each historical driving route, the feature data including vehicle speed-acceleration transition probability matrix, the correspondence between acceleration and latitude / longitude, and the correspondence between vehicle speed and latitude / longitude; the system sends the target model parameters and the driving route data to the multi-core heterogeneous controller on the vehicle terminal, so that the... The multi-core heterogeneous controller configures a preset second energy management model based on the target model parameters to obtain a target energy management model. The energy management planning steps are then executed by the application core in the multi-core heterogeneous controller. These steps include: determining a target driving route from the multiple historical driving routes based on route data and the vehicle's current operating condition data; obtaining control parameters for controlling the vehicle to drive on the target driving route based on the target energy management model and the feature data corresponding to the target driving route; sending the control parameters to the execution core in the multi-core heterogeneous controller; and sending the control parameters to the execution component through the execution core.

[0015] In some embodiments, training a preset first energy management model to obtain target model parameters includes: dividing the historical driving data into a training set and a test set; training the first energy management model based on the training set to obtain a candidate energy management model; verifying the candidate energy management model based on the test set, and determining the model parameters of the candidate energy management model as the target model parameters if the verification is successful.

[0016] In some embodiments, obtaining the driving route data based on the historical driving data includes: Based on the latitude and longitude information in the historical driving data, a clustering algorithm is used to determine multiple target locations for the vehicle, where the vehicle's dwell time exceeds a preset duration. Based on these multiple target locations, route data for multiple historical driving routes are obtained. For each historical driving route, based on the latitude and longitude information, acceleration, and vehicle speed of the historical driving route, a Markov chain model is used to obtain the corresponding vehicle speed-acceleration transition probability matrix, as well as the correspondence between acceleration and latitude and longitude, and the correspondence between vehicle speed and latitude and longitude.

[0017] A third aspect of this application provides a vehicle energy management system, including a vehicle terminal and a cloud terminal. The vehicle terminal includes a multi-core heterogeneous controller, which includes an application core and an execution core. The multi-core heterogeneous controller is communicatively connected to the cloud terminal. The multi-core heterogeneous controller is used to send vehicle identification information to the cloud terminal after the vehicle is powered on. The cloud terminal is used to obtain historical driving data of the vehicle based on the vehicle identification information, and to obtain driving route data based on the historical driving data, and to train a preset first energy management model to obtain target model parameters. The driving route data includes route data for multiple historical driving routes and feature data corresponding to each historical driving route. The feature data includes a vehicle speed-acceleration transition probability matrix, a correspondence between acceleration and latitude / longitude, and a correspondence between vehicle speed and latitude / longitude. The multi-core heterogeneous controller is used to receive the target model parameters and the driving route data sent from the cloud terminal, and to configure a preset second energy management model based on the target model parameters to obtain a target energy management model. The energy management planning steps are executed through the application core. These steps include: determining a target driving route from the multiple historical driving routes based on route data and the vehicle's current operating condition data; obtaining control parameters for controlling the vehicle to drive on the target driving route based on the target energy management model and the feature data corresponding to the target driving route; sending the control parameters to the execution core; and sending the control parameters to an execution component through the execution core, so that the execution component controls the vehicle based on the control parameters. The execution component includes at least an engine controller, a motor controller, and a thermal management controller.

[0018] In a fourth aspect of this application, a vehicle is provided, the vehicle including a multi-core heterogeneous controller, the multi-core heterogeneous controller being used to execute the steps of the energy management method of the embodiments of this application.

[0019] The vehicle energy management method provided in this embodiment is applied to a multi-core heterogeneous controller on the vehicle side. The multi-core heterogeneous controller includes an application core and an execution core. First, after the vehicle is powered on, the cloud obtains the vehicle's historical driving data based on the vehicle identification information, and obtains driving route data based on the historical driving data, and trains a preset first energy management model to obtain target model parameters. The driving route data includes route data of multiple historical driving routes and feature data corresponding to each historical driving route. The feature data includes the vehicle speed acceleration transition probability matrix, the correspondence between acceleration and latitude and longitude, and the correspondence between vehicle speed and latitude and longitude. By training the first energy management model on the cloud, it is possible to avoid consuming and occupying a large amount of vehicle-side resources.

[0020] Then, the multi-core heterogeneous controller receives the target model parameters and driving route data from the cloud, and configures the preset second energy management model based on the target model parameters to obtain the target energy management model. The application core then executes the energy management planning steps, which include: determining the target driving route from multiple historical driving routes based on route data and the vehicle's current operating condition data; and obtaining control parameters for controlling the vehicle to drive on the target driving route based on the target energy management model and the corresponding feature data. These control parameters are then sent to the execution core, enabling the execution component to control the vehicle based on these parameters. Even in network anomaly scenarios, the application core and execution core on the vehicle side can still independently complete the entire energy management task from route planning to control execution based on the target energy management model and target driving data. Furthermore, the application core and execution core collaborate to optimize task configuration. With the application core responsible for energy management planning and the execution core responsible for high real-time, high-reliability control command distribution, mutual interference between the application core and execution core is effectively avoided. Attached Figure Description

[0021] To more clearly illustrate the technical solutions of the embodiments of this application, the drawings used in the description of the embodiments of this application will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0022] Figure 1 This is a flowchart illustrating the steps of a vehicle energy management method provided in an embodiment of this application; Figure 2 This is a schematic diagram of a sub-model of a first energy model and a second energy model provided in an embodiment of this application; Figure 3 This is a schematic diagram of a vehicle energy management architecture provided in an embodiment of this application; Figure 4 This is a flowchart illustrating the steps of another vehicle energy management method provided in this application embodiment; Figure 5 This is a flowchart of the steps for implementing an energy management method provided in an embodiment of this application. Detailed Implementation

[0023] Exemplary embodiments of this application will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of this application are shown in the drawings, it should be understood that this application may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided to enable a more thorough understanding of this application and to fully convey the scope of this application to those skilled in the art.

[0024] Currently, energy management in the three-electric system (power battery, drive motor, and electronic control system) mostly adopts rule-based fixed strategies, which are difficult to adapt to complex and ever-changing road environments and users' personalized driving habits, resulting in the inability to achieve global energy consumption optimization. The development of traditional energy management strategies usually relies on engine and powertrain simulation software (such as GT-Power, or Gamma Technologies-Power) or vehicle dynamics simulation tools (such as CarSim, or Car Simulation) to initially determine parameters through dynamic and economic simulations. However, the model parameter calibration process in this type of method is time-consuming, and once deployed to the vehicle, it is fixed, making it impossible to dynamically adjust according to actual driving conditions to continuously optimize energy consumption.

[0025] Furthermore, while deploying the model entirely in the cloud allows for large-scale simulation and model training based on massive amounts of vehicle data and offers powerful storage capabilities, its long communication links, insufficient real-time performance, and susceptibility to vehicle-side function degradation or failure during network anomalies make it vulnerable. Conversely, deploying the model entirely on the vehicle's multi-core controller offers advantages such as fast control response, data not leaving the vehicle, and independence from the network, but it is limited by the vehicle's computing power, making it difficult to support complex, large models. Moreover, it can only perform limited model training based on single-vehicle data, hindering cross-vehicle knowledge fusion and generalization.

[0026] In view of this, the vehicle energy management method provided in this embodiment, by deploying a multi-core heterogeneous controller with application cores and execution cores on the vehicle end, can perform model training and knowledge mining on large-scale historical driving data in the cloud, generate high-precision target model parameters and user driving feature data and send them to the vehicle end; the application core uses these target model parameters and driving route data sent from the cloud, combined with the current operating conditions of the vehicle, to determine the target driving data and control parameters of the vehicle; the execution core is responsible for sending the control parameters to each execution component in high real-time and high reliability, effectively overcoming the inherent defects of traditional fixed strategies and single deployment architectures.

[0027] The energy management method for a vehicle provided in this embodiment will be described in detail below.

[0028] Figure 1 This is a flowchart illustrating the steps of a vehicle energy management method provided in an embodiment of this application. Figure 1 As can be seen from this, the multi-core heterogeneous controller applied to the vehicle end includes an application core and an execution core, and the method includes: Step S101: After the vehicle is powered on, vehicle identification information is sent to the cloud so that the cloud can obtain the vehicle's historical driving data based on the vehicle identification information, and obtain driving route data based on the historical driving data, and train a preset first energy management model to obtain target model parameters; wherein, the driving route data includes route data of multiple historical driving routes and feature data corresponding to each historical driving route, the feature data includes vehicle speed acceleration transition probability matrix, the correspondence between acceleration and latitude and longitude, and the correspondence between vehicle speed and latitude and longitude.

[0029] In this embodiment, after the vehicle is powered on, the vehicle terminal can automatically send vehicle identification information, such as the Vehicle Identification Number (VIN), to the cloud. The VIN is used to uniquely identify the vehicle and index its dedicated database. After receiving the vehicle's VIN, the cloud obtains the vehicle's historical driving data. The historical driving data can be the vehicle's historical driving data within the past six months or three months. The historical driving data can include timestamps, latitude and longitude, vehicle speed, acceleration, driving route, battery charge changes, driver's driving style, destination of the driving route, etc. Then, based on the vehicle's historical driving data, the cloud can determine the route data of multiple historical driving routes of the vehicle within the historical time period, as well as the feature data corresponding to each historical driving route. The route data can include vehicle speed, acceleration, destination, etc., and the feature data includes the vehicle speed-acceleration transition probability matrix, the correspondence between acceleration and latitude and longitude, and the correspondence between vehicle speed and latitude and longitude. The vehicle speed-acceleration transition probability matrix represents the probability that a vehicle will transition from its current driving state (e.g., medium speed, slow acceleration) to the next driving state (e.g., high speed, constant speed) on the driving route. The correspondence between acceleration and latitude and longitude represents the acceleration state at each position of the vehicle on the driving route. The correspondence between vehicle speed and latitude and longitude represents the probability distribution of the vehicle speed (low speed, medium speed, high speed) at each position of the vehicle on the driving route.

[0030] Secondly, the first energy management model is a model that predicts the overall energy consumption of a vehicle while it is driving on a route. The overall energy consumption of a vehicle can include driving energy consumption, accessory energy consumption, battery energy consumption, driving mode energy consumption, etc. Based on this, the cloud can also train the first energy management model according to historical driving data (vehicle speed, acceleration, latitude and longitude (used to calculate slope), ambient temperature, battery status (voltage, current, temperature), accessory power (air conditioning, high voltage load), timestamp, etc.) to obtain target model parameters. The target model parameters are the set of internal weights, coefficients and thresholds that the cloud obtains after training the first energy management model, which enables the first energy model to accurately reproduce or predict the energy consumption characteristics of the vehicle.

[0031] Step S102: Receive the target model parameters and the driving route data sent from the cloud, and configure the preset second energy management model based on the target model parameters to obtain the target energy management model.

[0032] In this embodiment, the application core is used to execute energy management planning and generate optimized control parameters. The application core deploys a second energy management model with the same architecture as the first energy management model. The vehicle-side application core can receive target model parameters and driving route data sent from the cloud, and update the model parameters of the second energy management model according to the target model parameters. Specifically, it configures each sub-module (drive energy consumption sub-model, high-voltage accessory energy consumption sub-model, low-voltage accessory energy consumption sub-model, battery thermal management sub-model, and driver style sub-model) in the second energy management model to obtain the target energy management model. The target energy management model can predict and output the energy consumption curve, charge state curve, and power battery power limit curve of the entire target driving route based on the input target driving data.

[0033] Step S103: Execute the energy management planning step through the application core. The energy management planning step includes: determining the target driving route from the multiple historical driving routes based on the route data of the multiple historical driving routes and the current operating condition data of the vehicle; and obtaining control parameters for controlling the vehicle to drive on the target driving route based on the target energy management model and the feature data corresponding to the target driving route, and sending the control parameters to the execution core.

[0034] In this embodiment, the application kernel obtains target driving data based on the driving route data and the vehicle's current operating conditions. The vehicle's current operating conditions can specifically include its current location, current time, and the current state of its power battery. The target driving data is the vehicle's predicted future driving information based on its current location and time, i.e., the target driving route to be traveled, as well as the corresponding speed-acceleration transition probability matrix, the correspondence between acceleration and latitude / longitude, and the correspondence between vehicle speed and latitude / longitude. The target driving data is then input into the target energy management model to generate predicted energy consumption curves, state-of-charge trajectory curves, and power limitation curves. Finally, the control parameters with the lowest energy consumption are selected based on these curves to control the vehicle's driving steps on the target driving route.

[0035] S104: The control parameters are sent to the execution component through the execution core, so that the execution component controls the vehicle based on the control parameters; wherein the execution component includes at least an engine controller, a motor controller and a thermal management controller.

[0036] In this embodiment, after acquiring the control parameters, the application core sends the control parameters to the execution core, which then sends the control parameters to the corresponding execution components, enabling the execution components to control the vehicle based on the control parameters. The execution components include at least an engine controller, a motor controller, and a thermal management controller. Therefore, the control parameters may include the required output power of the range extender, the engine's most efficient operating speed range, the maximum output torque of the drive motor, the power level of the motor for energy recovery during coasting or braking, the on / off times of the battery heating or cooling system during the trip, the set temperature of the passenger compartment air conditioning, and the maximum operating power of the electric compressor, etc.

[0037] Figure 2 This is a schematic diagram of a sub-model of a first energy model and a second energy model provided in an embodiment of this application. Figure 2 It can be seen that the first energy model and the second energy model include at least: the driving energy consumption sub-model, the high-pressure accessory energy consumption sub-model, the low-pressure accessory energy consumption model, the low-pressure accessory energy consumption model, and the driver style sub-model.

[0038] In this embodiment, the driving energy consumption sub-model represents the predicted vehicle speed, acceleration, and road gradient as the vehicle travels on the target route, as well as the power output required by the drive motor and engine. The inputs to the driving energy consumption sub-model can be vehicle speed, gradient, and vehicle speed variance, and the output is driving power. The driving energy consumption sub-model can employ polynomial fitting, as detailed in the following formula:

[0039] in, For vehicle speed; For vehicle speed variance; Drive power (W); The coefficients are polynomial coefficients, which are dimensionless. denoted as , and b as the slope resistance correction factor.

[0040] The high-voltage accessory energy consumption sub-model calculates and accumulates the power consumption of the passenger compartment air conditioning and battery circuit thermal management actuators during the vehicle's journey along the target route, based on the initial, transition, and steady-state phases defined by the vehicle's thermal management strategy. The sub-model primarily considers the power consumption of the passenger compartment air conditioning thermal management and battery circuit thermal management actuators. Furthermore, the accessory power consumption calculation is based on the vehicle's thermal management control strategy, dividing the functional operation into three phases: the initial phase (air conditioning power is a constant, and power consumption is constant * duration), the transition phase (air conditioning power is intelligently adjusted based on target airflow, target passenger compartment temperature, actual passenger compartment temperature, actual passenger compartment humidity, ambient temperature, climate, and machine learning to identify current energy consumption), and the steady-state phase (current energy consumption is obtained through linear fitting of ambient temperature and target passenger compartment temperature). Power consumption is calculated separately for each phase and then accumulated. A machine learning algorithm is used to train a model relating the dependent variable (passenger compartment air conditioning energy consumption) to the independent variables (ambient temperature, wind speed, air humidity, weather conditions (sunny, cloudy, rainy, snowy), actual passenger compartment temperature, and target set temperature). The battery circuit thermal management power consumption is calculated separately for the transient and steady-state phases. A multinomial model is used to train the dependent variable battery thermal management energy consumption and the independent variables ambient temperature, vehicle speed, highest single cell temperature, and lowest single cell temperature.

[0041] The low-voltage accessory energy consumption sub-model calculates the total power consumption of all low-voltage electrical appliances as the vehicle travels along the target route, based on ambient temperature, intelligent driving activation indicator, low-voltage air conditioning activation indicator, seat heating, seat massage, and headlight activation indicators. The high-voltage accessory energy consumption sub-model and the low-voltage accessory energy consumption sub-model can take the on / off status of each onboard electrical appliance, vehicle speed, ambient temperature, and airflow as inputs, and output the accessory power.

[0042] The low-voltage accessory energy consumption sub-model calculates the power consumption required by the battery thermal management system for the vehicle as it travels along the target route, based on ambient temperature, vehicle speed, and the highest and lowest individual cell temperatures, in both transient and steady-state phases. The inputs to the low-voltage accessory energy consumption sub-model can be the highest and lowest cell temperatures, ambient temperature, battery discharge or charging power, vehicle speed, and inlet water temperature; the outputs can be the highest and lowest cell temperatures.

[0043] The driver style sub-model quantifies the personalized influencing factors on drive and regenerative power requirements based on the driving habits identified through cluster analysis of historical driving data as the vehicle travels along the target route. The inputs to the driver style factor model can be accelerator pedal opening, brake opening, vehicle acceleration, and vehicle aggression; the outputs can be driver style (aggressive, conservative) and energy consumption correction factors.

[0044] In summary, since the first energy model and the second energy model have the same architecture, the parameters of the target model trained in the cloud can be seamlessly compatible with and directly applied to the second energy management model of the application core. This avoids complex model training on the vehicle side, reduces the consumption of computing resources on the vehicle side, and ensures the real-time response capability of the vehicle control system.

[0045] In step S102, the current operating condition data includes the vehicle's current position, and the route data includes the start position, end position, and number of trips for each historical driving route. Based on the route data of multiple historical driving routes and the vehicle's current operating condition data, a target driving route is determined from multiple historical driving routes, including: determining the vehicle's target position based on the current position and the start position; and determining the target driving route from multiple historical driving routes based on the target position and the number of trips. The target driving route is the route with the highest driving frequency between the target position and the current position.

[0046] In this embodiment, the current operating condition data includes the vehicle's current location, and the route data includes the start point, end point, and number of trips for each historical driving route. The process for determining the target driving route is as follows: First, the vehicle's real-time positioning coordinates are acquired as the current operating condition data input to determine the vehicle's current position. Then, based on the starting position of each historical driving route, the endpoint position corresponding to the starting position that is the same as the current position is determined as the target position. Since the target driving route is the route with the highest driving frequency between the target position and the current position, the route with the most driving frequency can be selected from multiple historical driving routes based on the target position and the number of driving times. Finally, the target driving route, as well as the corresponding vehicle speed and acceleration transition probability matrix, the correspondence between acceleration and latitude and longitude, and the correspondence between vehicle speed and latitude and longitude, are determined as the target driving data.

[0047] In step S102, the current operating condition data also includes the vehicle's current time, and the route data also includes the time period of each historical driving route. Based on the current location and the starting location, the target location of the vehicle is determined, including: based on the current location, selecting a first driving route from multiple historical driving routes whose starting location matches the current location; based on the current time, selecting a second driving route from the first driving route whose current time corresponds to the time period; obtaining at least one target location corresponding to the second driving route, and determining the target location with the highest vehicle dwell frequency among the at least one target location as the target location.

[0048] In this embodiment, the current location is the real-time latitude and longitude coordinates obtained by the vehicle via GPS, the current time is the real-time time obtained by the vehicle, which may include the day of the week and the time of day, and the driving route data is data stored in the cloud, including multiple driving paths from target location A to target location B. Then, based on the current location, multiple first driving routes matching the current location are filtered from the driving route data, that is, the set of driving routes of all possible destinations of the vehicle from the current location. For example, starting from near home, possible destinations include the company, gym, supermarket, etc., but never a point in another city. Then, based on the multiple first driving routes and the current time, at least one second driving route is selected from the set of destinations the driver is most accustomed to going from the current location, which may be in the same time period as the current time. For example, starting from home: Monday morning at 7:00: the destination of the second driving route is likely to be the company; Saturday afternoon at 3:00: the destination of the second driving route is likely to be the gym or shopping mall. Finally, multiple target locations corresponding to each of the at least one second driving route are obtained, and the target location with the highest vehicle stopping frequency among the multiple target locations is determined as the target location. For example, if in all historical records of leaving home on Monday mornings, the number of times you go to the office is much higher than other locations, then the destination for this trip is predicted to be the office.

[0049] For example, firstly, a matching radius (e.g., 500 meters) is established centered on the vehicle's current location. From multiple historical driving route data, all historical driving routes with starting points within this area are selected to form a first route set. Then, using the time period characteristics of the current moment (e.g., weekday morning rush hour) as a filtering condition, a second route set with the same historical driving records during this time period is further extracted from the first route set. Then, the frequency of occurrence of all endpoints (target locations) in the second route set is statistically analyzed. Finally, the target location with the highest statistical frequency (e.g., when departing from near home at 8 am on Monday, historical records show that 70% of the trips ended at the company campus) is determined as the target location of the target driving route.

[0050] In step S103, the application kernel obtains the control parameters for controlling the vehicle to travel on the target driving route, including: determining the energy management control parameters of the vehicle on the target driving route based on the target energy management model, the route data of the target driving path, and the feature data. The energy management control parameters include the energy consumption curve, the state of charge trajectory curve, the power battery drive power limit curve, and the power battery recovery power limit curve. Based on the energy management control parameters, as well as the preset vehicle hybrid control strategy and thermal management temperature strategy, the control parameters are determined.

[0051] In this embodiment, the target route data includes the vehicle's target driving route. The characteristic data of the target route data includes the vehicle's speed-acceleration transition probability matrix, the correspondence between latitude and longitude and acceleration, and the correspondence between latitude and longitude and vehicle speed. Therefore, the vehicle's target driving data can be directly input into the target energy management model. The target energy management model predicts the vehicle's energy demand and state changes throughout the target driving route through forward simulation calculations, generating energy management control parameters. These energy management control parameters include energy consumption curves, state of charge trajectory curves, power battery drive power limit curves, and power battery recovery power limit curves. The total power demand of the vehicle changes throughout the entire target driving route; the power battery drive power limit curve is the maximum output power boundary of the power battery under different operating conditions; the power battery recovery power limit curve is the maximum feedback power boundary of the power battery under different operating conditions; the vehicle hybrid control strategy calculates the power generation power and power generation speed of the hybrid system based on the target state of charge of the power battery, the actual state of charge of the power battery, the driving mode, the vehicle speed, and the driver's power demand; the thermal management temperature strategy calculates the vehicle thermal management request mode and the thermal management heating or cooling power based on the highest cell temperature, the lowest cell temperature, the ambient temperature, and the driving mode.

[0052] Then, based on the energy management control parameters, as well as the preset vehicle hybrid control strategy and thermal management temperature strategy, a dynamic programming algorithm is used to optimize and solve the problem in the global time domain to generate the vehicle control sequence that minimizes total energy consumption. This control sequence is then converted into real-time control parameters for each execution component. Based on the difference between the state of charge trajectory curve and the real-time state of charge, and the real-time power demand determined by the energy consumption curve, the power generation power and power generation speed are dynamically adjusted through the vehicle hybrid control strategy. Based on the power battery temperature state and environmental conditions, the thermal management system's operating mode and power level are precisely controlled through the thermal management temperature strategy, ultimately obtaining control parameters including the engine target power, motor torque command, and thermal management system operating mode.

[0053] For example, taking a plug-in hybrid electric vehicle with a battery state of charge of 80% commuting on a Monday morning in a 5°C environment as an example, firstly, based on the current location and the time characteristics at 7:50, the destination is predicted to be the company from historical driving data, and a high-frequency driving route is selected. Then, through the target energy management model, a forward simulation is performed on the route, including typical operating conditions such as acceleration at the highway entrance, high-speed cruising, and deceleration at the off-ramp, to generate the optimal energy consumption trajectory that requires the state of charge to drop to 25% upon arrival, the power demand curves for each road segment, and the battery power safety boundary. Based on this, according to the principle of optimal energy distribution throughout the journey, pure electric drive is used to meet the 80kW power demand at the highway entrance segment. After entering high-speed cruising, the range extender is activated and maintained at an efficient operating point of 2000rpm-15kW. At the off-ramp segment, a maximum of 25kW of energy recovery is implemented. At the same time, 2kW of battery heating is activated to ensure battery performance in the low-temperature environment. Finally, the above requirements are converted into control parameters that can control the vehicle, which are then distributed to the corresponding execution components through the execution core. The engine controller receives precise parameters: a target power output of 15kW and a target speed of 2000rpm to ensure optimal operation during high-speed cruising. Dynamic commands are sent to the motor controller for the maximum drive torque at the high-speed entrance and the energy recovery intensity of 25kW at the off-ramp, achieving an optimal balance between power output and energy recovery. Simultaneously, a continuous control signal is transmitted to the thermal management controller to maintain the battery heater at 25°C.

[0054] In step S103, the application kernel determines control parameters based on energy management control parameters and preset vehicle hybrid control and thermal management temperature strategies. These parameters include: determining the total energy consumption demand of the vehicle on the target driving route based on the energy consumption curve; determining the total change in battery charge on the target driving route based on the state of charge trajectory curve; determining the power constraint conditions of the power battery based on the power battery drive power limit curve and the power battery recovery power limit curve; and determining the control parameters for the vehicle's operation on the target driving route based on the total energy consumption demand, the total change in battery charge, the power constraint conditions of the power battery, and the vehicle hybrid control and thermal management temperature strategies. The control parameters include at least the generator's power output and speed, the start / stop time and power level of the power battery thermal management controller, and the air conditioning temperature.

[0055] In this embodiment, the application kernel can determine the control parameters based on the energy management control parameters, as well as the preset vehicle control strategy and thermal management temperature strategy. Specifically, by establishing a multi-objective optimization problem between total energy consumption demand, total change in battery charge and power constraints, the application kernel dynamically plans the energy allocation with the goal of global energy optimization while meeting the safety boundary of the power battery.

[0056] For example, based on the energy consumption curve, assuming the total energy consumption demand for the target driving route is calculated to be 15kWh, and based on the state of charge trajectory curve, it is determined that the state of charge trajectory requirement is to decrease from 80% to 25%, i.e., consume 55% of the electricity. At the same time, based on the power battery drive power limit curve and the power battery recovery power limit curve, it is determined that the battery power constraint limits the drive power to no more than 90kW and the recovery power to no more than 25kW. Based on the above boundary conditions, the vehicle hybrid control strategy calculates that the generator should remain off (0kW of power generation) in the high-speed entry section, and the battery alone provides 80kW of drive power. When entering the high-speed cruising section, according to the balance requirement of total energy consumption and power change, it is calculated that the generator needs to output 15kW of power generation at 2000rpm. At the same time, the thermal management temperature strategy determines that the battery heater should be turned on and maintained at 2kW power level in the 5th minute of the trip according to the battery temperature status, and the air conditioning temperature is set to 22℃ to balance comfort and energy consumption.

[0057] In some embodiments, the application core sends control parameters to the execution core, including: The application kernel sends control parameters to the execution kernel through internal inter-process communication.

[0058] In this embodiment, Figure 3 This is a schematic diagram of a vehicle energy management architecture provided in an embodiment of this application. Figure 3 As can be seen, the application core and the execution core are connected through an internal inter-process communication line (IPCF), the execution core and the execution component are connected through a controller area network bus, the cloud is connected to the multi-core heterogeneous device, and the execution core is connected to the execution component through a controller area network bus.

[0059] In this embodiment, internal inter-process communication provides a high-speed data exchange channel between cores, ensuring that the control parameters generated by the complex planning algorithms of the application core can be transmitted to the dedicated execution core in a timely and reliable manner. This also maintains the system's memory isolation and functional safety. The application core securely transmits the optimized control parameter sequence (including generator target power / speed, thermal management system start / stop times, etc.) to the execution core via internal inter-process communication. The execution core then issues power control commands to the motor controller and engine controller in real time via the controller area network bus, and simultaneously transmits temperature control parameters to the thermal management controller via the local interconnect network bus, thus constructing a complete control closed loop from intelligent decision-making to precise execution. Specifically, when the system plans to output 15kW of generator power at 2000rpm during the high-speed cruise segment, this control parameter is sent to the execution core via inter-core communication and then transmitted to the power system controller via the CAN bus in milliseconds. For timing control requirements such as the battery heater starting at the 5th minute and the air conditioning temperature setting at 22℃, reliable coordination with the thermal management system is achieved via the LIN bus. Ultimately, this hierarchical communication architecture ensures the precise and synchronized execution of the global energy management strategy across multiple systems, including power and thermal management.

[0060] In step S103, after the control parameters are sent to the execution component via the execution core so that the execution component controls the vehicle based on the control parameters, the method further includes: During the vehicle's journey along the target route, actual energy consumption data and predicted energy consumption data are acquired at each sampling time point. The predicted energy consumption data is determined based on the mileage and energy consumption curve corresponding to the sampling time point. If the vehicle travels a first preset mileage along the target route, and the first energy consumption error between the first actual energy consumption data and the first predicted energy consumption data at each sampling time point within the first preset mileage exceeds a preset threshold, and the trend of the first energy consumption error at each sampling time point meets a preset trend, then the energy consumption curve is offset compensation based on the offset between the actual energy consumption data and the predicted energy consumption data.

[0061] In this embodiment, actual energy consumption data refers to the energy actually consumed by the vehicle during its journey along the target route, while predicted energy consumption data refers to the energy consumed by the vehicle during its journey along the target route, predicted using an energy consumption curve. The first preset mileage is the preset mileage traveled by the vehicle along the target displayed route after the execution core sends control parameters to the execution component. The first actual energy consumption data is the energy consumed by the vehicle at each sampling time point during its journey along the first preset mileage. The first predicted energy consumption data is the energy consumed by the vehicle at each sampling time point determined from the energy consumption curve during its journey along the first preset mileage. Then, the first actual energy consumption data and the first predicted energy consumption data collected at each sampling time point are compared. The system checks whether the first energy consumption error between the first actual energy consumption data and the first predicted energy consumption data exceeds a preset threshold, and whether the trend of the first energy consumption error at each sampling point meets a preset trend. The preset trend is that the change pattern of the energy consumption error is consistent. If so, the energy consumption curve is compensated for based on the offset between the actual energy consumption data and the predicted energy consumption data. In this embodiment, the first energy consumption data and the first preset energy consumption data are selected and compared at multiple sampling time points within a preset mileage in order to avoid errors caused by temporary changes in road conditions or accidental driving behavior.

[0062] For example, assuming the predicted total energy consumption for the target driving route is 15kWh, and the plan is to instruct the engine to generate electricity at 15kW during the high-speed cruising segment. After the vehicle completes the first preset mileage, the application core immediately retrieves multiple first actual energy consumption data points collected at sampling time points during this process. Simultaneously, the application core extracts first predicted energy consumption data from the energy consumption curve corresponding to the first actual energy consumption data at the same sampling time point for comparison. If the first actual energy consumption at the first sampling time point is 20% higher than the first predicted energy consumption data, and the first energy consumption error exceeds a preset threshold, the application core will not immediately initiate compensation but will enter a continuous monitoring state. If the first energy consumption error at multiple subsequent consecutive sampling time points all exceed 20%, and the distribution trend of the first energy consumption error remains consistent, the application core determines that a systematic pattern exists. The reasons may include increased internal resistance due to battery performance degradation at low temperatures in winter, or higher-than-expected heating power consumption from the air conditioning.

[0063] Then, the application kernel initiates the vehicle-side compensation mechanism, namely offset compensation, to calculate the average energy consumption offset and compensate the energy consumption curve accordingly. In subsequent trip planning, the application kernel automatically adjusts the predicted energy consumption proportionally upwards and re-optimizes the control parameters based on the corrected energy consumption demand. For example, if the original plan was to generate 15kW of electricity during high-speed cruising, the application kernel will instruct the engine to output a correspondingly increased power after correction, thereby ensuring that the actual energy supply meets the demand while continuing to follow the optimal state of charge trajectory.

[0064] In step S103, after the application kernel performs offset compensation on the energy consumption curve based on the offset between the actual energy consumption data and the predicted energy consumption data, the method further includes: when the vehicle travels a distance of a second preset mileage along the target driving route, if the second energy consumption error between the second actual energy consumption data and the second predicted energy consumption data collected at each sampling time point within the second preset mileage exceeds a preset threshold, a self-learning flag is sent to the cloud so that the cloud can retrain the first energy management model after the vehicle is powered on again to update the target energy management model.

[0065] In this embodiment, the second preset mileage is the mileage traveled by the vehicle on the target route after replenishment. The second actual energy consumption data is the energy consumed by the vehicle at each sampling time point during the second preset mileage. The second predicted energy consumption data is the energy consumed by the vehicle at each sampling time point determined from the energy consumption curve during the second preset mileage. Then, the second actual energy consumption data and the second predicted energy consumption data collected at each sampling time point are compared. If the second energy consumption error between the second actual energy consumption data and the second predicted energy consumption data exceeds a preset threshold, a self-learning flag is sent to the cloud. The self-learning flag is a signal used to inform the cloud that the vehicle-side compensation mechanism is invalid and request retraining.

[0066] For example, when the second energy consumption error (18.4%) between the vehicle's actual energy consumption (4.5kWh) and the original predicted value (3.8kWh) within the second preset mileage continuously exceeds the preset threshold (5%), it indicates that the vehicle-side offset compensation mechanism can no longer correct the deep model bias. At this time, the application kernel sends a self-learning flag to the cloud. After receiving the flag for the vehicle, the cloud adds it to the retraining queue. When the vehicle is powered on again and transmits data, the cloud will collect the latest historical driving data, retrain the first energy management model, and obtain target model parameters that better match the vehicle's current state (such as battery health status and actual engine efficiency). After the vehicle is powered on again, the cloud will resend the updated target model parameters to the application kernel, reconfigure the model parameters of the second energy management model, and update the target energy management model, thereby solving the energy consumption prediction bias problem at its root.

[0067] In step S103, after the control parameters are sent to the execution component via the execution core so that the execution component controls the vehicle based on the control parameters, the method further includes: During the vehicle's journey along the target route, the actual energy consumption data and predicted energy consumption data of the vehicle are acquired at each sampling time point. The predicted energy consumption data is determined based on the mileage and energy consumption curve corresponding to the sampling time point. If the mileage traveled by the vehicle along the target route reaches a first preset mileage, and the first energy consumption error between the first actual energy consumption data and the first predicted energy consumption data collected at each sampling time point within the first preset mileage exceeds a preset threshold, a self-learning flag is sent to the cloud so that the cloud can retrain the first energy management model after the vehicle is powered on again to update the target energy management model.

[0068] In this embodiment, actual energy consumption data refers to the energy actually consumed by the vehicle while traveling on the target route, and predicted energy consumption data refers to the energy consumed by the vehicle while traveling on the target route, predicted by the energy consumption curve. The first preset mileage is the preset mileage traveled by the vehicle on the target displayed route after the execution core sends control parameters to the execution component. The first actual energy consumption data is the energy consumed by the vehicle at each sampling time point during the first preset mileage. The first predicted energy consumption data is the energy consumed by the vehicle at each sampling time point determined from the energy consumption curve during the first preset mileage. Then, the first actual energy consumption data collected at each sampling time point is compared with the first predicted energy consumption data. If the first energy consumption error between the first actual energy consumption data and the first predicted energy consumption data exceeds a preset threshold, the application core sends a self-learning flag to the cloud.

[0069] After receiving the vehicle's flag, it is added to the retraining queue. When the vehicle next powers on and transmits data, the cloud collects the latest historical driving data and retrains the first energy management model to obtain target model parameters that better reflect the vehicle's current state (such as battery health and actual engine efficiency). Upon the vehicle's next power-on, the cloud resends the updated target model parameters to the application core, reconfigures the second energy management model's parameters, and updates the target energy management model, thereby addressing the energy consumption prediction bias problem at its root.

[0070] Figure 4 This is a flowchart illustrating another method for vehicle energy management provided in this application embodiment, applied in the cloud, from... Figure 4 As can be seen from this, the steps of this method include: Step S401: Receive vehicle identification information sent by the vehicle terminal.

[0071] In this embodiment, the cloud receives vehicle identification information sent by the vehicle depot, specifically the vehicle's VIN. The VIN is used to uniquely identify the vehicle and index its dedicated database.

[0072] Step S402: Obtain the vehicle's historical driving data based on the vehicle identification information.

[0073] In this embodiment, after receiving the vehicle's VIN code, the cloud obtains the vehicle's historical driving data. The historical driving data can be the vehicle's historical driving data within the past six months or three months. The historical driving data can include timestamps, latitude and longitude, vehicle speed, acceleration, driving route, battery charge changes, driver's driving style, destination of the driving route, etc.

[0074] Step S403: Obtain historical driving data of the vehicle based on vehicle identification information, and obtain driving route data based on historical driving data, and train the preset first energy management model to obtain target model parameters; wherein, driving route data includes route data of multiple historical driving routes and feature data corresponding to each historical driving route, and feature data includes vehicle speed acceleration transition probability matrix, the correspondence between acceleration and latitude and longitude, and the correspondence between vehicle speed and latitude and longitude.

[0075] In this embodiment, the cloud can determine the route data of multiple historical driving routes of a vehicle within a historical time period based on the vehicle's historical driving data, as well as the feature data corresponding to each historical driving route. The route data may include vehicle speed, acceleration, destination, etc. The feature data includes a vehicle speed-acceleration transition probability matrix, the correspondence between acceleration and latitude / longitude, and the correspondence between vehicle speed and latitude / longitude. The vehicle speed-acceleration transition probability matrix represents the probability of a vehicle transitioning from its current driving state (e.g., medium speed, slow acceleration) to the next driving state (e.g., high speed, constant speed) on the driving route; the correspondence between acceleration and latitude / longitude represents the acceleration state corresponding to each position of the vehicle on the driving route; and the correspondence between vehicle speed and latitude / longitude represents the probability distribution of the vehicle speed (low speed, medium speed, high speed) corresponding to each position of the vehicle on the driving route.

[0076] Secondly, the first energy management model is a model that predicts the overall energy consumption of a vehicle while it is driving on a route. The overall energy consumption of a vehicle can include driving energy consumption, accessory energy consumption, battery energy consumption, driving mode energy consumption, etc. Based on this, the cloud can also train the first energy management model according to historical driving data (vehicle speed, acceleration, latitude and longitude (used to calculate slope), ambient temperature, battery status (voltage, current, temperature), accessory power (air conditioning, high voltage load), timestamp, etc.) to obtain the target model parameters.

[0077] Step S404: The target model parameters and driving route data are sent to the multi-core heterogeneous controller on the vehicle side, so that the multi-core heterogeneous controller can configure the preset second energy management model based on the target model parameters to obtain the target energy management model. The energy management planning steps are executed by the application core in the multi-core heterogeneous controller. The energy management planning steps include: determining the target driving route from multiple historical driving routes based on route data of multiple historical driving routes and the current operating condition data of the vehicle; and obtaining the control parameters for controlling the vehicle to drive on the target driving route based on the target energy management model and the feature data corresponding to the target driving route. The control parameters are sent to the execution core in the multi-core heterogeneous controller; and the execution core sends the control parameters to the execution component.

[0078] In this embodiment, the cloud sends the target model parameters and driving route data to the multi-core heterogeneous controller on the vehicle. The application core in the multi-core heterogeneous controller updates the model parameters of the second energy management model according to the target model parameters. Specifically, it configures each sub-module (drive energy consumption sub-model, high-voltage accessory energy consumption sub-model, low-voltage accessory energy consumption model, battery thermal management sub-model, and driver style sub-model) in the second energy management model to obtain the target energy management model.

[0079] Based on the vehicle's current operating conditions and route data from multiple historical driving routes, the target driving route is obtained. Furthermore, based on the target energy management model and the characteristic data corresponding to the target driving route, control parameters for controlling the vehicle to travel on the target driving route are obtained, and these control parameters are sent to the execution core in the multi-core heterogeneous controller; the execution core then sends the control parameters to the execution components.

[0080] In step S403, the cloud trains the preset first energy management model to obtain the target model parameters, including: dividing the historical driving data into a training set and a test set; training the first energy management model based on the training set to obtain a candidate energy management model; verifying the candidate energy management model based on the test set, and determining the model parameters of the candidate energy management model as the target model parameters if the verification is successful.

[0081] In this embodiment, the first energy management model is trained by first dividing historical driving data into a training set and a test set. Specifically, 80% of the historical driving data is divided into the training set and 20% into the test set. Then, based on the training set, the random forest regression algorithm is used to train each sub-model of the first energy management model (including the driving energy consumption sub-model, the high-voltage accessory energy consumption sub-model, etc.) in parallel. The model hyperparameters are optimized through ten-fold cross-validation to obtain candidate energy management models. Next, the test set is used to fully validate the trained candidate energy management models. The root mean square error and mean absolute percentage error of the energy consumption estimate and the actual measurement value in the forward prediction of the candidate energy management model are calculated. When the mean absolute percentage error is less than 3% and the root mean square error is less than a preset error threshold, the validation is deemed successful, and the model parameters corresponding to the candidate energy management model are determined as the target model parameters. If the mean absolute percentage error is less than 3% or the root mean square error is less than the preset error threshold, the training of the first energy management model continues until the validation is successful.

[0082] In step S403, the cloud obtains driving route data based on historical driving data, including: Based on the latitude and longitude information in historical driving data, a clustering algorithm is used to determine multiple target locations for the vehicle. The target locations are those where the vehicle stays for more than a preset time. Based on the multiple target locations, route data for multiple historical driving routes are obtained. For each historical driving route, based on the latitude and longitude information, acceleration, and vehicle speed of the historical driving route, a Markov chain model is used to obtain the corresponding vehicle speed and acceleration transition probability matrix, as well as the correspondence between acceleration and latitude and longitude, and the correspondence between vehicle speed and latitude and longitude.

[0083] In this embodiment, based on the latitude and longitude information in historical driving data, multiple target locations of the vehicle can be determined through clustering algorithms. Target locations are places where the vehicle has stayed for more than a preset time in the historical driving data (such as company, home, and shopping mall), or places frequently visited by the vehicle. Then, the driving trajectory is segmented using these locations as start and end points, and different driving routes (such as highway routes and urban routes) are distinguished through trajectory clustering. Finally, for each historical driving route, a Markov chain model is used to analyze its historical vehicle speed and acceleration sequence, generating a vehicle speed and acceleration transition probability matrix. At the same time, a correspondence between latitude and longitude and vehicle speed and acceleration status is established, ultimately forming route data containing multiple historical driving routes, as well as the corresponding feature data (vehicle speed and acceleration transition probability matrix, correspondence between acceleration and latitude and longitude, and correspondence between vehicle speed and latitude and longitude) for each historical driving route.

[0084] For example, Figure 5 This is a flowchart of an energy management method provided in an embodiment of this application, which will be discussed below. Figures 1-4 The energy management method implemented in this embodiment will be described in detail: Step S501: Power on the vehicle.

[0085] Step S502: Obtain the vehicle identification code from the cloud, download the vehicle's historical driving data, train the first energy model, and obtain driving route data. The historical driving data includes vehicle speed, temperature, latitude and longitude, acceleration, etc. The first energy model includes a driving energy consumption sub-model, a battery thermal management sub-model, a high-voltage accessory energy consumption sub-model, a low-voltage accessory energy consumption model, and a driver style sub-model, etc.

[0086] Step S503: Verify whether the accuracy of the first energy model meets the preset accuracy. If yes, send the first energy parameters and driving route data to the application core and proceed to step S504. If no, proceed to step S502.

[0087] Step S504: The application kernel receives the first energy parameter and the driving route data, replaces the second energy parameter of the second energy model parameter to obtain the target energy model, and combines the vehicle's current operating conditions and driving route data to obtain the target driving data.

[0088] Step S505: The application core determines the control parameters that minimize the energy consumption of the vehicle under the target driving route based on the target driving data, and sends the control parameters to the execution core.

[0089] Step S506: The execution core distributes the control parameters to the corresponding execution components, which then control the engine controller, motor controller, and thermal management controller.

[0090] Step S507: After driving the first preset mileage, determine that the first energy consumption error between the first actual energy consumption and the first predicted energy consumption exceeds the preset threshold, and the change trend of the first energy consumption error meets the preset trend. If so, proceed to step S508. If it is greater than the preset error and the error does not meet the preset rule, send a flag bit and proceed to step S502.

[0091] Step S508: Start offset compensation.

[0092] Step S509: After starting vehicle-mounted compensation, after driving the second preset mileage, determine whether the second energy consumption error between the second actual energy consumption and the second predicted energy consumption exceeds the preset threshold. If not, send a flag bit and proceed to step S502; if yes, proceed to step S505.

[0093] This embodiment also provides a vehicle energy management system, which includes a vehicle end and a cloud. The vehicle end includes a multi-core heterogeneous controller, which includes an application core and an execution core. The multi-core heterogeneous controller is communicatively connected to the cloud. The multi-core heterogeneous controller is used to send vehicle identification information to the cloud after the vehicle is powered on; The cloud is used to obtain historical driving data of vehicles based on vehicle identification information, and based on the historical driving data, to obtain driving route data, and to train the preset first energy management model to obtain target model parameters; among them, the driving route data includes route data of multiple historical driving routes and feature data corresponding to each historical driving route. The feature data includes vehicle speed acceleration transition probability matrix, the correspondence between acceleration and latitude and longitude, and the correspondence between vehicle speed and latitude and longitude. The multi-core heterogeneous controller is used to receive target model parameters and driving route data sent from the cloud, and configure the preset second energy management model based on the target model parameters to obtain the target energy management model; The energy management planning steps are executed by the application kernel. The energy management planning steps include: determining the target driving route from multiple historical driving routes based on route data and the vehicle's current operating condition data; and obtaining control parameters for controlling the vehicle to drive on the target driving route based on the target energy management model and the feature data corresponding to the target driving route, and sending the control parameters to the execution kernel. The execution core sends control parameters to the execution components, enabling the execution components to control the vehicle based on the control parameters; wherein the execution components include at least the engine controller, the motor controller, and the thermal management controller.

[0094] This embodiment also provides a vehicle, which includes a multi-core heterogeneous controller for executing the steps of the energy management method of this embodiment.

[0095] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The same or similar parts between the various embodiments can be referred to each other.

[0096] Embodiments of the present invention are described with reference to flowchart illustrations and / or block diagrams of methods and apparatus according to embodiments of the present invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing terminal device to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing terminal device, generate instructions for implementing the flowchart illustrations. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0097] Although preferred embodiments of the present invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of the embodiments of the present invention.

[0098] Finally, it should be noted that in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or terminal device that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or terminal device. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or terminal device that includes said element.

[0099] The present invention has provided a detailed description of a vehicle energy management method, system, and vehicle. Specific examples have been used to illustrate the principles and implementation methods of the present invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of the present invention. At the same time, those skilled in the art will recognize that there will be changes in the specific implementation methods and application scope based on the ideas of the present invention. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. An energy management method of a vehicle, characterized by, A multi-core heterogeneous controller applied to a vehicle end, the multi-core heterogeneous controller comprising an application core and an execution core, the method comprising: After the vehicle is powered on, vehicle identification information is sent to the cloud end, so that the cloud end obtains historical driving data of the vehicle based on the vehicle identification information, obtains driving route data based on the historical driving data, and trains a preset first energy management model to obtain target model parameters; wherein the driving route data comprises route data of a plurality of historical driving routes and feature data corresponding to each historical driving route, and the feature data comprises a speed acceleration transition probability matrix, a corresponding relationship between acceleration and latitude and longitude, and a corresponding relationship between speed and latitude and longitude; The target model parameters and the driving route data issued by the cloud end are received, and a preset second energy management model is configured based on the target model parameters to obtain a target energy management model; An energy management planning step is performed by the application core, the energy management planning step comprising: determining a target driving route from the plurality of historical driving routes based on the route data of the plurality of historical driving routes and current working condition data of the vehicle, and obtaining control parameters for controlling the vehicle to drive on the target driving route based on the target energy management model and the feature data corresponding to the target driving route, and sending the control parameters to the execution core; The control parameters are sent to an execution component by the execution core, so that the execution component controls the vehicle based on the control parameters; wherein the execution component at least comprises an engine controller, a motor controller and a thermal management controller.

2. The energy management method of claim 1, wherein, The first energy model and the second energy model at least comprise a driving energy consumption sub-model, a high-voltage accessory energy consumption sub-model, a low-voltage accessory energy consumption sub-model, a battery thermal management sub-model and a driver style sub-model.

3. The energy management method of claim 1, wherein, The current working condition data comprises a current position of the vehicle, the route data comprises a starting position, an ending position and a driving frequency of each historical driving route, and the target driving route is determined from the plurality of historical driving routes based on the route data of the plurality of historical driving routes and the current working condition data of the vehicle, comprising: determining a target position of the vehicle based on the current position and the starting position; determining a target driving route from the plurality of historical driving routes based on the target position and the driving frequency; wherein the target driving route is a route with the highest driving frequency between the target position and the current position.

4. The energy management method of claim 3, wherein, The current working condition data further comprises a current time of the vehicle, and the route data further comprises a time period of each historical driving route, and the target position of the vehicle is determined based on the current position and the starting position, comprising: filtering out a first driving route with a starting position matching the current position from the plurality of historical driving routes based on the current position; filtering out a second driving route with a current time in the time period from the first driving route based on the current time; Obtaining at least one target location corresponding to the second driving route, and determining a target location with the highest vehicle stay frequency among the at least one target location as the target position.

5. The energy management method of claim 1, wherein, The control parameter for controlling the vehicle to drive on the target driving route comprises: Based on the target energy management model, the route data and feature data of the target driving path, determining an energy management control parameter of the vehicle on the target driving route, the energy management control parameter comprising an energy consumption curve, a state of charge trajectory curve, a driving power limit curve of the power battery, and a recovery power limit curve of the power battery; Based on the energy management control parameter, and a preset vehicle hybrid control strategy and thermal management temperature strategy, determining the control parameter.

6. The energy management method of claim 5, wherein, The control parameter is determined based on the energy management control parameter, and a preset vehicle hybrid control strategy and thermal management temperature strategy, comprising: Based on the energy consumption curve, determining the total energy consumption demand of the vehicle on the target driving route; Based on the state of charge trajectory curve, determining the total battery power change of the vehicle on the target driving route; Based on the driving power limit curve of the power battery and the recovery power limit curve of the power battery, determining the power constraint condition of the power battery; Based on the total energy consumption demand, the total battery power change, the power constraint condition of the power battery, and the vehicle hybrid control strategy and thermal management temperature strategy, determining the control parameter of the vehicle driving on the target driving route, wherein the control parameter at least includes the power generation and speed of the generator, the start and stop time and power gear of the power battery thermal management controller, and the air conditioning temperature.

7. The energy management method of claim 1, wherein, The application core sends the control parameter to the execution core, comprising: The application core sends the control parameter to the execution core through internal inter-process communication.

8. The energy management method of claim 1, wherein, After the execution core sends the control parameter to the execution component to control the vehicle based on the control parameter, the method further comprises: During the driving of the vehicle according to the target driving route, obtaining actual energy consumption data and predicted energy consumption data of the vehicle at each sampling time point, the predicted energy consumption data being determined based on the driving distance corresponding to the sampling time point and the energy consumption curve; In the case that the driving distance of the vehicle according to the target driving route reaches a first preset distance, if the first energy consumption error between the first actual energy consumption data and the first predicted energy consumption data of each sampling time point collected within the first preset distance exceeds a preset threshold, and the change trend of the first energy consumption error of each sampling time point meets a preset trend, then the energy consumption curve is offset compensated based on the offset between the actual energy consumption data and the predicted energy consumption data.

9. The energy management method of claim 8, wherein, After the energy consumption curve is offset compensated based on the offset between the actual energy consumption data and the predicted energy consumption data, the method further comprises: In a case where a mileage of the vehicle driving according to the target driving route reaches a second preset mileage, if a second energy consumption error between the second actual energy consumption data and the second predicted energy consumption data of each sampling time point collected within the second preset mileage exceeds the preset threshold, a self-learning flag is sent to the cloud end, so that the cloud end re-trains the first energy management model to update the target energy management model after the vehicle is powered on next time.

10. The energy management method of claim 1, wherein, After the control parameter is sent to the execution component through the execution core to enable the execution component to control the vehicle based on the control parameter, the method further comprises: In the process of the vehicle driving according to the target driving route, actual energy consumption data and predicted energy consumption data of the vehicle at each sampling time point are obtained, the predicted energy consumption data being determined based on a driving mileage corresponding to the sampling time point and the energy consumption curve; In a case where a mileage of the vehicle driving according to the target driving route reaches a first preset mileage, if a first energy consumption error between the first actual energy consumption data and the first predicted energy consumption data of each sampling time point collected within the first preset mileage exceeds a preset threshold, a self-learning flag is sent to the cloud end, so that the cloud end re-trains the first energy management model to update the target energy management model after the vehicle is powered on next time.

11. An energy management method of a vehicle, characterized by, Applied to the cloud end, the method comprises: Receiving vehicle identification information sent by the vehicle end; Based on the vehicle identification information, historical driving data of the vehicle are obtained; Based on the vehicle identification information, historical driving data of the vehicle are obtained, and based on the historical driving data, driving route data are obtained, and a preset first energy management model is trained to obtain target model parameters; wherein the driving route data comprise route data of a plurality of historical driving routes and feature data corresponding to each historical driving route, the feature data comprising a vehicle speed acceleration transition probability matrix, a corresponding relationship between acceleration and latitude and longitude, and a corresponding relationship between vehicle speed and latitude and longitude; The target model parameters and the driving route data are sent to a multi-core heterogeneous controller of the vehicle end, so that the multi-core heterogeneous controller configures a preset second energy management model based on the target model parameters to obtain a target energy management model, and an application core in the multi-core heterogeneous controller executes an energy management planning step, the energy management planning step comprising: determining a target driving route from the plurality of historical driving routes based on the route data of the plurality of historical driving routes and current working condition data of the vehicle, and obtaining control parameters for controlling the vehicle to drive on the target driving route based on the target energy management model and feature data corresponding to the target driving route, and sending the control parameters to an execution core in the multi-core heterogeneous controller; the control parameters are sent to an execution component through the execution core.

12. The energy management method of claim 11, wherein, The training of the preset first energy management model to obtain the target model parameters comprises: The historical driving data are divided into a training set and a test set; training the first energy management model based on the training set to obtain a candidate energy management model; verifying the candidate energy management model based on the test set, and determining model parameters of the candidate energy management model as the target model parameters in a case where the verification passes.

13. The energy management method of claim 11, wherein, The driving route data is obtained based on the historical driving data, including: determining a plurality of target locations of the vehicle based on the latitude and longitude information in the historical driving data by a clustering algorithm, the target location being a location where the vehicle stays for a duration longer than a preset duration; obtaining route data of a plurality of historical driving routes based on the plurality of target locations; for each historical driving route, obtaining a speed-acceleration transition probability matrix corresponding to the historical driving route, and a corresponding relationship between the acceleration and the latitude and longitude and a corresponding relationship between the speed and the latitude and longitude based on the latitude and longitude information, the acceleration and the speed of the historical driving route by a Markov chain model.

14. An energy management system for a vehicle, characterized by including a vehicle end and a cloud end, the vehicle end including a multi-core heterogeneous controller, the multi-core heterogeneous controller including an application core and an execution core, the multi-core heterogeneous controller being in communication connection with the cloud end; the multi-core heterogeneous controller is configured to send vehicle identification information to the cloud end after the vehicle is powered on; the cloud end is configured to obtain historical driving data of the vehicle based on the vehicle identification information, obtain driving route data based on the historical driving data, and train a preset first energy management model to obtain target model parameters; wherein the driving route data includes route data of a plurality of historical driving routes and feature data corresponding to each historical driving route, the feature data including a speed-acceleration transition probability matrix, a corresponding relationship between the acceleration and the latitude and longitude, and a corresponding relationship between the speed and the latitude and longitude; the multi-core heterogeneous controller is configured to receive the target model parameters and the driving route data issued by the cloud end, and configure a preset second energy management model based on the target model parameters to obtain a target energy management model; the application core is configured to execute an energy management planning step, the energy management planning step including: determining a target driving route from the plurality of historical driving routes based on the route data of the plurality of historical driving routes and current working condition data of the vehicle, and obtaining control parameters for controlling the vehicle to drive on the target driving route based on the target energy management model and the feature data corresponding to the target driving route, and sending the control parameters to the execution core; the execution core is configured to send the control parameters to an execution component to control the vehicle based on the control parameters; wherein the execution component at least includes an engine controller, a motor controller and a thermal management controller.

15. A vehicle characterized by comprising: The vehicle includes a multi-core heterogeneous controller, and the multi-core heterogeneous controller is configured to execute the steps of the energy management method according to any one of claims 1-13.

Images