Sample construction method and control method for control model of hybrid vehicle

Abstract

The invention provides a sample construction method for a control model of a hybrid vehicle, and is applied to the technical field of data processing. The method comprises the steps that under the condition that a driving speed sequence of the hybrid power vehicle in a historical driving cycle is kept unchanged, a target decision sequence corresponding to the minimum weighted value of a plurality of driving cost indexes is determined, and the target decision sequence comprises target engine power corresponding to each driving speed; determining a vehicle driving parameter set corresponding to each driving speed according to the driving speed sequence and the multiple target engine powers; and according to the driving speed, the target engine power corresponding to the driving speed and the vehicle driving parameter set corresponding to the driving speed, a training sample used for training a control model of the hybrid vehicle is constructed. The invention further provides a control method of the hybrid vehicle.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and more specifically to a sample construction method and a control method of a hybrid vehicle control model. Background Art

[0002] In the assessment of greenhouse gas emissions and energy consumption in various fields, the proportion of the transportation sector is gradually increasing. The greenhouse gas generated by vehicle fuel consumption in the transportation sector has become one of the key factors affecting the environment. In order to deal with greenhouse gas emissions and energy consumption on the road, the use of hybrid vehicles has become an important solution.

[0003] In the process of realizing the concept of the present invention, there are at least the following problems in the relevant technology: Based on the characteristics of multi-energy coupling of hybrid vehicles, how to control the power ratio between multiple power devices during the driving process of hybrid vehicles to improve the hybrid performance of hybrid vehicles has become a problem that needs to be solved urgently. Summary of the invention

[0004] In view of the above problems, the present invention provides a sample construction method of a control model and a control method for a hybrid vehicle.

[0005] According to a first aspect of the present invention, a sample construction method for a control model of a hybrid vehicle is provided, comprising: determining a target decision sequence corresponding to a minimum weighted value of a plurality of driving cost indicators while keeping a driving speed sequence of the hybrid vehicle in a historical driving cycle unchanged, wherein the target decision sequence includes a target engine power corresponding to each driving speed; determining a vehicle driving parameter set corresponding to each driving speed based on the driving speed sequence and the plurality of target engine powers; and constructing a training sample for training a control model of the hybrid vehicle based on the driving speed, the target engine power corresponding to the driving speed, and the vehicle driving parameter set corresponding to the driving speed.

[0006] According to an embodiment of the present invention, while keeping the driving speed sequence of the hybrid vehicle in the historical driving cycle unchanged, a target decision sequence corresponding to the minimum weighted value of multiple driving cost indices is determined, including: while keeping the driving speed sequence of the hybrid vehicle in the historical driving cycle unchanged, using multiple weight coefficients as decision variables of an elite non-dominated genetic algorithm, minimizing the weighted sum of multiple driving cost indices as an optimization goal, determining driving cost index value sets and preferred decision sequences corresponding to multiple target weight coefficient sets respectively; determining a target driving cost index value set from multiple driving cost index value sets; and determining the preferred decision sequence corresponding to the target driving cost index value set as the target decision sequence.

[0007] According to an embodiment of the present invention, while keeping the driving speed sequence of the hybrid vehicle in the historical driving cycle unchanged, multiple weight coefficients are used as decision variables of the elite non-dominated genetic algorithm, minimizing the weighted sum of multiple driving cost indicators as the optimization goal, and determining the driving cost indicator value sets and the preferred decision sequence corresponding to the multiple target weight coefficient sets respectively, including: under the constraint conditions of the multiple weight coefficients, for each individual composed of the multiple weight coefficients in the p-th round population, while keeping the driving speed sequence unchanged, using the respective value ranges of the multiple vehicle operating parameters as constraints, minimizing the weighted sum of the multiple driving cost indicators in the historical driving cycle as the optimization goal, optimizing the engine power of the hybrid vehicle, and obtaining the optimal decision sequence corresponding to the driving speed. The method comprises the following steps: a decision sequence corresponding to a sequence, and a p-th driving cost index value set in a historical driving cycle, wherein p>0, and p is an integer, and the p-th driving cost index value set includes multiple p-th driving cost index values; according to the p-th driving cost index value set of each individual, determining the p-th fitness of each individual; based on the p-th fitness of each individual, selecting each individual to obtain a p-th wheel population; determining the (p+1)-th wheel population according to the p-th wheel population, and repeating the above operation until a preset iteration round P is reached; determining the P-th driving cost index value set corresponding to each individual in the P-th wheel population as multiple driving cost index value sets; determining the P-th decision sequence corresponding to each individual in the P-th wheel population as multiple preferred decision sequences.

[0008] According to an embodiment of the present invention, a target driving cost index value set is determined from a plurality of driving cost index value sets, including: constructing a decision matrix using the plurality of driving cost index value sets; normalizing the decision matrix to obtain a normalized matrix; determining an ideal optimal solution set and an ideal worst solution set based on the normalized matrix; calculating the relative proximity of each of the plurality of driving cost index value sets according to the ideal optimal solution and the ideal worst solution; and determining the target driving cost index value set from the plurality of driving cost index value sets according to the plurality of relative proximity.

[0009] According to an embodiment of the present invention, the method further includes: weighting each driving cost index value according to the index type of each driving cost index value in the normalized matrix to obtain a weighted normalized matrix, so as to determine an ideal optimal solution set and an ideal worst solution set based on the weighted normalized matrix.

[0010] According to an embodiment of the present invention, each driving cost index value is weighted according to the index type of each driving cost index value in the normalized matrix to obtain a weighted normalized matrix, including: constructing a judgment matrix according to the relative importance of multiple index types; using the judgment matrix to calculate the arithmetic mean values ​​corresponding to each index type; using each arithmetic mean value to weight each driving cost index value to obtain a weighted normalized matrix.

[0011] According to an embodiment of the present invention, the training samples include a first training sample and a second training sample; the vehicle driving parameter set includes motor power, vehicle power and state of charge; according to the driving speed, the target engine power corresponding to the driving speed and the vehicle driving parameter set corresponding to the driving speed, a training sample for training a control model of a hybrid vehicle is constructed, including: determining a driving mode of the hybrid vehicle based on the target engine power and the motor power; constructing a first training sample based on the driving mode, the vehicle power, the state of charge and the driving speed corresponding to the vehicle driving parameter set, wherein the driving mode is a label of the first training sample; constructing a second training sample corresponding to the first training sample based on the driving mode, the vehicle power, the state of charge, the driving speed corresponding to the vehicle driving parameter set and the target engine power corresponding to the driving speed, wherein the target engine power is a label of the second training sample.

[0012] According to an embodiment of the present invention, the value ranges of the plurality of vehicle operating parameters include an engine power interval, a state of charge interval and a motor power interval, and the engine power interval and the motor power interval are preset; the state of charge interval is determined by the following method: according to the (k+1)th state of charge and the engine power interval, a kth state of charge interval corresponding to the kth target engine power is determined, wherein K≥2, and K is an integer, k belongs to [1, K-1], and k is an integer; the maximum value with a smaller value in the first maximum value in the preset state of charge interval and the second maximum value in the kth state of charge interval is taken as the target maximum value; the minimum value with a larger value in the first minimum value in the preset state of charge interval and the second minimum value in the kth state of charge interval is taken as the target minimum value; the above steps are repeated until k=1; and the state of charge interval is determined based on the target minimum value and the target maximum value.

[0013] The second aspect of the present invention provides a control method for a hybrid vehicle, comprising: determining the vehicle power based on the collected driving speed and the power transmission model of the hybrid vehicle; inputting the driving speed, the vehicle power and the collected state of charge into a control model, and outputting the target engine power of the hybrid vehicle, wherein the control model is trained based on training samples corresponding to multiple driving speeds, and the training samples are constructed based on the sample construction method of the control model of the above-mentioned hybrid vehicle; determining the target motor power of the hybrid vehicle based on the target engine power, the vehicle power and the power transmission model; and controlling the operation of the hybrid vehicle based on the target motor power and the target motor power.

[0014] According to an embodiment of the present invention, the control model includes a first control module and a second control module, wherein the first control module is used to output the driving mode of the hybrid vehicle according to the driving speed, the state of charge, and the vehicle power; and the second control module is used to output the target engine power according to the driving mode, the driving speed, the state of charge, and the vehicle power.

[0015] According to an embodiment of the present invention, the driving speed of the hybrid vehicle in the historical driving cycle is taken as a constant value, and the driving speed sequence of the hybrid vehicle in the historical driving cycle is kept unchanged. The minimum weighted value of multiple driving cost indicators is used to optimize the power ratio between the engine and the motor in the historical driving cycle to determine the target decision sequence, and the driving speed of the hybrid vehicle and the target decision sequence are used to determine the vehicle driving parameter set corresponding to the driving speed. Then, the corresponding relationship between the driving state of the hybrid vehicle and the power ratio is constructed using the corresponding relationship between the driving speed, the target engine power corresponding to the driving speed, and the vehicle driving parameter set corresponding to the driving speed. Based on the corresponding relationship, a training sample for training a control model is constructed, and the control model is trained using multiple training samples, so that the control model can make decisions that are beneficial to optimizing energy management and improving overall performance during the driving of the hybrid vehicle. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] The above contents and other objects, features and advantages of the present invention will become more apparent through the following description of the embodiments of the present invention with reference to the accompanying drawings, in which:

[0017] Figure 1 An application scenario diagram of a sample construction method of a control model for a hybrid vehicle according to an embodiment of the present invention is shown.

[0018] Figure 2 A flow chart showing a sample construction method of a control model for a hybrid vehicle according to an embodiment of the present invention is shown.

[0019] Figure 3 A schematic diagram showing the principle of determining a decision sequence according to an embodiment of the present invention is shown.

[0020] Figure 4 A schematic diagram of determining a target driving cost indicator set according to an embodiment of the present invention is shown.

[0021] Figure 5 A flow chart showing a control method of a hybrid vehicle according to an embodiment of the present invention is shown.

[0022] Figure 6 A block diagram of an electronic device suitable for implementing a sample construction method of a control model for a hybrid vehicle according to an embodiment of the present invention is shown. DETAILED DESCRIPTION

[0023] Below, embodiments of the present invention will be described with reference to the accompanying drawings. However, it should be understood that these descriptions are exemplary only and are not intended to limit the scope of the present invention. In the following detailed description, for ease of explanation, many specific details are set forth to provide a comprehensive understanding of embodiments of the present invention. However, it is apparent that one or more embodiments may also be implemented without these specific details. In addition, in the following description, descriptions of known structures and technologies are omitted to avoid unnecessary confusion of concepts of the present invention.

[0024] The terms used herein are only for describing specific embodiments and are not intended to limit the present invention. The terms "comprise", "include", etc. used herein indicate the existence of the features, steps, operations and / or components, but do not exclude the existence or addition of one or more other features, steps, operations or components.

[0025] All terms (including technical and scientific terms) used herein have the meanings commonly understood by those skilled in the art unless otherwise defined. It should be noted that the terms used herein should be interpreted as having a meaning consistent with the context of this specification and should not be interpreted in an idealized or overly rigid manner.

[0026] When using expressions such as "at least one of A, B, and C, etc.", they should generally be interpreted according to the meaning of the expression commonly understood by those skilled in the art (for example, "a system having at least one of A, B, and C" should include but is not limited to a system having A alone, B alone, C alone, A and B, A and C, B and C, and / or A, B, C, etc.).

[0027] It should be noted that in the embodiments of the present application, certain software, components, models and other existing solutions in the industry may be mentioned, and they should be regarded as exemplary. Their purpose is only to illustrate the feasibility of implementing the technical solution of the present application, but it does not mean that the applicant has or will necessarily use the solution.

[0028] In the scenario of using personal information for automated decision-making, the methods, devices, and systems provided by the embodiments of the present invention provide users with corresponding operation portals for users to choose to agree or reject the automated decision-making results; if the user chooses to reject, the expert decision-making process will be entered. The expression "automated decision-making" here refers to the activity of automatically analyzing and evaluating an individual's behavioral habits, interests and hobbies, or economic, health, credit status, etc. through computer programs, and making decisions. The expression "expert decision-making" here refers to the activity of making decisions by people who specialize in a certain field, have specialized experience, knowledge and skills, and have reached a certain level of professionalism.

[0029] With the progress of society and the growth of energy demand, the problem of global warming has prompted countries to strengthen the control of greenhouse gas emissions. In this context, in order to reduce greenhouse gas emissions and energy consumption in the transportation sector, the use of hybrid vehicles has become a key solution to reduce carbon emissions from road transportation. Hybrid vehicles have improved the traditional use of engine fuel to provide driving force for the vehicle, and can partially or completely replace the engine to provide driving force for the vehicle through the motor.

[0030] Hybrid vehicles, due to their multi-energy coupling characteristics, have shown significant advantages in energy conservation and emission reduction compared to traditional single energy systems. This system intelligently controls each power unit, distributes power according to changes in operating conditions, and ensures that each power unit operates in the optimal working range, thereby improving the overall performance of the hybrid vehicle.

[0031] Since hybrid vehicles introduce new components such as batteries and motors, and there are multiple power ratios between motors and engines, the performance evaluation of hybrid vehicles must not only focus on fuel consumption, but also consider multiple objectives such as battery life, pollutant emissions, and engine operating conditions. Therefore, intelligent and efficient multi-objective energy management strategies have become an important direction for hybrid vehicle research.

[0032] The control strategy of the motor and engine in a hybrid vehicle requires a trade-off between multiple objectives such as power performance, cost, emissions, and energy efficiency. For example, in order to improve energy efficiency, battery life may be sacrificed or emissions may increase. The complexity of these issues makes traditional single-objective optimization methods difficult to apply. Therefore, an optimization method that can objectively and reasonably balance various objectives is needed to achieve the optimization of hybrid vehicle strategies.

[0033] An embodiment of the present invention provides a sample construction method for a control model of a hybrid vehicle, comprising: determining a target decision sequence corresponding to a minimum weighted value of a plurality of driving cost indicators while keeping a driving speed sequence of the hybrid vehicle in a historical driving cycle unchanged, wherein the target decision sequence includes a target engine power corresponding to each driving speed; determining a vehicle driving parameter set corresponding to each driving speed based on the driving speed sequence and the plurality of target engine powers; and constructing a training sample for training a control model of the hybrid vehicle based on the driving speed, the target engine power corresponding to the driving speed, and the vehicle driving parameter set corresponding to the driving speed.

[0034] Figure 1 An application scenario diagram of a sample construction method of a control model for a hybrid vehicle according to an embodiment of the present invention is shown.

[0035] like Figure 1As shown, the application scenario according to this embodiment may include a first hybrid vehicle 101 , a second hybrid vehicle 102 , a third hybrid vehicle 103 and a server 104 .

[0036] The first hybrid vehicle 101, the second hybrid vehicle 102, and the third hybrid vehicle 103 support a conventional fuel engine combined with one or more motors to drive the vehicle forward. The first hybrid vehicle 101, the second hybrid vehicle 102, and the third hybrid vehicle 103 are equipped with a system for controlling vehicle driving, a control model, and a device for detecting vehicle driving speed and battery remaining power.

[0037] Server 104 may be a server that provides various services for the first hybrid vehicle 101, the second hybrid vehicle 102, and the third hybrid vehicle 103. Various model parameters in the control models of the first hybrid vehicle 101, the second hybrid vehicle 102, and the third hybrid vehicle 103 may be obtained by training on server 104, and a sample construction method of the control model of the hybrid vehicle may be executed in server 104.

[0038] It should be understood that Figure 1 The number and variety of hybrid vehicles and servers in the world are just...

[0039] The following will be based on Figure 1 The scene described by Figure 2~Figure 4 A sample construction method of a control model for a hybrid vehicle according to an embodiment of the present invention is described in detail.

[0040] Figure 2 A flow chart showing a sample construction method of a control model for a hybrid vehicle according to an embodiment of the present invention is shown.

[0041] like Figure 2 As shown, the sample construction method of the control model of the hybrid vehicle of this embodiment includes: operation S210 to operation S230.

[0042] In operation S210 , a target decision sequence corresponding to a minimum weighted value of a plurality of driving cost indicators is determined while keeping a driving speed sequence of the hybrid vehicle in a historical driving cycle unchanged, wherein the target decision sequence includes a target engine power corresponding to each driving speed.

[0043] In operation S220 , a vehicle driving parameter set corresponding to each driving speed is determined according to the driving speed sequence and a plurality of target engine powers.

[0044] In operation S230, a training sample for training a control model of a hybrid vehicle is constructed according to a driving speed, a target engine power corresponding to the driving speed, and a vehicle driving parameter set corresponding to the driving speed.

[0045] According to an embodiment of the present invention, the historical driving cycle is a continuous driving process of the hybrid vehicle in a historical time period. The driving speed sequence can be obtained by sampling the driving speed of the hybrid vehicle at multiple time points in the historical driving cycle.

[0046] For example, during the 60 minutes of continuous driving of a hybrid vehicle, the driving speed change data of the hybrid vehicle over time is recorded. The driving speed at 60 time points from 1 to 60 minutes can be collected in a 1-minute cycle as the driving speed sequence of the hybrid vehicle.

[0047] According to an embodiment of the present invention, when collecting the driving speed of a hybrid vehicle in a historical driving cycle, two time interval methods can be used: a fixed time interval and a non-fixed time interval. A fixed time interval refers to maintaining the same time period between each time point, such as recording the driving speed once per second; while a non-fixed time interval allows the interval between time points to change flexibly according to actual conditions, such as shortening the interval when the driving speed changes greatly, and extending the interval when the driving speed is stable. These two methods can be selected according to actual needs.

[0048] According to an embodiment of the present invention, the driving cost index is an index that needs to be optimized during the driving of a hybrid vehicle. The multiple driving cost indexes may include fuel consumption, power consumption, battery attenuation, engine start and stop times, etc. Different driving cost indexes have different energy-saving and emission-reduction effects.

[0049] For example, fuel consumption refers to the amount of fuel consumed by a hybrid vehicle during driving. This indicator can be used to measure the fuel economy of a hybrid vehicle, that is, the amount of fuel consumed per unit distance (such as per 100 kilometers). The lower the fuel consumption, the higher the fuel efficiency of the hybrid vehicle, the lower the operating cost, and also means less carbon emissions.

[0050] For example, power consumption refers to the amount of electrical energy consumed by a hybrid vehicle while driving. This indicator reflects the efficiency of the hybrid vehicle's energy utilization through the motor to generate power. The lower the power consumption, the more energy-efficient the motor is, which can extend the driving range and reduce dependence on fuel, thereby reducing operating costs and emissions.

[0051] For example, battery degradation refers to the phenomenon that the capacity of a battery gradually decreases during use, which can be expressed as a percentage of the battery life reduction. Battery degradation is an important factor affecting the performance and cost of hybrid vehicles. The slower the degradation, the longer the battery life and the lower the maintenance and replacement costs. It also means that hybrid vehicles can maintain good electric performance for a longer time and reduce the amount of fuel used by the vehicle.

[0052] For example, the number of engine starts and stops refers to the number of times the engine starts and stops during vehicle driving. This indicator is closely related to the vehicle's fuel economy and emissions. Frequent starts and stops may increase fuel consumption and emissions. Optimizing the engine start-stop strategy can reduce unnecessary starts and stops, thereby improving fuel economy and reducing emissions.

[0053] According to an embodiment of the present invention, the weights of multiple driving cost indicators can be determined based on expert experience. Based on the Bellman optimal principle, the optimization problem in the entire historical driving cycle is converted into a backward optimization sequence, and each discrete stage is regarded as a sub-problem. The optimal solution of each sub-problem is solved from back to front and recorded, and it is iterated to the initial state according to the stage sequence, and then an optimal decision path is determined according to the initial state.

[0054] For example, in a historical driving cycle of 120 minutes, the driving speed of the hybrid vehicle in the historical driving cycle is collected in a 2-minute period, and the 120 minutes can be divided into 60 time periods, namely, 118-120, 116-118, ... 0-2. From back to front, while keeping the driving speed sequence of the hybrid vehicle unchanged, the target engine power corresponding to the minimum weighted value of multiple driving cost indicators is determined, thereby determining the target decision sequence.

[0055] According to an embodiment of the present invention, the vehicle driving parameter set may include the state of charge, vehicle power, driving mode, etc. of the hybrid vehicle. The vehicle driving parameter set of the hybrid vehicle may be determined by using the dynamic model of the hybrid vehicle according to the driving speed and target engine power of the hybrid vehicle. The dynamic model of the hybrid vehicle is a complex system for simulating and analyzing the linkage relationship including the internal combustion engine, the electric motor, the battery and other components. For example, the vehicle power of the hybrid model may be determined in the dynamic model by the driving speed.

[0056] According to an embodiment of the present invention, the training sample includes input data and labels of the input control model, the input data may include a driving speed and a vehicle driving parameter set corresponding to the driving speed, and the label may be a target engine power corresponding to the driving speed.

[0057] According to an embodiment of the present invention, a vehicle driving parameter set may include the motor power of a hybrid vehicle, and the data input into a control model may include a driving speed, a state of charge corresponding to the driving speed, a vehicle power, a driving mode, etc. The label may be the motor power corresponding to the driving speed, and the motor power is the battery power in a hybrid vehicle.

[0058] According to an embodiment of the present invention, the driving speed of the hybrid vehicle in the historical driving cycle is taken as a constant value, and the driving speed sequence of the hybrid vehicle in the historical driving cycle is kept unchanged. The minimum weighted value of multiple driving cost indicators is used to optimize the power ratio between the engine and the motor in the historical driving cycle to determine the target decision sequence, and the driving speed of the hybrid vehicle and the target decision sequence are used to determine the vehicle driving parameter set corresponding to the driving speed. Then, the corresponding relationship between the driving state of the hybrid vehicle and the power ratio is constructed using the corresponding relationship between the driving speed, the target engine power corresponding to the driving speed, and the vehicle driving parameter set corresponding to the driving speed. Based on the corresponding relationship, a training sample for training a control model is constructed, and the control model is trained using multiple training samples, so that the control model can make decisions that are beneficial to optimizing energy management and improving overall performance during the driving of the hybrid vehicle.

[0059] According to an embodiment of the present invention, while keeping the driving speed sequence of the hybrid vehicle in the historical driving cycle unchanged, a target decision sequence corresponding to the minimum weighted value of multiple driving cost indices is determined, including: while keeping the driving speed sequence of the hybrid vehicle in the historical driving cycle unchanged, using multiple weight coefficients as decision variables of an elite non-dominated genetic algorithm, minimizing the weighted sum of multiple driving cost indices as an optimization goal, determining driving cost index value sets and preferred decision sequences corresponding to multiple target weight coefficient sets respectively; determining a target driving cost index value set from multiple driving cost index value sets; and determining the preferred decision sequence corresponding to the target driving cost index value set as the target decision sequence.

[0060] According to an embodiment of the present invention, the weight coefficient can be set as an unknown quantity, and the search ability of the elite non-dominated genetic algorithm in the optimization target composed of multiple driving cost indicators is utilized. Through iterative calculation and genetic operations such as selection, crossover, and mutation, the optimal solution is gradually approached, and the weight coefficient is finally determined. These weight coefficients reflect the importance of different driving cost indicators in the optimization process. In the "selection" process, multiple weight coefficients in the weight coefficient set are used as known quantities. By minimizing the weighted sum of multiple driving cost indicators, the driving cost indicator value set and the preferred decision sequence corresponding to the weight coefficient set are determined, and selection is made from multiple weight coefficient sets based on each driving cost indicator. Through multiple iterations, the Pareto optimal solution composed of multiple target weight coefficient sets, as well as the driving cost indicator value set and the preferred decision sequence corresponding to the multiple target weight coefficient sets are determined.

[0061] According to an embodiment of the present invention, the driving cost index value set is a set of multiple driving index values ​​generated by the hybrid vehicle in the historical driving cycle. The preferred decision sequence is the target engine power corresponding to each driving speed determined based on the corresponding weight coefficient set in the historical driving cycle.

[0062] According to an embodiment of the present invention, the target driving cost index value set is a driving cost index value set that best meets the minimum consumption cost of hybrid vehicles among multiple driving cost index value sets. The preferred decision sequence corresponding to the target driving cost index value set is determined as the target decision sequence.

[0063] According to an embodiment of the present invention, an elite non-dominated genetic algorithm with a powerful global search capability is used to find multiple groups of Pareto optimal solution sets, i.e., multiple driving cost index value sets, among multiple driving cost indexes, and a target driving cost index value set is determined from the multiple driving cost index value sets. The preferred decision sequence corresponding to the target driving cost index value set is determined as the target decision sequence, rather than directly using the weighted driving cost index value including the weight coefficient as the selection direction, so that the balanced optimization of multiple driving cost indexes can be fully considered, the limitations brought by the optimization of a single index are reduced, and the comprehensiveness and robustness of the target decision sequence are improved, so that the control model can be trained using the obtained more comprehensive and robust training samples to more efficiently control the hybrid vehicle.

[0064] According to an embodiment of the present invention, while keeping the driving speed sequence of the hybrid vehicle in the historical driving cycle unchanged, multiple weight coefficients are used as decision variables of the elite non-dominated genetic algorithm, minimizing the weighted sum of multiple driving cost indicators as the optimization goal, and determining the driving cost indicator value sets and the preferred decision sequence corresponding to the multiple target weight coefficient sets respectively, including: under the constraint conditions of the multiple weight coefficients, for each individual composed of the multiple weight coefficients in the p-th round population, while keeping the driving speed sequence unchanged, using the respective value ranges of the multiple vehicle operating parameters as constraints, minimizing the weighted sum of the multiple driving cost indicators in the historical driving cycle as the optimization goal, optimizing the engine power of the hybrid vehicle, and obtaining the optimal decision sequence corresponding to the driving speed. The method comprises the following steps: a decision sequence corresponding to a sequence, and a p-th driving cost index value set in a historical driving cycle, wherein p>0, and p is an integer, and the p-th driving cost index value set includes multiple p-th driving cost index values; according to the p-th driving cost index value set of each individual, determining the p-th fitness of each individual; based on the p-th fitness of each individual, selecting each individual to obtain a p-th wheel population; determining the (p+1)-th wheel population according to the p-th wheel population, and repeating the above operation until a preset iteration round P is reached; determining the P-th driving cost index value set corresponding to each individual in the P-th wheel population as multiple driving cost index value sets; determining the P-th decision sequence corresponding to each individual in the P-th wheel population as multiple preferred decision sequences.

[0065] According to an embodiment of the present invention, the constraints of the multiple weight coefficients may include: each weight coefficient is greater than 0, and the sum of the multiple weights is 1. To reduce the amount of calculation, a weight vector consisting of one less weight coefficient than the multiple weight coefficients may be used to characterize the individual.

[0066] For example, in the case of four weight coefficients, we can use As decision variables in elitist non-dominated genetic algorithm.

[0067] According to an embodiment of the present invention, each individual in the first round of population can be obtained by random initialization under the constraints of multiple weight coefficients. In the pth round of population, for each individual consisting of multiple weight coefficients, a dynamic programming algorithm can be used to optimize the engine power of the hybrid vehicle by taking the value range of each of the multiple vehicle operating parameters as a constraint and minimizing the weighted sum of the multiple driving cost indicators in the historical driving cycle as the optimization target while keeping the driving speed sequence unchanged.

[0068] According to an embodiment of the present invention, the cost at the end moment can be initialized, and the weighted sum of multiple driving cost indicators can be used as the objective function to find the optimal decision variable at each driving speed in each sampling period from back to front, and iterate until the initial moment, so as to obtain a global optimal decision path corresponding to the individual. In the case where the multiple driving cost indicators are fuel consumption, power consumption, battery attenuation and the number of engine starts and stops in the historical driving cycle, due to the different dimensions of each cost, multiple driving cost indicators can be normalized using multiple weight coefficients. Then the fuel consumption at each moment can be determined by formula (1):

[0069] (1);

[0070] in, for Fuel consumption over time, is the sampling step length, The engine power is The fuel consumption of the engine at time k is the sequence number represented by the sampling time point.

[0071] The power consumption at each moment can be determined by formula (2):

[0072] (2);

[0073] in, for Power consumption over time, is the motor power.

[0074] The battery degradation at each moment can be determined by formula (3):

[0075] (3);

[0076] in, for The battery degradation over time, is the pre-exponential factor, is the gas constant, is the thermodynamic temperature, is the battery current rate, It can be determined by formula (4):

[0077] (4);

[0078] in, is the battery open circuit voltage, is the battery capacity in Ah.

[0079] The number of engine starts and stops at each moment can be determined by formula (5):

[0080] (5);

[0081] in, The number of engine starts and stops within a sampling period.

[0082] According to an embodiment of the present invention, the weighted sum of multiple driving cost indicators is used as the objective function, and it is iterated from back to front. The weighted sum iterative process of the multiple driving cost indicators is shown in formula (6):

[0083] (6);

[0084] in, Represents the weighted sum of multiple driving cost indicators generated from the kth sampling time node to the end of the historical driving cycle; The weighted sum of multiple driving cost indicators that characterize the end of the historical driving cycle from the kth sampling time node to the (k+1)th sampling time node. Hybrid vehicles have different working modes, and the decision variables in each working mode are The costs generated are different. Only the minimum value of the weighted sum iteration of multiple driving cost indicators at each moment is taken to determine the decision sequence corresponding to the driving rate sequence when the weighted sum of multiple driving cost indicators of the entire historical driving cycle is the minimum. The decision sequence and the decision path formed by the motor power corresponding to each engine power in the decision sequence are the optimal control strategy for the state at the initial moment. Then, the optimal decision path determined by the initial state is forwarded to obtain the optimal solution of the cycle.

[0085] Figure 3 A schematic diagram showing the principle of determining a decision sequence according to an embodiment of the present invention is shown.

[0086] like Figure 3 As shown in the dynamic programming process, the decision variables u(1)~u(N) and the current cost J(k) are the same as the current cost J(k) in the above formula (6). The corresponding relationship between them is: the decision variable, namely the engine power u(k), and the current state variable, namely the state of charge s(k), jointly determine the next state variable s(k+1), where k belongs to 1~N, and the current cost J(k) is generated in this process.

[0087] Figure 3 The figure shows the change of state of charge over time, where t represents time. In dynamic programming, optimization usually involves the control of state variables. The arrows and labels in the figure are It shows the optimization direction of the state of charge SOC at different time points. For example, the arrow from SOC(t+1) to SOC(t) shows how the decision made at time t affects the state at the next time point.

[0088] According to an embodiment of the present invention, the p-th driving cost index value set includes multiple p-th driving cost index values, and the multiple p-th driving cost index values ​​are values ​​of each driving cost index without weight. For each individual in the p-th round population, the p-th driving cost index value set corresponding to the individual is recorded during the process of determining the decision sequence, so that the multiple p-th driving cost index values ​​included in the p-th driving cost index value set of each individual are used as the values ​​of multiple objective functions in the elite non-genetic algorithm to calculate the p-th fitness of each individual.

[0089] According to an embodiment of the present invention, based on the pth fitness of each individual, each individual is selected to obtain the pth round sub-population, until the number of population iterations reaches a preset value P. When the number of population iterations does not reach the preset value P, the pth round sub-population is crossovered and mutated to obtain a plurality of pth transformed individuals, and the plurality of pth transformed individuals are combined with the pth round sub-population to obtain the (p+1)th round population.

[0090] According to an embodiment of the present invention, the Pth driving cost index value set corresponding to multiple individuals in the Pth wheel subpopulation is determined as multiple driving cost index value sets, that is, multiple weight coefficient sets obtained by an elite non-genetic algorithm, driving cost index value sets corresponding to the multiple weight coefficient sets are determined, and multiple preferred decision sequences corresponding to the multiple weight coefficient sets are determined.

[0091] According to an embodiment of the present invention, the driving cost index value set generated during the optimization process is used as the basis for calculating fitness to determine the impact of different individuals on the final result. By using multiple unweighted driving cost index values ​​to evaluate individuals, multiple driving cost indexes can be better balanced, avoiding ignoring the impact of other driving cost indexes on the final result due to over-emphasis on a single goal, thereby improving the accuracy of the driving cost index value set and the preferred decision sequence.

[0092] According to an embodiment of the present invention, a target driving cost index value set is determined from a plurality of driving cost index value sets, including: constructing a decision matrix using the plurality of driving cost index value sets; normalizing the decision matrix to obtain a normalized matrix; determining an ideal optimal solution set and an ideal worst solution set based on the normalized matrix; calculating the relative proximity of each of the plurality of driving cost index value sets according to the ideal optimal solution and the ideal worst solution; and determining the target driving cost index value set from the plurality of driving cost index value sets according to the plurality of relative proximity.

[0093] According to an embodiment of the present invention, a decision matrix is ​​constructed in this study. It can be expressed by formula (7):

[0094] (7);

[0095] in, Indicates Among individuals The driving cost index value, is the number of individuals in the P-th round subpopulation, is the number of types of travel cost index values.

[0096] According to an embodiment of the present invention, the decision matrix is ​​normalized to obtain a normalized matrix. It can be expressed by formula (8):

[0097] (8);

[0098] in, is the number of individuals in the P-th wheel population The minimum value of the driving cost index value; is the number of individuals in the P-th wheel population According to an embodiment of the present invention, in the normalized matrix, the minimum value of the driving cost index value of each category is selected as the ideal optimal solution set; in the normalized matrix, the minimum value of the driving cost index value of each category is selected as the ideal worst solution set.

[0099] According to an embodiment of the present invention, the relative proximity of each of the plurality of driving cost index value sets is calculated based on the ideal optimal solution and the ideal worst solution. The relative proximity can be expressed by formula (9):

[0100] (9);

[0101] in, For the The relative closeness of the travel cost index value sets, For the The distance from the set of driving cost index values ​​to the ideal optimal solution set, For the The distance from a set of travel cost index values ​​to the ideal worst solution set.

[0102] According to an embodiment of the present invention, when the relative proximity is expressed by formula (9), multiple driving cost index value sets are sorted according to the relative proximity, and the driving cost index value set with the smallest relative proximity among the multiple driving cost index value sets is used as the target driving cost index value set.

[0103] According to an embodiment of the present invention, the numerator in formula (9) can also be changed to In this case, the driving cost index value set with the largest relative proximity among the multiple driving cost index value sets is taken as the target driving cost index value set.

[0104] According to an embodiment of the present invention, a target driving cost index value set is determined from multiple driving cost index value sets by calculating relative proximity to an ideal optimal solution and an ideal worst solution, wherein the multiple driving cost index value sets are sets of pre-selected solutions that satisfy a relative balance of multiple driving cost index values. In the multiple driving cost index value sets, the target driving cost index value set is determined based on multiple relative proximity. By calculating the relative proximity, solutions that are balanced and close to the ideal in multiple dimensions can be further accurately identified, thereby improving the reliability of the target driving cost index value set, and further improving the reliability of the training samples.

[0105] According to an embodiment of the present invention, the sample construction method of the control model of a hybrid vehicle also includes: weighting each driving cost index value according to the index type of each driving cost index value in the normalized matrix to obtain a weighted normalized matrix, so as to determine an ideal optimal solution set and an ideal worst solution set based on the weighted normalized matrix.

[0106] According to an embodiment of the present invention, each driving cost index value in the normalized matrix is ​​weighted. Specifically, a weight is assigned to each driving cost index value according to the importance of each index type to construct a weighted normalized matrix. Each element in the weighted normalized matrix can be expressed by formula (10):

[0107] (10);

[0108] in, For the The driving cost index value set The importance weight of the class travel cost index value, It can be determined based on expert experience. It is the index value after weighting the driving cost index value.

[0109] According to an embodiment of the present invention, after the weighted normalization matrix is ​​determined, the ideal optimal solution set and the ideal worst solution set are calculated to obtain the maximum and minimum values ​​of different indicators, wherein the ideal optimal solution set It can be expressed by formula (11):

[0110] (11);

[0111] Ideal worst solution set It can be expressed by formula (12):

[0112] (12);

[0113] in, is the maximum value of the driving cost index value m in the weighted normalized matrix, is the maximum value vector composed of the maximum values ​​of m travel cost indicators; is the minimum value of the driving cost index m in the weighted normalized matrix, It is the minimum value vector composed of the minimum values ​​of m indicators.

[0114] According to an embodiment of the present invention, the distances of each alternative solution to an ideal optimal solution set and an ideal worst solution set are calculated; The distance from a Pareto optimal solution to the ideal solution can be expressed by formula (13):

[0115] (13);

[0116] No. The distance from a Pareto optimal solution to the ideal solution can be expressed by formula (14):

[0117] (14);

[0118] According to an embodiment of the present invention, on the basis of the Pareto optimal solution, by constructing a decision matrix, normalizing, assigning weights, calculating the distance between the ideal solution and the worst solution and the relative proximity, the solution ranking and screening is finally achieved. In this way, multiple evaluation indicators can be comprehensively considered to avoid the one-sidedness of single indicator decision-making; quantitative evaluation can be objectively performed based on the relationship between the indicator value and the ideal solution and the worst solution to reduce the interference of subjective factors; further screening can be performed on the basis of the Pareto optimal solution to accurately locate the optimal solution with the best comprehensive performance among multiple excellent solutions, thereby providing a more scientific, accurate and efficient solution for decision-making to improve the quality and effect of decision-making.

[0119] According to an embodiment of the present invention, each driving cost index value is weighted according to the index type of each driving cost index value in the normalized matrix to obtain a weighted normalized matrix, including: constructing a judgment matrix according to the relative importance of multiple index types; using the judgment matrix to calculate the arithmetic mean values ​​corresponding to each index type; using each arithmetic mean value to weight each driving cost index value to obtain a weighted normalized matrix.

[0120] According to an embodiment of the present invention, the hierarchical analysis method is used to determine the relative importance of multiple indicator types. A judgment matrix can be constructed based on importance cognition, and a weight vector of each target can be obtained based on the judgment matrix. After passing the consistency test, the weight vector is the importance weight of each target.

[0121] According to an embodiment of the present invention, constructing the judgment matrix A can be expressed by formula (15):

[0122] (15);

[0123] is the driving cost index Driving cost index The importance of , its value can be determined according to the 1-9 scaling method, and the weight vector is calculated using the arithmetic mean method. The weight vector can be expressed by formula (16):

[0124] (16);

[0125] According to an embodiment of the present invention, a consistency check is performed on the judgment matrix and the weight vector to calculate the consistency index. The consistency ratio is calculated based on the average random consistency index. If the consistency ratio is less than 0.1, the judgment is considered to be accurate. The judgment matrix can be used to calculate the arithmetic mean corresponding to each indicator type. Each driving cost indicator value is weighted using each arithmetic mean to obtain a weighted normalized matrix.

[0126] According to the embodiment of the present invention, by constructing a judgment matrix based on the relative importance of multiple indicator types, it is possible to systematically quantify the important differences between different indicators, and use the calculated arithmetic mean to weight the driving cost indicator value, so that the final weighted normalized matrix more accurately reflects the value weight of each indicator in actual decision-making. This helps to highlight the impact of key driving cost indicators in determining multiple non-dominated Pareto optimal connections in the optimization of hybrid vehicle driving costs, thereby more scientifically and rationally allocating resources and formulating strategies, and improving the accuracy and effectiveness of decision-making.

[0127] Figure 4 A schematic diagram of determining a target driving cost indicator set according to an embodiment of the present invention is shown.

[0128] like Figure 4 As shown in the figure, it is assumed that the indicator types include driving cost indicator 1, driving cost indicator 2, driving cost indicator 3 and driving cost indicator 4. The decision maker makes a subjective judgment on the relative importance of multiple indicator types to construct a judgment matrix to quantify the relative importance of different types of driving cost indicators. The judgment matrix is ​​used to calculate the arithmetic mean corresponding to each indicator type; each arithmetic mean is used to weight each driving cost indicator value in the Pareto frontier to obtain a weighted normalized matrix.

[0129] According to the decision matrix, determine the ideal optimal solution and the ideal worst solution. Calculate the relative closeness of each alternative plan to the ideal optimal solution and the ideal worst solution in terms of driving cost index 1, driving cost index 2, driving cost index 3 and driving cost index 4. When the relative closeness is , use the numerator in formula (9) to change to When the formula is obtained, the driving cost index value set with increased relative proximity is used to determine the target driving cost index value set.

[0130] According to an embodiment of the present invention, the training samples include a first training sample and a second training sample; the vehicle driving parameter set includes motor power, vehicle power and state of charge; according to the driving speed, the target engine power corresponding to the driving speed and the vehicle driving parameter set corresponding to the driving speed, a training sample for training a control model of a hybrid vehicle is constructed, including: determining a driving mode of the hybrid vehicle based on the target engine power and the motor power; constructing a first training sample based on the driving mode, the vehicle power, the state of charge and the driving speed corresponding to the vehicle driving parameter set, wherein the driving mode is a label of the first training sample; constructing a second training sample corresponding to the first training sample based on the driving mode, the vehicle power, the state of charge, the driving speed corresponding to the vehicle driving parameter set and the target engine power corresponding to the driving speed, wherein the target engine power is a label of the second training sample.

[0131] According to an embodiment of the present invention, in the case where a hybrid vehicle is equipped with only one motor for providing power for vehicle pulling, the driving mode of the hybrid vehicle can be determined according to the target engine power and the motor power. Specifically, when the target engine power is not zero and the motor power is not zero, the driving mode of the hybrid vehicle is hybrid drive; when the target engine power is zero and the motor power is not zero, the driving mode of the hybrid vehicle is motor drive; when the target engine power is not zero and the motor power is zero, the driving mode of the hybrid vehicle is engine drive.

[0132] According to an embodiment of the present invention, the vehicle power, state of charge, driving speed corresponding to the vehicle driving parameter set and driving mode included in the vehicle driving parameter set are used to construct a first training sample, wherein the vehicle power, state of charge and driving speed corresponding to the vehicle driving parameter set can be used as the working condition of the hybrid vehicle as input data in training, and the driving mode is the label of the first training sample. The relationship between the vehicle power, the engine power and the motor power is related to the configuration of the hybrid vehicle, and is therefore expressed by the following formula (17):

[0133] (17);

[0134] in, is the battery power, It is the power required by the whole vehicle, that is, the vehicle power.

[0135] According to an embodiment of the present invention, based on the vehicle driving parameter set included in the vehicle power, state of charge, driving mode, driving speed corresponding to the vehicle driving parameter set and the target engine power corresponding to the driving speed, a second training sample corresponding to the first training sample is constructed, wherein the target engine power is the label of the second training sample, and the vehicle power, state of charge, driving speed corresponding to the vehicle driving parameter set and driving mode are used as input data during training.

[0136] According to an embodiment of the present invention, the driving mode of the hybrid vehicle is determined by using the target engine power and the motor power included in the vehicle driving parameter set, thereby constructing a first training sample and a second training sample, and the driving mode is used as an intermediate result, so that the driving mode as an intermediate variable can be used to more specifically assist the control model in predicting the target engine power, so as to improve the accuracy of the control model and thus improve the precision of controlling the hybrid vehicle.

[0137] According to an embodiment of the present invention, the value ranges of the plurality of vehicle operating parameters include an engine power interval, a state of charge interval and a motor power interval, and the engine power interval and the motor power interval are preset; the state of charge interval is determined by the following method: according to the (k+1)th state of charge and the engine power interval, a kth state of charge interval corresponding to the kth target engine power is determined, wherein K≥2, and K is an integer, k belongs to [1, K-1], and k is an integer; the maximum value with a smaller value in the first maximum value in the preset state of charge interval and the second maximum value in the kth state of charge interval is taken as the target maximum value; the minimum value with a larger value in the first minimum value in the preset state of charge interval and the second minimum value in the kth state of charge interval is taken as the target minimum value; the above steps are repeated until k=1; and the state of charge interval is determined based on the target minimum value and the target maximum value.

[0138] According to an embodiment of the present invention, the plurality of vehicle operating parameters may include engine power, state of charge, and motor power. The value ranges of the plurality of vehicle operating parameters include an engine power interval, a state of charge interval, and a motor power interval, and the engine power interval and the motor power interval are pre-set. The value ranges of the plurality of vehicle operating parameters may be expressed by formula (18):

[0139] (18);

[0140] in, is the minimum limit of the state of charge, is the maximum limit of the state of charge, is the engine power, Minimum engine power limit, Maximum engine power limit, is the motor power, is the minimum power limit of the motor, The present embodiment takes a single motor as an example, but may include multiple motors.

[0141] According to an embodiment of the present invention, the state of charge of the battery After discretization, it is used as a discrete state variable. for Moment , establish the state transfer equation, and use the discretized engine power as the discrete decision variable , for Engine power at the time, is the sampling time. A complete driving cycle consists of “time-vehicle speed” and the time range can be 0~K.

[0142] According to an embodiment of the present invention, the state transition at each moment is realized by decision variables. The state transition can be expressed by formula (19):

[0143] (19);

[0144] in, is the battery open circuit voltage, is the internal resistance of the battery, is the battery capacitance, is the sampling step, and the battery internal resistance and battery capacitance are fixed parameters.

[0145] According to an embodiment of the present invention, setting The changing boundary of Constraints to ensure that when the working condition is over In the reverse optimization process, by determining the maximum and minimum power of the battery generated in the decision-making process at the previous moment, the battery power at the previous moment is derived according to the state transfer equation. Minimum and maximum values, calculated backwards until The maximum and minimum values ​​reach the boundaries 0.3 and 0.8. The minimum boundary can be expressed by formula (20):

[0146] (20);

[0147] in, for time The minimum value of for time The minimum value of for At the moment of decision variable The change in power obtained under this condition can be understood by combining formula (17) and formula (19): The change in power can be expressed by formula (21):

[0148] (twenty one);

[0149] The maximum value boundary can be expressed by formula (22):

[0150] (20);

[0151] in, for time The maximum limit of for time The maximum limit of for At the moment of decision variable The change in power obtained.

[0152] Battery state of charge at the last discrete state , discrete decision variables .

[0153] According to an embodiment of the present invention, the state of charge interval can be dynamically defined in combination with the engine power interval and the state of charge at different stages. Compared with simply using a preset state of charge interval, it can more accurately reflect the reasonable range of the state of charge of the vehicle during actual operation. When determining the target engine power, a more accurate state of charge interval means that the determination of the target engine power can better match the current energy state of the vehicle, reducing the unreasonable setting of the target engine power due to inaccurate estimation of the value range of the state of charge, thereby improving the reliability and accuracy of the training samples.

[0154] With the development of machine learning technology, learning-based energy management strategies are expected to achieve online optimal control of hybrid vehicles. Among machine learning algorithms, the random forest algorithm has attracted attention due to its effectiveness in dealing with complex data and multi-objective optimization problems. Through these advanced technologies and strategies, hybrid vehicles can be more effectively managed and optimized to achieve more efficient and environmentally friendly energy utilization.

[0155] The second aspect of the present invention provides a control method for a hybrid vehicle, comprising: determining the vehicle power based on the collected driving speed and the power transmission model of the hybrid vehicle; inputting the driving speed, the vehicle power and the collected state of charge into a control model, and outputting the target engine power of the hybrid vehicle, wherein the control model is trained using multiple training samples corresponding to multiple driving speeds; determining the target motor power of the hybrid vehicle based on the target engine power, the vehicle power and the power transmission model; and controlling the operation of the hybrid vehicle based on the target motor power and the target motor power.

[0156] According to an embodiment of the present invention, training samples for training a control model of a hybrid vehicle can also be constructed by the following method: based on the vehicle power, state of charge, driving speed corresponding to the vehicle driving parameter set and target engine power corresponding to the driving speed included in the vehicle driving parameter set, training samples are constructed, wherein the target engine power is the label of the training sample.

[0157] According to an embodiment of the present invention, in order to increase the number of training samples, the sample construction method of the control model of the hybrid vehicle described above may be used to process multiple historical driving cycles respectively to obtain a larger number of training samples.

[0158] According to an embodiment of the present invention, the collected driving speed is used as a variable to calculate the vehicle demand power at the driving speed, that is, the vehicle power, using the power transmission model of the hybrid vehicle. The collected driving speed, the collected state of charge and the vehicle power are input into the control model to output the target engine power of the hybrid vehicle.

[0159] According to an embodiment of the present invention, the target engine power and vehicle power are input into a powertrain model to calculate the target motor power of the hybrid vehicle using the powertrain model; and the operation of the hybrid vehicle is controlled based on the target motor power and the target motor power.

[0160] According to an embodiment of the present invention, the engine control unit can adjust the fuel injection, ignition timing and other parameters of the engine to achieve the target motor power. The motor controller is required to adjust the speed and torque of the motor according to the target motor power to achieve the target motor power.

[0161] According to an embodiment of the present invention, the driving speed of a hybrid vehicle in a historical driving cycle is taken as a constant value, and the driving speed sequence of the hybrid vehicle in the historical driving cycle is kept unchanged. The power ratio between the engine and the motor in the historical driving cycle is optimized by using the minimum weighted value of multiple driving cost indicators to determine a target decision sequence, and the driving speed of the hybrid vehicle and the target decision sequence are used to determine a vehicle driving parameter set corresponding to the driving speed. Then, the corresponding relationship between the driving state of the hybrid vehicle and the power ratio is constructed by using the corresponding relationship between the driving speed, the target engine power corresponding to the driving speed, and the vehicle driving parameter set corresponding to the driving speed. Based on the corresponding relationship, a training sample for training a control model is constructed, and the control model is trained by using multiple training samples, so that the control model can make decisions that are beneficial to optimizing energy management and improving overall performance during the driving of the hybrid vehicle.

[0162] According to an embodiment of the present invention, the control model includes: a first control module and a second control module, wherein the first control module is used to output the driving mode of the hybrid vehicle according to the driving speed, the state of charge, and the vehicle power; the second control module is used to output the target engine power according to the driving mode, the driving speed, the state of charge, and the vehicle power.

[0163] According to an embodiment of the present invention, the first control module is responsible for outputting the driving mode of the hybrid vehicle according to the driving speed of the vehicle and the state of charge and vehicle power in the vehicle driving parameter set. The first control module can be trained by the first training sample, and the second control module is trained by the second training sample corresponding to the first training sample. Based on the driving mode determined by the first control module, the second control module further outputs the target engine power according to the driving speed, state of charge and vehicle power. This ensures that the power output of the engine matches the driving mode to achieve better fuel efficiency and performance.

[0164] Figure 5 A flow chart showing a control method of a hybrid vehicle according to an embodiment of the present invention is shown.

[0165] like Figure 5 As shown, the driving speed, state of charge, and vehicle power are input into the first control module 501, and the driving mode of the hybrid vehicle is output; the driving mode, driving speed, state of charge, and vehicle power are input into the second control module, and the target engine power of the hybrid vehicle is output.

[0166] According to an embodiment of the present invention, the first control module and the second control module can be obtained by training the algorithm model based on the random forest algorithm. During the training process, if the driving mode is correct, the parameters and algorithms related to the engine power calculation in the second control module are further checked; if the driving mode is wrong, the judgment logic of the first control module can be optimized. This sub-module feedback and adjustment mechanism can more efficiently improve the accuracy of the output results of the entire system.

[0167] According to the embodiment of the present invention, the driving mode is an intermediate result, which contains important information about the vehicle's operating state. Since the calculation method and influencing factors of the target engine power are different under different driving modes, the driving mode output by the first control module is used as one of the inputs of the second control module, so that the prediction of the target engine power can better adapt to different driving scenarios, so that the second control module can perform power prediction more targetedly, and improve the accuracy of the target engine power.

[0168] Figure 6 A block diagram of an electronic device suitable for implementing a sample construction method of a control model for a hybrid vehicle according to an embodiment of the present invention is shown.

[0169] like Figure 6 As shown, the electronic device 600 according to an embodiment of the present invention includes a processor 601, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 602 or a program loaded from a storage part 608 to a random access memory (RAM) 603. The processor 601 may include, for example, a general-purpose microprocessor (e.g., a CPU), an instruction set processor and / or a related chipset and / or a special-purpose microprocessor (e.g., an application-specific integrated circuit (ASIC)), etc. The processor 601 may also include an onboard memory for caching purposes. The processor 601 may include a single processing unit or multiple processing units for performing different actions of the method flow according to an embodiment of the present invention.

[0170] In RAM 603, various programs and data required for the operation of electronic device 600 are stored. Processor 601, ROM 602 and RAM 603 are connected to each other via bus 604. Processor 601 performs various operations of the method flow according to the embodiment of the present invention by executing the programs in ROM 602 and / or RAM 603. It should be noted that the program can also be stored in one or more memories other than ROM 602 and RAM 603. Processor 601 can also perform various operations of the method flow according to the embodiment of the present invention by executing the programs stored in one or more memories.

[0171] According to an embodiment of the present invention, the electronic device 600 may further include an input / output (I / O) interface 605, which is also connected to the bus 604. The electronic device 600 may further include one or more of the following components connected to the input / output (I / O) interface 605: an input portion 606 including a keyboard, a mouse, etc.; an output portion 607 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker, etc.; a storage portion 608 including a hard disk, etc.; and a communication portion 609 including a network interface card such as a LAP card, a modem, etc. The communication portion 609 performs communication processing via a network such as the Internet. A drive 610 is also connected to the input / output (I / O) interface 605 as needed. A removable medium 611, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 610 as needed, so that a computer program read therefrom is installed into the storage portion 608 as needed.

[0172] The present invention also provides a computer-readable storage medium, which may be included in the device / apparatus / system described in the above embodiment; or may exist independently without being assembled into the device / apparatus / system. The above computer-readable storage medium carries one or more programs, and when the above one or more programs are executed, the method according to the embodiment of the present invention is implemented.

[0173] According to an embodiment of the present invention, the computer-readable storage medium may be a non-volatile computer-readable storage medium, for example, it may include but is not limited to: a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In the present invention, the computer-readable storage medium may be any tangible medium containing or storing a program, which may be used by or in combination with an instruction execution system, an apparatus or a device. For example, according to an embodiment of the present invention, the computer-readable storage medium may include the ROM 602 and / or RAM 603 described above and / or one or more memories other than ROM 602 and RAM 603.

[0174] The embodiment of the present invention also includes a computer program product, which includes a computer program, and the computer program includes a program code for executing the method shown in the flowchart. When the computer program product is run in a computer system, the program code is used to enable the computer system to implement the sample construction method of the control model of the hybrid vehicle provided by the embodiment of the present invention.

[0175] The computer program executes the above functions defined in the system / device of the embodiment of the present invention when it is executed by the processor 601. According to the embodiment of the present invention, the system, device, module, unit, etc. described above can be implemented by a computer program module.

[0176] In one embodiment, the computer program may rely on tangible storage media such as optical storage devices, magnetic storage devices, etc. In another embodiment, the computer program may also be transmitted and distributed in the form of signals on a network medium, and downloaded and installed through the communication part 609, and / or installed from a removable medium 611. The program code contained in the computer program may be transmitted using any appropriate network medium, including but not limited to: wireless, wired, etc., or any suitable combination of the above.

[0177] In such an embodiment, the computer program can be downloaded and installed from the network through the communication part 609, and / or installed from the removable medium 611. When the computer program is executed by the processor 601, the above functions defined in the system of the embodiment of the present invention are performed. According to the embodiment of the present invention, the system, device, means, module, unit, etc. described above can be implemented by a computer program module.

[0178] According to an embodiment of the present invention, the program code for executing the computer program provided by the embodiment of the present invention can be written in any combination of one or more programming languages. Specifically, these computing programs can be implemented using high-level process and / or object-oriented programming languages, and / or assembly / machine languages. Programming languages ​​include, but are not limited to, Java, C++, python, "C" language or similar programming languages. The program code can be executed entirely on the user computing device, partially on the user device, partially on the remote computing device, or entirely on the remote computing device or server. In the case of a remote computing device, the remote computing device can be connected to the user computing device through any type of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computing device (e.g., using an Internet service provider to connect through the Internet).

[0179] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present invention. In this regard, each box in the flow chart or block diagram can represent a module, a program segment, or a part of a code, and the above-mentioned module, program segment, or a part of a code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram or flow chart, and the combination of the boxes in the block diagram or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0180] It will be appreciated by those skilled in the art that the features described in the various embodiments of the present invention may be combined and / or combined in various ways, even if such combinations or combinations are not explicitly described in the present invention. In particular, without departing from the spirit and teachings of the present invention, the features described in the various embodiments of the present invention may be combined and / or combined in various ways. All of these combinations and / or combinations fall within the scope of the present invention.

[0181] The embodiments of the present invention are described above. However, these embodiments are only for the purpose of illustration, and are not intended to limit the scope of the present invention. Although each embodiment is described above, it does not mean that the measures in each embodiment cannot be used in combination advantageously. Without departing from the scope of the present invention, those skilled in the art may make various substitutions and modifications, which should all fall within the scope of the present invention.

Claims

1. A sample construction method for a control model of a hybrid vehicle, characterized in that: The method comprises: Under the condition that the driving speed sequence of the hybrid vehicle in the historical driving cycle remains unchanged, determining a target decision sequence corresponding to the minimum weighted value of the plurality of driving cost indicators, wherein the target decision sequence includes a target engine power corresponding to each driving speed; Determining a vehicle driving parameter set corresponding to each of the driving speeds according to the driving speed sequence and the plurality of target engine powers; A training sample for training a control model of the hybrid vehicle is constructed based on the driving speed, the target engine power corresponding to the driving speed, and the vehicle driving parameter set corresponding to the driving speed.

2. The method according to claim 1, characterized in that The method of determining a target decision sequence corresponding to a minimum weighted value of a plurality of driving cost indicators while keeping the driving speed sequence of the hybrid vehicle in the historical driving cycle unchanged comprises: Under the condition that the driving speed sequence of hybrid vehicles in the historical driving cycle remains unchanged, multiple weight coefficients are used as decision variables of the elite non-dominated genetic algorithm, and the weighted sum of multiple driving cost indicators is minimized as the optimization goal, and the driving cost indicator value sets and the optimal decision sequence corresponding to the multiple target weight coefficient sets are determined; determining a target driving cost index value set from the plurality of driving cost index value sets; The preferred decision sequence corresponding to the target driving cost index value set is determined as the target decision sequence.

3. The method according to claim 2, characterized in that The method comprises: taking multiple weight coefficients as decision variables of an elite non-dominated genetic algorithm while keeping the driving speed sequence of the hybrid vehicle in the historical driving cycle unchanged, minimizing the weighted sum of multiple driving cost indicators as the optimization target, and determining driving cost indicator value sets and optimal decision sequences corresponding to the multiple target weight coefficient sets respectively. Under the constraint conditions of the plurality of weight coefficients, for each individual in the p-th wheel population composed of the plurality of weight coefficients, while keeping the driving speed sequence unchanged, taking the value ranges of the plurality of vehicle operating parameters as constraints, minimizing the weighted sum of the plurality of driving cost indices in the historical driving cycle as an optimization target, optimizing the engine power of the hybrid vehicle, and obtaining a decision sequence corresponding to the driving speed sequence and a p-th driving cost index value set in the historical driving cycle, wherein p>0, and p is an integer, and the p-th driving cost index value set includes a plurality of p-th driving cost index values; Determining the pth fitness of each individual according to the pth travel cost index value set of each individual; Based on the p-th fitness of each individual, each individual is selected to obtain a p-th wheel population; Determine the (p+1)th round population according to the pth round sub-population, and repeat the above operation until a preset iteration round P is reached; Determine the Pth driving cost index value set corresponding to each individual in the Pth wheel subpopulation as a plurality of the driving cost index value sets; The P-th decision sequence corresponding to each individual in the P-th wheel sub-population is determined as a plurality of the preferred decision sequences.

4. The method according to claim 2, characterized in that: The step of determining a target driving cost index value set from the plurality of driving cost index value sets comprises: constructing a decision matrix using a plurality of the driving cost index value sets; Normalizing the decision matrix to obtain a normalized matrix; Determining an ideal optimal solution set and an ideal worst solution set based on the normalized matrix; Calculating the relative proximity of each of the plurality of driving cost index value sets according to the ideal optimal solution and the ideal worst solution; The target driving cost index value set is determined from a plurality of the driving cost index value sets according to the plurality of relative proximity degrees.

5. The method according to claim 4, characterized in that The method further comprises: Each driving cost index value in the normalized matrix is ​​weighted according to its index type to obtain a weighted normalized matrix, so as to determine an ideal optimal solution set and an ideal worst solution set based on the weighted normalized matrix.

6. The method according to claim 5, characterized in that The step of weighting each driving cost index value according to the index type of each driving cost index value in the normalized matrix to obtain a weighted normalized matrix includes: Constructing a judgment matrix according to the relative importance of the multiple indicator types; Calculate the arithmetic mean values ​​corresponding to each indicator type by using the judgment matrix; Each of the driving cost index values ​​is weighted using each of the arithmetic mean values ​​to obtain the weighted normalized matrix.

7. The method according to claim 1, characterized in that The training samples include a first training sample and a second training sample; the vehicle driving parameter set includes motor power, vehicle power and state of charge; The step of constructing a training sample for training a control model of the hybrid vehicle according to the driving speed, the target engine power corresponding to the driving speed, and the vehicle driving parameter set corresponding to the driving speed comprises: determining a driving mode of the hybrid vehicle based on the target engine power and the motor power; constructing a first training sample based on the driving mode, the vehicle power, the state of charge, and the driving speed corresponding to the vehicle driving parameter set, wherein the driving mode is a label of the first training sample; Based on the driving mode, the vehicle power, the state of charge, the driving speed corresponding to the vehicle driving parameter set, and the target engine power corresponding to the driving speed, a second training sample corresponding to the first training sample is constructed, wherein the target engine power is a label of the second training sample.

8. The method according to claim 3, characterized in that The respective value ranges of the plurality of vehicle operating parameters include an engine power interval, a state of charge interval and a motor power interval, and the engine power interval and the motor power interval are pre-set; The state of charge interval is determined by the following method: Determine a kth state of charge interval corresponding to the kth target engine power according to the (k+1)th state of charge and the engine power interval, wherein K≥2, K is an integer, k belongs to [1, K-1], and k is an integer; The maximum value of the first maximum value in the preset state of charge interval and the second maximum value in the kth state of charge interval, whichever has a smaller value, is used as the target maximum value; The first minimum value in the preset state of charge interval and the second minimum value in the kth state of charge interval, the minimum value with a larger value is used as the target minimum value; Repeat the above steps until k=1; The state of charge range is determined based on the target minimum value and the target maximum value.

9. A control method for a hybrid vehicle, characterized in that: The method comprises: determining vehicle power based on the acquired driving speed and a powertrain model of the hybrid vehicle; Input the driving speed, the vehicle power and the collected state of charge into a control model, and output a target engine power of the hybrid vehicle, wherein the control model is trained based on training samples corresponding to a plurality of driving speeds, and the training samples are constructed according to the sample construction method of the control model of the hybrid vehicle according to claims 1 to 8; determining a target motor power of a hybrid vehicle according to the target engine power, the vehicle power and the powertrain model; The hybrid vehicle is controlled to operate based on the target motor power and the target motor power.

10. The method according to claim 9, characterized in that The control model includes: a first control module and a second control module, wherein the first control module is used to output the driving mode of the hybrid vehicle according to the driving speed, the state of charge, and the vehicle power; the second control module is used to output the target engine power according to the driving mode, the driving speed, the state of charge, and the vehicle power.

Images