Dual-motor coupling driving method, system and equipment for electric vehicle and medium

Abstract

The invention relates to the technical field of new energy vehicles, and discloses an electric vehicle dual-motor coupling driving method, system and device and a medium. The method comprises the steps that the vehicle state of the electric vehicle is obtained in real time; inputting the vehicle state of the electric vehicle at each moment into the trained deep Q network, dynamically obtaining the driving mode of the electric vehicle at each moment, and if the driving mode of the electric vehicle is dual-motor coupling driving, determining the driving mode of the electric vehicle according to the vehicle state of the electric vehicle at each moment through the trained deep deterministic strategy gradient network. The torque distribution proportion of the two motors of the electric vehicle at each moment is dynamically output; and according to the driving capacity required by the vehicle state of the electric vehicle at each moment and the torque distribution proportion of the two motors of the electric vehicle at each moment, the torque is dynamically distributed to the two motors of the electric vehicle, so that the two motors are driven by the distributed torque coupling, and the dual-motor torque distribution under the complex dynamic working condition is realized.

Description

Technical Field

[0001] The present invention relates to the technical field of new energy vehicles, and in particular to a dual-motor coupling driving method, system, equipment and medium for an electric vehicle. Background Art

[0002] With the rapid development of the new energy vehicle industry, electric drive systems have gradually become one of the key technologies that determine vehicle performance. The popularity of electric vehicles is not only due to their zero-emission and environmentally friendly characteristics, but also to their significant advantages in energy efficiency, power output and driving comfort. Against this background, the traditional single-motor drive system has gradually exposed certain limitations, especially in complex driving environments, and it is difficult to meet the increasingly diverse needs in terms of power output, energy efficiency and vehicle stability. This is mainly manifested in: the energy efficiency management of the single-motor system is relatively simple, and it is impossible to flexibly adjust the output characteristics of the vehicle under different working conditions, resulting in energy waste, insufficient endurance, and insufficient vehicle response.

[0003] In contrast, the dual-motor drive system has gradually become an important development direction for future electric vehicle drive systems due to its multiple advantages in power output, energy efficiency management, and dynamic performance control. The dual-motor system can not only improve the power redundancy and stability of the vehicle, but also flexibly switch working modes under different working conditions to achieve better energy efficiency management. This makes the dual-motor drive technology not only limited to the application of high-performance models, but also likely to be widely popularized in various levels of new energy vehicles in the future.

[0004] In the dual-motor drive system, the two motors work together on the vehicle's transmission system, and by reasonably distributing the motor's output torque, they work together to achieve more efficient power output, energy recovery, and vehicle power balance regulation. The system can intelligently adjust the working state of each motor according to the vehicle's operating conditions, such as vehicle speed, load, road conditions, etc. Through the coordinated work of the dual motors, the vehicle can save energy by relying on a single motor when starting at low speed, and the two motors work together at high speed or high load to ensure the vehicle's power output. At the same time, when the vehicle brakes or goes downhill, the dual motors can convert the vehicle's kinetic energy into electrical energy through the energy recovery mode and feed it back to the battery system, thereby effectively extending the vehicle's cruising range.

[0005] However, the core challenge of the dual-motor drive system lies in how to achieve coordinated control between the motors, that is, how to intelligently adjust the output of the two motors according to the real-time working conditions of the vehicle, so that it can achieve the optimization of energy efficiency management while ensuring the power. At present, most of the existing dual-motor coupling control technologies adopt strategies based on feedback control. This method usually relies on pre-set control rules or working conditions to distribute torque to the motor through simple thresholds or conditional judgments. Although it can provide a certain response speed and system stability in specific scenarios, its core problem lies in the lack of sufficient intelligence and adaptability. This means that when the vehicle is running under complex dynamic conditions, it is difficult for the system to respond flexibly. For example, in scenarios of uphill and downhill with different slopes, continuous speed change or variable load, it is impossible to dynamically adjust the motor output according to the actual situation, which can easily lead to problems such as unreasonable energy distribution, insufficient power redundancy or excessive consumption of electricity. Summary of the invention

[0006] The purpose of the present invention is to provide a dual-motor coupled drive method, system, device and medium for an electric vehicle, which can realize dual-motor torque distribution under complex dynamic conditions.

[0007] In order to solve the above technical problems, an embodiment of the present invention provides a dual-motor coupled driving method for an electric vehicle, comprising the following steps:

[0008] Obtain the vehicle status of electric vehicles in real time;

[0009] The vehicle state of the electric vehicle at each moment is respectively input into the trained deep Q network DQN to dynamically obtain the driving mode of the electric vehicle at each moment, wherein the driving mode is single motor driving or dual motor coupling driving;

[0010] The DQN network is trained by estimating the first Q value of different driving modes in each vehicle state based on a preset first Q value function, and training with the goal of maximizing the first Q value of the driving mode adopted in each vehicle state; the first Q value is used to indicate that the adopted driving mode has the minimum energy consumption of the electric vehicle under the premise of satisfying the driving capacity required by the vehicle state;

[0011] If the driving mode of the electric vehicle is dual-motor coupling drive, the trained deep deterministic policy gradient DDPG network dynamically outputs the torque distribution ratio of the two motors of the electric vehicle at each moment according to the vehicle state of the electric vehicle at each moment;

[0012] The DDPG network is trained by estimating the second Q value of different torque distribution ratios in each vehicle state based on a preset second Q value function, and training with the second Q value of the torque distribution ratio in each vehicle state as the maximum as the goal; the second Q value is used to indicate that the energy consumption of the two motors corresponding to the adopted torque distribution ratio is minimized;

[0013] According to the driving capacity required to meet the vehicle state of the electric vehicle at each moment and the torque distribution ratio of the two motors of the electric vehicle at each moment, torque is dynamically distributed to the two motors of the electric vehicle so that the two motors are driven with the distributed torque coupling.

[0014] Optionally, the DDPG network includes an Actor network and a Critic network;

[0015] The Actor network is used to generate the torque distribution ratio corresponding to the two motors according to the input vehicle state;

[0016] The critic network is used to estimate a second Q value of the torque distribution ratio generated under the input vehicle state based on the second Q value function;

[0017] The DDPG network updates the network parameters of the Actor network with the goal of maximizing the second Q value estimated by the Critic network.

[0018] Optionally, the network parameters of the Critic network are updated based on a temporal difference error method by the difference between the second Q value of the current vehicle state and the second Q value of the next vehicle state.

[0019] Optionally, the DDPG network adopts a replay buffer mechanism, which stores experience consisting of several corresponding current vehicle states, torque distribution ratios, immediate rewards adopted by the second Q-value function, and the next vehicle state through the replay buffer; the network parameters of the Actor network and the Critic network are updated by randomly extracting experience from the replay buffer.

[0020] Optionally, the vehicle state includes: speed, acceleration, state of charge and torque of two motors of the electric vehicle.

[0021] Optionally, the torque distribution ratio of the two motors output by the DDPG network has a value range of [0, 1], and the motors of the electric vehicle include a first motor and a second motor;

[0022] The method dynamically distributes torque to the two motors of the electric vehicle by satisfying the driving capacity required by the vehicle state of the electric vehicle at each moment and the torque distribution ratio of the two motors of the electric vehicle at each moment, including:

[0023] According to the driving capacity required to meet the vehicle state of the electric vehicle at each moment, the vehicle torque required to meet the vehicle state is determined;

[0024] When the torque distribution ratio of the two motors is 1, all the vehicle torque required to meet the vehicle state is distributed to the first motor;

[0025] When the torque distribution ratio of the two motors is 0, all the vehicle torque required to meet the vehicle state is distributed to the second motor;

[0026] When the torque distribution ratio of the two motors is between [0, 1], torque is distributed to the first motor and the second motor according to the vehicle torque and torque distribution ratio required to meet the vehicle state.

[0027] Optionally, the method further includes:

[0028] If the current driving mode of the electric vehicle is driven by any one motor, all the vehicle torque required to meet the vehicle state will be allocated to any one motor, so that any one motor can be driven alone.

[0029] An embodiment of the present invention further provides an electric vehicle dual-motor coupled drive system, comprising:

[0030] A status acquisition module is used to acquire the vehicle status of the electric vehicle in real time;

[0031] A mode selection module is used to input the vehicle state of the electric vehicle at each moment into the trained deep Q network DQN, and dynamically obtain the driving mode of the electric vehicle at each moment, wherein the driving mode is a single motor drive or a dual motor coupling drive;

[0032] The DQN network is trained by estimating the first Q value of different driving modes in each vehicle state based on a preset first Q value function, and training with the goal of maximizing the first Q value of the driving mode adopted in each vehicle state; the first Q value is used to indicate that the adopted driving mode has the minimum energy consumption of the electric vehicle under the premise of satisfying the driving capacity required by the vehicle state;

[0033] The torque distribution module is used to dynamically output the torque distribution ratio of the two motors of the electric vehicle at each moment according to the vehicle state of the electric vehicle at each moment through the trained deep deterministic policy gradient DDPG network when the driving mode of the electric vehicle is dual-motor coupling drive;

[0034] The DDPG network is trained by estimating the second Q value of different torque distribution ratios in each vehicle state based on a preset second Q value function, and training with the second Q value of the torque distribution ratio in each vehicle state as the maximum as the goal; the second Q value is used to indicate that the energy consumption of the two motors corresponding to the adopted torque distribution ratio is minimized;

[0035] The motor drive module is used to dynamically distribute torque to the two motors of the electric vehicle according to the driving capacity required to meet the vehicle state of the electric vehicle at each moment and the torque distribution ratio of the two motors of the electric vehicle at each moment, so that the two motors are driven with the distributed torque coupling.

[0036] An embodiment of the present invention also provides a computer device, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the above-mentioned electric vehicle dual-motor coupled drive method.

[0037] An embodiment of the present invention further provides a computer-readable storage medium storing a computer program, wherein the computer program, when executed by a processor, implements the above-mentioned electric vehicle dual-motor coupled driving method.

[0038] The electric vehicle dual-motor coupling driving method provided by the present invention has at least the following beneficial effects:

[0039] Firstly, the DQN network is used to select the driving mode of the electric vehicle according to the vehicle state of the electric vehicle to determine whether any one motor is driven alone or the dual motors are coupled. If it is dual motor coupled, the DDPG network is used to output the torque distribution ratio of the two motors of the electric vehicle in the dual motor coupling driving mode to control the torque distribution of the two motors.

[0040] This method uses real-time feedback of the vehicle state and hybrid control of the DQN network and DDPG network to dynamically select the driving mode and torque distribution ratio of the electric vehicle, thereby realizing the intelligent driving mode selection and torque distribution of the electric vehicle dual-motor drive system. In addition, by coupling the dual motors of the electric vehicle through this method, the energy consumed by the two motors of the electric vehicle can be optimized under the premise of meeting the driving capacity required by the vehicle state, thereby realizing the dynamic adjustment of the output of the dual motors according to the actual vehicle conditions (including dynamic working conditions or multiple working conditions), reasonably distributing the motor energy, and avoiding insufficient vehicle power redundancy or excessive consumption of electric energy. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] One or more embodiments are exemplarily described by the pictures in the corresponding drawings, and these exemplary descriptions do not constitute limitations on the embodiments.

[0042] Figure 1 is a flow chart of a dual-motor coupling driving method for an electric vehicle provided according to an embodiment of the present invention;

[0043] Figure 2 is a network architecture diagram provided according to an embodiment of the present invention;

[0044] Figure 3 A network training flow chart is provided according to an embodiment of the present invention. DETAILED DESCRIPTION

[0045] In order to make the purpose, technical scheme and advantages of the embodiments of the present invention clearer, the embodiments of the present invention will be described in detail below in conjunction with the accompanying drawings. However, it will be appreciated by those skilled in the art that in the embodiments of the present invention, many technical details are proposed in order to enable the reader to better understand the present invention. However, even without these technical details and various changes and modifications based on the following embodiments, the technical scheme claimed in the present invention can be implemented. The division of the following embodiments is for the convenience of description and should not constitute any limitation on the specific implementation of the present invention. The various embodiments can be combined and referenced with each other without contradiction.

[0046] At present, the motor coordination control method used in the dual-motor drive system generally adopts a rule-based control method, which means that a set of control rules and thresholds are preset to directly control the torque distribution of the dual motors according to the vehicle's operating status, environmental conditions and other information. These rules are usually based on engineering experience, experimental data or simulation analysis, have a clear logical structure, and can quickly implement vehicle control decisions. However, this type of method has the following disadvantages:

[0047] 1. Poor flexibility: Since the preset rules are relatively fixed, it is difficult to adapt to complex or changing working conditions. The system cannot adaptively adjust the control strategy according to the actual driving scenario, making it difficult to achieve optimal energy efficiency management.

[0048] 2. Limited optimization capability: The control rules rely on empirical settings and lack global optimization capabilities, which may lead to suboptimal control strategies, especially in nonlinear and multi-operating conditions.

[0049] 3. Poor scalability: When the driving environment or working conditions change, the system needs to readjust or expand the rule set, making it difficult to automatically learn new driving modes or working conditions.

[0050] Although the rule-based control method is simple and easy to implement, it is difficult to provide the optimal control strategy under complex dynamic conditions due to the fixed nature of the rules and the lack of adaptive and global optimization capabilities. It is usually suitable for relatively simple or relatively fixed conditions. For more complex and dynamically changing conditions, more advanced control strategies may be required.

[0051] In addition, existing dual-motor control technologies are usually based on linear or simple logic control methods, and the limitations of their control strategies are particularly evident under complex nonlinear conditions. For example, the energy efficiency curve of a vehicle under different loads or speeds does not change linearly, but has highly complex and time-varying characteristics. Traditional control algorithms often cannot fully capture these changing patterns, resulting in poor performance in energy efficiency optimization under multiple conditions.

[0052] Therefore, the current dual-motor coupling control technology urgently needs a more intelligent solution, which not only needs to find a balance between power output and energy recovery, but also needs to have the ability to adapt to complex working conditions. The ideal dual-motor control system should be able to monitor and analyze the dynamic working conditions of the vehicle in real time, and automatically adjust the output characteristics of the dual motors based on these working condition information, so as to achieve more efficient energy management and better driving performance.

[0053] In recent years, with the development of artificial intelligence technology, intelligent control technology based on machine learning, especially reinforcement learning, has shown great potential in the control of complex nonlinear systems. Through reinforcement learning, the system can autonomously learn the best control strategy without supervision, and no longer rely on preset control rules. The reinforcement learning algorithm can continuously optimize the coupling control strategy of the motor according to the feedback data under different working conditions, so as to achieve dynamic optimization of energy efficiency management while ensuring power and stability. This provides a new technical path for dual-motor coupling control, which can significantly improve the intelligence level and adaptability of the system. This intelligent dual-motor coupling control system can not only improve the overall energy efficiency of the vehicle, but also optimize the power response of the vehicle under complex driving conditions by dynamically adjusting the output torque and speed, thereby providing the driver with a smoother and more comfortable driving experience. In the future, intelligent dual-motor coupling control technology is expected to become one of the core technologies in the drive system of new energy vehicles, promoting electric vehicles to achieve a new leap in performance, endurance and user experience.

[0054] An embodiment of the present invention relates to a dual-motor coupled driving method for an electric vehicle. The implementation details of the dual-motor coupled driving method for an electric vehicle of this embodiment are specifically described below. The following content is only the implementation details provided for easy understanding and is not necessary for implementing this solution.

[0055] The specific process of the electric vehicle dual-motor coupling driving method of this embodiment is as follows: Figure 1 As shown, including:

[0056] Step 101, obtaining the vehicle status of the electric vehicle in real time.

[0057] Among them, the vehicle status includes: the speed, acceleration, state of charge SOC and torque of the two motors of the electric vehicle.

[0058] Step 102, the vehicle state of the electric vehicle at each moment is respectively input into the trained deep Q network DQN, and the driving mode of the electric vehicle at each moment is dynamically obtained, and the driving mode is any one motor driving alone or dual motor coupling driving; wherein the DQN network is trained in the following manner: based on a preset first Q value function, the first Q value of different driving modes adopted in each vehicle state is estimated, and the training is performed with the goal of maximizing the first Q value of the driving mode adopted in each vehicle state; the first Q value is used to indicate that the adopted driving mode minimizes the energy consumption of the electric vehicle under the premise of satisfying the driving capability required by the vehicle state.

[0059] Specifically, the core of the Deep Q-Network (DQN) network is to use a neural network to approximate the Q-value table in traditional Q learning. The Q-value function Q(s,a) represents the expected return that can be obtained by taking action a in state s. By training the neural network, the DQN network can estimate the Q value of each state-action pair in the high-dimensional state space, thereby guiding the agent to choose the best action. The basic idea is to use the target network (TargetNetwork) and the estimation network (Q-Network), and break the correlation between data through experience replay, thereby improving the stability and efficiency of training.

[0060] The estimation network is used to calculate the Q value in the current state. In each state s, the network outputs the Q value of each possible action in that state. By comparing these Q values, the action corresponding to the maximum Q value is selected as the best action for the agent in that state.

[0061] The structure of the target network is exactly the same as that of the estimation network, except that its parameters are updated more slowly. The target network is used to calculate the target Q value, thereby guiding the learning of the estimation network. The existence of the target network is to improve the stability of training and prevent the network from updating too quickly, which will cause the training process to diverge.

[0062] The network structure of the estimated network and the target network:

[0063] Input layer: The input layer accepts the feature vector of the current state s. For the control task of the electric vehicle dual-motor system, the input layer receives state variables such as vehicle speed, acceleration, SOC, motor torque, etc.

[0064] Hidden layer: The hidden layer is usually composed of several fully connected layers and uses the ReLU activation function. The hidden layer is responsible for extracting the deep features of the input state and helping the network learn the complex relationships between different states.

[0065] Output layer: The number of neurons in the output layer is equal to the size of the action space. In the dual-motor drive mode control, each neuron in the output layer corresponds to a possible drive mode, and the output value is the Q value of the action.

[0066] The Q-value function Q(s,a) is a value function of a state-action pair. It represents the expected return (total reward) that the agent can obtain after selecting action a in state s. The formula is as follows:

[0067]

[0068] In the formula, r t is the reward obtained by the agent after performing action a starting from state s; γ is the discount factor, which measures the degree of decay of future rewards (usually 0<γ<1), so that the agent pays more attention to recent returns; E represents the expected value of all possible paths.

[0069] The goal of the agent is to choose action a that maximizes Q(s,a) * , thereby obtaining the highest cumulative return in the long run:

[0070]

[0071] Therefore, the core task of the DQN network is to approximate Q(s,a) through a neural network and find the optimal action a that maximizes the Q value in each state. * .

[0072] The DQN network updates the Q value through the temporal difference (TD) method. The basic idea of ​​TD error is to use the difference between the current Q value and the target Q value as a learning signal to update the parameters of the neural network so that it gradually approaches the true Q value. The target Q value (TD target) is expressed by the following formula:

[0073]

[0074] Where r is the immediate reward obtained after executing action a in the current state s; s′ is the next state after executing the action; θ - It is the parameter of the target network, which is used to calculate the target Q value; It is the maximum Q value that can be obtained among all possible actions in state s′, representing the maximum expected reward that can be obtained by taking the best action in the next state.

[0075] By calculating the TD error, the agent can measure the gap between the current network's Q value and the target Q value. The formula for the TD error is:

[0076] d=(yQ(s,a;θ)) 2

[0077] Where y is the target value Q; Q(s,a;θ) is the estimated Q value of action a when the network is in the current state s.

[0078] Therefore, the DQN network approximates the Q value function of each state-action pair through a neural network. The neural network learns the relationship between the state s and the expected return that can be obtained from the corresponding action a, and continuously updates the network weights so that the Q value function can more accurately reflect the long-term return. Maximizing the Q value selects the optimal action. In each state s, the DQN network calculates the Q value Q(s,a) of all possible actions and selects the action a that maximizes the Q value. * This process ensures that the agent can choose the best action in each state to obtain the highest long-term reward. At the same time, the DQN network introduces a target network, which slowly updates the parameters of the target network to avoid instability caused by Q value fluctuations during network training. The target network ensures that during the learning process, the estimated network can gradually approach the real Q value, ensuring the rationality and reliability of the output action.

[0079] Step 103, if the driving mode of the electric vehicle is dual-motor coupled drive, the torque distribution ratio of the two motors of the electric vehicle at each moment is dynamically output through the trained deep deterministic policy gradient DDPG network according to the vehicle state of the electric vehicle at each moment; wherein the DDPG network is trained in the following manner: based on a preset second Q value function, a second Q value of different torque distribution ratios is estimated under each vehicle state, and training is performed with the goal of maximizing the second Q value of the torque distribution ratio adopted under each vehicle state; the second Q value is used to indicate that the energy consumption of the two motors corresponding to the adopted torque distribution ratio is minimized.

[0080] Specifically, the Deterministic Policy Gradient (DDPG) network combines the ideas of policy gradient and Q-learning, and makes decisions through two neural networks:

[0081] Actor network, used to generate continuous actions a=π(s|θ π ), indicating that in state s, an action a is selected, θ π is the parameter of the policy network. This network directly outputs a continuous action, which is suitable for tasks that require precise control, such as torque ratio distribution.

[0082] Critic network, used to evaluate the quality of the action selected by the Actor network, that is, to estimate the state-action value function Q(s,a|θ Q ), θ Q is the parameter of the Critic network. The Critic network gives the Q value of taking action a in the current state s, which represents the expected return that can be obtained from the action.

[0083] In order to maintain the stability of the learning process, DDPG also uses two target networks:

[0084] Target Actor Network: Same parameters as the Actor Network, but updated more slowly to ensure a smooth transition of strategies. Target Critic Network: Same parameters as the Critic Network, also updated more slowly to ensure the stability of Q-value estimation.

[0085] The goal of the DDPG network is to generate the optimal action a=π(s|θ π ), and evaluate the Q value of the action through the Critic network. This process can be expressed by the following formula:

[0086] Actor network generates actions:

[0087] a=π(s|θ π )

[0088] In each state s, the Actor network generates a continuous action a according to the input state. This action can be expressed as selecting a certain motor torque ratio or distribution ratio in the current state.

[0089] Critic network evaluates Q value:

[0090] Q(s,a|θ Q )

[0091] The Critic network receives the state s and the action a generated by the Actor network, and calculates the Q value of taking the action in the state. The Q value represents the cumulative reward that can be obtained in the future, so the higher the Q value, the better the action.

[0092] Policy update: The goal is to maximize the Q value evaluated by the Critic network, so the update goal of the Actor network is:

[0093]

[0094] By adjusting the parameters θ of the Actor network π , so that the actions generated by the Actor network can maximize the Q value given by the Critic network.

[0095] The update of the Critic network is similar to Q-learning, which adjusts the parameters θ of the Critic network by minimizing the TD (Temporal Difference) error. Q The target Q value is calculated using the target network and is expressed as:

[0096] y=r+gQ′(s′,π′(s′|θ π′ )|θ Q′ )

[0097] Where r is the immediate reward, which indicates the feedback obtained by the agent after performing the action; γ is the discount factor, which indicates the importance of future rewards; Q′(s′,π′(s′|θ π′ )|θ Q′ ) is the Q value corresponding to the next state s′ calculated by the target network; θ π′ and θ Q′ These are the parameters of the target Actor and target Critic networks.

[0098] Then, the parameters of the Critic network are updated by minimizing the TD error:

[0099] L(θ Q )=E(yQ(s,a|θ Q )) 2

[0100] The Critic network learns the Q value of the current state and action, enabling it to evaluate the quality of different actions, thereby helping the Actor network generate the optimal action.

[0101] In order to improve the stability and efficiency of learning, DDPG adopts the following two mechanisms:

[0102] Experience replay: A replay buffer is used to store the agent’s experience (s, a, r, s′) in the environment. During each update, a small batch of experience is randomly sampled to update the Actor and Critic networks, breaking the temporal correlation between samples and improving the stability of training.

[0103] Target network: DDPG uses target actor and target critic networks. These networks have the same structure as the main network, but the parameters are updated more slowly (using soft update). By controlling the update rate of the target network, the drastic fluctuation of Q value estimation can be avoided and the stability of training can be improved. The update formula of the target network is:

[0104] tθ+(1-t)θ′→θ′

[0105] Where t is a very small value (such as 0.001), which represents the update rate of the target network parameters.

[0106] Therefore, the Actor network of the DDPG network will generate a continuous action a according to the vehicle state of the electric vehicle (such as speed, SOC, motor torque, etc.), indicating the torque distribution ratio of the two motors. This ratio will be directly used to control the output of the motor. For example, the torque distribution ratio a∈0,1 output by the Actor network can represent the power distribution relationship between motor 1 and motor 2. The Critic network is responsible for evaluating whether the torque distribution strategy output by the Actor network is reasonable and whether the strategy can achieve optimal energy management and power output under the current state. The Critic network helps the Actor network generate a better torque distribution strategy by evaluating the Q value. The advantage of the DDPG network is that it can handle continuous action space, making the torque distribution of the motor more refined and flexible. At the same time, combined with the Critic network for Q value evaluation, the adjustment direction of the strategy is clearer, and ultimately the optimal energy management and torque distribution are achieved.

[0107] Step 103 , dynamically distributes torque to the two motors of the electric vehicle according to the driving capacity required to meet the vehicle state of the electric vehicle at each moment and the torque distribution ratio of the two motors of the electric vehicle at each moment, so that the two motors are driven by coupling with the distributed torque.

[0108] In a specific implementation, the torque distribution ratio of the two motors output by the DDPG network has a value range of [0, 1], and the motors of the electric vehicle include a first motor and a second motor. Therefore, in the process of dynamically distributing torque to the two motors of the electric vehicle by satisfying the vehicle state required by the vehicle at each moment and the torque distribution ratio of the two motors of the electric vehicle at each moment, when the torque distribution ratio of the two motors is 1, the vehicle torque required to satisfy the vehicle state is fully distributed to the first motor; when the torque distribution ratio of the two motors is 0, the vehicle torque required to satisfy the vehicle state is fully distributed to the second motor; when the torque distribution ratio of the two motors is between [0, 1], torque is distributed to the first motor and the second motor according to the vehicle torque required to satisfy the vehicle state and the torque distribution ratio.

[0109] In one example, if the driving mode of the electric vehicle is driven by any one motor, the entire vehicle torque required to meet the vehicle state is allocated to any one motor, so that any one motor can be driven alone.

[0110] In this embodiment, the DQN network is first used to select the driving mode of the electric vehicle according to the vehicle state of the electric vehicle, and it is determined whether any motor is driven alone or the dual motors are coupled. If it is a dual motor coupled drive, the DDPG network is used to output the torque distribution ratio of the two motors of the electric vehicle in the dual motor coupled drive mode to control the torque distribution of the two motors. The method uses the DQN network and the DDPG network hybrid control to dynamically select the driving mode and torque distribution ratio of the electric vehicle through real-time feedback of the vehicle state, thereby realizing the intelligent driving mode selection and torque distribution of the dual motor drive system of the electric vehicle. In addition, by using this method to couple the dual motors of the electric vehicle, the energy consumed by the two motors can be optimized under the premise of meeting the driving capacity required by the vehicle state, thereby realizing the dynamic adjustment of the output of the dual motors according to the actual situation of the vehicle (including dynamic working conditions or multiple working conditions), reasonably distributing the motor energy, and avoiding insufficient power redundancy or excessive consumption of electric energy in the vehicle.

[0111] In one embodiment, the network architecture of the DQN network and the DDPG network in the present invention is as follows: Figure 2 As shown in Figure 3 As shown, the training process includes:

[0112] 1. System initialization:

[0113] (1) After the vehicle is started, the system initializes relevant parameters, including vehicle status (speed, acceleration, SOC, torque of the two motors, etc.).

[0114] (2) These states are input into the DQN network and DDPG network for subsequent mode selection and torque allocation.

[0115] 2. Status input:

[0116] (1) Input parameters include: speed (v), acceleration (a), SOC, T_motor1 (torque of motor 1), and T_motor2 (torque of motor 2).

[0117] (2) The state space state is input into the DQN and DDPG networks as the basis for subsequent action selection. The state definition is as follows:

[0118] state={v t ,T_motor it ,N_motor it |i={1,2},t={1,2,L n|n=steps max}}

[0119] 3. DQN network mode selection:

[0120] (1) The working mode is a discrete set of three options: 0 means that only motor 1 works, 1 means that only motor 2 works, and 2 means that both motors work together. The DQN action space is defined as follows:

[0121] action DQN ={0,1,2}

[0122] (2) The DQN network selects the vehicle’s driving mode mode based on the input state state.

[0123] mode=action DQN

[0124] (3) The DQN network selects an optimal driving mode based on the current state at each time step.

[0125] 4. DDPG network torque distribution:

[0126] (1) If the DQN selects the dual-motor mode (mode 2), the DDPG network is started to distribute the torque ratio.

[0127] action DDPG ∈0,1

[0128] (2) The DDPG network outputs a continuous distribution ratio value distribution, which is used to control the torque distribution ratio between motor 1 and motor 2.

[0129] distribution=action DDPG

[0130] (3) According to the torque distribution ratio, the torque borne by motor 1 is calculated as T_motor 1t , the torque borne by motor 2 is T_motor 2t .

[0131] T_motor 1t =T_req*distribution

[0132] T_motor 2t =T_req*(1-distribution)

[0133] 5. Environmental update and reward calculation:

[0134] (1) The system applies the decision results of DQN and DDPG to the model vehicle and performs vehicle status updates, including actual speed, SOC, torque, etc.

[0135] (2) Calculate the reward value of the current time step based on the updated state t, the reward value is used to guide network learning to optimize the energy management and driving efficiency of the motor.

[0136] reward t = -(a*|Dsoc|+b*|Dv|)

[0137]

[0138] 6. Experience storage:

[0139] (1) Store the current time step’s state, action, reward, next state, and whether the game is terminated in the experience pool (Replay Buffer).

[0140] (2) The data in the experience pool will be used for subsequent network updates to achieve strategy optimization.

[0141] 7. DQN network and DDPG network update:

[0142] (1) By sampling a batch of data (Batch) from the experience pool, the DQN network updates the Q value network and optimizes the working mode selection strategy.

[0143] (2) At the same time, the DDPG network uses the data in the experience pool to update the Actor and Critic networks to optimize the torque distribution strategy.

[0144] (3) The updates of the two networks are performed in parallel, ensuring that the system can gradually improve the performance of the strategy in each round of training.

[0145] 8. Termination condition check:

[0146] (1) After each execution of the control strategy, the system will check whether the current number of steps reaches the maximum step length requirement (equivalent to the end of a working cycle).

[0147] (2) If the maximum step length requirement is reached, a working cycle ends, the system stores the accumulated experience, and updates the network model.

[0148] (3) After the training rounds reach the maximum set rounds, the system will stop training and save the final model.

[0149] Based on the model obtained through this training process, driving mode selection and torque ratio distribution can be performed:

[0150] First, the system collects the current vehicle status information, including the vehicle speed (v), remaining battery power (soc), motor torque (T_motor), speed (N_motor), etc. This information is used as the input of the DQN and DDPG networks. The DQN network outputs a discrete action (mode) based on the current state, indicating the current working mode of the vehicle. There are three working modes: mode 0 (motor 1 drives alone), mode 1 (motor 2 drives alone) and mode 2 (dual motor coupling drive). In mode 0 and mode 1, the system only controls the torque output of one motor (the whole vehicle's required torque is all borne by the single motor), and there is no need to activate the DDPG network.

[0151] When the DQN network selects mode 2, the DDPG network will be activated and responsible for calculating the torque distribution ratio of the dual motors. The DDPG network outputs a continuous value distribution, whose value range is [0,1], which represents the torque distribution ratio of the two motors. Specifically, when distribution = 1, all torque is distributed to motor 1; when distribution = 0, all torque is distributed to motor 2; when distribution is an intermediate value, the two motors distribute torque proportionally (motor 1 is allocated to the part of the product of the total required torque and the proportional coefficient, and motor 2 is allocated to the part of the product of the total required torque and one minus the proportional coefficient). After that, the two motors output torque according to the torque they are responsible for and transmit it to the vehicle dynamics system to drive the vehicle forward. The specific distribution method is as follows:

[0152] Motor 1 bears torque: T_motor 1t =T_req*distribution

[0153] Motor 2 bears torque: T_motor 2t =T_req*(1-distribution)

[0154] Through this intelligent torque distribution strategy, the system can achieve optimal energy distribution under different working conditions, ensuring vehicle power output while maximizing energy efficiency.

[0155] In addition, at each time step, the system updates the vehicle state and calculates the reward value of the current step based on the results of the mode selection and torque distribution. Through the experience replay mechanism, the parameters of the DQN and DDPG networks are continuously updated to adapt to complex working conditions and gradually learn the optimal control strategy.

[0156] The electric vehicle dual-motor coupling driving method of the present invention has the following beneficial effects:

[0157] 1. Improve the ability to adapt to complex working conditions

[0158] (1) Introducing a more intelligent control method, combined with technology based on machine learning or deep reinforcement learning, automatically adjusts the torque distribution and working mode of the dual motors by learning the dynamic behavior of the vehicle and changes in the external environment in real time, thus solving the problem of poor adaptability of rule-based methods to different working conditions.

[0159] (2) Through real-time data-driven learning mechanisms, the system’s responsiveness and adaptive adjustment capabilities in complex and changing driving environments are improved.

[0160] 2. Realize the global optimal control strategy

[0161] (1) To solve the problem that rule-based control methods cannot achieve global optimization, my solution will combine adaptive optimization methods such as reinforcement learning to dynamically analyze driving conditions, energy status, etc., optimize multiple objectives (such as power, energy efficiency, comfort, etc.), and achieve global optimization among multiple objectives.

[0162] (2) In an environment with high system complexity and multiple variables, it is able to balance various objectives and seek the overall optimal control strategy instead of being limited to preset rules.

[0163] 3. Reduce computational complexity and improve real-time response capabilities

[0164] (1) To address the problems of high computational complexity and poor real-time performance of optimization algorithms, my solution will combine lightweight deep reinforcement learning strategies on the basis of traditional optimization algorithms, such as using neural network-based control strategies, to avoid tedious iterative calculations and improve the response speed and computational efficiency of the control system.

[0165] (2) By introducing technologies such as approximate dynamic programming and online learning algorithms, the high dependence on computing resources in real-time control can be reduced, and an effective balance between real-time performance and control accuracy can be achieved.

[0166] 4. Improve robustness to changes in the external environment

[0167] (1) The problem that the optimization algorithm is highly dependent on the system model and lacks the ability to respond to environmental changes will be solved by introducing data-driven deep learning technology. My solution will use real-time data to build a dynamic model, and combine it with external sensor information to dynamically update the model parameters, so that the system has stronger robustness when facing external environmental changes.

[0168] (2) The system can automatically adjust under complex conditions such as road slope, wind resistance changes, and vehicle speed fluctuations to avoid control failures caused by inaccurate models.

[0169] 5. Realize intelligent torque distribution and energy management

[0170] In view of the limitations of torque distribution and energy management strategies in existing solutions, my technical solution will learn the optimal dual-motor torque distribution strategy through a deep reinforcement learning method. Through intelligent control, efficient cooperation of motors in different driving modes can be achieved, energy recovery and battery utilization efficiency can be maximized, and overall energy efficiency can be improved.

[0171] 6. Provide scalability and system upgrade capabilities

[0172] It is not only applicable to the current dual-motor coupling system, but also has strong scalability and can be applied to more types of drive systems in the future (such as three-motor or hybrid systems). Through modular design, it can be flexibly expanded to different types of electric vehicle architectures and reserve space for subsequent system upgrades.

[0173] The step division of the various methods above is only for clear description. When implemented, they can be combined into one step or some steps can be split and decomposed into multiple steps. As long as they include the same logical relationship, they are all within the protection scope of the present invention. Adding insignificant modifications or introducing insignificant designs to the algorithm or process without changing the core design of the algorithm and process are all within the protection scope of the invention.

[0174] Another embodiment of the present invention relates to an electric vehicle dual-motor coupled drive system. The implementation details of the electric vehicle dual-motor coupled drive system of this embodiment are specifically described below. The following content is only for the convenience of understanding the implementation details, and is not necessary for the implementation of this solution. The electric vehicle dual-motor coupled drive system of this embodiment includes:

[0175] A status acquisition module is used to acquire the vehicle status of the electric vehicle in real time;

[0176] The mode selection module is used to input the vehicle state of the electric vehicle at each moment into the trained deep Q network DQN, and dynamically obtain the driving mode of the electric vehicle at each moment. The driving mode is either a single motor drive or a dual motor coupling drive.

[0177] The DQN network is trained by estimating the first Q value of different driving modes in each vehicle state based on a preset first Q value function, and training with the goal of maximizing the first Q value of the driving mode adopted in each vehicle state; the first Q value is used to indicate that the adopted driving mode has the minimum energy consumption of the electric vehicle under the premise of satisfying the driving capacity required by the vehicle state;

[0178] The torque distribution module is used to dynamically output the torque distribution ratio of the two motors of the electric vehicle at each moment according to the vehicle state of the electric vehicle at each moment through the trained deep deterministic policy gradient DDPG network when the driving mode of the electric vehicle is dual-motor coupling drive;

[0179] The DDPG network is trained by estimating the second Q value of different torque distribution ratios in each vehicle state based on a preset second Q value function, and training with the second Q value of the torque distribution ratio in each vehicle state as the maximum as the goal; the second Q value is used to indicate that the energy consumption of the two motors corresponding to the adopted torque distribution ratio is minimized;

[0180] The motor drive module is used to dynamically distribute torque to the two motors of the electric vehicle according to the driving capacity required to meet the vehicle state of the electric vehicle at each moment and the torque distribution ratio of the two motors of the electric vehicle at each moment, so that the two motors are driven with the distributed torque coupling.

[0181] It is not difficult to find that this embodiment is a system embodiment corresponding to the above method embodiment, and this embodiment can be implemented in conjunction with the above method embodiment. The relevant technical details and technical effects mentioned in the above embodiment are still valid in this embodiment, and in order to reduce repetition, they are not repeated here. Accordingly, the relevant technical details mentioned in this embodiment can also be applied in the above embodiment.

[0182] It is worth mentioning that all modules involved in this embodiment are logic modules. In practical applications, a logic unit can be a physical unit, a part of a physical unit, or a combination of multiple physical units. In addition, in order to highlight the innovative part of the present invention, this embodiment does not introduce units that are not closely related to solving the technical problem proposed by the present invention, but this does not mean that there are no other units in this embodiment.

[0183] Another embodiment of the present invention relates to a computer device, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the electric vehicle dual-motor coupled drive method in the above-mentioned embodiments.

[0184] Among them, the memory and the processor are connected in a bus manner, and the bus may include any number of interconnected buses and bridges, and the bus connects various circuits of one or more processors and memories together. The bus can also connect various other circuits such as peripherals, voltage regulators, and power management circuits, which are well known in the art and are therefore not further described herein. The bus interface provides an interface between the bus and the transceiver. The transceiver can be one element or multiple elements, such as multiple receivers and transmitters, providing a unit for communicating with various other devices on a transmission medium. The data processed by the processor is transmitted on a wireless medium via an antenna, and further, the antenna also receives data and transmits the data to the processor.

[0185] The processor is responsible for managing the bus and general processing, and can also provide various functions, including timing, peripheral interfaces, voltage regulation, power management, and other control functions. Memory can be used to store data used by the processor when performing operations.

[0186] Another embodiment of the present invention relates to a computer-readable storage medium storing a computer program, which implements the above method embodiment when executed by a processor.

[0187] That is, those skilled in the art can understand that all or part of the steps in the above-mentioned embodiment method can be completed by instructing the relevant hardware through a program, and the program is stored in a storage medium, including a number of instructions for a device (which can be a single-chip microcomputer, chip, etc.) or a processor (processor) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (Read-Only Memory, referred to as: ROM), random access memory (Random Access Memory, referred to as: RAM), disk or optical disk and other media that can store program codes.

[0188] Those skilled in the art will appreciate that the above embodiments are specific embodiments for implementing the present invention, and in actual applications, various changes may be made thereto in form and detail without departing from the spirit and scope of the present invention.

Claims

1. A dual-motor coupling driving method for an electric vehicle, characterized in that: include: Obtain the vehicle status of electric vehicles in real time; The vehicle state of the electric vehicle at each moment is respectively input into the trained deep Q network DQN to dynamically obtain the driving mode of the electric vehicle at each moment, wherein the driving mode is single motor driving or dual motor coupling driving; The DQN network is trained by estimating the first Q value of different driving modes in each vehicle state based on a preset first Q value function, and training with the goal of maximizing the first Q value of the driving mode adopted in each vehicle state; the first Q value is used to indicate that the adopted driving mode has the minimum energy consumption of the electric vehicle under the premise of satisfying the driving capacity required by the vehicle state; If the driving mode of the electric vehicle is dual-motor coupling drive, the trained deep deterministic policy gradient DDPG network dynamically outputs the torque distribution ratio of the two motors of the electric vehicle at each moment according to the vehicle state of the electric vehicle at each moment; The DDPG network is trained by estimating the second Q value of different torque distribution ratios in each vehicle state based on a preset second Q value function, and training with the second Q value of the torque distribution ratio in each vehicle state as the maximum as the goal; the second Q value is used to indicate that the energy consumption of the two motors corresponding to the adopted torque distribution ratio is minimized; According to the driving capacity required to meet the vehicle state of the electric vehicle at each moment and the torque distribution ratio of the two motors of the electric vehicle at each moment, torque is dynamically distributed to the two motors of the electric vehicle so that the two motors are driven with the distributed torque coupling.

2. The electric vehicle dual-motor coupled driving method according to claim 1, characterized in that: The DDPG network includes an Actor network and a Critic network; The Actor network is used to generate the torque distribution ratio corresponding to the two motors according to the input vehicle state; The critic network is used to estimate a second Q value of the torque distribution ratio generated under the input vehicle state based on the second Q value function; The DDPG network updates the network parameters of the Actor network with the goal of maximizing the second Q value estimated by the Critic network.

3. The electric vehicle dual-motor coupled driving method according to claim 2, characterized in that: The network parameters of the Critic network are updated based on the temporal difference error method by the difference between the second Q value of the current vehicle state and the second Q value of the next vehicle state.

4. The electric vehicle dual-motor coupled driving method according to claim 3, characterized in that: The DDPG network adopts a replay buffer mechanism, which stores experiences consisting of several corresponding current vehicle states, torque distribution ratios, immediate rewards adopted by the second Q-value function, and the next vehicle state through the replay buffer; the network parameters of the Actor network and the Critic network are updated by randomly extracting experiences from the replay buffer.

5. The electric vehicle dual-motor coupled driving method according to any one of claims 1 to 4, characterized in that: The vehicle status includes: speed, acceleration, state of charge of the electric vehicle and torque of two motors.

6. The electric vehicle dual-motor coupled driving method according to claim 1, characterized in that: The torque distribution ratio of the two motors output by the DDPG network has a value range of [0, 1], and the motors of the electric vehicle include a first motor and a second motor; The method dynamically distributes torque to the two motors of the electric vehicle according to the driving capacity required to meet the vehicle state of the electric vehicle at each moment and the torque distribution ratio of the two motors of the electric vehicle at each moment, including: By satisfying the driving capacity required by the vehicle state of the electric vehicle at each moment, the vehicle torque required to satisfy the vehicle state is determined; When the torque distribution ratio of the two motors is 1, all the vehicle torque required to meet the vehicle state is distributed to the first motor; When the torque distribution ratio of the two motors is 0, all the vehicle torque required to meet the vehicle state is distributed to the second motor; When the torque distribution ratio of the two motors is between [0, 1], torque is distributed to the first motor and the second motor according to the vehicle torque and torque distribution ratio required to meet the vehicle state.

7. The electric vehicle dual-motor coupled driving method according to claim 1, characterized in that: The method further comprises: If the driving mode of the electric vehicle is driven by any one motor, all the vehicle torque required to meet the vehicle state will be distributed to any one motor, so that any one motor can be driven alone.

8. An electric vehicle dual-motor coupled drive system, characterized in that: include: A status acquisition module is used to acquire the vehicle status of the electric vehicle in real time; A mode selection module is used to input the vehicle state of the electric vehicle at each moment into the trained deep Q network DQN, and dynamically obtain the driving mode of the electric vehicle at each moment, wherein the driving mode is a single motor drive or a dual motor coupling drive; The DQN network is trained by estimating the first Q value of different driving modes in each vehicle state based on a preset first Q value function, and training with the goal of maximizing the first Q value of the driving mode adopted in each vehicle state; the first Q value is used to indicate that the adopted driving mode has the minimum energy consumption of the electric vehicle under the premise of satisfying the driving capacity required by the vehicle state; The torque distribution module is used to dynamically output the torque distribution ratio of the two motors of the electric vehicle at each moment according to the vehicle state of the electric vehicle at each moment through the trained deep deterministic policy gradient DDPG network when the driving mode of the electric vehicle is dual-motor coupling drive; The DDPG network is trained by estimating the second Q value of different torque distribution ratios in each vehicle state based on a preset second Q value function, and training with the second Q value of the torque distribution ratio in each vehicle state as the maximum as the goal; the second Q value is used to indicate that the energy consumption of the two motors corresponding to the adopted torque distribution ratio is minimized; The motor drive module is used to dynamically distribute torque to the two motors of the electric vehicle according to the driving capacity required to meet the vehicle state of the electric vehicle at each moment and the torque distribution ratio of the two motors of the electric vehicle at each moment, so that the two motors are driven with the distributed torque coupling.

9. A computer device, characterized in that: include: at least one processor; And, a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the electric vehicle dual-motor coupled driving method as described in any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the electric vehicle dual-motor coupled driving method according to any one of claims 1 to 7 is implemented.

Images