Vehicle torque control method, device and vehicle

Abstract

This application provides a vehicle torque control method, device, and vehicle, relating to the field of vehicle torque control, to address the problems of insufficient collaborative optimization between trajectory tracking and battery SOC, and the inability to adaptively adjust battery weights in vehicle torque control, leading to low control accuracy and poor battery safety. The method includes: constructing a cumulative cost function in the prediction time domain based on acquired real-time vehicle state information, real-time battery state of charge, and the control torque command variable to be solved; wherein the cumulative cost function includes vehicle state trajectory error cost, control input penalty cost, and battery state of charge deviation cost; the weight of the battery state of charge deviation cost in the cumulative cost function is determined based on the predicted battery state of charge; solving for the cumulative cost function with the objective of minimizing it, to obtain the target control torque command sequence corresponding to the prediction time domain, wherein the control torque command corresponding to the first control cycle in this sequence controls the operation of the drive motor.

Description

Technical Field

[0001] This application relates to the field of vehicle torque control, specifically to a vehicle torque control method, device, and vehicle. Background Technology

[0002] With the rapid development of the new energy vehicle industry, vehicle power control technology is constantly iterating and upgrading. Precise distribution and efficient control of drive torque have become crucial means to improve vehicle power, economy, and driving smoothness. Vehicle drive torque is the torsional torque output by the drive system and applied to the wheels, directly determining the vehicle's acceleration, climbing ability, and load-bearing capacity. Currently, most related technologies generate torque commands based on basic driving information such as accelerator pedal opening and vehicle speed to control the drive system and achieve wheel drive.

[0003] However, with the improvement of vehicle intelligence, the state of the battery, as the core energy source, directly affects the overall vehicle performance. Among these factors, the battery's state of charge (SBC) is gradually becoming an important factor affecting torque distribution efficiency. However, related technologies have not yet effectively incorporated the SBC into the decision boundary when performing torque control and distribution. This results in the torque control strategy still following conventional logic for power output or energy recovery during vehicle operation, especially when the battery's SBC is high, failing to dynamically adjust the drive mode based on the battery's charge level.

[0004] Therefore, how to achieve coordinated optimization of vehicle torque control and battery status to improve the overall performance of vehicle control has become a technical problem that urgently needs to be solved in this field. Summary of the Invention

[0005] One of the purposes of this application is to provide a vehicle torque control method, device, and vehicle to solve the problem that the relevant vehicle torque control cannot simultaneously take into account vehicle state tracking and battery state of charge optimization, and the control weight of battery state of charge cannot be adaptively adjusted, resulting in poor vehicle control accuracy and battery safety.

[0006] To achieve the above objectives, the technical solution adopted in this application is as follows: Firstly, a vehicle torque control method is provided. The vehicle includes a drive motor. The method includes: acquiring real-time vehicle state information and real-time battery state of charge; constructing a cumulative cost function in the prediction time domain based on the real-time vehicle state information, the real-time battery state of charge, and the control torque command variable to be solved; wherein the prediction time domain is used to represent N control cycles starting from the current moment, and N is a positive integer; wherein the cumulative cost function includes vehicle state trajectory error cost, control input penalty cost, and battery state of charge deviation cost; the vehicle state trajectory error cost is used to evaluate the degree of deviation between the predicted vehicle state information and the desired state information; the control input penalty cost is used to suppress the control torque command. Fluctuation amplitude; Battery state of charge deviation cost is used to evaluate the deviation between the predicted and expected battery state of charge; The weight of the battery state of charge deviation cost in the cumulative cost function is determined based on the predicted battery state of charge; Vehicle predicted state information and battery predicted state of charge are determined based on the control torque command variable to be solved, vehicle real-time state information, and battery real-time state of charge; The cumulative cost function is minimized to obtain the target control torque command sequence corresponding to the predicted time domain; The target control torque command sequence includes N control torque commands ordered according to N control cycles; The drive motor is controlled to operate based on the control torque command corresponding to the first control cycle in the target control torque command sequence.

[0007] Based on the aforementioned technical means, a vehicle torque control method is provided. This method integrates the real-time vehicle state and the real-time battery state of charge (SOC). In the prediction time domain, it constructs a cumulative cost function that includes vehicle state trajectory error cost, control input penalty cost, and battery SOC deviation cost. During the optimization process of minimizing the cumulative cost function, the weight of the battery SOC deviation cost is dynamically determined based on the predicted battery SOC, achieving multi-objective optimization control of the vehicle's driving torque. This method can track the desired vehicle state using predicted vehicle state information and control torque output smoothly. Compared to traditional control methods using fixed weights, it dynamically adjusts the emphasis on battery protection based on the predicted battery SOC. When the predicted battery SOC deviates from the expected value, the control weights are adjusted to prioritize battery protection. This effectively extends battery life and improves overall vehicle energy efficiency while ensuring vehicle power performance.

[0008] Furthermore, the cumulative cost function in the prediction time domain is constructed as follows: The deviation between the predicted vehicle state information and the desired state information in the first control cycle is weighted to determine the vehicle state trajectory error cost in the first control cycle; the first control cycle is any control cycle within the prediction time domain; the deviation between the predicted battery state of charge and the desired battery state of charge in the first control cycle is weighted to determine the battery state of charge deviation cost in the first control cycle; the amplitude of the control torque command in the first control cycle is weighted to determine the control input penalty cost in the first control cycle; under the first control cycle, the vehicle state trajectory error cost, the battery state of charge deviation cost, and the control input penalty cost are summed to determine the periodic cost function; based on the periodic cost function of the first control cycle, the cumulative cost function in the prediction time domain is determined.

[0009] Based on the aforementioned technical methods, the vehicle state trajectory error cost, battery state of charge deviation cost, and control input penalty cost for each control cycle in the prediction time domain are weighted and summed to obtain a cycle cost function. Then, a cumulative cost function is constructed based on each cycle cost function. This method decomposes the multi-objective optimization problem into each control cycle, enabling the model predictive controller to balance vehicle state tracking accuracy, battery charge maintenance, and torque output smoothness cycle by cycle. It provides a clear cost accumulation structure for dynamic weight adjustment and rolling optimization, effectively improving solution efficiency and control real-time performance.

[0010] Furthermore, the weights used when weighting the deviation between the predicted state of charge (SOC) and the desired state of charge (SOC) of the battery in the first control cycle are exponentially positively correlated with the absolute value of the deviation between the predicted SOC and the desired SOC.

[0011] Based on the aforementioned technical methods, the weighted average of the cost of battery state of charge deviation is exponentially positively correlated with the absolute value of the deviation between the predicted and expected state of charge. Compared to linear weighting or fixed weighting, this mechanism can maintain a low weight when the state of charge is normal, without interfering with driving smoothness, while drastically increasing the weight when the state of charge approaches the boundary, forcing the controller to prioritize battery protection. This achieves adaptive power management of "no intervention in the normal zone and strong pull-back in the boundary zone," significantly extending battery cycle life and improving the overall vehicle energy economy.

[0012] Furthermore, the weights used when weighting the deviation between the predicted state of charge and the desired state of charge of the battery in the first control cycle satisfy the following relationship:

[0013] in, The weighting coefficient represents the degree of deviation between the predicted state of charge and the expected state of charge of the battery in the k-th control cycle. These are the basic weighting coefficients, used to set the basic magnitude of the weights; This is an adjustment factor used to adjust the rate at which the weight changes with the deviation of the battery's state of charge. Predict the state of charge of the battery in the k-th control cycle; This is the preset state of charge.

[0014] Based on the aforementioned technical methods, a specific mathematical relationship for the weighting of battery state of charge (SOC) deviation costs is provided. An exponential function is used to make the weights increase exponentially as the deviation between the predicted SOC and the expected value increases. This method enables the generation of a strong penalty signal for the battery SOC when it approaches a critical value, prioritizing the return of the battery SOC to the expected range and achieving rapid battery protection response. Simultaneously, the smoothness of the exponential function avoids abrupt changes in control commands caused by weight jumps, ensuring smooth vehicle operation.

[0015] Furthermore, based on the periodic cost function of the first control cycle, the cumulative cost function in the prediction time domain is determined, including: summing up the periodic cost functions corresponding to all first control cycles in the prediction time domain and adding the terminal cost term to determine the cumulative cost function in the prediction time domain. The terminal cost term is used to characterize the total additional loss in the prediction time domain.

[0016] Based on the aforementioned technical means, the cumulative cost function is obtained by summing the periodic cost functions of all control cycles within the prediction time domain and adding the terminal cost term. The summation ensures the controller's forward-looking optimization for multiple future cycles, enabling the torque command sequence to predict future state changes. The introduction of the terminal cost term compensates for the boundary effects caused by the truncation of the prediction time domain, avoiding state deviations or abrupt changes in control quantities at the end of the time domain, thereby ensuring the closed-loop stability and long-term convergence performance of the control system.

[0017] Furthermore, the cumulative cost function in the prediction time domain satisfies the following relationship:

[0018] in, This represents the cumulative cost function value within the prediction time domain. The state vector for the k-th control cycle includes vehicle predicted state information and battery predicted state of charge. The reference state vector includes desired state information and desired state of charge. The vehicle state trajectory error cost and battery state of charge deviation cost are represented by weighting based on the state error weight matrix Q; Q is the state error weight matrix, and the weight coefficients in the state error weight matrix corresponding to the battery predicted state of charge are functions of the battery predicted state of charge, so that the weight of the battery state of charge deviation cost is determined based on the battery predicted state of charge. Characterizes the control input penalty cost weighted based on the control input weight matrix R; The predicted control torque command for the k-th control cycle; The input weight matrix is ​​used to adjust the penalty intensity of the control torque command. This is the terminal cost term, used to characterize the additional loss in the prediction time domain; To predict the length of the time domain.

[0019] Based on the aforementioned technical methods, a specific mathematical form of the cumulative cost function is provided. A quadratic expression is used to unify the vehicle state trajectory error cost and the battery state of charge deviation cost into the state tracking term. Furthermore, by using the corresponding diagonal elements of the state error weight matrix as functions of the battery's predicted state of charge, the organic integration of dynamic weights and the cost function is achieved. This mathematical expression not only facilitates engineering implementation and real-time solution but also provides a clear mathematical foundation for subsequent constraint processing and optimization algorithm design, improving the computational efficiency and real-time performance of the vehicle torque control system.

[0020] Furthermore, the method in the first aspect also includes: after controlling the drive motor to run based on the control torque command corresponding to the first control cycle in the target control torque command sequence, obtaining the real-time vehicle status information and the real-time battery charge status at the time of the first control cycle, and updating the second control cycle in the target control torque command sequence to the first control cycle of the new prediction time domain, so as to redetermine the control torque command corresponding to the first control cycle of the new prediction time domain.

[0021] Based on the aforementioned technical means, a method is provided that after completing the torque control of the first control cycle, new real-time state information is acquired and the optimization time domain is rolled forward to re-solve the torque command for the next control cycle. This rolling time domain optimization strategy realizes closed-loop feedback of the control system, which can promptly correct control deviations caused by factors such as model mismatch and external disturbances, giving the vehicle torque control method adaptive capability and robustness; at the same time, each control cycle executes only the first command in the sequence, greatly reducing the computational burden.

[0022] Furthermore, the vehicle predicted state information and the battery predicted state of charge are determined as follows: Based on the control torque command variable to be solved, the real-time vehicle state information, and the real-time battery state of charge, the vehicle predicted state information and the battery predicted state of charge for the first control cycle in the prediction time domain are determined through a prediction model; based on the vehicle predicted state information and the battery predicted state of charge for the previous control cycle, the vehicle predicted state information and the battery predicted state of charge for the next control cycle are predicted through the prediction model, until the vehicle predicted state information and the battery predicted state of charge for each control cycle in the prediction time domain are obtained; the prediction model includes a vehicle dynamics model and a battery model; the vehicle dynamics model is used to characterize the periodic change relationship of the vehicle predicted state information between two adjacent control cycles; the battery model is used to characterize the periodic change relationship of the battery predicted state of charge between two adjacent control cycles.

[0023] Based on the above technical means, the synergistic effect of the vehicle dynamics model and the battery model can accurately depict the periodic changes in the vehicle's motion state and the battery's state of charge, respectively. This enables synchronous iterative prediction of the vehicle and battery states in each control cycle within the prediction time domain, providing accurate and effective data support for the subsequent optimization of the cumulative cost function and ensuring the rationality of torque control.

[0024] Furthermore, when solving the problem with the objective of minimizing the cumulative cost function, at least one of the following preset constraints is included: vehicle state trajectory data constraints, control input data constraints, or battery state of charge constraints; wherein, the vehicle state trajectory data constraints include: the predicted vehicle state information is greater than or equal to a first trajectory threshold and less than or equal to a second trajectory threshold, and the first trajectory threshold is less than the second trajectory threshold; the control input data constraints include: the control torque command is greater than or equal to a first control threshold and less than or equal to a second control threshold, and the first control threshold is less than the second control threshold; the battery state of charge constraints include: the predicted battery state of charge is greater than or equal to a first charge threshold and less than or equal to a second charge threshold, and the first charge threshold is less than the second charge threshold.

[0025] Based on the aforementioned technical methods, preset constraints are introduced during the optimization process, including constraints on vehicle state trajectory data, control input data, and battery state of charge. These constraints ensure that the obtained control torque command sequence always meets the vehicle's physical limits, motor capability limitations, and battery safety boundaries. These constraints confine the optimization problem to the feasible region, preventing the generation of control commands exceeding physical limits in pursuit of minimum cost. This ensures vehicle driving stability, drive motor operational safety, and battery lifespan, thereby improving the reliability and safety of the overall vehicle torque control.

[0026] Furthermore, the real-time vehicle status information includes at least one of the following: vehicle longitudinal acceleration, vehicle lateral acceleration, vehicle longitudinal velocity, vehicle lateral velocity, vehicle yaw rate, and wheel speed of each wheel.

[0027] Based on the aforementioned technical methods, the specific content of the vehicle's predicted state information has been clarified, including key state quantities such as longitudinal acceleration, lateral acceleration, longitudinal velocity, lateral velocity, yaw rate, and wheel speeds. These state quantities comprehensively cover the core information required for vehicle motion control, providing a data foundation for accurately assessing the deviation between the vehicle's current state and the desired state. This makes torque control more refined and accurate, improving the vehicle's handling stability, ride comfort, and trajectory tracking capabilities.

[0028] Secondly, this application provides a vehicle torque control device, comprising: an acquisition module, a construction module, a solution module, and a control module; the acquisition module is used to acquire real-time vehicle state information and real-time battery state of charge; the construction module is used to construct a cumulative cost function in the prediction time domain based on the real-time vehicle state information, the real-time battery state of charge, and the control torque command variable to be solved; wherein, the prediction time domain is used to represent N control cycles starting from the current moment, and N is a positive integer; wherein, the cumulative cost function includes vehicle state trajectory error cost, control input penalty cost, and battery state of charge deviation cost; the vehicle state trajectory error cost is used to evaluate the degree of deviation between the predicted vehicle state information and the desired state information; the control input penalty cost is used to suppress control... The fluctuation range of the control torque command; the battery state of charge deviation cost is used to evaluate the deviation between the predicted and expected battery state of charge; the weight of the battery state of charge deviation cost in the cumulative cost function is determined based on the predicted battery state of charge; the vehicle predicted state information and the battery predicted state of charge are determined based on the control torque command variable to be solved, the real-time vehicle state information, and the real-time battery state of charge; the solution module is used to: solve for the target control torque command sequence corresponding to the predicted time domain with the objective of minimizing the cumulative cost function; the target control torque command sequence includes N control torque commands sorted according to N control cycles; the control module is used to: control the drive motor to operate based on the control torque command corresponding to the first control cycle in the target control torque command sequence.

[0029] Thirdly, this application provides an electronic device comprising: a processor and a memory; the memory storing processor-executable instructions. When the processor is configured to execute the instructions, the electronic device implements the method described in the first aspect.

[0030] Fourthly, this application provides a vehicle that includes the electronic equipment described in the third aspect.

[0031] Fifthly, this application provides a computer-readable storage medium that, when the instructions in the computer-readable storage medium are executed by a vehicle's processor, enables the vehicle to perform the methods described in the first aspect and any of their possible implementations.

[0032] Sixthly, this application provides a computer program product including computer instructions that, when executed on a vehicle, cause the vehicle to perform the method described in the first aspect and any possible implementation thereof.

[0033] It should be noted that the technical effects of any of the implementation methods in aspects two through six can be found in the technical effects of the corresponding implementation methods in aspect one, and will not be repeated here.

[0034] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and do not limit this application. Attached Figure Description

[0035] Figure 1 This application provides a schematic diagram of the structure of a vehicle torque control system; Figure 2 A schematic flowchart of a vehicle torque control method provided in this application; Figure 3 A schematic diagram of a two-degree-of-freedom model provided in this application; Figure 4 A schematic diagram of the structure of a battery model provided in this application; Figure 5 This application provides a schematic diagram of the composition of a vehicle torque control device; Figure 6 This is a schematic diagram of the composition of an electronic device provided in this application. Detailed Implementation

[0036] The embodiments of this application will be described below with reference to the accompanying drawings and preferred embodiments. Those skilled in the art can easily understand other advantages and effects of this application from the content disclosed in this specification. This application can also be implemented or applied through other different specific embodiments, and various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of this application. It should be understood that the preferred embodiments are only for illustrating this application and are not intended to limit the scope of protection of this application.

[0037] It should be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of this application. Therefore, the drawings only show the components related to this application and are not drawn according to the actual number, shape and size of the components in the actual implementation. In the actual implementation, the form, quantity and proportion of each component can be arbitrarily changed, and the layout of the components may also be more complex.

[0038] With the continuous development of new energy vehicle technology and the increasing demands of users for overall vehicle performance, vehicle control systems need to ensure power response while also considering energy management and battery safety. During actual driving, the torque output of the drive motor directly affects the battery's charging and discharging power and state of charge (SOC). If only the vehicle's motion state is tracked and optimized, ignoring the dynamic changes in the battery's SOC, problems such as deviations from the reasonable SOC range, increased energy consumption, or unstable battery operation can easily occur, potentially even affecting battery life and driving safety. Therefore, how to achieve coordinated optimization of vehicle torque control and battery SOC to improve the overall performance of vehicle control has become a pressing technical problem to be solved in this field.

[0039] To address the aforementioned technical problems, this application provides a vehicle torque control method, device, and vehicle. The method constructs a cumulative cost function including vehicle state trajectory error, control input penalty, and battery state of charge deviation. By introducing a weighting factor dynamically adjusted according to the predicted battery state of charge, the battery management objective is embedded in the torque control decision-making process. After solving for the optimal control torque command sequence in the prediction time domain, the drive motor is controlled by the command of the first control cycle. In this way, this application achieves a combination of vehicle power control and battery energy management, enabling real-time tracking of the driver's desired state while actively adjusting the battery charging and discharging depth and rate, effectively avoiding overcharging and over-discharging, reducing energy loss, and improving the vehicle's economy, smoothness, and safety.

[0040] like Figure 1 As shown, this application proposes a vehicle torque control system, including: a sensor group 101, a drive motor 102, and a controller 103. The controller 103 is connected to both the sensor group 101 and the drive motor 102.

[0041] The controller 103 is used to construct a cumulative cost function in the prediction time domain based on the real-time vehicle status information and real-time battery charge status collected by the sensor group 101. The controller obtains the target control torque command sequence through optimization and outputs the control torque command corresponding to the first control cycle in the target control torque command sequence to control the drive motor 102 to run, thereby achieving precise and energy-saving control of vehicle torque.

[0042] In one possible implementation, the controller 103 first acquires the vehicle's real-time status information and the battery's real-time state of charge. Based on the vehicle's real-time status information, the battery's real-time state of charge, and the control torque command variable to be solved, it constructs a cumulative cost function in the prediction time domain. The controller 103 optimizes the solution with the objective of minimizing the cumulative cost function, obtaining a target control torque command sequence corresponding to the prediction time domain. The target control torque command sequence includes multiple control torque commands ordered according to multiple control cycles. Based on the control torque command corresponding to the first control cycle in the target control torque command sequence, the controller 103 controls the drive motor 102 to run. After execution, it updates the vehicle's real-time status information and the battery's real-time state of charge, redetermines a new prediction time domain, and continuously optimizes, continuously outputting the corresponding control torque command.

[0043] In some embodiments, the sensor group 101 is used to collect real-time vehicle operating information and battery status information, including vehicle longitudinal acceleration, lateral acceleration, yaw rate, wheel speed of each wheel and battery state of charge, and send the above information to the controller 103 in real time.

[0044] As one possible implementation, the sensor group 101 includes a 6D IMU sensor, a wheel speed sensor, and a battery status sensor.

[0045] Among them, the 6D IMU sensor is used to collect vehicle dynamic state information such as longitudinal acceleration, lateral acceleration, and yaw rate; the wheel speed sensor is used to collect the wheel speed of each wheel to supplement the real-time state information of the vehicle; the battery state sensor is used to collect parameters such as real-time battery state of charge, battery current, and battery terminal voltage, providing basic data for calculating the cost of battery state of charge deviation and determining its weight.

[0046] In some embodiments, the drive motor 102 is used to receive the control torque command output by the controller 103, convert electrical energy into mechanical energy, drive the vehicle wheels to rotate, and realize the vehicle's movement; the control command it executes is the control torque command corresponding to the first control cycle in the target control torque command sequence.

[0047] In one possible implementation, the drive motor 102 includes at least one drive motor, each drive motor driving a corresponding wheel. For example, there are four drive motors 102, corresponding to the four wheels of the vehicle: the left front wheel, the right front wheel, the left rear wheel, and the right rear wheel, forming a distributed drive structure. In this case, the control torque command is a four-wheel control torque command variable. The controller 103 can achieve torque vector control by independently controlling the torque of each drive motor, balancing path tracking accuracy and energy saving.

[0048] It should be pointed out that, Figure 1The structure shown does not constitute a limitation on the vehicle torque control system, which may include fewer or more components than shown, or combine certain components, or adopt different component arrangements. This application embodiment does not impose any limitations in this regard.

[0049] In some embodiments, the vehicle torque control method of this application can be applied to the controller in the above-described vehicle torque control system.

[0050] like Figure 2 As shown, this application proposes a vehicle torque control method, including: S201. Obtain real-time vehicle status information and real-time battery charge status.

[0051] Real-time vehicle status information is a set of various parameters that reflect the current dynamic characteristics and motion state of the vehicle. It provides basic data support for the controller to construct the cumulative cost function and optimize the solution of control torque commands.

[0052] In one possible implementation, the real-time vehicle status information includes at least one of the following: vehicle longitudinal acceleration, vehicle lateral acceleration, vehicle longitudinal velocity, vehicle lateral velocity, vehicle yaw rate, and wheel speed of each wheel.

[0053] As one possible implementation, real-time vehicle status information is acquired through a sensor array. Specifically, the vehicle's longitudinal acceleration, lateral acceleration, and yaw rate are acquired through a six-dimensional inertial measurement unit; the wheel speeds of each wheel are acquired through wheel speed sensors; and the vehicle's longitudinal and lateral velocities are obtained by integrating the acceleration using a state estimator and fusing it with the wheel speed information.

[0054] In one possible implementation, the vehicle's longitudinal acceleration is used to reflect the intensity of acceleration and deceleration in the direction of travel. Changes in longitudinal acceleration are directly affected by the total torque output by the drive motor and are an important basis for evaluating the control input penalty cost. Excessive longitudinal acceleration implies higher energy consumption, which corresponds to a higher control input penalty cost in the cumulative cost function.

[0055] One possible implementation is to use lateral acceleration to reflect the lateral forces acting on the vehicle during cornering. Lateral acceleration, influenced by the yaw moment resulting from differences in torque distribution among the wheels, is a key indicator for assessing vehicle stability. When constructing the vehicle trajectory error cost, the deviation between the lateral acceleration and the expected value will be included in the cost function.

[0056] One possible implementation is that the vehicle's longitudinal velocity is used to reflect the vehicle's current speed. Longitudinal velocity is the basis for predicting the future vehicle position and a core state variable for evaluating trajectory tracking error. The longitudinal velocity of the current cycle will serve as the initial value for the prediction model, and together with the torque command to be solved, it will determine the velocity changes in future cycles.

[0057] One possible implementation is to use the vehicle's lateral velocity to reflect the degree of lateral slip. Lateral velocity is affected by tire lateral forces, which are related to the torque distribution between the wheels. Accurate estimation of lateral velocity is crucial for yaw stability control during high-speed cornering.

[0058] In one possible implementation, the vehicle yaw rate is used to reflect the vehicle's rotational rate about its vertical axis. The yaw rate is directly affected by the yaw moment generated by the torque difference between the left and right wheels and is the core feedback quantity for achieving torque vector control. In the cumulative cost function, the deviation between the yaw rate and the desired value is included as part of the trajectory error cost.

[0059] One possible implementation is that the wheel speed of each wheel is used to reflect the rotational speed of each drive wheel. Wheel speed is the basis for calculating wheel slip ratio, which directly reflects the efficiency of torque command transmission between the tire and the ground. An excessive slip ratio means energy loss, which will be indirectly reflected in the penalty cost of control input.

[0060] Real-time battery state of charge (SOC) includes the percentage of the battery's current remaining charge relative to its rated capacity. It directly reflects the available energy and guides the controller to optimize torque command distribution, balancing battery safety and vehicle energy efficiency.

[0061] As one possible implementation, the real-time state of charge of the battery is obtained through the battery management system.

[0062] As one possible implementation, the remaining battery charge percentage determines the baseline value of the cost of deviation from the battery's state of charge. The lower the current charge, the greater the weight of deviation from the desired charge in the cumulative cost function, thereby prompting the selection of a more energy-efficient torque command.

[0063] S202. Based on the real-time vehicle status information, the real-time battery state of charge, and the control torque command variable to be solved, a cumulative cost function in the prediction time domain is constructed.

[0064] The prediction time domain is used to represent N control cycles starting from the current time, where N is a positive integer.

[0065] The control torque command variable to be solved is the core variable used by the controller to control the operation of the drive motor. It corresponds to the predicted control torque command for each control cycle in the prediction time domain, and is used to characterize the magnitude of the torque required by the drive motor in each control cycle. Its value range must comply with the physical performance constraints of the drive motor and the safety output constraints of the battery. One possible implementation is to define the control input vector as the torque command of the four-wheel motor: .in, , , , These represent the control torque commands for the left front wheel, right front wheel, left rear wheel, and right rear wheel, respectively. In the prediction time domain, the control torque command variables to be solved form a sequence. Each All are four-dimensional vectors containing the torque of the four wheels.

[0066] It should be understood that the real-time vehicle status information and the real-time battery state of charge are data acquired at the current moment in the prediction time domain, and are the basic input data for constructing the cumulative cost function. The control torque command variable to be solved is an unknown variable in each control cycle in the prediction time domain. The controller uses a preset prediction model, combined with the real-time vehicle status information and the real-time battery state of charge at the current moment, to iteratively predict the predicted vehicle status information and the predicted battery state of charge for each control cycle. Then, based on the deviation between the prediction result and the expected target, the cumulative cost function is constructed to achieve the optimized solution of the control torque command variable.

[0067] In this embodiment, the cumulative cost function includes vehicle state trajectory error cost, control input penalty cost, and battery state of charge deviation cost.

[0068] The vehicle state trajectory error cost is used to evaluate the degree of deviation between the predicted and expected vehicle state information. The smaller the deviation, the higher the vehicle trajectory tracking accuracy.

[0069] The control input penalty is used to suppress fluctuations in the control torque command. This prevents sudden changes in torque command that could reduce vehicle ride smoothness and cause sudden changes in the load on the drive motor and battery.

[0070] The weight of the battery state-of-charge (POC) deviation cost in the cumulative cost function is determined based on the battery's predicted POC. The POC deviation cost is used to assess the deviation between the predicted and expected POC. A smaller deviation ensures safer battery operation and improves energy efficiency.

[0071] In some embodiments, the cumulative cost function in the prediction time domain is constructed as follows: (1) Weight the deviation between the predicted state information and the expected state information of the vehicle in the first control cycle to determine the vehicle state trajectory error cost in the first control cycle.

[0072] The first control period is any control period within the prediction time domain.

[0073] As one possible implementation, the vehicle predicted state information includes, but is not limited to, predicted values ​​of vehicle longitudinal acceleration, vehicle lateral acceleration, vehicle yaw rate, left front wheel speed, right front wheel speed, left rear wheel speed, and right rear wheel speed. The desired state information is provided by the controller (such as an autonomous driving planning module or a driver intent parsing module) and includes the expected values ​​corresponding to the above predicted information. The desired state information can be obtained from any database, and this application does not impose any specific restrictions on this.

[0074] One possible approach is to obtain vehicle predictive state information by using a vehicle dynamics model (e.g., a three-degree-of-freedom single-vehicle model or a two-degree-of-freedom model) to predict the current vehicle state based on real-time vehicle state information (e.g., longitudinal velocity, yaw rate, etc. collected by onboard sensors) and the control torque command variable to be solved. Specifically, the control torque command is input into the vehicle dynamics model, and the state variables for each future control cycle are recursively calculated.

[0075] As one possible implementation, the weights of the vehicle state trajectory error cost are determined by a state error weight matrix, where the diagonal elements correspond to the weight coefficients of different state variables. These coefficients are typically pre-calibrated based on the vehicle's dynamic characteristics and control requirements. For example, the longitudinal velocity error is given a higher weight to ensure tracking accuracy, while the yaw rate error is given a higher weight to ensure stability.

[0076] In this embodiment, by weighting the vehicle state trajectory error cost, the optimal control torque command obtained can drive the vehicle to accurately track the desired trajectory. At the same time, control priorities can be allocated according to the importance of different state variables, thereby improving the accuracy of the vehicle's dynamic response and driving safety.

[0077] (2) The weight used when weighting the deviation between the predicted state of charge and the expected state of charge of the battery in the first control cycle is used to determine the cost of deviation of the battery state of charge in the first control cycle.

[0078] Predicted state of charge (SOC) refers to the percentage of battery charge remaining in a future control cycle predicted by a battery model; the expected SOC is a preset target charge value, usually set to a value within the battery safety window (e.g., 50%).

[0079] One possible approach is to obtain the predicted state of charge (SOC) of the battery by using a battery model (e.g., a first-order RC equivalent circuit model or an ampere-hour integral model) based on the current real-time SOC of the battery and the control torque command variable to be solved. The control torque command affects the motor power, which in turn affects the battery's discharge or charging current, and the battery model calculates the change in SOC accordingly.

[0080] As one possible implementation, the weight of the cost of battery state of charge deviation is not a fixed value, but is dynamically adjusted with the SOC deviation. As another possible implementation, the weight used when weighting the deviation between the predicted and desired state of charge in the first control cycle is exponentially positively correlated with the absolute value of the deviation between the predicted and desired state of charge.

[0081] It should be understood that when the SOC deviation is small, the weight is close to the baseline value and does not significantly interfere with normal driving; when the SOC deviation is close to the safety boundary, the weight increases exponentially, forcing the controller to adjust the torque to recharge or consume the battery.

[0082] As one possible implementation, the weights used when weighting the deviation between the predicted state of charge and the desired state of charge of the battery in the first control cycle are:

[0083] in, The weighting coefficient represents the cost of the battery state of charge deviation in the k-th control cycle; These are the basic weighting coefficients, used to set the basic magnitude of the weights; This is an adjustment factor used to adjust the rate at which the weight changes with the deviation of the battery's state of charge. Predict the state of charge of the battery in the k-th control cycle; This is the preset state of charge.

[0084] As one implementation method, this is achieved by adjusting the basic weight coefficients. The minimum protection level can be set by adjusting the factor. It can be flexibly adapted to the discharge characteristics of different battery types (such as lithium iron phosphate and ternary lithium) or different driving modes (economy / sport), making the control strategy portable and adjustable.

[0085] It is understandable that the weight of the battery state-of-charge deviation cost in the cumulative cost function implements a non-linear adjustment mechanism, when the battery predicts its state of charge... With preset state of charge When close, weight Approaching the base value ;when Deviation At that time, weight The growth is exponential, and the greater the deviation, the faster the growth. When the predicted state of charge of the battery is lower than the preset state of charge, the weighting coefficient increases nonlinearly with the increase of the deviation, making the optimization tend to select more energy-efficient torque commands; when the predicted state of charge of the battery is higher than the preset state of charge, the weighting coefficient remains at a low level, so that the optimization can prioritize ensuring the trajectory tracking accuracy.

[0086] The implementation method of this application achieves a power management strategy of "low intervention in the normal zone and strong pull-back in the boundary zone" through exponential adaptive weights. While ensuring driving smoothness, it effectively prevents battery over-discharge or over-charge, extends battery life and improves the energy economy of the whole vehicle.

[0087] (3) Weight the amplitude of the control torque command in the first control cycle to determine the control input penalty cost of the first control cycle.

[0088] The amplitude of the control torque command refers to the absolute value of the target torque applied to the drive motor, measured in Newton-meters (Nm). This command is the optimization variable to be solved.

[0089] One possible implementation is to obtain the control torque command as follows: during the optimization process, the control torque command is used as a decision variable, and the optimal sequence is obtained by minimizing the cumulative cost function. The torque command amplitude for each control cycle is directly read from the optimization variables.

[0090] As one possible implementation, the weights of the control input penalty cost are determined by a control input weight matrix. This matrix can be a positive scalar or positive definite matrix; the larger its value, the stronger the penalty on the torque amplitude, thereby suppressing drastic fluctuations in torque command. The values ​​of this matrix are pre-calibrated according to the desired driving comfort requirements.

[0091] For example, the specific calculation of the control input penalty cost is as follows: (For single torque input) or (For multiple inputs), where For the first The torque command for the control cycle, where R is the control input weight matrix.

[0092] In this embodiment, by weighting the torque command amplitude, abrupt changes in control are avoided, resulting in smooth motor torque output, improved driver and passenger comfort, and reduced impact load on the transmission system.

[0093] (4) In the first control cycle, the vehicle state trajectory error cost, battery state of charge deviation cost, and control input penalty cost are summed to determine the cycle cost function.

[0094] In this embodiment, the three costs are linearly added together within the same control cycle to form the total cost of the first control cycle. This structure allows the optimization problem to simultaneously balance three objectives: trajectory tracking accuracy, battery maintenance, and driving smoothness.

[0095] (5) Based on the periodic cost function of the first control period, determine the cumulative cost function in the prediction time domain.

[0096] One possible implementation involves predicting multiple control cycles within the time domain (e.g., N future cycles starting from the current moment). Starting from the current control cycle, the cycle cost function for each cycle is calculated sequentially.

[0097] As one possible implementation, determining the cumulative cost function in the prediction time domain based on the periodic cost function of the first control cycle includes: summing up the periodic cost functions corresponding to all first control cycles in the prediction time domain and adding the terminal cost term to determine the cumulative cost function in the prediction time domain.

[0098] The terminal cost term is used to characterize the total additional loss after the prediction time domain ends, such as the state deviation penalty or remaining energy loss at the end of the prediction time domain.

[0099] As one possible implementation, the terminal cost item includes at least one of the following: the deviation between the vehicle state at the end of the predicted time domain and the expected state (such as speed error, yaw angle error), the remaining deviation between the battery state of charge and the expected state of charge, and the equivalent loss due to motor energy consumption or battery aging.

[0100] In this implementation, the control effects of multiple future cycles are uniformly incorporated into the optimization objective by summing, giving the controller a forward-looking capability. The introduction of the terminal cost term ensures the system stability after the prediction time domain is truncated, preventing boundary effects from causing a decline in control performance.

[0101] Combining the above (1)-(5), the cycle cost function is obtained by weighting and summing the vehicle state trajectory error cost, battery state of charge deviation cost, and control input penalty cost for each control cycle in the prediction time domain, and then constructing a cumulative cost function based on the cycle cost function. This method decomposes the multi-objective optimization problem into each control cycle, and can balance the vehicle state tracking accuracy, battery charge maintenance, and torque output smoothness cycle by cycle. It provides a clear cost accumulation structure for dynamic weight adjustment and rolling optimization, effectively improving the solution efficiency and control real-time performance.

[0102] As one possible implementation, the cumulative cost function satisfies the following relationship:

[0103] in, This represents the cumulative cost function value within the prediction time domain. The state vector for the k-th control cycle includes vehicle predicted state information and battery predicted state of charge. The reference state vector includes desired state information and desired state of charge. The vehicle state trajectory error cost and battery state of charge deviation cost are represented by weighting based on the state error weight matrix Q; Q is the state error weight matrix, and the weight coefficients in the state error weight matrix corresponding to the battery predicted state of charge are functions of the battery predicted state of charge, so that the weight of the battery state of charge deviation cost is determined based on the battery predicted state of charge. Characterizes the control input penalty cost weighted based on the control input weight matrix R; The predicted control torque command for the k-th control cycle; The input weight matrix is ​​used to adjust the penalty intensity of the control torque command. This is the terminal cost term, used to characterize the additional loss in the prediction time domain; To predict the length of the time domain.

[0104] Understandably, the aforementioned cumulative cost function achieves multi-objective optimization by comprehensively considering vehicle trajectory tracking accuracy, control torque command stability, and battery state of charge safety. This ensures that the vehicle can accurately track the target trajectory, suppress control torque fluctuations, protect battery safety, and also takes into account the overall vehicle energy-saving effect, making the control strategy more practical and reliable.

[0105] One possible implementation is a state vector. The specific form can be defined as:

[0106] in, The predicted longitudinal acceleration of the vehicle during the k-th control cycle; The predicted value of the vehicle's lateral acceleration during the k-th control cycle; The predicted value of the vehicle yaw rate in the kth control cycle; , , , These are the predicted wheel speeds of the left front wheel, right front wheel, left rear wheel, and right rear wheel during the k-th control cycle. The predicted state of charge of the battery in the k-th control cycle.

[0107] The state vector represents the predicted state for each control cycle within the prediction time domain. Its physical meaning corresponds to the real-time data type collected by the sensor group, namely the real-time vehicle status information and the real-time battery state of charge. This allows the optimization problem to be solved based on physical quantities of the same dimension as the measured data, ensuring the consistency between subsequent predictions and real data.

[0108] In this embodiment, the vehicle predicted state information and the battery predicted state of charge are determined based on the control torque command variable to be solved, the real-time vehicle state information, and the real-time battery state of charge.

[0109] In some embodiments, based on the control torque command variable to be solved, the real-time vehicle status information, and the real-time battery state of charge, a prediction model is used to determine the vehicle predicted status information and the battery predicted state of charge for the first control cycle in the prediction time domain. Based on the vehicle predicted status information and the battery predicted state of charge for the previous control cycle, the prediction model is used to predict the vehicle predicted status information and the battery predicted state of charge for the next control cycle, until the vehicle predicted status information and the battery predicted state of charge for each control cycle in the prediction time domain are obtained.

[0110] The prediction model includes a vehicle dynamics model and a battery model. The vehicle dynamics model characterizes the periodic changes in the predicted vehicle state information between two adjacent control cycles. The battery model characterizes the periodic changes in the predicted battery state of charge between two adjacent control cycles. Specific embodiments of the vehicle dynamics model and battery model are not shown here and do not affect the logical understanding of the scheme. Specific embodiments can be found in the following detailed embodiments.

[0111] Based on multi-rigid-body vehicle dynamics, and combined with wheel dynamics, tire mechanics, and drive motor output characteristics, a state-space model of the vehicle system in the continuous time domain is constructed, forming a nonlinear state transition equation:

[0112] in, For continuous time The system state vector includes the vehicle's longitudinal acceleration, lateral acceleration, yaw rate, wheel speeds, and battery state of charge. For continuous time The control input vector is the control torque command to be solved. This is a nonlinear state transition function constructed based on vehicle dynamics and battery models, used to describe the dynamic evolution of the system state as the control input changes.

[0113] To adapt to the execution characteristics of discrete control cycles, the Euler method is used to discretize the above continuous-time state transition equations. The Euler method is a first-order numerical discretization method that utilizes the current state variables and the rate of change of state at a fixed step size. The internal approximation recursively derives the state variables for the next time step, and the specific discretization relation satisfies:

[0114] in, The time step for a single control cycle; For the first The system state vector of the control cycle; For the first The rate of change of system state during the control period; For the first The system state vector of the control cycle.

[0115] Substituting the continuous state transition equation into the discretization relation above, we can obtain a discrete-time state recursion model suitable for prediction iteration:

[0116] This discrete iterative model describes the process described by the first... The system state and control torque command of the control cycle are recursively used to obtain the first... The mathematical relationship of the system state during the control cycle. Based on this model, starting from the real-time vehicle state information and the real-time battery state of charge at the current moment, and combining the control torque command variable to be solved, the predicted vehicle state information and the predicted battery state of charge for all future control cycles in the prediction time domain can be obtained through iterative calculation cycle by cycle. This provides a complete sequence of predicted states for the construction and optimization of the cumulative cost function.

[0117] In some embodiments, the weight of the battery state of charge deviation cost in the cumulative cost function is determined based on the predicted battery state of charge. Specifically, the predicted battery state of charge is compared with a preset state of charge, and the weight value is dynamically adjusted according to the comparison result. Unlike traditional control methods that use fixed weights or simple linear adjustments based solely on the current battery charge level, this application introduces a nonlinear weight adjustment mechanism: when the predicted battery state of charge is lower than the preset state of charge, the weight increases significantly, causing the optimization to favor more energy-efficient torque commands; when the predicted battery state of charge is higher than the preset state of charge, the weight is relatively smaller, allowing the optimization to prioritize trajectory tracking accuracy. This adjustable weight design achieves a dynamic balance between driving performance and energy-saving requirements under different battery charge levels.

[0118] As one possible implementation, the weights used when weighting the deviation between the predicted and desired states of charge (SOC) of the battery in the first control cycle are exponentially positively correlated with the absolute value of the deviation between the predicted and desired SOC (see the foregoing embodiments for details). In one possible implementation, the weights for the aforementioned battery SOC deviation costs... It is directly embedded into the state error weight matrix Q of the cumulative cost function.

[0119] For example, the state error weight matrix Q is related to the battery's predicted state of charge. The corresponding diagonal element is set to ,like: Wherein, the q-related values ​​are the corresponding weight coefficients for each physical quantity in the vehicle's predicted state information, while the weights corresponding to the battery's predicted state of charge are determined by... Dynamic determination. In this way, the nonlinear weight adjustment mechanism is seamlessly integrated into the cumulative cost function, enabling the controller to automatically adjust the emphasis on the cost of deviation from the battery's predicted state of charge based on changes in the battery's predicted state of charge during optimization. This achieves a dynamic trade-off between trajectory tracking accuracy, control stability, and battery safety and energy saving within the cumulative cost function.

[0120] As can be seen, step S202 employs a nonlinear model predictive control (nMPC) structure to uniformly model and optimize the entire prediction time domain, rather than performing local optimization only for a single control cycle. Compared to traditional control methods, this step achieves a global trade-off for the next N control cycles by comprehensively considering the vehicle state trajectory error cost, control input penalty cost, and battery state of charge deviation cost. This allows for advance prediction of vehicle driving needs and battery energy state, effectively suppressing abrupt changes in control torque commands while ensuring accurate tracking of the target trajectory and improving driving smoothness. Furthermore, by dynamically adjusting the weight of the battery state of charge deviation cost, priority is given to battery safety and energy saving in low-charge scenarios, and to driving performance in high-charge scenarios. This achieves a multi-objective dynamic trade-off between trajectory tracking accuracy, control stability, battery safety, and vehicle energy consumption, significantly improving the robustness, adaptability, and overall energy efficiency of the control strategy.

[0121] S203. Solve the problem by minimizing the cumulative cost function to obtain the target control torque command sequence corresponding to the predicted time domain.

[0122] The target control torque command sequence includes N control torque commands ordered according to N control cycles.

[0123] For example, Each of them This is the optimized control torque command corresponding to the control cycle.

[0124] Understandably, this application transforms the problem of solving the cumulative cost function into a constrained optimization problem, which is then solved using a numerical optimization algorithm. Specifically, the cumulative cost function is used as the objective function, and the control torque command variable to be solved is... Given the decision variable, and under the premise of satisfying the preset constraints, the optimal solution that minimizes the cumulative cost function is found through iterative optimization.

[0125] As one possible implementation, the process of obtaining the target control torque command sequence includes the following steps: First, based on the real-time vehicle status information and the real-time battery state of charge at the current moment, and combined with the control torque command variable to be solved, the predicted vehicle status information and the predicted battery state of charge for each control cycle in the prediction time domain are calculated iteratively through the prediction model.

[0126] Secondly, the predicted state information is substituted into the cumulative cost function to calculate the cumulative cost function value under different control torque command variable values.

[0127] Then, optimization algorithms (such as quadratic programming, sequential quadratic programming, or interior point method) are used to search the feasible solution space to find the combination of control torque command variables that minimizes the cumulative cost function value.

[0128] Finally, the optimized control torque command variable sequence is output as the target control torque command sequence for subsequent control execution.

[0129] As one possible implementation, when solving for the goal of minimizing the cumulative cost function, at least one of the following preset constraints is included: vehicle state trajectory data constraints, control input data constraints, or battery state of charge constraints.

[0130] The vehicle state trajectory data constraints include: the predicted vehicle state information is greater than or equal to a first trajectory threshold and less than or equal to a second trajectory threshold, where the first trajectory threshold is less than the second trajectory threshold. This constraint is used to ensure that the vehicle's motion state does not exceed the safe operating range, such as limiting the maximum longitudinal speed and maximum yaw rate, to prevent vehicle instability.

[0131] The control input data constraints include: the control torque command is greater than or equal to a first control threshold and less than or equal to a second control threshold, where the first control threshold is less than the second control threshold. This constraint corresponds to the physical performance limits of the drive motor, ensuring that the output torque command is within the motor's executable range, while also considering the battery's instantaneous power output capability.

[0132] For example, the control input data constraint can be represented as: ,in To control the lower limit of the torque command, i.e., the first control threshold, This is to control the upper limit of the torque command, i.e., the second control threshold.

[0133] The battery state of charge (SOC) constraint includes: the predicted SOC is greater than or equal to a first SOC threshold and less than or equal to a second SOC threshold, where the first SOC threshold is less than the second SOC threshold. This constraint is used to protect the battery from overcharging or over-discharging, extend battery life, and ensure that the vehicle has sufficient charge to complete the predetermined trajectory within the predicted time domain.

[0134] For example, the battery state of charge constraint can be expressed as: ,in This is the lower limit of the battery's state of charge, i.e., the first charge threshold. This is the upper limit of the battery's state of charge, i.e., the second charge threshold.

[0135] As one possible implementation, a quadratic programming algorithm is used for solving the problem. First, the cumulative cost function is transformed into the standard form of a quadratic programming algorithm in the prediction time domain: ;

[0136] in: For prediction of the time domain The decision variable vector is composed of the changes in control torque command in each control cycle. Indicates the first The increment of the control torque command relative to the previous cycle; The Hessian matrix (quadratic term matrix) is determined by the control input weight matrix. and identity matrix of order The Kronecker product is obtained, that is This reflects the secondary impact of decision variables on cumulative costs; The gradient vector is a linear term, consisting of the first-order partial derivative of the cumulative cost function with respect to the control torque command, specifically calculated from the deviation between the current state and the reference trajectory. Inequality constraints are used to describe constraints on the magnitude of control torque command, control increment, and vehicle predicted state information (such as battery state of charge constraints, wheel slip ratio constraints, etc.). This represents an equality constraint used to describe the consistency conditions of the system's dynamic model, ensuring that the predicted state satisfies the recursive relationship between the vehicle dynamics model and the battery model.

[0137] By solving the above quadratic programming problem, the optimal control increment sequence is obtained. By combining the control torque commands actually executed in the previous control cycle, the control torque command sequence that minimizes the cumulative cost function can be obtained through accumulation:

[0138] in, This refers to the control torque command actually executed in the previous control cycle.

[0139] One possible implementation is to use a sequential quadratic programming algorithm when the cumulative cost function or constraints exhibit strong nonlinear characteristics. This algorithm transforms the original problem into a series of quadratic programming subproblems that are solved iteratively. In each iteration, a quadratic approximation is performed on the nonlinear objective function and constraints at the current iteration point to obtain the search direction by solving the quadratic programming subproblems. The step size is then determined through line search until convergence to the optimal solution.

[0140] Another possible implementation is to use the interior-point method. The interior-point method introduces obstacle functions into the objective function to handle inequality constraints, transforming the original constrained optimization problem into a series of unconstrained or equality-constrained subproblems. It gradually approximates the optimal solution from within the feasible region and is suitable for solving large-scale optimization problems.

[0141] It can be seen that step S203 adopts a unified solution for the entire prediction time domain, and determines the vehicle prediction state information and battery prediction state of charge corresponding to each control cycle through iterative calculation, thereby obtaining the complete cumulative cost value, rather than performing local optimization only for a single control cycle.

[0142] Compared to traditional control methods that only focus on local optima at the current moment, this step, through global optimization over the next N control cycles, can predict vehicle driving needs and battery energy state in advance. While ensuring the vehicle accurately tracks the target trajectory, it effectively suppresses abrupt changes in control torque commands, improving driving smoothness. Furthermore, by dynamically adjusting the weight of the cost of battery state of charge deviation, it prioritizes battery safety and energy saving in low-charge scenarios and prioritizes driving performance in high-charge scenarios. This achieves a multi-objective dynamic trade-off between trajectory tracking accuracy, control stability, battery safety, and vehicle energy consumption, significantly improving the robustness and adaptability of the control strategy. It provides an advanced torque control solution for distributed drive electric vehicles that balances high performance and high energy efficiency.

[0143] S204. Control the drive motor to run based on the control torque command corresponding to the first control cycle in the target control torque command sequence.

[0144] In one possible implementation, after controlling the drive motor to run based on the control torque command corresponding to the first control cycle in the target control torque command sequence, the real-time vehicle status information and the real-time battery charge status at the time of the first control cycle are obtained, and the second control cycle in the target control torque command sequence is updated to the first control cycle of the new prediction time domain, so as to redetermine the control torque command corresponding to the first control cycle of the new prediction time domain.

[0145] Understandably, employing rolling execution and rolling optimization allows control commands to be adjusted in real time to keep pace with the vehicle's latest state, preventing error accumulation. Simultaneously, executing only the currently optimal command ensures smoother drive motor operation and more economical battery consumption. Furthermore, continuously updating the prediction time domain enables the vehicle to stably track the target trajectory even under complex driving conditions, resulting in higher safety and reliability.

[0146] In some embodiments, when constructing the cumulative cost function, the cost is essentially a quantitative assessment of the deviation between the predicted data and the expected data. Specifically, the vehicle state trajectory error cost in the cumulative cost function is used to measure the degree of deviation between the predicted vehicle state information and the expected state information. Therefore, the accurate acquisition of the expected state information directly determines the rationality and effectiveness of the cumulative cost function in evaluating the control effect.

[0147] In one possible implementation, obtaining the desired state information includes: generating the vehicle's target trajectory based on the driver's operating intention or autonomous driving trajectory planning; converting the target trajectory into desired state information for each control cycle in the prediction time domain; the desired state information includes at least one of the following: desired longitudinal velocity, desired lateral velocity, desired yaw rate, desired longitudinal acceleration, desired lateral acceleration, and desired position.

[0148] In one implementation, in manual driving mode, the controller collects driver operation signals such as steering wheel angle and accelerator / brake pedal opening, and combines them with the current vehicle state to analyze the driver's steering intention and acceleration / deceleration intention in real time, generating the vehicle's target trajectory, including the desired longitudinal speed and desired yaw angle, to provide reference input for subsequent control.

[0149] In one implementation, in autonomous driving mode, the path planning module generates a target driving trajectory that meets safety and efficiency requirements based on a high-precision map, real-time traffic information, and the vehicle's current location. This trajectory consists of a series of time-series location points and is used to guide the vehicle to travel along a predetermined path.

[0150] As one possible implementation, based on the generated target trajectory, the target trajectory is converted into the expected state information of each control cycle in the prediction time domain.

[0151] Based on the target trajectory, the controller converts the continuous trajectory into discrete desired state information for each control cycle in the prediction time domain using numerical differentiation or interpolation methods. Specifically, the first-order derivative of the target position trajectory yields the desired longitudinal and lateral velocities; the second-order derivative yields the desired longitudinal and lateral accelerations; and the derivative of the target yaw angle trajectory yields the desired yaw rate. This ultimately forms the desired state vector for each control cycle in the prediction time domain, which is used to calculate the vehicle state trajectory error cost in the cumulative cost function.

[0152] As one possible implementation, the desired state information includes at least one of the following: desired longitudinal velocity, desired lateral velocity, desired yaw rate, desired longitudinal acceleration, desired lateral acceleration, and desired position.

[0153] Specifically, the expected longitudinal velocity is used to guide the controller in adjusting the drive torque so that the vehicle's actual longitudinal velocity tracks the target velocity, ensuring driving efficiency; the expected lateral velocity is used to constrain the vehicle's lateral movement to avoid instability risks such as sideslip and fishtailing; the expected yaw rate is used to correct the vehicle's steering posture, improving path tracking accuracy and steering stability; the expected longitudinal acceleration is used to predict the vehicle's acceleration and deceleration needs, adjust the torque command in advance, and improve driving smoothness; the expected lateral acceleration is used to constrain the vehicle's roll state to ensure smooth steering; and the expected position serves as the core reference for path tracking, ensuring that the vehicle travels along the predetermined trajectory.

[0154] In some embodiments, the prediction model mentioned in step S202 includes a vehicle dynamics model and a battery model, which together form the basis for iteratively predicting the state information of future control cycles. Functionally, the vehicle dynamics model describes the dynamic evolution of the vehicle's motion state as a function of the control torque command, while the battery model describes the energy transfer relationship as the battery's state of charge changes with the motor's power demand. From an implementation perspective, after discretization, the vehicle dynamics model characterizes the recursive relationship of the predicted vehicle state information between two adjacent control cycles, and the battery model characterizes the recursive relationship of the predicted battery state of charge between two adjacent control cycles. Together, they form the core computational basis for iterative prediction.

[0155] The following will provide a detailed explanation of the construction methods, mathematical expressions, and discretization methods of the vehicle dynamics model and the battery model, using specific embodiments as examples.

[0156] As one possible implementation, the vehicle dynamics model is constructed using a three-degree-of-freedom model, including longitudinal, lateral, and yaw degrees of freedom. This model describes the mapping relationship between the vehicle's motion state and the control input through the following dynamic equations:

[0157] in, The total mass of the vehicle; Let be the moment of inertia of the vehicle about its vertical axis; , These are the longitudinal and lateral velocities of the vehicle's center of gravity, respectively. The vehicle's yaw rate; , These are the longitudinal forces on the front and rear axle wheels, respectively; , These are the lateral forces on the front and rear axle wheels, respectively; , These are the steering angles of the front and rear wheels, respectively. Wheelbase; , These are the distances from the vehicle's center of gravity to the front and rear axles, respectively. air density; This refers to the air drag coefficient; The frontal area of ​​the vehicle.

[0158] The aforementioned dynamic equations establish a nonlinear dynamic mapping relationship between the control torque command (manifested through tire force) and the vehicle's motion state (longitudinal velocity, lateral velocity, and yaw rate), providing a theoretical basis for state prediction.

[0159] As another possible implementation, please refer to the appendix. Figure 3 The two-degree-of-freedom model shown is Figure 3 It includes: a global coordinate system with X and Y axes, used to describe the vehicle's absolute position in the plane; and longitudinal velocity. Let be the longitudinal velocity of the vehicle's center of mass along the u-axis of the vehicle's coordinate system, and be the lateral velocity. The lateral velocity of the vehicle's center of gravity along the v-axis of the vehicle's coordinate system; the sideslip angle of the center of gravity. Yaw angle is the angle between the direction of the vehicle's center of gravity velocity and the longitudinal direction of the vehicle body. The yaw rate is the angle between the vehicle coordinate system and the X-axis of the global coordinate system. Angular velocity of the vehicle about its center of mass; distance from the center of mass to the front axle. and distance from center of mass to rear axle These are the distances from the vehicle's center of gravity to the front and rear axles, respectively; the front wheel steering angle. The steering angle of the front wheels relative to the vehicle body in the longitudinal direction; the longitudinal force of the front wheels. Rear wheel longitudinal force Generated by driving / braking torque, directed along the tire plane; front wheel lateral force Rear wheel lateral force Generated by the contact between the tire and the ground, it provides lateral support during steering; front wheel slip angle. Rear wheel slip angle Let be the angle between the tire velocity direction and the tire plane. This two-degree-of-freedom model clearly describes the relationship between the lateral and yaw motions of the vehicle under steering conditions, providing a simplified and efficient prediction basis for the controller.

[0160] As one possible implementation method, please refer to Figure 4 The battery model is constructed using a first-order RC equivalent circuit model (such as the Thevenin model). This model consists of an open-circuit voltage source. Battery current Series ohmic internal resistance and polarized RC branch (polarization resistor) and polarization capacitors The system consists of three parts, which can effectively reflect the steady-state output characteristics and transient response behavior of the battery. Its dynamic behavior is described by the following three sets of equations: The battery terminal voltage equation is:

[0161] in, This refers to the output voltage at the battery terminal. The open-circuit voltage varies with the state of charge; The internal resistance is ohmic; This is the battery current (positive during discharge). This is the polarization voltage.

[0162] The voltage dynamic behavior of the polarization branch satisfies a first-order differential equation:

[0163] in, Polarization resistor; The polarization capacitor is represented by this equation, which describes the delayed response characteristics of the battery.

[0164] The relationship between the battery's state of charge and time is as follows:

[0165] in: The coulomb efficiency is used to account for energy loss during charging and discharging. This refers to the battery's rated capacity. It represents the ratio of remaining power to total capacity.

[0166] There is a defined mapping relationship between battery current and the control torque command of the drive motor: when the motor outputs drive torque, the battery discharges to provide electrical energy; when the motor performs energy recovery, the battery charges to absorb electrical energy. Through this mapping relationship, the control torque command variable to be solved can be converted into the input current of the battery model, thereby realizing iterative prediction of the battery state of charge.

[0167] In some implementations, to adapt to the execution characteristics of the discrete control cycle of the controller, the Euler method is used to discretize the above continuous-time model. The Euler method is a first-order numerical discretization method that uses the state variables at the current moment and the rate of change of the state to approximately recursively deduce the state variables at the next moment within a fixed step size.

[0168] As one possible implementation, for the vehicle dynamics model, discretization forms a corresponding state recursive relationship, which allows the vehicle motion state for the next control cycle to be directly calculated based on the vehicle motion state and control torque command of the current control cycle. The vehicle motion state includes longitudinal velocity, lateral velocity, yaw rate, and wheel speeds.

[0169] As one possible implementation, for the battery model, discretization also forms a corresponding state recursive relationship, which can calculate the battery state of charge and polarization voltage for the next control cycle based on the battery state of charge, polarization voltage, and battery current in the current control cycle. The battery current is determined by the control torque command through the motor power conversion relationship.

[0170] As one possible implementation, the discretization relationships between the vehicle dynamics model and the battery model are combined to obtain a unified discrete-time state recursive model (as described above). This model can take the current cycle's system state and control torque command as input and directly output the overall system state for the next cycle. The system state includes both vehicle motion state information and battery state information.

[0171] As one possible implementation, based on this discrete recursive model, starting from the latest real-time vehicle state information and battery state of charge, and combining all control torque command variables to be solved in the prediction time domain, the system state corresponding to each future control cycle is calculated iteratively cycle by cycle to gradually derive the system state corresponding to each future control cycle. Through cycle-by-cycle iteration, the predicted vehicle state information and predicted battery state of charge corresponding to all control cycles in the prediction time domain are finally obtained, providing complete and continuous predicted state data for the subsequent construction and optimization of the cumulative cost function.

[0172] The foregoing mainly describes the solutions provided by the embodiments of this application from a methodological perspective. To achieve the above functions, the vehicle torque control device includes hardware structures and / or software modules corresponding to the execution of each function. Those skilled in the art should readily recognize that, based on the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein, this application can be implemented in hardware or a combination of hardware and computer software. Whether a function is executed in hardware or by computer software driving hardware depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0173] This application embodiment can, based on the above method, exemplarily divide a vehicle torque control device or electronic device into functional modules. For example, a vehicle battery level testing device or electronic device may include functional modules corresponding to each functional division, or two or more functions may be integrated into a single processing module. The integrated module can be implemented in hardware or as a software functional module. It should be noted that the module division in this application embodiment is illustrative and only represents one logical functional division; in actual implementation, other division methods may be used.

[0174] Please see Figure 5 The vehicle torque control device provided in this application embodiment includes: an acquisition module 501, a construction module 502, a solution module 503, and a control module 504.

[0175] The acquisition module 501 is used to acquire real-time vehicle status information and real-time battery charge status.

[0176] The construction module 502 is used to construct a cumulative cost function in the prediction time domain based on the real-time vehicle status information, the real-time battery state of charge, and the control torque command variable to be solved. The prediction time domain represents N control cycles starting from the current moment, where N is a positive integer. The cumulative cost function includes vehicle state trajectory error cost, control input penalty cost, and battery state of charge deviation cost. The vehicle state trajectory error cost is used to evaluate the degree of deviation between the predicted and desired vehicle state information. The control input penalty cost is used to suppress the fluctuation amplitude of the control torque command. The battery state of charge deviation cost is used to evaluate the deviation between the predicted and desired battery state of charge. The weight of the battery state of charge deviation cost in the cumulative cost function is determined based on the predicted battery state of charge. The predicted vehicle state information and the predicted battery state of charge are determined based on the control torque command variable to be solved, the real-time vehicle status information, and the real-time battery state of charge.

[0177] The solver module 503 is used to solve for the target control torque command sequence corresponding to the predicted time domain with the objective of minimizing the cumulative cost function; the target control torque command sequence includes N control torque commands sorted according to N control cycles.

[0178] The control module 504 is used to control the operation of the drive motor based on the control torque command corresponding to the first control cycle in the target control torque command sequence.

[0179] In some embodiments, the construction module 502 is specifically used to construct the cumulative cost function in the prediction time domain in the following ways: weighting the deviation between the predicted vehicle state information and the desired state information in the first control cycle to determine the vehicle state trajectory error cost in the first control cycle; the first control cycle is any control cycle in the prediction time domain; weighting the deviation between the predicted battery state of charge and the desired battery state of charge in the first control cycle to determine the battery state of charge deviation cost in the first control cycle; weighting the amplitude of the control torque command in the first control cycle to determine the control input penalty cost in the first control cycle; summing the vehicle state trajectory error cost, the battery state of charge deviation cost, and the control input penalty cost in the first control cycle to determine the periodic cost function; and determining the cumulative cost function in the prediction time domain based on the periodic cost function of the first control cycle.

[0180] In some embodiments, the weights used when weighting the deviation between the predicted state of charge and the desired state of charge of the battery in the first control cycle are exponentially positively correlated with the absolute value of the deviation between the predicted state of charge and the desired state of charge.

[0181] In some embodiments, the weights used when weighting the deviation between the predicted state of charge and the desired state of charge of the battery in the first control cycle satisfy the following relationship:

[0182] in, The weighted parameter represents the degree of deviation between the predicted state of charge and the expected state of charge of the battery in the k-th control cycle. These are the basic weighting coefficients, used to set the basic magnitude of the weights; This is an adjustment factor used to adjust the rate at which the weight changes with the deviation of the battery's state of charge. Predict the state of charge of the battery in the k-th control cycle; This is the preset state of charge.

[0183] In some embodiments, the construction module 502 is specifically used to determine the cumulative cost function in the prediction time domain based on the periodic cost function of the first control period, including: summing up the periodic cost functions corresponding to all first control periods in the prediction time domain and adding the terminal cost term to determine the cumulative cost function in the prediction time domain, wherein the terminal cost term is used to characterize the total additional loss in the prediction time domain.

[0184] In some embodiments, the cumulative cost function in the prediction time domain satisfies the following relationship:

[0185] in, This represents the cumulative cost function value within the prediction time domain. The state vector for the k-th control cycle includes vehicle predicted state information and battery predicted state of charge. The reference state vector includes desired state information and desired state of charge. The vehicle state trajectory error cost and battery state of charge deviation cost are represented by weighting based on the state error weight matrix Q; Q is the state error weight matrix, and the weight coefficients in the state error weight matrix corresponding to the battery predicted state of charge are functions of the battery predicted state of charge, so that the weight of the battery state of charge deviation cost is determined based on the battery predicted state of charge. Characterizes the control input penalty cost weighted based on the control input weight matrix R; The predicted control torque command for the k-th control cycle; The input weight matrix is ​​used to adjust the penalty intensity of the control torque command. This is the terminal cost term, used to characterize the additional loss in the prediction time domain; To predict the length of the time domain.

[0186] In some embodiments, the solving module 503 is further configured to, after controlling the drive motor to run based on the control torque command corresponding to the first control cycle in the target control torque command sequence, obtain the real-time vehicle status information and the real-time battery charge status at the time of the first control cycle, and update the second control cycle in the target control torque command sequence to the first control cycle of the new prediction time domain, so as to redetermine the control torque command corresponding to the first control cycle of the new prediction time domain.

[0187] In some embodiments, the construction module 502 is specifically used to determine the vehicle predicted state information and the battery predicted state of charge in the following manner: based on the control torque command variable to be solved, the real-time vehicle state information, and the real-time battery state of charge, the vehicle predicted state information and the battery predicted state of charge for the first control cycle in the prediction time domain are determined through a prediction model; based on the vehicle predicted state information and the battery predicted state of charge for the previous control cycle, the vehicle predicted state information and the battery predicted state of charge for the next control cycle are predicted through a prediction model, until the vehicle predicted state information and the battery predicted state of charge for each control cycle in the prediction time domain are obtained; the prediction model includes a vehicle dynamics model and a battery model; the vehicle dynamics model is used to characterize the periodic change relationship of the vehicle predicted state information between two adjacent control cycles; the battery model is used to characterize the periodic change relationship of the battery predicted state of charge between two adjacent control cycles.

[0188] In some embodiments, the vehicle dynamics model is determined based on a three-degree-of-freedom model or a two-degree-of-freedom model.

[0189] In some embodiments, the battery model is determined based on a first-order RC equivalent circuit model.

[0190] In some embodiments, the construction module 502 is further configured to, when solving for the objective of minimizing the cumulative cost function, include at least one of the following preset constraints: vehicle state trajectory data constraints, control input data constraints, or battery state of charge constraints; wherein, the vehicle state trajectory data constraints include: the predicted vehicle state information is greater than or equal to a first trajectory threshold and less than or equal to a second trajectory threshold, and the first trajectory threshold is less than the second trajectory threshold; the control input data constraints include: the control torque command is greater than or equal to a first control threshold and less than or equal to a second control threshold, and the first control threshold is less than the second control threshold; the battery state of charge constraints include: the predicted battery state of charge is greater than or equal to a first charge threshold and less than or equal to a second charge threshold, and the first charge threshold is less than the second charge threshold. In some embodiments, the real-time vehicle status information includes at least one of the following: vehicle longitudinal acceleration, vehicle lateral acceleration, vehicle longitudinal velocity, vehicle lateral velocity, vehicle yaw rate, and wheel speed of each wheel.

[0191] like Figure 6 As shown, the electronic device 600 provided in this application embodiment includes, but is not limited to, a processor 601 and a memory 602.

[0192] The memory 602 described above is used to store the executable instructions of the processor 601. It is understood that the processor 601 is configured to execute instructions to implement the methods in the above embodiments.

[0193] It should be noted that those skilled in the art will understand that Figure 6 The electronic device structure shown does not constitute a limitation on electronic device 600; electronic device may include, but is not limited to, other electronic devices. Figure 6 This may indicate more or fewer components, or combinations of certain components, or different component arrangements.

[0194] Processor 601 is the control center of electronic device 600. It connects various parts of the electronic device via various interfaces and lines. By running or executing software programs and / or modules stored in memory 602, and by calling data stored in memory 602, it performs various functions and processes data of electronic device 600, thereby providing overall monitoring of electronic device 600. Processor 601 may include one or more processing units. Optionally, processor 601 may integrate an application processor and a modem processor. The application processor mainly handles the operating system, user interface, and applications, while the modem processor mainly handles wireless communication. It is understood that the modem processor may not be integrated into processor 601.

[0195] The memory 602 can be used to store software programs and various data. The memory 602 may primarily include a program storage area and a data storage area. The program storage area may store the operating system, application programs required by at least one functional module (such as a determination unit, processing unit, etc.), etc. Furthermore, the memory 602 may include high-speed random access memory, and may also include non-volatile memory, such as at least one disk storage device, flash memory device, or other volatile solid-state storage device.

[0196] In an exemplary embodiment, a vehicle is also provided, including the electronic equipment described above.

[0197] In an exemplary embodiment, a computer-readable storage medium including instructions is also provided, such as a memory 602 including instructions, which can be executed by a processor 601 of an electronic device 600 to implement the methods in the above embodiments.

[0198] In actual implementation, Figure 5 The functions of each module can be provided by Figure 6 The processor 601 calls the computer program stored in the memory 602 to implement the process. The specific execution process can be found in the description of the method section in the previous embodiment, and will not be repeated here.

[0199] Optionally, the computer-readable storage medium may be a non-transitory computer-readable storage medium, such as a read-only memory (ROM), random access memory (RAM), CD-ROM, magnetic tape, floppy disk, and optical data storage device.

[0200] In an exemplary embodiment, this application also provides a computer program product including one or more instructions, which can be executed by the processor 601 of the electronic device 600 to perform the methods described above.

[0201] It should be noted that when one or more instructions in the computer-readable storage medium or computer program product are executed by the processor of an electronic device, they implement the various processes of the above method embodiments and achieve the same technical effect as the above method. To avoid repetition, they will not be described again here.

[0202] Through the above description of the embodiments, those skilled in the art can clearly understand that, for the sake of convenience and brevity, only the division of the above functional modules is used as an example. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above.

[0203] In the several embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of modules or units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another apparatus, or some features may be ignored or not executed. Furthermore, the mutual coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0204] The units described as separate components may or may not be physically separate. A component shown as a unit can be one or more physical units; that is, it can be located in one place or distributed in multiple different locations. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0205] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0206] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a readable storage medium. Based on this understanding, the technical solutions of the embodiments of this application, essentially, or the parts that contribute to the prior art, or all or part of the technical solutions, can be embodied in the form of a software product. This software product is stored in a storage medium and includes several instructions to cause a device (which may be a microcontroller, chip, etc.) or processor to execute all or part of the steps of the methods of the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, ROM, RAM, magnetic disks, or optical disks.

[0207] The above embodiments are merely preferred embodiments provided to fully illustrate this application, and the scope of protection of this application is not limited thereto. Equivalent substitutions or modifications made by those skilled in the art based on this application are all within the scope of protection of this application.

Claims

1. A vehicle torque control method, characterized in that, Applied to a vehicle, the vehicle including a drive motor, the method includes: Obtain real-time vehicle status information and real-time battery state of charge; Based on the real-time vehicle status information, the real-time battery state of charge, and the control torque command variable to be solved, a cumulative cost function in the prediction time domain is constructed; wherein, the prediction time domain is used to represent N control cycles starting from the current moment, and N is a positive integer; The cumulative cost function includes vehicle state trajectory error cost, control input penalty cost, and battery state of charge deviation cost. The vehicle state trajectory error cost is used to evaluate the degree of deviation between the predicted and desired vehicle state information. The control input penalty cost is used to suppress the fluctuation amplitude of the control torque command. The battery state of charge deviation cost is used to evaluate the deviation between the predicted and desired battery state of charge. The weight of the battery state of charge deviation cost in the cumulative cost function is determined based on the predicted battery state of charge. The predicted vehicle state information and the predicted battery state of charge are determined based on the control torque command variable to be solved, the real-time vehicle state information, and the real-time battery state of charge. The objective is to minimize the cumulative cost function to obtain the target control torque command sequence corresponding to the predicted time domain; the target control torque command sequence includes N control torque commands sorted according to the N control cycles; The drive motor is controlled to operate based on the control torque command corresponding to the first control cycle in the target control torque command sequence.

2. The method according to claim 1, characterized in that, The cumulative cost function in the prediction time domain is constructed in the following way: The deviation between the predicted vehicle state information and the expected state information in the first control cycle is weighted to determine the vehicle state trajectory error cost in the first control cycle; the first control cycle is any control cycle within the prediction time domain. The deviation between the predicted state of charge and the expected state of charge of the battery in the first control cycle is weighted to determine the battery state of charge deviation cost in the first control cycle. The amplitude of the control torque command in the first control cycle is weighted to determine the control input penalty cost of the first control cycle; In the first control cycle, the vehicle state trajectory error cost, battery state of charge deviation cost, and control input penalty cost are summed to determine the cycle cost function; based on the cycle cost function of the first control cycle, the cumulative cost function in the prediction time domain is determined.

3. The method according to claim 2, characterized in that, The weights used when weighting the deviation between the predicted state of charge and the desired state of charge of the battery in the first control cycle are exponentially positively correlated with the absolute value of the deviation between the predicted state of charge and the desired state of charge.

4. The method according to claim 3, characterized in that, The weights used when weighting the deviation between the predicted state of charge and the desired state of charge of the battery in the first control cycle satisfy the following relationship: in, The weighting coefficient represents the degree of deviation between the predicted state of charge and the expected state of charge of the battery in the k-th control cycle. These are the basic weighting coefficients, used to set the basic magnitude of the weights; This is an adjustment factor used to adjust the rate at which the weight changes with the deviation of the battery's state of charge. Predict the state of charge of the battery in the k-th control cycle; This is the preset state of charge.

5. The method according to claim 2, characterized in that, The determination of the cumulative cost function in the prediction time domain based on the periodic cost function of the first control period includes: The cumulative cost function corresponding to all the first control cycles within the prediction time domain is summed and the terminal cost term is added to determine the cumulative cost function within the prediction time domain. The terminal cost term is used to characterize the total additional loss in the prediction time domain.

6. The method according to claim 1, characterized in that, The cumulative cost function in the prediction time domain satisfies the following relationship: in, This represents the cumulative cost function in the prediction time domain. The state vector for the k-th control cycle includes the vehicle predicted state information and the battery predicted state of charge. The reference state vector includes desired state information and desired state of charge. The vehicle state trajectory error cost and the battery state of charge deviation cost are represented by weighting based on the state error weight matrix Q; Q is the state error weight matrix, and the weight coefficients in the state error weight matrix corresponding to the battery predicted state of charge are functions of the battery predicted state of charge, so that the weight of the battery state of charge deviation cost is determined based on the battery predicted state of charge. The control input penalty cost is represented by a weighting based on the control input weight matrix R; The predicted control torque command for the k-th control cycle; The input weight matrix is ​​used to adjust the penalty intensity of the control torque command. This is the terminal cost term, used to characterize the additional loss in the prediction time domain; To predict the length of the time domain.

7. The method according to claim 1, characterized in that, The method further includes: After controlling the drive motor to operate based on the control torque command corresponding to the first control cycle in the target control torque command sequence, the real-time vehicle status information and the real-time battery charge status at the time of the first control cycle are obtained, and the second control cycle in the target control torque command sequence is updated to the first control cycle of the new prediction time domain, so as to redetermine the control torque command corresponding to the first control cycle of the new prediction time domain.

8. The method according to claim 1, characterized in that, The vehicle predicted state information and the battery predicted state of charge are determined in the following manner: Based on the control torque command variable to be solved, the real-time vehicle status information, and the real-time battery state of charge, the predicted vehicle status information and the predicted battery state of charge for the first control cycle in the prediction time domain are determined by the prediction model. Based on the vehicle predicted state information and battery predicted state of charge of the previous control cycle, the vehicle predicted state information and battery predicted state of charge of the next control cycle are predicted by the prediction model until the vehicle predicted state information and battery predicted state of charge of the vehicle in each control cycle in the prediction time domain are obtained; the prediction model includes a vehicle dynamics model and a battery model; the vehicle dynamics model is used to characterize the periodic change relationship of the vehicle predicted state information between two adjacent control cycles. The battery model is used to characterize the periodic change relationship of the predicted state of charge of the battery between two adjacent control cycles.

9. The method according to claim 1, characterized in that, When solving the problem with the objective of minimizing the cumulative cost function, at least one of the following preset constraints is included: Vehicle state trajectory data constraints, control input data constraints, or battery state of charge constraints; The vehicle state trajectory data constraints include: the vehicle predicted state information is greater than or equal to a first trajectory threshold and less than or equal to a second trajectory threshold, wherein the first trajectory threshold is less than the second trajectory threshold; The control input data constraints include: the control torque command is greater than or equal to a first control threshold and less than or equal to a second control threshold, wherein the first control threshold is less than the second control threshold; The battery state of charge constraint includes: the predicted state of charge of the battery is greater than or equal to a first charge threshold and less than or equal to a second charge threshold, wherein the first charge threshold is less than the second charge threshold.

10. The method according to claim 1, characterized in that, The real-time vehicle status information includes at least one of the following: vehicle longitudinal acceleration, vehicle lateral acceleration, vehicle longitudinal velocity, vehicle lateral velocity, vehicle yaw rate, and wheel speed of each wheel.

11. A vehicle torque control device, characterized in that, The vehicle torque control device includes: an acquisition module, a construction module, a solution module, and a control module; The acquisition module is used to: acquire real-time vehicle status information and real-time battery state of charge; The construction module is used to: construct a cumulative cost function in the prediction time domain based on the vehicle's real-time state information, the battery's real-time state of charge, and the control torque command variable to be solved; where the prediction time domain represents N control cycles starting from the current moment, and N is a positive integer; the cumulative cost function includes vehicle state trajectory error cost, control input penalty cost, and battery state of charge deviation cost; the vehicle state trajectory error cost is used to evaluate the degree of deviation between the predicted and desired state information of the vehicle; the control input penalty cost is used to suppress the fluctuation amplitude of the control torque command; the battery state of charge deviation cost is used to evaluate the deviation between the predicted and desired state of charge of the battery; the weight of the battery state of charge deviation cost in the cumulative cost function is determined based on the predicted state of charge of the battery; the vehicle's predicted state information and the battery's predicted state of charge are determined based on the control torque command variable to be solved, the vehicle's real-time state information, and the battery's real-time state of charge; The solution module is used to: solve for the target control torque command sequence corresponding to the predicted time domain with the objective of minimizing the cumulative cost function; the target control torque command sequence includes N control torque commands sorted according to N control cycles; The control module is used to control the drive motor based on the control torque command corresponding to the first control cycle in the target control torque command sequence.

12. A vehicle, characterized in that, The vehicle includes a processor that stores a computer program; when the computer program is executed by the processor, it implements the method as described in any one of claims 1-10.

Images