A sensorless full-speed control method for an in-wheel permanent magnet motor

Abstract

The application relates to a position sensor-free full-speed control method for an inlaid permanent magnet wheel hub motor. The method comprises the following steps: obtaining an operation signal of a target motor; determining a current operation condition of the target motor according to the operation signal and a preset operation condition judgment model; the operation condition comprises a zero-low-speed condition and a medium-high-speed condition; determining a rotor angle estimation position of the target motor by using a corresponding operation mode according to the operation condition; and controlling the operation of the target motor according to the rotor angle estimation position. The method can realize position sensor-free full-speed control of the wheel hub motor under different operation conditions.

Description

Technical Field

[0001] This application relates to the field of hub motor control technology, and in particular to a motor control method, device, equipment and storage medium. Background Technology

[0002] In recent years, hub motors have been widely used due to their advantages such as simple structure, reliable operation, low noise, no need for position sensors, and high speed limits. Embedded permanent magnet hub motors are considered the most promising hub motors. The biggest characteristic of hub motors is that each wheel motor can operate independently and in coordination, so they can be applied to the field of electric vehicles to ensure safe driving.

[0003] To ensure the normal operation of embedded permanent magnet hub motors, it is urgent to solve the problem of providing a sensorless full-speed control method for hub motors under different operating conditions. Summary of the Invention

[0004] Therefore, it is necessary to provide a motor control method, device, equipment, and storage medium that can achieve full-speed control of hub motors without position sensors under different working conditions, in order to address the above-mentioned technical problems.

[0005] Firstly, this application provides a motor control method. The method includes:

[0006] Acquire the operating signal of the target motor;

[0007] The current operating condition of the target motor is determined based on the operating signal and the preset operating condition judgment model; the operating condition includes zero low speed condition and medium high speed condition;

[0008] Based on the operating conditions, the estimated position of the rotor angle of the target motor is determined using the corresponding calculation method;

[0009] The target motor is controlled to operate based on the estimated position of the rotor angle.

[0010] In one embodiment, the estimated rotor angle position of the target motor is determined by a corresponding calculation method based on the operating condition, including: determining the estimated rotor angle position based on the high-frequency pulse voltage injection method when the operating condition is a zero-speed condition; and determining the estimated rotor angle position based on the sliding film observation method when the operating condition is a medium-speed condition.

[0011] In one embodiment, determining the estimated rotor angle position based on the high-frequency pulse voltage injection method includes: obtaining the high-frequency response current signal of the target motor in the synchronous rotating coordinate system when a pulse high-frequency voltage signal is injected into the target motor; modulating the high-frequency response current signal of the q-axis corresponding to the high-frequency response current signal in the synchronous rotating coordinate system and then performing low-pass filtering to obtain the processed high-frequency response current signal of the q-axis; inputting the processed high-frequency response current signal of the q-axis to the position observer and controlling the processed high-frequency response current signal of the q-axis to be zero to obtain the estimated rotor angle position.

[0012] In one embodiment, determining the estimated rotor angle position based on the sliding diaphragm observation method includes: determining the parameters of the sliding diaphragm observer according to the current equation of the target motor, the parameters including the sliding diaphragm surface and the sliding diaphragm control law; determining the estimated extended back electromotive force of the target motor according to the sliding diaphragm observer after determining the parameters; obtaining the initial estimated rotor angle position by taking the arctangent of the estimated extended back electromotive force; and adding the initial estimated rotor angle position to the target angle compensation value to obtain the estimated rotor angle position.

[0013] In one embodiment, controlling the target motor to operate based on the estimated rotor angle position includes: performing differential processing on the estimated rotor angle position to obtain the estimated rotor speed of the target motor; and controlling the target motor to operate based on a PI dual closed-loop control method according to the estimated rotor angle position and the estimated rotor speed.

[0014] In one embodiment, determining the current operating condition of the target motor based on the operating signal and a preset operating condition judgment model includes: inputting the operating signal into the preset operating condition judgment model to obtain the initial state result output by the preset operating condition judgment model; if the initial state result is less than a preset state threshold, then determining the current operating condition of the target motor as the zero-low speed condition; if the initial state result is greater than the preset state threshold, then determining the current operating condition of the target motor as the medium-high speed condition.

[0015] In one embodiment, the preset operating condition judgment model is obtained by pre-training a fully connected neural network model. The training process of the fully connected neural network model includes: acquiring historical operating signals during the driving process of the target vehicle; normalizing the historical operating signals to obtain an initial training dataset; removing outliers from the initial training dataset to obtain a target training dataset; training the fully connected neural network model based on the target training dataset; and using the trained fully connected neural network model as the preset operating condition judgment model.

[0016] In one embodiment, the target motor is a hub motor used in a vehicle, and the operating signals include the vehicle speed, the torque of the target motor, the tire swerve angle of the vehicle, and the yaw rate of the vehicle.

[0017] Secondly, this application also provides a motor control device. The device includes:

[0018] The first acquisition module is used to acquire the operating signal of the target motor;

[0019] The first determining module is used to determine the current operating condition of the target motor based on the operating signal and the preset operating condition judgment model; the operating condition includes zero low speed condition and medium high speed condition;

[0020] The second determining module is used to determine the estimated position of the rotor angle of the target motor by using the corresponding calculation method according to the operating condition.

[0021] The motor control module is used to control the operation of the target motor based on the estimated position of the rotor angle.

[0022] In one embodiment, the second determining module is specifically used to: determine the estimated rotor angle position based on the high-frequency pulse voltage injection method when the operating condition is a zero-speed condition; and determine the estimated rotor angle position based on the sliding film observation method when the operating condition is a medium-speed condition.

[0023] In one embodiment, the second determining module is specifically configured to: obtain the high-frequency response current signal of the target motor in the synchronous rotating coordinate system when the target motor is injected with a pulsating high-frequency voltage signal; modulate the high-frequency response current signal of the q-axis corresponding to the high-frequency response current signal in the synchronous rotating coordinate system and then perform low-pass filtering to obtain the processed high-frequency response current signal of the q-axis; input the processed high-frequency response current signal of the q-axis to the position observer and control the processed high-frequency response current signal of the q-axis to be zero to obtain the estimated position of the rotor angle.

[0024] In one embodiment, the second determining module is specifically used to: determine the parameters of the sliding diaphragm observer based on the current equation of the target motor, the parameters including the sliding diaphragm surface and the sliding diaphragm control law; determine the estimated extended back electromotive force of the target motor based on the sliding diaphragm observer after the parameters are determined; obtain the initial rotor angle estimated position by taking the arctangent of the estimated extended back electromotive force; and add the initial rotor angle estimated position to the target angle compensation value to obtain the rotor angle estimated position.

[0025] In one embodiment, the motor control module is specifically used to: perform differential processing on the estimated rotor angle position to obtain the estimated rotor speed of the target motor; and control the operation of the target motor based on the estimated rotor angle position and the estimated rotor speed using a PI dual closed-loop control method.

[0026] In one embodiment, the first determining module is specifically used to: input the operating signal into the preset operating condition judgment model to obtain the initial state result output by the preset operating condition judgment model; if the initial state result is less than the preset state threshold, then determine that the current operating condition of the target motor is the zero low-speed condition; if the initial state result is greater than the preset state threshold, then determine that the current operating condition of the target motor is the medium-high speed condition.

[0027] In one embodiment, the preset working condition judgment model is obtained by pre-training a fully connected neural network model, and the device further includes:

[0028] The second acquisition module is used to acquire historical operating signals of the target vehicle during its driving process;

[0029] The processing module is used to normalize the historical running signal to obtain the initial training dataset;

[0030] The outlier removal module is used to remove outliers from the initial training dataset to obtain the target training dataset.

[0031] The training module is used to train the fully connected neural network model based on the target training dataset, and to use the trained fully connected neural network model as the preset working condition judgment model.

[0032] In one embodiment, the target motor is a hub motor used in a vehicle, and the operating signals include the vehicle speed, the torque of the target motor, the tire swerve angle of the vehicle, and the yaw rate of the vehicle.

[0033] Thirdly, this application also provides a computer device, including a memory and a processor, the memory storing a computer program, the processor executing the computer program to implement the steps of the method described in any of the first aspects above.

[0034] Fourthly, this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method described in any one of the first aspects above.

[0035] The aforementioned motor control method, device, equipment, and storage medium acquire the operating signal of the target motor and determine the current operating condition of the target motor based on the operating signal and a preset operating condition judgment model. This operating condition includes a zero-low speed condition and a medium-high speed condition. Then, based on the operating condition, a corresponding calculation method is used to determine the estimated rotor angle position of the target motor, and the target motor is controlled based on this estimated rotor angle position. In this way, the current real-time operating condition of the target motor can be determined based on the real-time operating signal of the target motor and the preset operating condition judgment model. Since the estimation accuracy of the motor rotor angle varies under different operating conditions, and the accuracy of the rotor angle estimation affects the control accuracy of the target motor, a corresponding calculation method is used to determine the current estimated rotor angle position of the target motor under different operating conditions. This allows for the determination of an accurate estimated rotor angle position, thereby achieving precise control of the target motor. Thus, even for target motors without position sensors, the position of the rotor in the target motor can be directly determined, achieving accurate control and reducing motor costs while achieving accurate motor control. Attached Figure Description

[0036] Figure 1 This is a flowchart illustrating a motor control method in one embodiment;

[0037] Figure 2 This is a flowchart illustrating the process of determining operating conditions in one embodiment;

[0038] Figure 3 This is a schematic diagram of hysteresis control in one embodiment;

[0039] Figure 4 This is a flowchart illustrating the model training process in one embodiment;

[0040] Figure 5 This is a schematic diagram of the coordinate system relationship in one embodiment;

[0041] Figure 6 This is a flowchart illustrating the process of determining the estimated rotor angle position in one embodiment;

[0042] Figure 7 This is a schematic diagram of another process for determining the estimated position of the rotor angle in one embodiment;

[0043] Figure 8 This is a schematic diagram of the process for controlling the operation of the target motor in one embodiment;

[0044] Figure 9 This is a flowchart illustrating a sensorless full-speed control method for an embedded permanent magnet hub motor in one embodiment.

[0045] Figure 10This is a structural block diagram of the motor control device in one embodiment;

[0046] Figure 11 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation

[0047] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0048] In recent years, hub motors have been widely used and rapidly developed in various fields, especially in the field of electric vehicles, due to their advantages such as simple structure, low cost, reliable operation, small torque ripple, low noise, no need for position sensors, and high speed limit.

[0049] Hub motors are a type of permanent magnet synchronous motor. Hub motors are mostly distributed drive systems, allowing each wheel to steer independently. Compared to traditional centralized drives, they offer advantages such as high transmission efficiency, high precision, and ease of integration and expansion. Permanent magnet synchronous motors are small in size, lightweight, and can improve power density and motor operating efficiency, making them increasingly the most widely used type of motor. Based on the placement of the permanent magnets, permanent magnet synchronous motors can be mainly classified into surface-mounted, surface-inserted, and embedded types. Surface-mounted motors are salient-pole rotors, exhibiting no salient-pole effect. Surface-inserted and embedded types are salient-pole motors, exhibiting a salient-pole effect. Although embedded permanent magnet synchronous motors have a more complex manufacturing process, they can achieve higher limiting speeds, thus offering a wider speed range and greater torque.

[0050] Therefore, embedded permanent magnet synchronous motors are considered the most promising in-wheel motors. In the electric vehicle field, the most significant characteristic of in-wheel motors is that each wheel motor can operate independently and coordinately to ensure safe vehicle operation. To achieve high performance with embedded permanent magnet in-wheel motors at low cost, reducing the number of position sensors is a good solution. The accuracy of the rotor's angular position directly affects the performance of the in-wheel motor, and sensorless embedded permanent magnet in-wheel motors require position estimation. However, due to the characteristics of permanent magnet synchronous motors, rotor position estimation needs to be studied separately under different operating conditions. Therefore, researching a full-speed, sensorless in-wheel motor control is highly significant. Based on this, how to achieve sensorless full-speed control of embedded permanent magnet in-wheel motors is an urgent problem to be solved.

[0051] In view of this, embodiments of this application provide a motor control method to solve the above problems and realize the control of hub motors under different operating conditions.

[0052] In one embodiment, such as Figure 1As shown, a motor control method is provided. This application embodiment illustrates the method applied to a terminal. It is understood that this method can also be applied to a server, and further to a system including both a terminal and a server, and is implemented through interaction between the terminal and the server. The terminal can be a personal computer, laptop, smartphone, or tablet, etc.; the server can be a single server or a server cluster consisting of multiple servers. In this embodiment, the method includes the following steps:

[0053] Step 101: Obtain the operating signal of the target motor.

[0054] The target motor refers to the motor that needs to be monitored and controlled. The target motor can be a permanent magnet synchronous motor or a hub motor. Optionally, the target motor can be a sensorless embedded permanent magnet hub motor; no specific limitation is made here. Alternatively, the target motor can refer to a sensorless hub motor applied to the target vehicle.

[0055] The operating signal is the signal generated during the operation of the target motor. This operating signal may include a single signal or multiple signals related to the operation of the target motor; the specific number of signals included in the operating signal is not specifically limited here. Optionally, if the target motor is applied to a target vehicle, the operating signal may include signals related to the target motor and various parts of the vehicle during operation. It should be noted that the operating signal may consist of multiple signals related to the rotor angular position of the target motor; the specific signal types are not specifically limited here and can be determined according to the actual situation.

[0056] Understandably, the terminal can periodically acquire the current operating signal of the target motor, or it can acquire the operating signal of the target motor in real time.

[0057] Step 102: Determine the current operating condition of the target motor based on the operating signal and the preset operating condition judgment model. The operating conditions include zero-low speed condition and medium-high speed condition.

[0058] The terminal is equipped with a preset operating condition judgment model, which is used to determine the operating condition of the target motor based on the operating signal. Specifically, after acquiring the operating signal, the operating signal is input into the preset operating condition judgment model to obtain the judgment result output by the preset operating condition judgment model. Based on the judgment result, the current operating condition of the target motor can be determined.

[0059] The operating condition can be used to characterize the current speed range of the target motor. If the target motor's operating condition is zero-speed (low speed), it means the current speed is close to zero or within the low-speed range. If the target motor's operating condition is medium-speed (medium-speed) or high-speed, it means the current speed is within the medium-speed or high-speed range. The division into low-speed, medium-speed, and high-speed ranges can be determined based on actual conditions and is not specifically limited here.

[0060] Step 103: Based on the operating conditions, determine the estimated position of the rotor angle of the target motor using the corresponding calculation method.

[0061] When controlling a target motor, control is required based on the rotor's angular position. Traditional motors output this rotor angular position using a position sensor, but the presence of a position sensor increases cost. Therefore, sensorless hub motors have been developed. Correspondingly, the rotor angular position of the hub motor needs to be estimated, and control of the hub motor is based on this estimated rotor angular position.

[0062] In this embodiment, the performance differences of hub motors under different operating conditions are taken into account. Therefore, different calculation methods are used to determine the corresponding rotor angle estimation position for the target motor under different operating conditions. This rotor angle estimation position is also the estimated rotor angle position of the target motor. It can be understood that, based on the corresponding calculation method, the rotor angle estimation position with zero or minimal error compared to the actual rotor angle position can be determined.

[0063] The terminal is pre-configured with calculation methods for both low-speed and medium-speed operating conditions. After determining the operating condition of the target motor, the estimated rotor angle position is calculated based on the corresponding calculation method.

[0064] Step 104: Estimate the position and control the operation of the target motor based on the rotor angle.

[0065] Since the error between the estimated rotor angle position and the actual rotor angle position is zero, the estimated rotor angle position can directly characterize the rotor angle position in the target motor. Therefore, controlling the operation of the target motor directly based on this estimated rotor angle position can achieve accurate control of the target motor. Because this operating condition includes zero-speed and medium-speed conditions, it also means that accurate control of the target motor can be achieved across the entire speed range.

[0066] It is understood that the operating conditions of the target motor can change in real time. Therefore, the terminal switches the corresponding calculation method in real time to determine the estimated rotor angle position based on the real-time operating conditions. Optionally, in this embodiment, the estimated rotor angle position is determined by switching the corresponding calculation method in real time based on a hysteresis smooth switching control strategy.

[0067] The aforementioned motor control method acquires the operating signal of the target motor and determines its current operating condition based on this signal and a preset operating condition judgment model. This operating condition includes zero-speed (low speed) and medium-speed (medium-speed) conditions. Then, based on this operating condition, a corresponding calculation method is used to determine the estimated rotor angle position of the target motor, and the motor is controlled according to this estimated rotor angle position. This allows the real-time operating condition of the target motor to be determined based on its real-time operating signal and the preset operating condition judgment model. Since the estimation accuracy of the motor rotor angle varies under different operating conditions, and this accuracy affects the control accuracy of the target motor, a corresponding calculation method is used to determine the estimated rotor angle position of the target motor under different operating conditions. This allows for accurate rotor angle estimation and precise control of the target motor. Furthermore, even for target motors without position sensors, the rotor position can be directly determined, achieving accurate control while reducing motor costs.

[0068] In one embodiment, the target motor is a hub motor used in a vehicle, and the operating signals include the vehicle speed, the torque of the target motor, the vehicle's tire swerve angle, and the vehicle's yaw rate.

[0069] Optionally, the terminal can connect to sensors or other data acquisition devices to collect vehicle speed, torque, tire steering angle, and yaw rate, and transmit these data to the terminal. It is understood that, depending on the specific circumstances, this operating signal may also include other signals to improve the accuracy of motor control. Optionally, when the target motor is applied in other fields, this operating signal is determined based on the operating information related to the applied equipment. The process of determining the operating condition of the target motor based on this operating signal will be explained below.

[0070] In one embodiment, such as Figure 2 The diagram illustrates a flowchart of a method for determining operating conditions according to an embodiment of this application. The method involves determining the current operating condition of the target motor based on operating signals and a preset operating condition judgment model, including:

[0071] Step 201: Input the running signal into the preset working condition judgment model to obtain the initial state result output by the preset working condition judgment model.

[0072] Step 202: If the initial state result is less than the preset state threshold, then the current operating condition of the target motor is determined to be zero low speed condition.

[0073] Step 203: If the initial state result is greater than the preset state threshold, then the current operating condition of the target motor is determined to be medium-high speed condition.

[0074] The operating signal is input into the preset operating condition judgment model, and deep learning is used to determine the current operating condition of the target motor.

[0075] Specifically, the method for determining the operating condition of the target motor involves setting a preset state threshold, which is obtained through simulation training based on historical data. This preset state threshold is stored in the terminal. The initial state result is compared with the preset state threshold. When the initial state result output by the preset operating condition judgment model is less than the preset state threshold, it indicates that the target motor is operating in the zero-low speed domain, making real-time estimation of the rotor angle position suitable based on the high-frequency pulse voltage injection method. When the initial state result output by the preset operating condition judgment model is greater than or equal to the preset state threshold, it indicates that the target motor is operating in the medium-high speed domain, making real-time estimation of the rotor angle position suitable based on the sliding film observation method.

[0076] In this embodiment, due to uncertainties in the characteristics of the controller and the actual operating conditions of the hub motor, the output value of the preset operating condition judgment model may frequently change around the preset state threshold before stabilizing, or may oscillate within a small range to stabilize. This situation leads to frequent switching of the estimation strategy, thereby increasing the computational load of the system and potentially causing system failure. Therefore, this embodiment introduces a hysteresis strategy. After determining the operating conditions of the target motor, a smooth switching control strategy based on hysteresis is used to achieve smooth switching of the estimation strategy during the operation of the target motor. That is, based on the hysteresis smooth switching control strategy, the estimation method for the rotor angle estimation position is smoothly switched, thereby achieving full-speed control of the motor. Figure 3 The diagram illustrates a hysteresis control scheme provided in this application embodiment, where △A represents the invalid region for state transitions, and A is a preset state threshold. Y represents the operating condition of the target motor, with Y1 representing the zero-speed condition and Y2 representing the medium-speed condition. In this way, the influence of environmental fluctuations is accommodated by the hysteresis of the response, thereby achieving a smooth switching of the signal identification strategy.

[0077] Optionally, the preset operating condition judgment model can be constructed based on a fully connected neural network model. The process of obtaining this preset operating condition judgment model is explained below. It should be noted that this preset operating condition judgment model can also be constructed based on other types of network models, as long as it can obtain accurate operating condition judgment results for the target motor. A complete example is not provided here.

[0078] In one embodiment, the preset working condition judgment model is obtained by pre-training a fully connected neural network model. For example... Figure 4 The diagram illustrates a flowchart of a model training process provided in an embodiment of this application. The training process of this fully connected neural network model includes:

[0079] Step 401: Obtain historical operating signals of the target vehicle during its driving process.

[0080] In-wheel motors, due to their inherent advantages, have been applied to various types of vehicles. Therefore, the required feature dataset can be obtained using existing vehicle operating data. Selecting important features as model inputs is fundamental to obtaining reliable outputs. For ease of measurement and consideration of importance, this embodiment uses vehicle speed v, torque T, and tire steering angle as inputs. And the yaw rate α is used as a characteristic quantity. Therefore, the historical operating signal can at least include multiple sets of vehicle speed v, torque T, and tire yaw rate generated during the target vehicle's historical driving process. And the yaw rate 'a'. As mentioned above, historical operating signals can be collected through sensors and transmitted to the terminal.

[0081] Step 402: Normalize the historical running signals to obtain the initial training dataset.

[0082] To improve the training speed and prediction accuracy of the model, the collected historical running signals need to be normalized, that is, the data in the historical running signals need to be normalized to the interval [0,1]. Optionally, the normalization process can be carried out in the following ways:

[0083]

[0084] Where, x * The data is normalized, and x is the original data; x min and x max These are the maximum and minimum values ​​in a set of original data, respectively.

[0085] Step 403: Perform outlier removal on the initial training dataset to obtain the target training dataset.

[0086] Data collected from multiple sensors may contain some noisy data, also known as outliers, which can affect subsequent training and prediction results. Therefore, to ensure training accuracy, the data collected by each sensor can be cross-referenced to remove some obviously unacceptable outliers. Appropriate values ​​can then be inserted based on the preceding and following data to obtain the target training dataset.

[0087] Optionally, the outlier removal process can be as follows: determine a threshold W, and use the absolute value averaging method to determine the value of W, that is:

[0088]

[0089] Where p is an empirical parameter, Z i Let be the i-th data point in the dataset to be processed, and n be the number of data points in the dataset. For each data point in the dataset to be processed, if the absolute value of the data point is greater than W, that is, |Z|... i When |≥W, then Z is determined. i These are outlier points and should be removed or replaced.

[0090] Step 404: Train a fully connected neural network model based on the target training dataset, and use the trained fully connected neural network model as the preset working condition judgment model.

[0091] In a fully connected neural network model, for layers n-1 and n, any node in layer n-1 is connected to all nodes in layer n. That is, when each node in layer n performs computation, the input to the activation function is a weighted sum of the inputs from all nodes in layer n-1.

[0092] A fully connected neural network model mainly consists of an input layer, hidden layers, and an output layer. It typically has multiple hidden layers. Adding hidden layers can better separate the features of the data, but too many hidden layers can also increase training time and lead to overfitting.

[0093] In this embodiment, the number of neurons in the input layer of the fully connected neural network model represents the number of features. Four feature values ​​(x1, x2, x3, x4) are used to represent the input variables, where x1 represents the vehicle speed v, x2 represents the torque T, and x3 represents the tire angle. x4 represents the yaw rate 'a'. The output of the fully connected neural network model represents the operating conditions of the hub motor.

[0094] Specifically, the input data undergoes non-linear transformations between each layer of the fully connected neural network model. When the original input data is linearly inseparable, the fully connected neural network model generates a non-linear output through activation functions, such as the Sigmoid function, Tanh function, and ReLU function. For the output value of the j-th node in the i-th layer of the fully connected neural network... It can be represented as follows:

[0095]

[0096] Where f i (·) is the activation function. This represents the output value of the j-th node in the (i-1)-th layer. This is true when the (i-1)-th layer is the input layer. Let N be the input vector of the j-th node. i-1 Let i be the number of neurons in the (i-1)th layer. Let be the weight matrix of the j-th node in the i-th layer. Let L be the bias vector of the j-th node in the i-th layer. Its value is adjusted during training to fit the optimal model. L is the number of layers in the fully neural network model.

[0097] Optionally, the training objective of a fully connected neural network model can be to minimize the mean squared error. Hyperparameters of a fully connected neural network model include the learning rate, regularization parameter, number of layers, number of neurons in each hidden layer, number of training epochs, choice of objective function, and neuron activation functions. The appropriate selection of these hyperparameters determines the model's performance.

[0098] Based on this, by inputting each data point in the target training dataset into the fully connected neural network model and continuously adjusting the parameters of the fully connected neural network model according to the output results, an accurate fully connected neural network model can be obtained, thus completing the model training process. The trained fully connected neural network model can then be used as the preset working condition judgment model.

[0099] In this embodiment, an initial training dataset is obtained by normalizing historical operating signals, and a target training dataset is obtained by removing outliers from the initial training dataset. This allows for the training of an accurate preset operating condition judgment model based on the target training dataset. A fully connected neural network model is used to construct this preset operating condition judgment model, improving the efficiency and accuracy of operating condition judgment.

[0100] Once the operating conditions are determined, the corresponding calculation method can be used to determine the estimated position of the rotor angle.

[0101] In one embodiment, the rotor angle estimation position of the target motor is determined by using a corresponding calculation method according to the operating conditions, including: determining the rotor angle estimation position based on the high-frequency pulse voltage injection method when the operating conditions are zero low speed; and determining the rotor angle estimation position based on the sliding film observation method when the operating conditions are medium and high speed.

[0102] Specifically, in this embodiment, under zero-low-speed operating conditions, the rotor angle position of the target motor can be estimated based on a high-frequency pulsed voltage injection method. This method mainly utilizes the salient pole characteristics of the motor itself. When a high-frequency voltage or current signal is injected into the three-phase stator windings, the injected high-frequency voltage or current will generate corresponding high-frequency response signals in the three-phase windings. These high-frequency response signals contain rotor angle position information. Therefore, by detecting the high-frequency response signals and performing appropriate signal separation and extraction, the rotor position information and rotational speed can be obtained. In other words, in this embodiment, under zero-low-speed operating conditions, the estimated rotor angle position can be determined based on the high-frequency response signal of the target motor.

[0103] When the motor operates at medium to high speeds, its back electromotive force is relatively high. Therefore, using the same rotor angle position estimation method as for zero-speed operation would place higher demands on the motor. In this embodiment, under medium to high speed conditions, the rotor angle position can be estimated based on the motor's fundamental voltage and current. Specifically, the motor's current mode can be determined, and the sliding film observation method can be further used to estimate the rotor angle position.

[0104] The following section explains the process of calculating the rotor angle estimation position under the condition of zero-speed operation. First, a theoretical explanation will be given, using an embedded permanent magnet synchronous hub motor as an example.

[0105] The mathematical model of the embedded permanent magnet synchronous hub motor in the synchronous rotating coordinate system (dq coordinate system) is shown below:

[0106]

[0107] Among them, u d For the d-axis high-frequency voltage component; u q i represents the q-axis high-frequency voltage component; d i represents the d-axis high-frequency current component; q R represents the q-axis high-frequency current component. s For stator resistance; L d For d-axis inductance; L q ω is the q-axis inductance; ω is the rotor electrical angle. is the flux linkage of the permanent magnet; p is the differential operator.

[0108] Since the injected high-frequency signal frequency is much higher than the fundamental frequency, the voltage drop across the inductor is much greater than the voltage drop across the resistor. Ignoring the voltage drop across the resistor, R... s *i d With R s *i qThe value is 0. The control scenario is a zero-speed condition, where the rotational voltage and back electromotive force related to the rotor electrical angle are negligible. Therefore, the mathematical model under high-frequency excitation can be simplified as follows:

[0109]

[0110] To accurately estimate the rotor angular position of the motor, it is necessary to establish an estimated rotor synchronous rotation coordinate system and observe its relationship with the actual rotor synchronous rotation coordinate system, such as... Figure 5 The diagram illustrates a coordinate system relationship provided in an embodiment of this application. Wherein, d and q represent the actual synchronous rotation coordinate system of the rotor. The coordinate system is used to estimate the synchronous rotation of the rotor; α and β are the stationary coordinate systems.

[0111] Depend on Figure 5 It can be seen that the rotor angle estimation error is: θ is the actual rotor angular position; The estimated rotor angular position;

[0112] Because the injected high-frequency excitation is based on the estimated rotor synchronous rotation coordinate system The current is injected downwards, but the current response is caused in the actual axis dq, therefore the following transformation is required:

[0113]

[0114] and These are the coordinate systems for estimating the synchronous rotation of the rotor. The high-frequency voltage components of the d-axis and q-axis.

[0115] Since the response current is in the actual dq axis system, while the estimated response current... for In the axis system, the following inverse transformation is therefore required:

[0116]

[0117] Combining Equations 4, 5, and 6 above, we can obtain the high-frequency current equation for the estimated shaft system:

[0118]

[0119] Simplifying Equation 7, we get:

[0120]

[0121] Formula 8 characterizes the high-frequency response current under high-frequency excitation in the estimated shaft system. Where L is the average inductance. ΔL is the half-difference inductance.

[0122] As can be seen from Equation 8, this high-frequency response current is related to the rotor angle estimation error. Relatedly, if the rotor angle estimation error is caused by the high-frequency response current... If the value is zero, the estimated rotor angle position can directly represent the actual rotor angle position for motor operation control. Therefore, in this embodiment, the estimated rotor angle position can be determined based on the high-frequency response current signal of the target motor in the synchronous rotating coordinate system. Specifically, as follows.

[0123] In one embodiment, such as Figure 6 The diagram illustrates a flowchart of a method for determining the estimated rotor angle position according to an embodiment of this application. The method for determining the estimated rotor angle position based on a high-frequency pulse voltage injection method includes:

[0124] Step 601: Obtain the high-frequency response current signal of the target motor in the synchronous rotating coordinate system when a pulsating high-frequency voltage signal is injected into the target motor.

[0125] Optionally, the high-frequency excitation can be a pulsed high-frequency voltage signal, such as a high-frequency sinusoidal voltage signal.

[0126] The estimated high-frequency sinusoidal voltage signal of the dq axis in the synchronous rotating coordinate system obtained based on the pulsed high-frequency voltage injection method is:

[0127]

[0128] Among them, u in ω represents the amplitude of the injected high-frequency voltage. in The frequency of the injected high-frequency voltage.

[0129] Correspondingly, the obtained high-frequency response current is:

[0130]

[0131] As shown in Formula 10, if the inductances of the d-axis and q-axis are different (ΔL≠0), then in the synchronous axis system, The amplitude of the high-frequency current component of the shaft and the rotor observation angle error Related to the rotor observation angle error At that time, the estimated response current signal of the d-axis Estimated response current signal along the q-axis Therefore, for ease of processing, the high-frequency response current signal of the q-axis can be used. The signal is processed and used as the signal input for the position observer, thereby controlling... This causes the rotor angle estimation error This is used to obtain the rotor angle estimate position, which represents the actual rotor angle position.

[0132] Step 602: Modulate the high-frequency response current signal of the q-axis in the synchronous rotating coordinate system corresponding to the high-frequency response current signal, and then perform low-pass filtering to obtain the processed high-frequency response current signal of the q-axis.

[0133] Step 603: Input the processed high-frequency response current signal of the q-axis to the position observer, and control the processed high-frequency response current signal of the q-axis to be zero to obtain the estimated rotor angle position.

[0134] To extract the rotor angle estimation position from the high-frequency signal, the high-frequency response current of the q-axis is first analyzed. Modulation processing is performed. This modulation processing may include: making... Multiply by sinω in t doubles the frequency, making it easier for the low-pass filter to filter it out, that is:

[0135]

[0136] The modulated signal is then low-pass filtered to obtain the low-pass filtered q-axis high-frequency response current signal. The following is an expression:

[0137]

[0138] LPF stands for low-pass filter. Optional, due to the typical rotor angle estimation error... The error signal is relatively small, therefore it can be linearized, that is, by making... Therefore, formula 12 can be simplified as follows:

[0139]

[0140] in:

[0141] As can be seen from Formula 13, the high-frequency response current signal of the q-axis after adjustment and control processing... When, the estimation error is That is, the rotor angular position converges to the actual angular position. At this point, the estimated rotor angular position can be directly determined, which represents the actual rotor angular position.

[0142] The following section will explain the process of calculating the rotor angle estimation position under medium-to-high-speed operating conditions. First, a theoretical explanation will be provided.

[0143] The voltage equation for the embedded permanent magnet hub motor in the actual dq axis system is as follows:

[0144]

[0145] Traditional sliding membrane observers are mostly based on mathematical models of motors in the stationary coordinate system. Therefore, by transforming the voltage equation in the dq coordinate system, we obtain the following equation:

[0146]

[0147] Among them, u α For the α-axis stator voltage component; u β For the β-axis stator voltage component; i α i represents the α-axis stator current component; β For the β-axis stator current component; L α For α-axis inductance; L β L is a beta-axis inductance. α =L + ΔLcos2θ, L β =L - ΔLcos2θ, L αβ =ΔLcos2θ. L is the average inductance, and ΔL is the half-difference inductance. R s ω is the stator resistance; ω is the rotor electrical angle. is the flux linkage of the permanent magnet; p is the differential operator.

[0148] As shown in Equation 15, in the α-β axis system, both the back electromotive force term in the motor voltage equation and the inductance matrix contain rotor angular position information θ. Since the rotor position information contained in the inductance is 2θ, it will affect the estimation result. Therefore, Equation 15 is rewritten as the dq axis voltage equation:

[0149]

[0150] in, To extend the back electromotive force, The back electromotive force generated by the permanent magnet, (L d -L q )(ωi d -pi q ) is the back electromotive force generated by the convex mechanism.

[0151] Simplifying Equation 16 back to the stationary coordinate system, we get:

[0152]

[0153] in, To extend the back electromotive force, E α E is the α-axis back electromotive force component; β This represents the β-axis back electromotive force component.

[0154] Since the extended back EMF expression of the embedded permanent magnet synchronous motor already includes the rotor angular position information, the voltage equation is rewritten as a current equation to facilitate the calculation of the rotor angular position information:

[0155]

[0156] in,

[0157] Therefore, the extended back electromotive force (EMF) can be determined based on this current equation, and since the extended back EMF is related to the rotor angular position, the rotor angular position can be determined. Similarly, in the embodiments of this application, based on the sliding diaphragm observer and combined with this current equation, the estimated value of the extended back EMF can be determined, that is, the extended back EMF can be estimated, and then the estimated rotor angular position can be further determined based on the estimated extended back EMF.

[0158] Please refer to Figure 7 This illustrates a flowchart of another method for determining the estimated rotor angle position according to an embodiment of this application. Determining the estimated rotor angle position based on the sliding film observation method includes:

[0159] Step 701: Determine the parameters of the sliding diaphragm observer based on the current equation of the target motor. The parameters include the sliding diaphragm surface and the sliding diaphragm control law.

[0160] To obtain an estimate of the extended back electromotive force, the slug observer is designed with the following equation:

[0161]

[0162] in, and u is an estimated value for the stator current. α and u β To control the input.

[0163] Subtracting formula 18 from formula 19 will give us:

[0164]

[0165] To make the stator currents on the α and β axes the same, the sliding surface can be designed as follows:

[0166]

[0167] For the extended back electromotive force (EMF), the sliding control law aims to return the current to the sliding surface, thus eliminating the error between the estimated and actual current. Consequently, the error between the estimated and actual extended back EMF is also zero. Based on the estimated extended back EMF, the accurate rotor angle estimation position can be determined. Therefore, the sliding control law is designed as follows:

[0168]

[0169] Where k is the rate of approaching the glial surface / switching surface, and sgn is the sign function.

[0170] Step 702: Determine the estimated extended back electromotive force of the target motor based on the sliding diaphragm observer after the parameters have been determined.

[0171] After setting the parameters, the estimated extended back electromotive force is the sliding control rate, which is also the estimated extended back electromotive force:

[0172]

[0173] Step 703: Calculate the arctangent of the estimated extended back electromotive force to obtain the estimated position of the initial rotor angle.

[0174] The initial rotor angle estimation position can be obtained by taking the arctangent of the estimated extended back electromotive force.

[0175]

[0176] Step 704: Add the initial rotor angle estimated position to the target angle compensation value to obtain the rotor angle estimated position.

[0177] To reduce the rotor angle position estimation error caused by low-pass filtering to zero or extremely low, a target angle compensation value θ can be added to the value estimated in Equation 24. com That is, the estimated position of the rotor angle obtained. for:

[0178]

[0179] Therefore, in this embodiment of the application, for the operating conditions of the target motor, the corresponding calculation method can accurately determine the rotor angle estimation position. Since the error is zero, the rotor angle estimation position can directly represent the actual rotor angle position. Therefore, the operation of the target motor can be further controlled based on the rotor angle estimation position. For motors without position sensors, precise control of the motor can be achieved.

[0180] The motor control process will be explained below.

[0181] Please refer to Figure 8 This document illustrates a flowchart of a method for controlling the operation of a target motor according to an embodiment of this application. Controlling the target motor's operation based on rotor angle-estimated position includes:

[0182] Step 801: Differentiate the rotor angle estimation position to obtain the estimated rotor speed of the target motor.

[0183] Step 802: Based on the estimated position and estimated speed of the rotor, the target motor is controlled to run using a PI dual closed-loop control method.

[0184] Among them, PI dual-loop control is a control strategy for DC motors. The outer loop is the speed loop, and the inner loop is the current loop. It uses a PI controller as the regulator and has good starting and regulation characteristics. Specifically, based on the rotor angle to estimate the position and the rotor speed, the torque and speed of the target motor can be adjusted according to the PI dual-loop control method to meet the control requirements and achieve flexible motor operation control.

[0185] In one embodiment, such as Figure 9 The diagram illustrates a flow chart of a sensorless full-speed control method for an embedded permanent magnet hub motor according to an embodiment of this application. The method includes:

[0186] Step 901: Obtain the operating signals of the hub motor. The hub motor is used in the vehicle, and the operating signals include the vehicle speed, the torque of the target motor, the vehicle's tire steering angle, and the vehicle's yaw rate.

[0187] Step 902: Input the running signal into the fully connected neural network model to obtain the initial state result output by the fully connected neural network model.

[0188] Step 903: If the initial state result is less than the preset state threshold, the current operating condition of the hub motor is determined to be zero low speed condition; if the initial state result is greater than the preset state threshold, the current operating condition of the hub motor is determined to be medium high speed condition.

[0189] Step 904: Under the condition of zero low speed operation, determine the high frequency response current signal of the q-axis based on the high frequency response current signal of the hub motor in the synchronous rotating coordinate system.

[0190] Step 905: Modulate the high-frequency response current signal of the q-axis and then perform low-pass filtering to obtain the processed high-frequency response current signal of the q-axis.

[0191] Step 906: Input the processed high-frequency response current signal of the q-axis to the position observer, control the processed high-frequency response current signal of the q-axis to be zero, and obtain the estimated position of the rotor angle.

[0192] Step 907: Under medium-to-high speed operating conditions, determine the parameters of the diaphragm observer based on the current equation of the hub motor. The parameters include the diaphragm surface and the diaphragm control rate.

[0193] Step 908: Determine the estimated extended back electromotive force of the hub motor based on the sliding diaphragm observer after the parameters have been determined.

[0194] Step 909: Calculate the arctangent of the estimated extended back electromotive force to obtain the estimated initial rotor angle position, and add the estimated initial rotor angle position to the target angle compensation value to obtain the estimated rotor angle position.

[0195] Step 910: Differentiate the rotor angle estimation position to obtain the estimated rotor speed of the hub motor.

[0196] Step 911: Based on the estimated position and estimated speed of the rotor, the hub motor is controlled to run using the PI dual closed-loop control method.

[0197] The sensorless full-speed control method for an embedded permanent magnet hub motor provided in this application obtains the current operating conditions of the vehicle through a fully connected neural network, and then adopts an appropriate rotor angle position estimation strategy. To prevent frequent strategy switching due to environmental fluctuations, a hysteresis strategy is introduced to achieve smooth switching of the estimation strategy. This achieves sensorless full-speed control of the embedded permanent magnet hub motor.

[0198] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0199] Based on the same inventive concept, this application also provides a motor control device for implementing the motor control method described above. The solution provided by this device is similar to the solution described in the above method; therefore, the specific limitations in one or more motor control device embodiments provided below can be found in the limitations of the motor control method described above, and will not be repeated here.

[0200] In one embodiment, such as Figure 10 As shown, a motor control device 1000 is provided, comprising: a first acquisition module 1001, a first determination module 1002, a second determination module 1003, and a motor control module 1004, wherein:

[0201] The first acquisition module 1001 is used to acquire the operating signal of the target motor;

[0202] The first determining module 1002 is used to determine the current operating condition of the target motor based on the operating signal and the preset operating condition judgment model; the operating condition includes zero low speed condition and medium high speed condition;

[0203] The second determining module 1003 is used to determine the estimated position of the rotor angle of the target motor according to the operating conditions and by using the corresponding calculation method.

[0204] The motor control module 1004 is used to control the operation of the target motor by estimating the position based on the rotor angle.

[0205] In one embodiment, the second determining module is specifically used to: determine the estimated rotor angle position based on the high-frequency pulse voltage injection method when the operating condition is zero low speed; and determine the estimated rotor angle position based on the sliding film observation method when the operating condition is medium high speed.

[0206] In one embodiment, the second determining module 1003 is specifically used to: obtain the high-frequency response current signal of the target motor in the synchronous rotating coordinate system when a pulsating high-frequency voltage signal is injected into the target motor; modulate the high-frequency response current signal of the q-axis in the synchronous rotating coordinate system corresponding to the high-frequency response current signal and then perform low-pass filtering to obtain the processed high-frequency response current signal of the q-axis; input the processed high-frequency response current signal of the q-axis to the position observer and control the processed high-frequency response current signal of the q-axis to be zero to obtain the rotor angle estimated position.

[0207] In one embodiment, the second determining module 1003 is specifically used for: determining the parameters of the sliding diaphragm observer according to the current equation of the target motor, the parameters including the sliding diaphragm surface and the sliding diaphragm control law; determining the estimated extended back electromotive force of the target motor according to the sliding diaphragm observer after determining the parameters; obtaining the initial rotor angle estimated position by taking the arctangent of the estimated extended back electromotive force; and adding the initial rotor angle estimated position to the target angle compensation value to obtain the rotor angle estimated position.

[0208] In one embodiment, the motor control module 1004 is specifically used to: perform differential processing on the rotor angle estimation position to obtain the rotor estimation speed of the target motor; and control the operation of the target motor based on the rotor angle estimation position and the rotor estimation speed using a PI dual closed-loop control method.

[0209] In one embodiment, the first determining module 1001 is specifically used to: input the operating signal into the preset operating condition judgment model to obtain the initial state result output by the preset operating condition judgment model; if the initial state result is less than the preset state threshold, then determine that the current operating condition of the target motor is zero low speed condition; if the initial state result is greater than the preset state threshold, then determine that the current operating condition of the target motor is medium high speed condition.

[0210] In one embodiment, the preset working condition judgment model is obtained by pre-training a fully connected neural network model, and the device further includes:

[0211] The second acquisition module is used to acquire historical operating signals of the target vehicle during its driving process;

[0212] The processing module is used to normalize the historical running signals to obtain the initial training dataset;

[0213] The outlier removal module is used to remove outliers from the initial training dataset to obtain the target training dataset.

[0214] The training module is used to train a fully connected neural network model based on the target training dataset, and to use the trained fully connected neural network model as a preset working condition judgment model.

[0215] In one embodiment, the target motor is a hub motor used in a vehicle, and the operating signals include the vehicle speed, the torque of the target motor, the vehicle's tire swerve angle, and the vehicle's yaw rate.

[0216] Each module in the aforementioned motor control device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in the processor of a computer device in hardware form or independent of it, or stored in the memory of a computer device in software form, so that the processor can call and execute the operations corresponding to each module.

[0217] In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 11As shown, this computer device includes a processor, memory, input / output interfaces (I / O), and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides the environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The database stores motor control data. The I / O interfaces are used for exchanging information between the processor and external devices. The communication interface is used for communication with external terminals via a network connection. When the computer program is executed by the processor, it implements a motor control method.

[0218] Those skilled in the art will understand that Figure 11 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0219] In one embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above-described method embodiments.

[0220] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon that, when executed by a processor, implements the steps in the above method embodiments.

[0221] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above method embodiments.

[0222] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0223] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0224] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A motor control method, characterized in that, The method includes: Acquire the operating signal of the target motor; The current operating condition of the target motor is determined based on the operating signal and a preset operating condition judgment model. The operating conditions include zero-low speed and medium-high speed conditions. This includes: inputting the operating signal into the preset operating condition judgment model to obtain an initial state result output by the preset operating condition judgment model through deep learning; if the initial state result is less than a preset state threshold, the current operating condition of the target motor is determined to be the zero-low speed condition; if the initial state result is greater than the preset state threshold, the current operating condition of the target motor is determined to be the medium-high speed condition. The preset operating condition judgment model is constructed based on a fully connected neural network model. The fully connected neural network model includes an input layer, a hidden layer, and an output layer. The number of neurons in the input layer represents the number of features, and each feature value includes vehicle speed, torque, tire steering angle, and yaw rate. The input data undergoes nonlinear transformation between each layer in the fully connected neural network model, and when the original input data is linearly inseparable, the fully connected neural network model generates a nonlinear output through an activation function. The output of the fully connected neural network model represents the operating condition of the hub motor. Based on the operating conditions, the estimated position of the rotor angle of the target motor is determined using the corresponding calculation method; Controlling the target motor operation based on the estimated rotor angle position includes: performing differential processing on the estimated rotor angle position to obtain the estimated rotor speed of the target motor; and controlling the torque and speed of the target motor based on the estimated rotor angle position and the estimated rotor speed using a PI dual closed-loop control method. During the real-time changes in the operating conditions of the target motor, the calculation method corresponding to the current operating condition is switched in real time based on the hysteresis smooth switching control strategy to determine the estimated rotor angle position in real time. Under the hysteresis smooth switching control strategy, the preset state threshold is located between the first threshold reference value and the second threshold reference value. The first threshold reference value is obtained by subtracting the invalid state transition region from the preset state threshold, and the second threshold reference value is obtained by adding the invalid state transition region to the preset state threshold.

2. The method according to claim 1, characterized in that, The step of determining the estimated rotor angle position of the target motor using a corresponding calculation method based on the operating conditions includes: Under the condition of zero low-speed operation, the estimated position of the rotor angle is determined based on the high-frequency pulse voltage injection method. When the operating condition is medium to high speed, the estimated position of the rotor angle is determined based on the sliding film observation method.

3. The method according to claim 2, characterized in that, The method of determining the rotor angle estimation position based on high-frequency pulse voltage injection includes: When a pulsating high-frequency voltage signal is injected into the target motor, the high-frequency response current signal of the target motor in the synchronous rotating coordinate system is obtained. The high-frequency response current signal of the q-axis corresponding to the high-frequency response current signal in the synchronous rotating coordinate system is modulated and then low-pass filtered to obtain the processed high-frequency response current signal of the q-axis. The processed high-frequency response current signal of the q-axis is input to the position observer, and the processed high-frequency response current signal of the q-axis is controlled to be zero to obtain the estimated position of the rotor angle.

4. The method according to claim 2, characterized in that, The determination of the rotor angle estimation position based on the sliding film observation method includes: The parameters of the sliding diaphragm observer are determined based on the current equation of the target motor, and the parameters include the sliding diaphragm surface and the sliding diaphragm control law. The estimated extended back electromotive force of the target motor is determined based on the sliding diaphragm observer after the parameters are determined; The estimated rotor angle position is obtained by taking the arctangent of the estimated extended back electromotive force; The estimated rotor angle position is obtained by adding the initial rotor angle estimated position to the target angle compensation value.

5. The method according to claim 1, characterized in that, The training process of the fully connected neural network model aims to minimize the mean squared error. The hyperparameters of the fully connected neural network model include at least one of the following: learning rate, regularization parameter, number of layers in the neural network, number of neurons in each hidden layer, number of learning rounds, selection of objective function, and neuron activation function.

6. The method according to claim 1, characterized in that, Operating conditions are used to characterize the current speed range of the target motor.

7. The method according to claim 1, characterized in that, The preset working condition judgment model is obtained by pre-training a fully connected neural network model. The training process of the fully connected neural network model includes: Acquire historical operating signals of the target vehicle during its driving process; The historical running signals are normalized to obtain the initial training dataset; The initial training dataset is subjected to outlier removal to obtain the target training dataset; The fully connected neural network model is trained based on the target training dataset, and the trained fully connected neural network model is used as the preset working condition judgment model.

8. The method according to claim 1, characterized in that, The target motor is a hub motor used in a vehicle, and the operating signals include the vehicle speed, the torque of the target motor, the tire rotation angle of the vehicle, and the yaw rate of the vehicle.

9. A motor control device, characterized in that, The device includes: The first acquisition module is used to acquire the operating signal of the target motor; The first determining module is used to determine the current operating condition of the target motor based on the operating signal and a preset operating condition judgment model. The operating conditions include zero-low speed condition and medium-high speed condition. Specifically, the first determining module is used to input the operating signal into the preset operating condition judgment model to obtain the initial state result output by the preset operating condition judgment model through deep learning. If the initial state result is less than a preset state threshold, the current operating condition of the target motor is determined to be the zero-low speed condition. If the initial state result is greater than the preset state threshold, the current operating condition of the target motor is determined to be the medium-high speed condition. The preset operating condition judgment model is constructed based on a fully connected neural network model. The fully connected neural network model includes an input layer, a hidden layer, and an output layer. The number of neurons in the input layer represents the number of features, and each feature value includes vehicle speed, torque, tire rotation angle, and yaw rate. The input data undergoes nonlinear transformation between each layer in the fully connected neural network model. When the original input data is linearly inseparable, the fully connected neural network model generates a nonlinear output through an activation function. The output of the fully connected neural network model represents the operating condition of the hub motor. The second determining module is used to determine the estimated position of the rotor angle of the target motor according to the operating conditions and by using the corresponding calculation method. The motor control module is used to control the operation of the target motor according to the estimated rotor angle position; specifically, it is used to: perform differential processing on the estimated rotor angle position to obtain the estimated rotor speed of the target motor; and control the torque and speed of the target motor based on the estimated rotor angle position and the estimated rotor speed using a PI dual closed-loop control method. The switching module is used to switch the calculation mode corresponding to the current operating condition in real time based on the hysteresis smooth switching control strategy during the real-time change of the operating condition of the target motor, so as to determine the rotor angle estimation position in real time; wherein, under the hysteresis smooth switching control strategy, the preset state threshold is located between a first threshold reference value and a second threshold reference value, the first threshold reference value is obtained by subtracting the invalid region of state transformation from the preset state threshold, and the second threshold reference value is obtained by adding the invalid region of state transformation to the preset state threshold.

10. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 8.

Images