Motor and control method, device, equipment and motor system thereof

Abstract

The present invention relates to the field of motor control technology, specifically a motor and its control method, device, equipment and motor system. The present invention obtains the first multi-dimensional data, second data, third data and historical abnormal data of a sweeping robot, pre-processes and fuses the first multi-dimensional data, generates an environmental state vector, and calculates the power demand and speed of the motor based on the environmental state vector, and adjusts the output power and speed in real time. The second data is analyzed by the Fourier transform method to obtain its frequency domain component, and the signal and torque of the motor are adaptively adjusted. The abnormal recognition model is trained by the historical abnormal data, and the real-time data is recognized by the trained abnormal recognition model, and the similarity is calculated. When the similarity exceeds the preset threshold, the operating parameters of the motor are dynamically adjusted according to the abnormal type output by the model. The present invention effectively improves the intelligent control and abnormality handling capabilities of the motor through the motor control method.

Description

Technical Field

[0001] The present invention relates to the field of motor control technology, and in particular to a motor and a control method, device, equipment and motor system thereof. Background Art

[0002] With the rapid development of smart homes, sweeping robots have become one of the important household cleaning devices. Modern sweeping robots integrate multiple sensors, drive motors and intelligent algorithms to achieve automatic floor cleaning functions. However, there are still some problems in the motor control of the existing technology that need to be solved.

[0003] For example, existing sweeping robots usually rely on simple sensor feedback to adjust the power output of the motor, lack sophisticated processing of complex environments, and cannot fully adapt to the cleaning needs of different environmental conditions. In addition, existing noise reduction technologies mainly rely on the structural design of the motor and passive noise suppression methods, lacking intelligent control measures. Especially when running at high load or high speed, the motor noise and vibration increase significantly, affecting the daily life of users. These problems directly affect the performance of the sweeping robot, user experience, and the overall reliability of the equipment.

[0004] To this end, a motor and a control method, device, equipment and motor system are proposed. Summary of the invention

[0005] The object of the present invention is to provide a motor and a control method, device, equipment and motor system thereof. The present invention relates to the field of motor control technology, specifically a motor and a control method, device, equipment and motor system thereof. By obtaining the first multidimensional data, the second data, the third data and the historical abnormal data of the sweeping robot, the first multidimensional data is preprocessed and fused, an environmental state vector is generated, and the power demand and speed of the motor are calculated based on the environmental state vector, and the output power and speed are adjusted in real time. The second data is analyzed by the Fourier transform method to obtain its frequency domain component, and the signal and torque of the motor are adaptively adjusted. The abnormal recognition model is trained by the historical abnormal data, and the real-time data is recognized by the trained abnormal recognition model, and the similarity is calculated. When the similarity exceeds the preset threshold, the operating parameters of the motor are dynamically adjusted according to the abnormal type output by the model. The present invention effectively improves the intelligent control and abnormal handling capabilities of the motor through the motor control method.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] A motor control method, comprising:

[0008] S1. Using multiple sensors to collect the first multi-dimensional data of the sweeping robot in real time, the sensors include optical sensors, pressure sensors, accelerometers, current sensors and temperature sensors;

[0009] S2. Preprocessing the first multidimensional data collected to obtain first multidimensional detection data;

[0010] S3. Fusing the first multi-dimensional detection data to generate a first environmental state vector;

[0011] S4. Calculate the first motor power requirement and the first speed of the cleaning robot according to the first environmental state vector; and control the voltage and current of the motor in real time according to the first motor power requirement and the first speed;

[0012] S5. Obtain second data of the sweeping robot, analyze the second data using the Fourier transform method to obtain the frequency domain component of the second data, and adaptively obtain the first adjustment of the sweeping robot motor according to the second data and the frequency domain component. signal duty cycle and first adjusted torque;

[0013] S6. Obtain historical abnormal data and third data of the sweeping robot, construct a first abnormal data set through the historical abnormal data, construct a second abnormal recognition data set through the third data, and train an abnormal recognition model through the first abnormal data set to obtain a trained abnormal recognition model, and use the trained abnormal recognition model to recognize the second abnormal recognition data set to obtain a first similarity. If the first similarity is greater than a preset threshold, adjust the operating parameters of the motor in real time according to the abnormality type output by the output layer of the abnormal recognition model.

[0014] Preferably, the first multidimensional data includes ground material type, pressure distribution data, obstacle detection data, motion state data, current data and temperature data; the ground material type is obtained by analyzing the light intensity data and reflected light data obtained by the optical sensor; the pressure distribution data is obtained by measuring the pressure distribution of each point on the bottom of the sweeping robot through a pressure sensor; the obstacle detection data is obtained by measuring the instantaneous pressure value of the sweeping robot when it contacts an obstacle through a pressure sensor; the motion state data is obtained by measuring the three-axis acceleration of the sweeping robot in real time through an accelerometer; the current data is obtained by measuring the current of the motor in real time through a current sensor; the temperature data is obtained by measuring the working temperature of the motor in real time through a temperature sensor; the second data includes the noise data of the motor during operation collected by the acoustic sensor and the motor vibration data collected by the vibration sensor.

[0015] Preferably, the first environmental state vector The vector includes multiple components; each component is obtained by multiplying a data item in the first multidimensional detection data by a corresponding first impact weight coefficient; the first impact weight coefficient is used to represent the importance of the data item in the overall environmental state;

[0016] The first motor power requirement By the first environment state vector Each component in is combined with a second influence weight coefficient; the second influence weight coefficient is used to represent the influence of each component on the motor power demand;

[0017] The first speed According to the first environment state vector The third influence weight coefficient is used to represent the influence degree of each component on the motor speed.

[0018] Preferably, the frequency domain component includes a characteristic frequency component of a noise signal and a characteristic frequency component of a vibration signal;

[0019] The first adjustment The calculation process of the signal duty cycle is as follows: the characteristic frequency component of the noise signal is multiplied by the first signal adjustment parameter to obtain the first Signal correction term; by multiplying the characteristic frequency component of the vibration signal with the second signal adjustment parameter, a second signal correction term; and based on these two The signal correction term is the initial The signal duty cycle is corrected to obtain the first adjustment The first signal adjustment parameter and the second signal adjustment parameter represent the noise signal and the vibration signal respectively. The impact of signal conditioning;

[0020] The calculation process of the first adjustment torque is: a first torque correction term is obtained by multiplying the characteristic frequency component of the noise signal by the first torque adjustment parameter; a second torque correction term is obtained by multiplying the characteristic frequency component of the vibration signal by the second torque parameter; and the initial torque is corrected according to the two torque correction terms to obtain the first adjustment torque; the first torque adjustment parameter and the second torque adjustment parameter respectively represent the influence of the noise signal and the vibration signal on the torque adjustment.

[0021] Preferably, the first abnormal data set includes abnormal temperature data, abnormal vibration data, abnormal current data and abnormal rotation speed; the third data includes temperature data, vibration data, current data and rotation speed; the training process of the abnormal recognition model includes labeling the first abnormal data set with abnormal type labels, and analyzing the first abnormal data set through a convolutional neural network to obtain the correlation between historical abnormal data features and the abnormal type and the historical abnormal data features, and calculate the error between the predicted value and the actual label through a cross entropy loss function to finally obtain the trained abnormal recognition model.

[0022] Preferably, the calculation process of the first similarity is: inputting the second abnormality recognition data into the trained abnormality recognition model, and passing through the preprocessing layer, feature extraction layer and output layer of the model in sequence; the feature extraction layer is used to provide CNN for feature extraction to obtain a first abnormality feature, and according to the historical abnormal data features and the first abnormality features, a cosine similarity formula is established to obtain the first similarity, and if the first similarity is greater than the preset threshold, the operating parameters of the motor are dynamically adjusted according to the first similarity; the operating parameters include speed, torque and power.

[0023] A motor control device, the control device comprising:

[0024] A data acquisition unit, used for acquiring first multi-dimensional data, second data, third data and historical abnormal data of the sweeping robot;

[0025] An intelligent power and speed management unit, configured to pre-process the first multi-dimensional data to obtain first multi-dimensional detection data, fuse the first multi-dimensional detection data, and generate a first environment state vector; calculate a first motor power requirement and a first speed of the sweeping robot according to the first environment state vector; and control the voltage, current, and speed of the motor in real time according to the first motor power requirement and the first speed;

[0026] The adaptive noise reduction and vibration suppression unit is used to analyze the second data by Fourier transform method to obtain the frequency domain component of the second data, and obtain the first adjustment according to the second data and the frequency domain component. signal duty cycle and the first adjustment torque; and adaptively adjust the motor of the sweeping robot signal duty cycle and torque;

[0027] An abnormality identification and control unit is used to construct a first abnormality data set through the historical abnormality data, construct a second abnormality identification data set through the third data, and train an abnormality identification model through the first abnormality data set to obtain a trained abnormality identification model, and use the trained abnormality identification model to identify the second abnormality identification data set to obtain a first similarity. If the first similarity is greater than a preset threshold, the operating parameters of the motor are adjusted in real time according to the abnormality type output by the output layer of the abnormality identification model.

[0028] A motor system, the system comprising:

[0029] A data acquisition module, used to obtain first multi-dimensional data, second data, third data and historical abnormal data of the sweeping robot;

[0030] an intelligent power and speed management module, configured to pre-process the first multi-dimensional data to obtain first multi-dimensional detection data, fuse the first multi-dimensional detection data, and generate a first environment state vector; calculate a first motor power requirement and a first speed of the sweeping robot according to the first environment state vector; and control the voltage, current, and speed of the motor in real time according to the first motor power requirement and the first speed;

[0031] The adaptive noise reduction and vibration suppression module is used to analyze the second data by Fourier transform method to obtain the frequency domain component of the second data, and obtain the first adjustment according to the second data and the frequency domain component. signal duty cycle and the first adjustment torque; and adaptively adjust the motor of the sweeping robot signal duty cycle and torque;

[0032] An abnormality identification and control module is used to construct a first abnormality data set through the historical abnormality data, construct a second abnormality identification data set through the third data, and train an abnormality identification model through the first abnormality data set to obtain a trained abnormality identification model, and use the trained abnormality identification model to identify the second abnormality identification data set to obtain a first similarity. If the first similarity is greater than a preset threshold, the operating parameters of the motor are adjusted in real time according to the abnormality type output by the output layer of the abnormality identification model.

[0033] A motor control device comprises at least one processor and a memory, wherein the memory is used to store computer program instructions, and the processor is used to read and execute the computer program instructions to control the control device to implement any one of the motor control methods described above.

[0034] A motor comprises a motor control method as described in any one of the above items.

[0035] Compared with the prior art, the present invention has the following beneficial effects:

[0036] 1. The present invention obtains multi-dimensional data of the sweeping robot and generates an environmental state vector to obtain the power demand and speed of the motor, and controls the voltage and current of the motor in real time to ensure that the motor maintains the best operating state in a complex and changeable environment, thereby improving the intelligent control capability of the motor and thus improving the operating efficiency and safety of the motor.

[0037] 2. The present invention performs spectrum analysis on the noise data and vibration data of the sweeping robot by the Fourier transform method, and adaptively adjusts the operating parameters of the motor according to the time domain information and frequency domain information of the noise data and vibration data, thereby significantly reducing the noise and vibration during the operation of the motor and improving the intelligent control capability of the motor, which is beneficial to improving the operating efficiency of the motor and the user experience.

[0038] 3. The present invention obtains the historical abnormal data of the sweeping robot, constructs a first abnormal data set, and uses a convolutional neural network to train the data set to extract the correlation between the historical abnormal data features and the abnormal type. And through the trained abnormality recognition model, the operation abnormality of the motor is detected and predicted in real time to ensure that the operating parameters of the motor can be quickly adjusted when the abnormality occurs to avoid the impact of the abnormality on the motor performance. The present invention significantly improves the intelligent control capability of the motor, strengthens the ability to respond quickly to abnormal situations, and effectively ensures the operation safety of the motor and the overall reliability of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 A flow chart of a motor control method provided by an embodiment of the present invention;

[0040] Figure 2 A schematic diagram of a motor control device provided by an embodiment of the present invention;

[0041] Figure 3 A schematic diagram of a motor system provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0042] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0043] Embodiment 1

[0044] In order to improve the intelligent control efficiency of the motor of the sweeping robot, a motor control method is applied;

[0045] Reference Figure 1 , is a flow chart of a motor control method, comprising:

[0046] S1. Using multiple sensors to collect the first multi-dimensional data of the sweeping robot in real time, the sensors include optical sensors, pressure sensors, accelerometers, current sensors and temperature sensors;

[0047] S2. Preprocessing the first multidimensional data collected to obtain first multidimensional detection data;

[0048] Furthermore, the first multidimensional data includes ground material type, pressure distribution data, obstacle detection data, motion state data, current data and temperature data; the ground material type is obtained by analyzing the light intensity data and reflected light data obtained by the optical sensor; the pressure distribution data is obtained by measuring the pressure distribution of each point on the bottom of the sweeping robot through a pressure sensor; the obstacle detection data is obtained by measuring the instantaneous pressure value of the sweeping robot when it contacts an obstacle through a pressure sensor; the motion state data is obtained by measuring the three-axis acceleration of the sweeping robot in real time through an accelerometer; the current data is obtained by measuring the current of the motor in real time through a current sensor; the temperature data is obtained by measuring the working temperature of the motor in real time through a temperature sensor; the second data includes the noise data of the motor during operation collected by the acoustic sensor and the motor vibration data collected by the vibration sensor.

[0049] S3. Fusing the first multi-dimensional detection data to generate a first environmental state vector;

[0050] Furthermore, the first environment state vector The vector includes multiple components; each component is obtained by multiplying a data item in the first multidimensional detection data by a corresponding first impact weight coefficient; the first impact weight coefficient is used to represent the importance of the data item in the overall environmental state;

[0051] Furthermore, the first environment state vector is:

[0052] ;

[0053] in, represents the first environment state vector; Indicates the first multi-dimensional detection data individual data; Indicates The first impact weight coefficient;

[0054] S4. Calculate the first motor power requirement and the first speed of the cleaning robot according to the first environmental state vector; and control the voltage and current of the motor in real time according to the first motor power requirement and the first speed;

[0055] In this embodiment, the control idea is: by adjusting the voltage to ensure that the voltage meets the first speed requirement, and adjusting the current so that the product of the voltage and the current is equal to the first motor power requirement; to achieve the dual control goals of power and speed;

[0056] The first motor power requirement By the first environment state vector Each component in is combined with a second influence weight coefficient; the second influence weight coefficient is used to represent the influence of each component on the motor power demand;

[0057] Furthermore, the power requirement of the first motor is:

[0058] ;

[0059] in, represents the power requirement of the first motor; represents the first environmental state vector The second impact weight coefficient of the component; represents the first environmental state vector Quantity;

[0060] The first speed According to the first environment state vector The third influence weight coefficient is used to represent the influence degree of each component on the motor speed.

[0061] The first speed is:

[0062]

[0063] in, represents the first rotation speed; represents the first environmental state vector The third impact weight coefficient of the component; represents the first environmental state vector Quantity;

[0064] This embodiment obtains multi-dimensional data of the sweeping robot and generates an environmental state vector to obtain the power demand and speed of the motor, and controls the voltage and current of the motor in real time to ensure that the motor maintains the best operating state in a complex and changeable environment, thereby improving the intelligent control capability of the motor and thus improving the operating efficiency and safety of the motor.

[0065] S5. Obtain second data of the sweeping robot, analyze the second data using the Fourier transform method to obtain the frequency domain component of the second data, and adaptively obtain the first adjustment of the sweeping robot motor according to the second data and the frequency domain component. signal duty cycle and first adjusted torque;

[0066] Further, the frequency domain component includes a characteristic frequency component of a noise signal and a characteristic frequency component of a vibration signal;

[0067] The first adjustment The calculation process of the signal duty cycle is as follows: the characteristic frequency component of the noise signal is multiplied by the first signal adjustment parameter to obtain the first Signal correction term; by multiplying the characteristic frequency component of the vibration signal with the second signal adjustment parameter, a second signal correction term; and based on these two Signal correction term to the initial signal The duty cycle is corrected to obtain the first adjustment The first signal adjustment parameter and the second signal adjustment parameter represent the noise signal and the vibration signal respectively. The impact of signal conditioning;

[0068] Furthermore, the first adjustment The signal duty cycle is:

[0069] ;

[0070] in, Indicates the first adjustment Signal duty cycle; Indicates initial Signal duty cycle; Represents the first signal adjustment parameter, which is used to represent the noise signal The impact of signal conditioning; Represents the characteristic frequency component of the noise signal; Represents the characteristic frequency component of the vibration signal; Indicates the second signal adjustment parameter, used to indicate the vibration signal The impact of signal conditioning;

[0071] The calculation process of the first adjustment torque is as follows: a first torque correction term is obtained by multiplying a characteristic frequency component of a noise signal by a first torque adjustment parameter; a second torque correction term is obtained by multiplying a characteristic frequency component of a vibration signal by a second torque adjustment parameter; and the initial torque is corrected according to the two torque correction terms to obtain the first adjustment torque; the first torque adjustment parameter and the second torque adjustment parameter respectively represent the influence of the noise signal and the vibration signal on the torque adjustment;

[0072] Further, the first adjustment torque is:

[0073] ;

[0074] in, represents the first adjustment torque; Indicates the initial torque value; represents a first torque adjustment parameter, used to represent the influence of the noise signal on the torque adjustment; Represents the characteristic frequency component of the noise signal; Represents the characteristic frequency component of the vibration signal; represents a second torque adjustment parameter, used to represent the influence of the vibration signal on the torque adjustment;

[0075] This embodiment performs spectral analysis on the noise data and vibration data of the sweeping robot through the Fourier transform method, and adaptively adjusts the operating parameters of the motor according to the time domain information and frequency domain information of the noise data and vibration data, thereby significantly reducing the noise and vibration during the operation of the motor and improving the intelligent control capability of the motor, which is beneficial to improving the operating efficiency of the motor and user experience.

[0076] S6. Obtain historical abnormal data and third data of the sweeping robot, construct a first abnormal data set through the historical abnormal data, construct a second abnormal recognition data set through the third data, and train an abnormal recognition model through the first abnormal data set to obtain a trained abnormal recognition model, and use the trained abnormal recognition model to recognize the second abnormal recognition data set to obtain a first similarity. If the first similarity is greater than a preset threshold, adjust the operating parameters of the motor in real time according to the abnormality type output by the output layer of the abnormal recognition model;

[0077] Furthermore, the first abnormal data set includes abnormal temperature data, abnormal vibration data, abnormal current data and abnormal rotation speed; the third data includes temperature data, vibration data, current data and rotation speed; the training process of the abnormal recognition model includes labeling the first abnormal data set with abnormal type labels, and analyzing the first abnormal data set through a convolutional neural network to obtain the correlation between historical abnormal data features and the abnormal type and the historical abnormal data features, and calculate the error between the predicted value and the actual label through a cross entropy loss function to finally obtain the trained abnormal recognition model.

[0078] Furthermore, the calculation process of the first similarity is: inputting the second abnormality recognition data into the trained abnormality recognition model, and passing through the preprocessing layer, feature extraction layer and output layer of the model in sequence; the feature extraction layer is used to provide CNN for feature extraction to obtain a first abnormality feature, and according to the historical abnormal data features and the first abnormality features, a cosine similarity formula is established to obtain the first similarity, and if the first similarity is greater than the preset threshold, the operating parameters of the motor are dynamically adjusted according to the first similarity; the operating parameters include speed, torque and power.

[0079] The first similarity is:

[0080] ;

[0081] in, represents the first similarity; The first abnormal feature eigenvalues; The first feature values.

[0082] The abnormality types include overheating, vibration, excessive load, excessive current and abnormal speed. The operating parameters that need to be adjusted include voltage, speed, torque and current; the relationship between the abnormality type and the adjustment parameter is shown in Table 1;

[0083] Table 1 Relationship between abnormality types and adjustment parameters

[0084]

[0085] The above voltage, current, torque and The signal duty cycle adjustment formula is:

[0086] ;

[0087] ;

[0088] ;

[0089] ;

[0090] in, The distribution represents the adjusted voltage, current, torque and Signal duty cycle; The voltage, current, torque and Signal duty cycle; Indicates voltage regulation parameters; Indicates the current regulation parameter; Indicates torque adjustment parameter; express Signal duty cycle adjustment parameters; represents the first similarity; represents the preset threshold;

[0091] The present invention obtains the historical abnormal data of the sweeping robot, constructs a first abnormal data set, and uses a convolutional neural network to train the data set to extract the correlation between the historical abnormal data features and the abnormal type. And through the trained abnormality recognition model, the operation abnormality of the motor is detected and predicted in real time, ensuring that the operating parameters of the motor can be quickly adjusted when the abnormality occurs to avoid the impact of the abnormality on the motor performance. The present invention significantly improves the intelligent control capability of the motor, strengthens the ability to respond quickly to abnormal situations, and effectively ensures the operation safety of the motor and the overall reliability of the system.

[0092] This embodiment also includes a motor control device, which includes at least one processor and a memory, the memory is used to store computer program instructions, and the processor is used to read and execute the computer program instructions to control the control device to implement any one of the motor control methods described above.

[0093] This embodiment also includes a motor, including any one of the motor control methods described above.

[0094] Embodiment 2

[0095] In order to improve the intelligent control efficiency of the motor of the sweeping robot, a motor control device is applied;

[0096] Reference Figure 2 , is a schematic diagram of a motor control device, comprising:

[0097] A motor control device, the control device comprising:

[0098] A data acquisition unit, used for acquiring first multi-dimensional data, second data, third data and historical abnormal data of the sweeping robot;

[0099] an intelligent power and speed management unit, configured to preprocess the first multidimensional data to obtain first multidimensional detection data, fuse the first multidimensional detection data, and generate a first environment state vector; calculate a first motor power requirement and a first speed of the sweeping robot according to the first environment state vector; and control the voltage and current of the motor in real time according to the first motor power requirement and the first speed;

[0100] Furthermore, the first environment state vector is:

[0101] ;

[0102] in, represents the first environment state vector; Indicates the first multi-dimensional detection data individual data; Indicates The first impact weight coefficient;

[0103] Furthermore, the power requirement of the first motor is:

[0104] ;

[0105] in, represents the power requirement of the first motor; represents the first environmental state vector The second impact weight coefficient of the component; represents the first environmental state vector Quantity;

[0106] The first speed According to the first environment state vector The third influence weight coefficient is used to represent the influence degree of each component on the motor speed.

[0107] The first speed is:

[0108]

[0109] in, represents the first rotation speed; represents the first environmental state vector The third impact weight coefficient of the component; represents the first environmental state vector Quantity;

[0110] The adaptive noise reduction and vibration suppression unit is used to analyze the second data by Fourier transform method to obtain the frequency domain component of the second data, and obtain the first adjustment according to the second data and the frequency domain component. signal duty cycle and the first adjustment torque; and adaptively adjust the motor of the sweeping robot signal duty cycle and torque;

[0111] The first adjustment The calculation process of the signal duty cycle is as follows: the characteristic frequency component of the noise signal is multiplied by the first signal adjustment parameter to obtain the first Signal correction term; by multiplying the characteristic frequency component of the vibration signal with the second signal adjustment parameter, a second signal correction term; and based on these two The signal correction term is the initial The signal duty cycle is corrected to obtain the first adjustment The first signal adjustment parameter and the second signal adjustment parameter represent the noise signal and the vibration signal respectively. The impact of signal conditioning;

[0112] Furthermore, the first adjustment The signal duty cycle is:

[0113] ;

[0114] in, Indicates the first adjustment Signal duty cycle; Indicates initial Signal duty cycle; Represents the first signal adjustment parameter, which is used to represent the noise signal The impact of signal conditioning; Represents the characteristic frequency component of the noise signal; Represents the characteristic frequency component of the vibration signal; Indicates the second signal adjustment parameter, used to indicate the vibration signal The impact of signal conditioning;

[0115] The first torque The calculation process is as follows: the first torque correction term is obtained by multiplying the characteristic frequency component of the noise signal by the first torque adjustment parameter; the second torque correction term is obtained by multiplying the characteristic frequency component of the vibration signal by the second torque adjustment parameter; and the initial torque is corrected according to the two torque correction terms. Make corrections and obtain the ; The first torque adjustment parameter and the second torque adjustment parameter respectively represent the influence of the noise signal and the vibration signal on the torque adjustment;

[0116] Further, the first adjustment torque is:

[0117] ;

[0118] in, represents the first adjustment torque; Indicates the initial torque value; represents a first torque adjustment parameter, used to represent the influence of the noise signal on the torque adjustment; Represents the characteristic frequency component of the noise signal; Represents the characteristic frequency component of the vibration signal; represents a second torque adjustment parameter, used to represent the influence of the vibration signal on the torque adjustment;

[0119] An abnormality identification and control unit is used to construct a first abnormality data set through the historical abnormality data, construct a second abnormality identification data set through the third data, and train an abnormality identification model through the first abnormality data set to obtain a trained abnormality identification model, and use the trained abnormality identification model to identify the second abnormality identification data set to obtain a first similarity. If the first similarity is greater than a preset threshold, the operating parameters of the motor are adjusted in real time according to the abnormality type output by the output layer of the abnormality identification model.

[0120] Furthermore, the first abnormal data set includes abnormal temperature data, abnormal vibration data, abnormal current data and abnormal rotation speed; the third data includes temperature data, vibration data, current data and rotation speed; the training process of the abnormal recognition model includes labeling the first abnormal data set with abnormal type labels, and analyzing the first abnormal data set through a convolutional neural network to obtain the correlation between historical abnormal data features and the abnormal type and the historical abnormal data features, and calculate the error between the predicted value and the actual label through a cross entropy loss function to finally obtain the trained abnormal recognition model.

[0121] Furthermore, the calculation process of the first similarity is: inputting the second abnormality recognition data into the trained abnormality recognition model, and passing through the preprocessing layer, feature extraction layer and output layer of the model in sequence; the feature extraction layer is used to provide CNN for feature extraction to obtain a first abnormality feature, and according to the historical abnormal data features and the first abnormality features, a cosine similarity formula is established to obtain the first similarity, and if the first similarity is greater than the preset threshold, the operating parameters of the motor are dynamically adjusted according to the first similarity; the operating parameters include speed, torque and power.

[0122] The first similarity is:

[0123] ;

[0124] in, represents the first similarity; The first abnormal feature eigenvalues; The first feature values.

[0125] The abnormality types include overheating, vibration, excessive load, excessive current and abnormal speed. The operating parameters that need to be adjusted include voltage, speed, torque and current; the relationship between the abnormality type and the parameters that need to be adjusted is shown in Table 2;

[0126] Table 2 Relationship between abnormality types and parameters to be adjusted

[0127]

[0128] The above voltage, current, torque and The signal duty cycle adjustment formula is:

[0129] ;

[0130] ;

[0131] ;

[0132] ;

[0133] in, The distribution represents the adjusted voltage, current, torque and Signal duty cycle; The voltage, current, torque and Signal duty cycle; Indicates voltage regulation parameters; Indicates the current regulation parameter; Indicates torque adjustment parameter; express Signal duty cycle adjustment parameters; represents the first similarity; represents the preset threshold;

[0134] Embodiment 3

[0135] In order to improve the intelligent control efficiency of the motor of the sweeping robot, a motor system is applied;

[0136] Reference Figure 3 , is a schematic diagram of a motor system, a motor system, the system comprising:

[0137] A data acquisition module, used to obtain first multi-dimensional data, second data, third data and historical abnormal data of the sweeping robot;

[0138] an intelligent power and speed management module, configured to preprocess the first multidimensional data to obtain first multidimensional detection data, fuse the first multidimensional detection data, and generate a first environmental state vector; calculate a first motor power requirement and a first speed of the sweeping robot according to the first environmental state vector; and control the voltage and current of the motor in real time according to the first motor power requirement and the first speed;

[0139] The adaptive noise reduction and vibration suppression module is used to analyze the second data by Fourier transform method to obtain the frequency domain component of the second data, and obtain the first adjustment according to the second data and the frequency domain component. signal duty cycle and the first adjustment torque; and adaptively adjust the motor of the sweeping robot Signal Duty Cycle and Torque

[0140] An abnormality identification and control module is used to construct a first abnormality data set through the historical abnormality data, construct a second abnormality identification data set through the third data, and train an abnormality identification model through the first abnormality data set to obtain a trained abnormality identification model, and use the trained abnormality identification model to identify the second abnormality identification data set to obtain a first similarity. If the first similarity is greater than a preset threshold, the operating parameters of the motor are adjusted in real time according to the abnormality type output by the output layer of the abnormality identification model.

[0141] Furthermore, the first abnormal data set includes abnormal temperature data, abnormal vibration data, abnormal current data and abnormal rotation speed; the third data includes temperature data, vibration data, current data and rotation speed; the training process of the abnormal recognition model includes labeling the first abnormal data set with abnormal type labels, and analyzing the first abnormal data set through a convolutional neural network to obtain the correlation between historical abnormal data features and the abnormal type and the historical abnormal data features, and calculate the error between the predicted value and the actual label through a cross entropy loss function to finally obtain the trained abnormal recognition model.

[0142] Furthermore, the calculation process of the first similarity is: inputting the second abnormality recognition data into the trained abnormality recognition model, and passing through the preprocessing layer, feature extraction layer and output layer of the model in sequence; the feature extraction layer is used to provide CNN for feature extraction to obtain a first abnormality feature, and according to the historical abnormal data features and the first abnormality features, a cosine similarity formula is established to obtain the first similarity, and if the first similarity is greater than the preset threshold, the operating parameters of the motor are dynamically adjusted according to the first similarity; the operating parameters include speed, torque and power.

[0143] The first similarity is:

[0144] ;

[0145] in, represents the first similarity; The first abnormal feature eigenvalues; The first characteristic values.

[0146] The abnormality types include overheating, vibration, excessive load, excessive current and abnormal speed. The operating parameters that need to be adjusted include voltage, speed, torque and current; the relationship between the abnormality type and the adjustment parameter is shown in Table 1;

[0147] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A motor control method, characterized in that: include: S1. Using multiple sensors to collect the first multi-dimensional data of the sweeping robot in real time, the sensors include optical sensors, pressure sensors, accelerometers, current sensors and temperature sensors; S2. Preprocessing the first multidimensional data collected to obtain first multidimensional detection data; S3. Fusing the first multi-dimensional detection data to generate a first environmental state vector; S4. Calculate the first motor power requirement and the first speed of the cleaning robot according to the first environmental state vector; and control the voltage and current of the motor in real time according to the first motor power requirement and the first speed; S5. Obtain second data of the sweeping robot, analyze the second data using the Fourier transform method to obtain the frequency domain component of the second data, and adaptively obtain the first adjustment of the sweeping robot motor according to the second data and the frequency domain component. signal duty cycle and first adjusted torque; S6. Obtain historical abnormal data and third data of the sweeping robot, construct a first abnormal data set through the historical abnormal data, construct a second abnormal recognition data set through the third data, and train an abnormal recognition model through the first abnormal data set to obtain a trained abnormal recognition model, and use the trained abnormal recognition model to recognize the second abnormal recognition data set to obtain a first similarity. If the first similarity is greater than a preset threshold, adjust the operating parameters of the motor in real time according to the abnormality type output by the output layer of the abnormal recognition model.

2. A motor control method according to claim 1, characterized in that: The first multidimensional data includes ground material type, pressure distribution data, obstacle detection data, motion state data, current data and temperature data; the ground material type is obtained by analyzing the light intensity data and reflected light data obtained by the optical sensor; the pressure distribution data is obtained by measuring the pressure distribution of each point on the bottom of the sweeping robot through the pressure sensor; the obstacle detection data is obtained by measuring the instantaneous pressure value of the sweeping robot when it contacts an obstacle through the pressure sensor; the motion state data is obtained by measuring the three-axis acceleration of the sweeping robot in real time through the accelerometer; the current data is obtained by measuring the current of the motor in real time through the current sensor; the temperature data is obtained by measuring the working temperature of the motor in real time through the temperature sensor; the second data includes the noise data of the motor during operation collected by the acoustic sensor and the motor vibration data collected by the vibration sensor.

3. A motor control method according to claim 1, characterized in that: The first environment state vector The vector includes multiple components; each component is obtained by multiplying a data item in the first multidimensional detection data by a corresponding first impact weight coefficient; the first impact weight coefficient is used to represent the importance of the data item in the overall environmental state; The first motor power requirement By the first environment state vector Each component in is combined with a second influence weight coefficient; the second influence weight coefficient is used to represent the influence of each component on the motor power demand; The first speed According to the first environment state vector The third influence weight coefficient is used to represent the influence degree of each component on the motor speed.

4. A motor control method according to claim 1, characterized in that: The frequency domain components include characteristic frequency components of noise signals and characteristic frequency components of vibration signals; The first adjustment The calculation process of the signal duty cycle is as follows: the characteristic frequency component of the noise signal is multiplied by the first signal adjustment parameter to obtain the first Signal correction term; by multiplying the characteristic frequency component of the vibration signal with the second signal adjustment parameter, a second No. Amendment Item; According to these two The signal correction term is the initial The signal duty cycle is corrected to obtain the The first signal adjustment parameter and the second signal adjustment parameter represent the noise signal and the vibration signal respectively. The impact of signal conditioning; The calculation process of the first adjustment torque is: a first torque correction term is obtained by multiplying the characteristic frequency component of the noise signal by the first torque adjustment parameter; a second torque correction term is obtained by multiplying the characteristic frequency component of the vibration signal by the second torque parameter; The initial torque is corrected according to the two torque correction items to obtain the first adjustment torque; the first torque adjustment parameter and the second torque adjustment parameter respectively represent the influence of the noise signal and the vibration signal on the torque adjustment.

5. A motor control method according to claim 1, characterized in that: The first abnormal data set includes abnormal temperature data, abnormal vibration data, abnormal current data and abnormal rotation speed; the third data includes temperature data, vibration data, current data and rotation speed; the training process of the abnormal recognition model includes labeling the first abnormal data set with abnormal type labels, and analyzing the first abnormal data set through a convolutional neural network to obtain the correlation between historical abnormal data features and the abnormal type and the historical abnormal data features, and calculate the error between the predicted value and the actual label through a cross entropy loss function to finally obtain the trained abnormal recognition model.

6. A motor control method according to claim 5, characterized in that: The calculation process of the first similarity is: inputting the second abnormality recognition data into the trained abnormality recognition model, and passing through the preprocessing layer, feature extraction layer and output layer of the model in sequence; the feature extraction layer is used to provide CNN for feature extraction to obtain the first abnormality feature, and according to the historical abnormal data features and the first abnormality feature, a cosine similarity formula is established to obtain the first similarity; if the first similarity is greater than the preset threshold, the operating parameters of the motor are dynamically adjusted according to the first similarity.

7. A motor control device, characterized in that: The control device comprises: A data acquisition unit, used for acquiring first multi-dimensional data, second data, third data and historical abnormal data of the sweeping robot; an intelligent power and speed management unit, configured to preprocess the first multidimensional data to obtain first multidimensional detection data, fuse the first multidimensional detection data, and generate a first environment state vector; calculate a first motor power requirement and a first speed of the sweeping robot according to the first environment state vector; and control the voltage and current of the motor in real time according to the first motor power requirement and the first speed; The adaptive noise reduction and vibration suppression unit is used to analyze the second data by Fourier transform method to obtain the frequency domain component of the second data, and obtain the first adjustment according to the second data and the frequency domain component. signal duty cycle and the first adjustment torque; and adaptively adjust the motor of the sweeping robot signal duty cycle and torque; an abnormality identification and control unit, used to construct a first abnormality data set through the historical abnormal data, construct a second abnormality identification data set through the third data, and train an abnormality identification model through the first abnormality data set to obtain a trained abnormality identification model, and use the trained abnormality identification model to identify the second abnormality identification data set to obtain a first similarity. If the first similarity is greater than a preset threshold, the operating parameters of the motor are adjusted in real time according to the abnormality type output by the output layer of the abnormality identification model.

8. A motor control device, characterized in that: The control device includes at least one processor and a memory, the memory is used to store computer program instructions, and the processor is used to read and execute the computer program instructions to control the control device to implement a motor control method as described in any one of claims 1 to 6.

9. A motor, characterized in that: It comprises a motor control method as described in any one of claims 1-6.

Images