New energy special vehicle multi-drive motor cooperative control method and system

Abstract

The invention discloses a multi-drive-motor cooperative control method and system for a new energy special vehicle, and relates to the technical field of vehicle control. The method comprises the following steps: constructing a digital twin model of a special vehicle, carrying out analogue simulation on the digital twin model, constructing a road condition prediction model by using an LSTM neural network to output predicted road condition parameters, obtaining a driving anti-skid index, a yaw velocity deviation rate and a comprehensive stability coefficient of the special vehicle, obtaining a combined acceleration index, a power maintenance degree and a comprehensive dynamic coefficient of the special vehicle, and carrying out simulation on the road condition prediction model. Setting an expected stability coefficient and an expected dynamic coefficient, obtaining corresponding cooperative control parameters in the digital twin model, and constructing a cooperative control model of the special vehicle to perform cooperative control on the special vehicle; the motor control parameters with pertinence can be provided according to different road conditions, and the running stability and dynamic property of the vehicle can be guaranteed.

Description

Technical Field

[0001] The present invention relates to the field of vehicle control technology, and in particular to a method and system for coordinated control of multiple drive motors of a new energy special vehicle. Background Art

[0002] The coordinated control of multiple drive motors of new energy special vehicles is a highly integrated and complex technology. Through advanced hardware architecture, intelligent control algorithms, reliable communication mechanisms and comprehensive safety design, the vehicle achieves high efficiency, stability and dynamic performance under various working conditions. Different from the traditional special vehicle driving mode, new energy special vehicles generally adopt multi-motor driving control mode. This driving mode will lead to greater uncertainty between the longitudinal and lateral stability of the vehicle. How to achieve vehicle stability and dynamic control under multi-motor driving conditions is one of the current technical issues in multi-motor drive vehicle control. The prior art lacks a control method that can take into account both the driving stability and overall dynamics of special vehicles, and the existing technical means often ignore the impact of road conditions on motor drive, resulting in the inability to provide targeted motor control parameters. In view of the shortcomings of the prior art, the present invention provides a method and system for coordinated control of multiple drive motors of new energy special vehicles. Summary of the invention

[0003] The purpose of the present invention is to provide a method and system for coordinated control of multiple drive motors of new energy special vehicles.

[0004] The purpose of the present invention can be achieved through the following technical solutions: A new energy special vehicle multi-drive motor coordinated control system, comprising the following modules: The data acquisition module is used to obtain the basic information of special vehicles, build the corresponding digital twin model, and simulate the digital twin model; The road condition prediction module is used to construct a three-dimensional point cloud model of the road condition and obtain the corresponding road condition parameters, and use the LSTM neural network to build a road condition prediction model to output the predicted road condition parameters; The data evaluation module is used to obtain the driving anti-skid index and yaw rate deviation rate of the special vehicle, and obtain the corresponding comprehensive stability coefficient, and is also used to obtain the joint acceleration index and power maintenance of the special vehicle, and obtain the corresponding comprehensive power coefficient; The vehicle control module is used to set the expected stability coefficient and the expected power coefficient, and to obtain the corresponding collaborative control parameters by combining different predicted road condition parameters in the digital twin model. It is also used to build a collaborative control model for special vehicles and use the collaborative control model to perform collaborative control on special vehicles.

[0005] Furthermore, the basic information of the special vehicle is obtained and the corresponding digital twin model is constructed. The process of simulating the digital twin model includes: The basic information includes geometric parameters, mass parameters, power structure parameters, transmission structure parameters, suspension and steering parameters, tire and ground parameters of the special vehicle. The digital twin model of the special vehicle is constructed based on the acquired basic information using digital twin technology; The constructed digital twin model is simulated using simulation software to obtain the theoretical operating data of the special vehicle during operation. The theoretical operating data refers to all data generated when the special vehicle is running in the digital twin model.

[0006] Furthermore, the process of constructing a three-dimensional point cloud model of the road condition and obtaining corresponding road condition parameters, and constructing a road condition prediction model using an LSTM neural network to output predicted road condition parameters includes: A laser scanning unit and a visual acquisition unit are respectively set on the special vehicle to obtain point cloud data of the road in the direction of the special vehicle's advance, the point cloud data is registered using the ICP algorithm, and the point cloud data is converted into a three-dimensional point cloud model using a voxelization algorithm, and the corresponding road condition parameters are obtained in the three-dimensional point cloud model, including road slope, friction coefficient, and obstacle distribution; Acquire a road condition prediction set including point cloud data of common roads in different mines and construction sites and road condition parameters thereof, and divide the acquired road condition prediction set into a first training set and a first test set; The point cloud data of common roads in different mines and construction sites in the first training set are used as input data of the LSTM neural network, and the corresponding road condition parameters are used as output data of the LSTM neural network, and the LSTM neural network is trained to obtain an initial LSTM neural network; The initial LSTM neural network is verified using the first test set, and the initial LSTM neural network with a first test error threshold that is less than or equal to the preset value is output as a road condition prediction model. The point cloud data obtained by the laser scanning unit and the visual acquisition unit are continuously input into the road condition prediction model to output the corresponding predicted road condition parameters.

[0007] Furthermore, the process of obtaining the driving anti-skid index and the yaw rate deviation rate of the special vehicle and obtaining the corresponding comprehensive stability coefficient includes: The friction coefficient in the predicted road condition parameters is recorded as , the highest friction coefficient of the special vehicle on its driving road is recorded as ; The actual operation data of the special vehicle is obtained. The actual operation data refers to all the data generated when the special vehicle is running in the actual application scenario. The average angular velocity of each wheel in the actual operation data is recorded as ; In the digital twin model, the operation process of the special vehicle is simulated in combination with the predicted road condition parameters to obtain the theoretical operation data, and the theoretical angular velocity mean of each wheel in the theoretical operation data is recorded as , obtain the driving anti-skid index T of the special vehicle at the corresponding time c ;

[0008] The longitudinal speed, turning radius, and yaw rate of the special vehicle in the actual operation data are recorded as , R, , obtain the yaw rate deviation rate Y of the special vehicle at the corresponding moment r ;

[0009] is the preset expected yaw rate, is the driving anti-slip index T c and yaw rate deviation rate Y r Set the weight value Q respectively 1 and Q 2 , obtain the comprehensive stability coefficient S of the special vehicle at the corresponding time w ;

[0010] Furthermore, the process of obtaining the combined acceleration index and power maintenance of the special vehicle and obtaining the corresponding comprehensive power coefficient includes: The drive motors of special vehicles are numbered as i, i=1, 2, ..., n, where n is the number of drive motors. The actual acceleration of the i-th drive motor in the actual operation data is recorded as , get the acceleration variance of each drive motor at the same time ;

[0011] The theoretical maximum acceleration of the drive motor of a special vehicle is recorded as , obtain the joint acceleration index J of special vehicles at the corresponding time a , is the preset coordination parameter;

[0012] The actual power and real-time temperature of the i-th drive motor in the actual operation data are recorded as P i and W i , the rated power of the drive motor of the special vehicle is recorded as P r , obtain the power maintenance degree P of the special vehicle at the corresponding moments , is the preset temperature attenuation parameter, W b is the preset reference temperature;

[0013] is the joint acceleration index J a And power maintenance P s Set the weight value Q respectively 3 and Q 4 , obtain the comprehensive power coefficient S of the special vehicle at the corresponding time d , J ref is the joint acceleration index under ideal conditions;

[0014] Furthermore, the process of setting the expected stability coefficient and the expected power coefficient and obtaining the corresponding cooperative control parameters by combining different predicted road condition parameters in the digital twin model includes: The weight values ​​of the driving anti-slip index, yaw rate deviation rate, combined acceleration index, and power maintenance degree are all adjustable. According to actual needs, each weight value is adjusted to obtain the corresponding expected stability coefficient and expected power coefficient; In the digital twin model, the special vehicle is simulated in combination with different predicted road condition parameters to obtain the driving parameters of each drive motor when the expected stability coefficient and the expected power coefficient are achieved, including speed, torque, and direction, and the driving parameters of each drive motor are used as corresponding coordinated control parameters.

[0015] Furthermore, a collaborative control model for special vehicles is constructed. The process of using the collaborative control model to collaboratively control special vehicles includes: Generate a cooperative control set according to cooperative control parameters of the special vehicle under different expected stability coefficients, expected power coefficients, and predicted road condition parameters, and divide the obtained cooperative control set into a second training set and a second test set; Constructing a convolutional neural network, taking different expected stability coefficients, expected power coefficients, and predicted road condition parameters in the second training set as input data of the convolutional neural network, taking corresponding cooperative control parameters as output data of the convolutional neural network, and training the convolutional neural network to obtain an initial convolutional neural network; Using the second test set to perform model verification on the initial convolutional neural network, and outputting an initial convolutional neural network that is less than or equal to a preset second test error threshold as a collaborative control model; The predicted road condition parameters of the current special vehicle on its driving road are used as real-time road condition parameters, combined with the set expected stability coefficient and expected power coefficient, and input into the collaborative control model to output corresponding real-time control parameters, and each drive motor of the special vehicle is controlled according to the real-time control parameters.

[0016] A method for cooperatively controlling multiple drive motors of a new energy special vehicle comprises the following steps: Step S1: Obtain basic information of special vehicles, build a corresponding digital twin model, and simulate the digital twin model; Step S2: construct a three-dimensional point cloud model of the road condition and obtain the corresponding road condition parameters, and use the LSTM neural network to construct a road condition prediction model to output the predicted road condition parameters; Step S3: obtaining the driving anti-skid index and yaw rate deviation rate of the special vehicle, and obtaining the corresponding comprehensive stability coefficient, and obtaining the joint acceleration index and power maintenance degree of the special vehicle, and obtaining the corresponding comprehensive power coefficient; Step S4: Set the expected stability coefficient and the expected power coefficient, obtain the corresponding collaborative control parameters by combining different predicted road condition parameters in the digital twin model, and build a collaborative control model for special vehicles, and use the collaborative control model to perform collaborative control on special vehicles.

[0017] Compared with the prior art, the present invention has the following beneficial effects: The present invention constructs a digital twin model of a special vehicle and simulates it, so as to simulate the operation data of the special vehicle under ideal conditions, and constructs a road condition prediction model using an LSTM neural network, so as to directly output corresponding predicted road condition parameters according to the point cloud data of the road; By obtaining the driving anti-skid index, yaw rate deviation rate, joint acceleration index, and power maintenance of special vehicles, a comprehensive evaluation of the driving stability and overall dynamics of special vehicles can be achieved. The expected stability coefficient and expected power coefficient are set in the digital twin model, and the corresponding collaborative control parameters are obtained in combination with different predicted road condition parameters, and a collaborative control model is constructed. It can effectively combine user needs and real-time road conditions to output the driving parameters of each driving motor, and control it to operate according to the acquired real-time control parameters, which is conducive to providing targeted motor control parameters according to different road conditions and can ensure the stability and power of vehicle driving. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 It is a schematic diagram of the present invention. DETAILED DESCRIPTION

[0019] like Figure 1As shown, a new energy special vehicle multi-drive motor coordinated control system includes the following modules: The data acquisition module is used to obtain the basic information of special vehicles, build the corresponding digital twin model, and simulate the digital twin model; The road condition prediction module is used to construct a three-dimensional point cloud model of the road condition and obtain the corresponding road condition parameters, and use the LSTM neural network to build a road condition prediction model to output the predicted road condition parameters; The data evaluation module is used to obtain the driving anti-skid index and yaw rate deviation rate of the special vehicle, and obtain the corresponding comprehensive stability coefficient, and is also used to obtain the joint acceleration index and power maintenance of the special vehicle, and obtain the corresponding comprehensive power coefficient; The vehicle control module is used to set the expected stability coefficient and the expected power coefficient, and to obtain the corresponding collaborative control parameters by combining different predicted road condition parameters in the digital twin model. It is also used to build a collaborative control model for special vehicles and use the collaborative control model to perform collaborative control on special vehicles.

[0020] It should be further explained that, in the specific implementation process, the basic information of the special vehicle is obtained, and the corresponding digital twin model is constructed. The process of simulating the digital twin model includes: The special vehicles are vehicles specially designed and modified to meet specific work requirements and perform special tasks. They usually have unique functions, structures and performance characteristics and are used in military, engineering construction, medical and health, fire rescue, transportation and other technical fields; The basic information refers to the data necessary for building a digital twin model of a special vehicle, including geometric parameters, mass parameters, power structure parameters, transmission structure parameters, suspension and steering parameters, tire and ground parameters of the special vehicle. The digital twin model of the special vehicle is built based on the basic information obtained using digital twin technology; At this time, the digital twin model can only reflect the physical structure of the special vehicle. On this basis, the constructed digital twin model is simulated using simulation software. At this time, the digital twin model can obtain the theoretical operating data of the special vehicle during operation. The theoretical operating data refers to all data generated when the special vehicle is running in the digital twin model.

[0021] It should be further explained that, in the specific implementation process, the process of constructing a three-dimensional point cloud model of road conditions and obtaining corresponding road condition parameters, and using an LSTM neural network to construct a road condition prediction model to output predicted road condition parameters includes: A laser scanning unit and a visual acquisition unit are respectively arranged on the special vehicle, and the point cloud data of the road in the direction of advance of the special vehicle are respectively acquired by the laser scanning unit and the visual acquisition unit, and the point cloud data acquired by the two are registered by the ICP algorithm; Extract various feature information from the point cloud data after point cloud registration, including but not limited to surface normals, edges, geometric structures, etc. Modeling is performed based on the point cloud data and the extracted feature information, and the point cloud data is converted into a three-dimensional point cloud model using a voxelization algorithm. The corresponding road condition parameters are obtained from the three-dimensional point cloud model, including road slope, friction coefficient, obstacle distribution, etc.; Acquire a road condition prediction set, the road condition prediction set comprising point cloud data of common roads in different mines and construction sites and their corresponding road condition parameters, and divide the acquired road condition prediction set into a first training set and a first test set; The point cloud data of common roads in different mines and construction sites in the first training set are used as input data of the LSTM neural network, and the corresponding road condition parameters are used as output data of the LSTM neural network, and the LSTM neural network is trained to obtain an initial LSTM neural network; The initial LSTM neural network is verified using the first test set, and the initial LSTM neural network with an output value less than or equal to the preset first test error threshold is used as the corresponding road condition prediction model. The point cloud data obtained by the laser scanning unit and the visual acquisition unit are continuously input into the road condition prediction model to output the corresponding predicted road condition parameters.

[0022] It should be further explained that, in the specific implementation process, the process of obtaining the driving anti-skid index and yaw rate deviation rate of the special vehicle and obtaining the corresponding comprehensive stability coefficient includes: In an embodiment of the present invention, the driving anti-slip index and the yaw rate deviation rate of the special vehicle during its operation are evaluated respectively, and then the driving stability of the special vehicle is comprehensively evaluated based on the above two parameters; The driving anti-skid index is used to evaluate the ability of special vehicles to prevent wheel slip. The friction coefficient in the predicted road condition parameter is recorded as , the highest friction coefficient obtained by the special vehicle on its driving road is recorded as ; The actual operation data of the special vehicle is obtained. The actual operation data refers to all the data generated when the special vehicle is running in the actual application scenario. The average angular velocity of each wheel in the actual operation data is recorded as ; In the digital twin model, the operation process of the special vehicle is simulated in combination with the predicted road condition parameters to obtain the theoretical operation data, and the theoretical angular velocity mean of each wheel in the theoretical operation data is recorded as , obtain the driving anti-skid index of the special vehicle at the corresponding time, recorded as T c ;

[0023] The yaw rate deviation rate is used to evaluate the stability of the special vehicle when turning, where the yaw rate refers to the rotation speed of the special vehicle around the vertical axis, which is recorded as , the longitudinal speed and turning radius of the special vehicle in the actual operation data are recorded as and R, obtain the yaw rate deviation rate of the special vehicle at the corresponding moment, recorded as Y r ;

[0024] in, is the preset expected yaw angular velocity; is the obtained driving anti-slip index T c and yaw rate deviation rate Y r Set the corresponding weight values ​​respectively, recorded as Q 1 and Q 2 , obtain the comprehensive stability coefficient S of the special vehicle at the corresponding time w ;

[0025] It should be further explained that, in the specific implementation process, the process of obtaining the combined acceleration index and power maintenance degree of the special vehicle and obtaining the corresponding comprehensive power coefficient includes: In the embodiment of the present invention, the combined acceleration index, power maintenance degree, and maximum climbing angle dynamic margin of the special vehicle during its operation are evaluated respectively, and then the overall dynamic performance of the special vehicle is comprehensively evaluated based on the above three parameters; The combined acceleration index is used to evaluate the comprehensive acceleration performance of each drive motor when working in coordination. The drive motors of special vehicles are numbered as i, i=1, 2, ..., n, where n is the number of drive motors. The actual acceleration of the i-th drive motor in the actual operation data is recorded as , obtain the acceleration variance of each drive motor at the same time, recorded as ;

[0026] Get the theoretical maximum acceleration of the drive motor of the special vehicle. The drive motors of the same special vehicle are assumed to be of the same model and have the same theoretical maximum acceleration, which is recorded as , obtain the joint acceleration index of special vehicles at the corresponding time, recorded as J a ;

[0027] in, is the preset coordination parameter, with a value range of 0.5-1.2; The power maintenance degree is used to evaluate the ability of each drive motor to maintain stable power output when the load changes. The actual power and real-time temperature of the i-th drive motor in the actual operation data are recorded as P i and W i , get the rated power of the drive motor of the special vehicle, denoted as P r , obtain the power maintenance degree of the special vehicle at the corresponding time, denoted as P s ;

[0028] in, is the preset temperature attenuation parameter, with a value range of 0.02-0.05, W b is the base temperature, which is 25 degrees Celsius; is the obtained joint acceleration index J a And power maintenance P s Set the corresponding weight values ​​respectively, recorded as Q 3 and Q 4 , obtain the comprehensive power coefficient S of the special vehicle at the corresponding time d ;

[0029] Among them, J ref is the joint acceleration index under ideal conditions.

[0030] It should be further explained that, in the specific implementation process, the process of setting the expected stability coefficient and the expected power coefficient and obtaining the corresponding collaborative control parameters by combining different predicted road condition parameters in the digital twin model includes: In the process of obtaining the comprehensive stability coefficient, the weight values ​​corresponding to the two parameters contained therein are adjustable. If the application scenario of the special vehicle is a highway, the weight value of the driving anti-skid index is increased. If the application scenario of the special vehicle is a mountain road, the weight value of the yaw rate deviation rate is increased. The sum of the weight values ​​of the two is always equal to 1. In the process of obtaining the comprehensive power coefficient, the weight values ​​corresponding to the two parameters contained therein are also adjustable. If the user expects the special vehicle to have strong power, the weight value of the combined acceleration index is increased. If the user expects the special vehicle to have a long endurance, the weight value of the power maintenance degree is increased. The sum of the weight values ​​of the two is always equal to 1. The user adjusts each weight value according to actual needs, and uses the comprehensive stability coefficient and comprehensive power coefficient obtained after adjusting the weight values ​​as the expected stability coefficient and expected power coefficient of the special vehicle, respectively. In the digital twin model, the goal is to achieve the expected stability coefficient and expected power coefficient; The special vehicle is simulated in combination with different predicted road condition parameters to obtain the driving parameters of each driving motor when the expected stability coefficient and the expected power coefficient are achieved, wherein the driving parameters include speed, torque, direction, etc., and the driving parameters of each driving motor are used as corresponding coordinated control parameters.

[0031] It should be further explained that, in the specific implementation process, the process of constructing a collaborative control model for special vehicles and using the collaborative control model to collaboratively control special vehicles includes: Generate a corresponding cooperative control set according to cooperative control parameters of the special vehicle under different expected stability coefficients, expected power coefficients, and predicted road condition parameters, and divide the obtained cooperative control set into a second training set and a second test set; Constructing a convolutional neural network, taking different expected stability coefficients, expected power coefficients, and predicted road condition parameters in the second training set as input data of the convolutional neural network, taking corresponding cooperative control parameters as output data of the convolutional neural network, and training the convolutional neural network to obtain an initial convolutional neural network; Using the second test set to perform model verification on the initial convolutional neural network, outputting an initial convolutional neural network that is less than or equal to a preset second test error threshold as the corresponding collaborative control model; The predicted road condition parameters of the current special vehicle on its driving road are used as real-time road condition parameters, combined with the expected stability coefficient and expected power coefficient set by the user, and input into the collaborative control model, and the corresponding real-time control parameters are output, and each drive motor of the special vehicle is controlled according to the real-time control parameters.

[0032] The embodiment of the present invention also includes a method for cooperatively controlling multiple drive motors of a new energy special vehicle, comprising the following steps: Step S1: Obtain basic information of special vehicles, build a corresponding digital twin model, and simulate the digital twin model; Step S2: construct a three-dimensional point cloud model of the road condition and obtain the corresponding road condition parameters, and use the LSTM neural network to construct a road condition prediction model to output the predicted road condition parameters; Step S3: obtaining the driving anti-skid index and yaw rate deviation rate of the special vehicle, and obtaining the corresponding comprehensive stability coefficient, and obtaining the joint acceleration index and power maintenance degree of the special vehicle, and obtaining the corresponding comprehensive power coefficient; Step S4: Set the expected stability coefficient and the expected power coefficient, obtain the corresponding collaborative control parameters by combining different predicted road condition parameters in the digital twin model, and build a collaborative control model for special vehicles, and use the collaborative control model to perform collaborative control on special vehicles.

[0033] The above embodiments are only used to illustrate the technical method of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical method of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical method of the present invention.

Claims

1. A new energy special vehicle multi-drive motor coordinated control system, characterized in that: Includes the following modules: The data acquisition module is used to obtain the basic information of special vehicles, build the corresponding digital twin model, and simulate the digital twin model; The road condition prediction module is used to construct a three-dimensional point cloud model of the road condition and obtain the corresponding road condition parameters, and use the LSTM neural network to build a road condition prediction model to output the predicted road condition parameters; The data evaluation module is used to obtain the driving anti-skid index and yaw rate deviation rate of the special vehicle, and obtain the corresponding comprehensive stability coefficient, and is also used to obtain the joint acceleration index and power maintenance of the special vehicle, and obtain the corresponding comprehensive power coefficient; The vehicle control module is used to set the expected stability coefficient and the expected power coefficient, and to obtain the corresponding collaborative control parameters by combining different predicted road condition parameters in the digital twin model. It is also used to build a collaborative control model for special vehicles and use the collaborative control model to perform collaborative control on special vehicles.

2. A new energy special vehicle multi-drive motor coordinated control system according to claim 1, characterized in that: The process of building a digital twin model of a special vehicle and simulating it includes: The basic information includes geometric parameters, mass parameters, power structure parameters, transmission structure parameters, suspension and steering parameters, tire and ground parameters of the special vehicle. The digital twin model of the special vehicle is constructed based on the acquired basic information using digital twin technology; The constructed digital twin model is simulated using simulation software to obtain the theoretical operating data of the special vehicle during operation. The theoretical operating data refers to all data generated when the special vehicle is running in the digital twin model.

3. A new energy special vehicle multi-drive motor coordinated control system according to claim 2, characterized in that: The process of building a traffic condition prediction model to output predicted traffic condition parameters includes: A laser scanning unit and a visual acquisition unit are respectively set on the special vehicle to obtain point cloud data of the road in the direction of the special vehicle's advance, the point cloud data is registered using the ICP algorithm, and the point cloud data is converted into a three-dimensional point cloud model using a voxelization algorithm, and the corresponding road condition parameters are obtained in the three-dimensional point cloud model, including road slope, friction coefficient, and obstacle distribution; Acquire a road condition prediction set including point cloud data of common roads in different mines and construction sites and road condition parameters thereof, and divide the acquired road condition prediction set into a first training set and a first test set; The point cloud data of common roads in different mines and construction sites in the first training set are used as input data of the LSTM neural network, and the corresponding road condition parameters are used as output data of the LSTM neural network, and the LSTM neural network is trained to obtain an initial LSTM neural network; The initial LSTM neural network is verified using the first test set, and the initial LSTM neural network with a first test error threshold that is less than or equal to the preset value is output as a road condition prediction model. The point cloud data obtained by the laser scanning unit and the visual acquisition unit are continuously input into the road condition prediction model to output the corresponding predicted road condition parameters.

4. A new energy special vehicle multi-drive motor coordinated control system according to claim 3, characterized in that: The process of obtaining the comprehensive stability coefficient of special vehicles includes: The friction coefficient in the predicted road condition parameters is recorded as , the highest friction coefficient of the special vehicle on its driving road is recorded as ; The actual operation data of the special vehicle is obtained. The actual operation data refers to all the data generated when the special vehicle is running in the actual application scenario. The average angular velocity of each wheel in the actual operation data is recorded as ; In the digital twin model, the operation process of the special vehicle is simulated in combination with the predicted road condition parameters to obtain the theoretical operation data, and the theoretical angular velocity mean of each wheel in the theoretical operation data is recorded as , obtain the driving anti-skid index T of the special vehicle at the corresponding time c ; The longitudinal speed, turning radius, and yaw rate of the special vehicle in the actual operation data are recorded as , R, , obtain the yaw rate deviation rate Y of the special vehicle at the corresponding moment r ; is the preset expected yaw rate, is the driving anti-slip index T c and yaw rate deviation rate Y r Set the weight values ​​Q1 and Q2 respectively to obtain the comprehensive stability coefficient S of the special vehicle at the corresponding time w ; 。 5. A new energy special vehicle multi-drive motor coordinated control system according to claim 4, characterized in that: The process of obtaining the comprehensive power coefficient of special vehicles includes: The drive motors of special vehicles are numbered as i, i=1, 2, ..., n, where n is the number of drive motors. The actual acceleration of the i-th drive motor in the actual operation data is recorded as , get the acceleration variance of each drive motor at the same time ; The theoretical maximum acceleration of the drive motor of a special vehicle is recorded as , obtain the joint acceleration index J of special vehicles at the corresponding time a , is the preset coordination parameter; The actual power and real-time temperature of the i-th drive motor in the actual operation data are recorded as P i and W i , the rated power of the drive motor of the special vehicle is recorded as P r , obtain the power maintenance degree P of the special vehicle at the corresponding moment s , is the preset temperature attenuation parameter, W b is the preset reference temperature; is the joint acceleration index J a And power maintenance P s Set the weight values ​​Q3 and Q4 respectively to obtain the comprehensive power coefficient S of the special vehicle at the corresponding moment d , J ref is the joint acceleration index under ideal conditions; 。 6. A new energy special vehicle multi-drive motor coordinated control system according to claim 5, characterized in that: The process of obtaining collaborative control parameters in the digital twin model includes: The weight values ​​of the driving anti-slip index, yaw rate deviation rate, combined acceleration index, and power maintenance degree are all adjustable. According to actual needs, each weight value is adjusted to obtain the corresponding expected stability coefficient and expected power coefficient; In the digital twin model, the special vehicle is simulated in combination with different predicted road condition parameters to obtain the driving parameters of each drive motor when the expected stability coefficient and the expected power coefficient are achieved, including speed, torque, and direction, and the driving parameters of each drive motor are used as corresponding coordinated control parameters.

7. A new energy special vehicle multi-drive motor coordinated control system according to claim 6, characterized in that: The process of using the collaborative control model to collaboratively control special vehicles includes: Generate a cooperative control set according to cooperative control parameters of the special vehicle under different expected stability coefficients, expected power coefficients, and predicted road condition parameters, and divide the obtained cooperative control set into a second training set and a second test set; Constructing a convolutional neural network, taking different expected stability coefficients, expected power coefficients, and predicted road condition parameters in the second training set as input data of the convolutional neural network, taking corresponding cooperative control parameters as output data of the convolutional neural network, and training the convolutional neural network to obtain an initial convolutional neural network; Using the second test set to perform model verification on the initial convolutional neural network, and outputting an initial convolutional neural network that is less than or equal to a preset second test error threshold as a collaborative control model; The predicted road condition parameters of the current special vehicle on its driving road are used as real-time road condition parameters, combined with the set expected stability coefficient and expected power coefficient, and input into the collaborative control model to output corresponding real-time control parameters, and each drive motor of the special vehicle is controlled according to the real-time control parameters.

8. A method for cooperative control of multiple drive motors of a new energy special vehicle, which is implemented based on the cooperative control system of multiple drive motors of a new energy special vehicle according to any one of claims 1 to 7, characterized in that: The method comprises: Step S1: Obtain basic information of special vehicles, build a corresponding digital twin model, and simulate the digital twin model; Step S2: construct a three-dimensional point cloud model of the road condition and obtain the corresponding road condition parameters, and use the LSTM neural network to construct a road condition prediction model to output the predicted road condition parameters; Step S3: obtaining the driving anti-skid index and yaw rate deviation rate of the special vehicle, and obtaining the corresponding comprehensive stability coefficient, and obtaining the joint acceleration index and power maintenance degree of the special vehicle, and obtaining the corresponding comprehensive power coefficient; Step S4: Set the expected stability coefficient and the expected power coefficient, obtain the corresponding collaborative control parameters by combining different predicted road condition parameters in the digital twin model, and build a collaborative control model for special vehicles, and use the collaborative control model to perform collaborative control on special vehicles.

Images