An adaptive telescopic control system applied to a driving wheel of an electric tricycle

Abstract

The present application relates to the technical field of telescopic control, in particular to a self-adaptive telescopic control system applied to a driving wheel of an electric tricycle. The tricycle posture extraction module acquires data of the tricycle data acquisition module and processes the data to obtain a turning posture state matrix, the turning posture state matrix including a first state vector, a second state vector, a lateral state vector and a third state vector; the suspension parameter generation module includes an observation state space unit and an RNN-TD3 model unit, wherein the observation state space unit acquires information of a static data acquisition unit and information of the tricycle posture extraction module and processes the information to obtain an observation state space; the RNN-TD3 model unit acquires the observation state space and processes the observation state space to obtain a suspension parameter adjustment value vector. The suspension parameter adjustment module acquires the suspension parameter adjustment value vector and controls the suspension parameter adjustment value vector through an incremental PID algorithm to realize smooth adjustment of the self-adaptive telescopic control of the driving wheel of the electric tricycle.

Description

Technical Field

[0001] This invention relates to the field of telescopic control technology, specifically an adaptive telescopic control system for the drive wheels of an electric tricycle. Background Technology

[0002] When a tricycle is transporting goods or carrying passengers, the load distribution may be uneven, causing changes in the pressure on the drive wheels, which affects the driving force and suspension performance. This is especially true when making sharp turns or driving on slopes, where the vehicle's center of gravity may shift, potentially causing the drive wheels to lose effective contact.

[0003] Due to their three-point support structure, tricycles have poor lateral stability when turning, making them prone to rollover. Failure to adjust the suspension system of the drive wheels in time can exacerbate this risk. By adjusting the suspension height of the drive wheels to lower the vehicle's center of gravity and increase the support of the outer wheels when a tricycle is turning, and by combining real-time monitoring of the vehicle's lateral status with suspension control, rollover accidents can be avoided.

[0004] Therefore, an adaptive telescopic control system for the drive wheels of electric tricycles is proposed. Summary of the Invention

[0005] The purpose of this invention is to provide an adaptive telescopic control system for the drive wheels of an electric tricycle to improve the stability and maneuverability of the tricycle when turning. First, the tricycle attitude extraction module acquires and processes data from the tricycle data acquisition module to construct a turning attitude state matrix, which measures the state of the electric tricycle when turning. The suspension parameter generation module processes the static data and the turning attitude state matrix of the tricycle using an RNN-TD3 model to obtain a suspension parameter adjustment value vector. The suspension parameter adjustment module then uses an incremental PID algorithm to smoothly adjust the parameters, ultimately achieving adaptive telescopic control of the drive wheels of the electric tricycle when turning.

[0006] To achieve the above objectives, the present invention provides the following technical solution:

[0007] An adaptive telescopic control system for the drive wheels of an electric tricycle includes:

[0008] The tricycle data acquisition module is used to collect data from tricycles, including a static data acquisition unit and a dynamic data acquisition unit;

[0009] Furthermore, the static data acquisition unit includes:

[0010] A static data sensor is installed on the electric tricycle to collect static data, including the vehicle body length, width, and weight. The static data is acquired, preprocessed, and integrated to obtain a static data vector F1.

[0011] Furthermore, the dynamic data acquisition unit includes:

[0012] A dynamic data sensor is installed on the electric tricycle to collect dynamic data, which includes data collected from the front wheel, the left wheel, and the right wheel.

[0013] The front wheel data collected includes front wheel angle, front wheel load, front wheel speed, and front wheel friction. The front wheel angle is the rotation angle of the front wheel monitored in real time by an angle sensor. The front wheel load is the load change of the front wheel monitored in real time by a load sensor. The front wheel friction is the friction data between the front wheel and the ground obtained by a traction sensor.

[0014] The data collected from the front wheels is acquired, preprocessed, and integrated to obtain the front wheel data vector F2.

[0015] Furthermore, the data collected from the left wheel and the right wheel includes:

[0016] The left wheel data collection includes left wheel speed, left wheel load, left wheel suspension travel, left wheel vibration data, left wheel traction force, left wheel braking force, and left tire pressure; the left wheel data collection is acquired, preprocessed, and integrated to obtain the left wheel data vector F3;

[0017] The data collected from the right wheel includes right wheel speed, right wheel load, right wheel suspension travel, right wheel vibration data, right wheel traction force, right wheel braking force, and right tire pressure; the right wheel data is acquired, preprocessed, and integrated to obtain the right wheel data vector F4;

[0018] The left wheel data vector and the right wheel data vector have the same data format.

[0019] The tricycle posture extraction module is used to acquire and process the data from the dynamic data acquisition unit to obtain a turning posture state matrix. The turning posture state matrix includes a first state vector, a second state vector, a lateral state vector, and a third state vector. The first state vector is obtained based on the front wheel data vector and the left wheel data vector. The second state vector is obtained based on the front wheel data vector and the right wheel data vector. The lateral state vector is obtained based on the left wheel data vector and the right wheel data vector. The third state vector is obtained based on the power data of the electric tricycle.

[0020] Furthermore, the specific steps for constructing the turning posture state matrix of the electric tricycle based on the front wheel data vector, left wheel data vector, and right wheel data vector include:

[0021] The first state vector f1 is constructed based on the front wheel data vector and the left wheel data vector. The first state vector includes first state data and second state data. The first state data includes data of different properties, and the second state data includes data of the same properties.

[0022] The second state vector f2 is constructed based on the front wheel data vector and the right wheel data vector. The second state vector includes third state data and fourth state data. The third state data includes data of different properties, and the fourth state data includes data of the same property.

[0023] The lateral state vector f3 is constructed based on the difference between the left wheel data vector and the right wheel data vector;

[0024] The power data of the electric tricycle is acquired and processed to obtain the third state vector f4, wherein the power data includes the output power of the electric motor;

[0025] The first state vector, the second state vector, the lateral state vector, and the third state vector are obtained and integrated to obtain the turning posture state matrix of the electric tricycle.

[0026] The suspension parameter generation module is used to generate a suspension parameter adjustment value vector in real time based on the observation state space. It includes an observation state space unit and an RNN-TD3 model unit. The observation state space unit acquires data from the static data acquisition unit and the tricycle attitude extraction module, and integrates them to obtain the observation state space. The RNN-TD3 model unit acquires the observation state space and processes it using the RNN-TD3 model to obtain the suspension parameter adjustment value vector, which includes suspension stiffness and suspension travel.

[0027] Furthermore, the observation state space unit includes:

[0028] The static data vector and the turning attitude state matrix are acquired and integrated to obtain the first observation state space. The data of the suspension parameter adjustment value vector of the electric tricycle at the previous moment are acquired to obtain the second observation state space. The data of the first observation state space and the second observation state space are acquired and integrated to obtain the observation state space.

[0029] Furthermore, the RNN-TD3 model includes:

[0030] A first action network and a first target action network, the first action network and the first target action network having the same action network structure, the action network structure including an input layer, an RNN layer, a fully connected layer, an action normalization layer and an output layer;

[0031] The evaluation network comprises a first evaluation network, a first target evaluation network, a second evaluation network, and a second target evaluation network. The first evaluation network, the first target evaluation network, the second evaluation network, and the second target evaluation network have the same evaluation network structure, which includes an input layer, an RNN layer, a fully connected layer, and an output layer.

[0032] Furthermore, the training process of the RNN-TD3 model includes:

[0033] Step S1: Initialize the entire network;

[0034] Step S2: Input the observed state space and the historical observed state space into the first action network, output the suspension parameter adjustment value vector, and obtain the suspension parameter reward and the next state of the electric tricycle corresponding to the suspension parameter adjustment value vector; construct training samples based on the observed state space, the suspension parameter adjustment value vector, the reward and the next state, and store them in the experience replay pool;

[0035] Step S3: Sample training samples from the experience replay pool, and update the parameters of the first evaluation network and the second evaluation network based on the training samples;

[0036] Step S4: After updating the parameters of the first evaluation network and the second evaluation network at intervals of D, update the parameters of the first action network;

[0037] Step S5: Update the parameters of the first target action network, the first target evaluation network, and the second target evaluation network according to the soft update rules;

[0038] Step S6: Repeat steps S3 to S5 until the training round is reached.

[0039] Furthermore, the suspension parameter bonus includes:

[0040] R=ω1·(-(|θ′|+|θ″|))+ω2·(-|ΔG|)+ω3·(-E)+ω4·(-T);

[0041] Wherein, R represents the suspension parameter bonus, ω1 represents the stability weighting coefficient, θ′ represents the lateral tilt angle of the electric tricycle, θ″ represents the longitudinal tilt angle of the electric tricycle, ω2 represents the weighting coefficient of the center of gravity offset ΔG, ω3 represents the weighting coefficient of the suspension system adjustment energy consumption E, and ω4 represents the weighting coefficient of the suspension system response time T.

[0042] The suspension parameter adjustment module is used to obtain the suspension parameter adjustment value vector and control it using an incremental PID algorithm.

[0043] Furthermore, the specific steps for control using the incremental PID algorithm include:

[0044] Obtain the suspension parameter adjustment value vector ΔE output by the RNN-TD3 model unit and adjust the parameters according to the incremental PID algorithm. The calculation formula is as follows:

[0045]

[0046] in, K represents the control variable. p Let K represent the proportional gain, ΔE(t) represent the suspension parameter adjustment vector at time t, ΔE(t-1) represent the suspension parameter adjustment vector at time t-1, ΔE(t-2) represent the suspension parameter adjustment vector at time t-2, and K... i K represents the integral gain. d This represents the differential gain.

[0047] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0048] 1. The first and second state vectors, combined with data from the front and left wheels and the front and right wheels respectively, can capture the cooperative relationship between the front wheels and the drive wheels, and distinguish the dynamic characteristics of the left and right sides. The lateral state vector, by calculating the data difference between the left and right drive wheels, directly reflects the lateral state and lateral stability characteristics during turning, providing an important basis for optimizing the turning posture. After combining the third state vector with the power data, the turning posture state matrix can comprehensively evaluate the coupling relationship between the drive system and the turning motion, realizing comprehensive monitoring and quantification of various key characteristics of the tricycle during the turning process.

[0049] 2. The RNN-TD3 model can make full use of the historical state sequence information of the electric tricycle in the turning scene provided by the observation state space unit, such as vibration and body posture changes, so that the suspension parameter adjustment is more in line with the actual state; the reward function integrates multiple indicators such as lateral tilt angle, longitudinal tilt angle, center of gravity offset, energy consumption and response time, so that the model can improve the performance of the suspension system while taking into account energy consumption and real-time performance.

[0050] 3. By controlling the suspension parameter adjustment value vector output by the RNN-TD3 model through the incremental PID algorithm, the real-time performance, stability, energy efficiency optimization and driving comfort of the electric tricycle suspension system can be comprehensively improved, the robustness and adaptability of the control system can be enhanced, and the computational complexity and system resource requirements can be reduced. Efficient and stable extension and retraction control of the drive wheel can be achieved with limited hardware resources. Attached Figure Description

[0051] Figure 1This is a schematic diagram of the structure of an adaptive telescopic control system for the drive wheels of an electric tricycle, provided in an embodiment of the present invention.

[0052] Figure 2 This is a flowchart for calculating the turning attitude state matrix provided in an embodiment of the present invention;

[0053] Figure 3 This is a flowchart for calculating the suspension parameter adjustment value vector provided in an embodiment of the present invention;

[0054] Figure 4 This is a schematic diagram of the action network structure provided in an embodiment of the present invention;

[0055] Figure 5 This is a schematic diagram of the evaluation network structure provided in an embodiment of the present invention. Detailed Implementation

[0056] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0057] Example 1

[0058] To improve the operability and safety of its electric tricycles, a tricycle manufacturer introduced an adaptive telescopic control system for the drive wheels of electric tricycles, as provided in this invention. The system structure is as follows: Figure 1 As shown, the specific implementation method is as follows:

[0059] The tricycle data acquisition module is used to collect data from tricycles, including a static data acquisition unit and a dynamic data acquisition unit;

[0060] Furthermore, the static data acquisition unit includes:

[0061] Static data sensors, including distance sensors and load sensors, are installed on the electric tricycle to collect static data, such as vehicle length, vehicle width, and vehicle weight. The static data is acquired, preprocessed, and integrated to obtain a static data vector F1.

[0062] Through the static data acquisition unit, the adaptive telescopic control system of the electric tricycle can acquire the vehicle's basic structural parameters in real time, providing reliable data and initial conditions for subsequent suspension adjustments and intelligent decision-making, and further significantly improving the accuracy, real-time performance and environmental adaptability of the suspension system.

[0063] Furthermore, the dynamic data acquisition unit includes:

[0064] A dynamic data sensor is installed on the electric tricycle to collect dynamic data, which includes data collected from the front wheel, the left wheel, and the right wheel.

[0065] The front wheel data collected includes front wheel angle, front wheel load, front wheel speed, and front wheel friction. The front wheel angle is the rotation angle of the front wheel monitored in real time by an angle sensor. The front wheel load is the load change of the front wheel monitored in real time by a load sensor. The front wheel friction is the friction data between the front wheel and the ground obtained by a traction sensor.

[0066] The front wheel data was acquired, preprocessed, and integrated to obtain the front wheel data vector F2. Table 1 shows a partial collection of the front wheel data, where positive values ​​for the front wheel steering angle are the rightward steering angle and negative values ​​are the leftward steering angle.

[0067] Table 1. Partially Collected Front Wheel Data

[0068] Number of collections Front wheel steering angle front wheel load Front wheel speed 129th -13.77° 102.5kg 56.28 rpm 130th -12.12° 105.6kg 59.12 rpm 131st -6.56° 101.9kg 61.58 rpm 132nd time -2.45° 108.3kg 65.29 rpm 133rd time -1.29° 105.2kg 76.23 rpm

[0069] The dynamic data acquisition unit collects and processes front wheel data in real time, providing accurate and real-time input data to the suspension system and ensuring that the suspension system can respond quickly under various road conditions and turning conditions.

[0070] Furthermore, the data collected from the left wheel and the right wheel includes:

[0071] The left wheel data collection includes left wheel speed, left wheel load, left wheel suspension travel, left wheel vibration data, left wheel traction force, left wheel braking force, and left tire pressure; the left wheel data collection is acquired, preprocessed, and integrated to obtain the left wheel data vector F3;

[0072] The data collected from the right wheel includes right wheel speed, right wheel load, right wheel suspension travel, right wheel vibration data, right wheel traction force, right wheel braking force, and right tire pressure; the right wheel data is acquired, preprocessed, and integrated to obtain the right wheel data vector F4;

[0073] Furthermore, the left wheel data vector and the right wheel data vector have the same data format. Table 2 shows a partial collection of left wheel data.

[0074] Table 2. Data collected from some revolvers

[0075] Number of collections Left wheel speed Left wheel load Left wheel suspension travel Left wheel traction Left tire pressure 129th 55.12 rpm 112.3kg 12.23mm 229.2N 301.5 kPa 130th 58.34 rpm 115.2kg 10.35mm 230.5N 298.3 kPa 131st 60.38 rpm 108.6kg 8.12mm 215.4N 290.2 kPa 132nd time 64.19 rpm 111.9kg 5.93mm 212.9N 286.8 kPa 133rd time 76.02 rpm 108.4kg 3.11mm 210.3N 282.6 kPa

[0076] Through high-precision sensors and standardized data processing procedures, and by ensuring that the data formats of the left and right wheels are consistent, detailed and reliable dynamic data support is provided for the suspension control system of electric tricycles.

[0077] The tricycle posture extraction module is used to acquire and process the data from the dynamic data acquisition unit to obtain a turning posture state matrix. The turning posture state matrix includes a first state vector, a second state vector, a lateral state vector, and a third state vector. The first state vector is obtained based on the front wheel data vector and the left wheel data vector. The second state vector is obtained based on the front wheel data vector and the right wheel data vector. The lateral state vector is obtained based on the left wheel data vector and the right wheel data vector. The third state vector is obtained based on the power data of the electric tricycle.

[0078] Furthermore, the specific steps for constructing the turning posture state matrix of the electric tricycle based on the front wheel data vector, left wheel data vector, and right wheel data vector are as follows: Figure 2 As shown, it includes:

[0079] The first state vector f1 is constructed based on the front wheel data vector and the left wheel data vector. The first state vector includes first state data and second state data. The first state data includes data of different properties, and the second state data includes data of the same properties.

[0080] Furthermore, the first state data includes data of different natures, for example, the front wheel steering angle data in the front wheel data vector does not exist in the left wheel data vector, so it is data of different natures and is included in the first state data; the front wheel load data in the front wheel data vector and the left wheel load data in the left wheel data vector are data of the same nature, so they can be further processed, such as by subtraction or averaging, and are included in the second state data.

[0081] The second state vector f2 is constructed based on the front wheel data vector and the right wheel data vector. The second state vector includes third state data and fourth state data. The third state data includes data of different properties, and the fourth state data includes data of the same property.

[0082] Furthermore, the third state data is similar to the first state data, and the fourth state data is similar to the second state data.

[0083] The lateral state vector f3 is constructed based on the difference between the left wheel data vector and the right wheel data vector;

[0084] Furthermore, since the left wheel data vector and the right wheel data vector have the same data structure, their data properties are also the same, and they can be further processed, such as weighted averaging, difference calculation, or feature extraction from time series data, to obtain the horizontal state vector.

[0085] The power data of the electric tricycle is acquired and processed to obtain the third state vector f4, wherein the power data includes the output power of the electric motor;

[0086] The first state vector, the second state vector, the lateral state vector, and the third state vector are obtained and integrated to obtain the turning posture state matrix of the electric tricycle.

[0087] By integrating the first state vector, the second state vector, the lateral state vector, and the third state vector, a turning posture state matrix of the electric tricycle is formed, which comprehensively reflects the dynamic state of the tricycle when turning, providing efficient and accurate data support for the suspension control system.

[0088] The suspension parameter generation module is used to generate a suspension parameter adjustment value vector in real time based on the observation state space. It includes an observation state space unit and an RNN-TD3 model unit. The observation state space unit acquires data from the static data acquisition unit and the tricycle attitude extraction module, and integrates them to obtain the observation state space. The RNN-TD3 model unit acquires the observation state space and processes it using the RNN-TD3 model to obtain the suspension parameter adjustment value vector, which includes several suspension-related parameters such as suspension stiffness and suspension travel.

[0089] Furthermore, the calculation process for the suspension parameter adjustment value vector is as follows: Figure 3 As shown.

[0090] Furthermore, the observation state space unit includes:

[0091] The static data vector and the turning attitude state matrix are acquired and integrated to obtain the first observation state space. The data of the suspension parameter adjustment value vector of the electric tricycle at the previous moment are acquired to obtain the second observation state space. The data of the first observation state space and the second observation state space are acquired and integrated to obtain the observation state space.

[0092] The first observation state space combines the vehicle's static characteristics and dynamic behavior, providing comprehensive input data for the adaptive extension and retraction control of the drive wheels; the second observation state space reflects the historical adjustment effects of the suspension system, providing a basis for the next step of control optimization; by merging the first and second observation state spaces, a comprehensive, accurate and real-time observation state space is constructed.

[0093] Furthermore, the RNN-TD3 model includes:

[0094] A first action network and a first target action network, wherein the first action network and the first target action network have the same action network structure, the action network structure being as follows: Figure 4 As shown, it includes an input layer, an RNN layer, a fully connected layer, an action normalization layer, and an output layer;

[0095] A first evaluation network, a first target evaluation network, a second evaluation network, and a second target evaluation network, wherein the first evaluation network, the first target evaluation network, the second evaluation network, and the second target evaluation network have the same evaluation network structure, as shown in the figure. Figure 5 As shown, it includes an input layer, an RNN layer, a fully connected layer, and an output layer.

[0096] The RNN-TD3 model, by introducing a dual structure of action network and evaluation network and combining the time series feature extraction capability of recurrent neural network, constructs an efficient, stable and accurate suspension control mechanism. This design significantly improves the adaptive extension and retraction control capability of the drive wheels of electric tricycles in cornering conditions.

[0097] Furthermore, the training process of the RNN-TD3 model includes:

[0098] Step S1: Initialize the entire network;

[0099] Step S2: Input the observed state space and the historical observed state space into the first action network, output the suspension parameter adjustment value vector, and obtain the suspension parameter reward and the next state of the electric tricycle corresponding to the suspension parameter adjustment value vector; construct training samples based on the observed state space, the suspension parameter adjustment value vector, the reward and the next state, and store them in the experience replay pool;

[0100] Step S3: Sample training samples from the experience replay pool, and update the parameters of the first evaluation network and the second evaluation network based on the training samples;

[0101] Step S4: After updating the parameters of the first evaluation network and the second evaluation network at intervals of D, update the parameters of the first action network;

[0102] Step S5: Update the parameters of the first target action network, the first target evaluation network, and the second target evaluation network according to the soft update rules;

[0103] Step S6: Repeat steps S3 to S5 until the training round is reached.

[0104] The RNN layer enables efficient extraction of time-series features, improving training efficiency and enhancing the model's generalization ability. Through the reinforcement learning framework, the model can generate dynamically optimized suspension parameter adjustment strategies based on different driving states, effectively reducing the vehicle's lateral tilt, longitudinal tilt, and center of gravity shift, and providing precise adaptive extension and retraction control for the electric tricycle's drive wheels.

[0105] Furthermore, the suspension parameter bonus includes:

[0106] R=ω1·(-(|θ′|+|θ″|))+ω2·(-|ΔG|)+ω3·(-E)+ω4·(-T);

[0107] Wherein, R represents the suspension parameter bonus, ω1 represents the stability weighting coefficient, θ′ represents the lateral tilt angle of the electric tricycle, θ″ represents the longitudinal tilt angle of the electric tricycle, ω2 represents the weighting coefficient of the center of gravity offset ΔG, ω3 represents the weighting coefficient of the suspension system adjustment energy consumption E, and ω4 represents the weighting coefficient of the suspension system response time T.

[0108] The design of the suspension parameter bonus not only focuses on the dynamic stability of the vehicle, but also takes into account factors such as energy consumption and response speed, optimizing the stability of the tricycle, improving the handling of the tricycle, and achieving efficient and energy-saving suspension adjustment.

[0109] The suspension parameter adjustment module is used to obtain the suspension parameter adjustment value vector and control it using an incremental PID algorithm.

[0110] Furthermore, the specific steps for control using the incremental PID algorithm include:

[0111] Obtain the suspension parameter adjustment value vector ΔE output by the RNN-TD3 model unit and adjust the parameters according to the incremental PID algorithm. The calculation formula is as follows:

[0112]

[0113] in, K represents the control variable. p Let K represent the proportional gain, ΔE(t) represent the suspension parameter adjustment vector at time t, ΔE(t-1) represent the suspension parameter adjustment vector at time t-1, ΔE(t-2) represent the suspension parameter adjustment vector at time t-2, and K... iK represents the integral gain. d This represents the differential gain.

[0114] By controlling the suspension parameter adjustment vector output by the RNN-TD3 model using an incremental PID algorithm, comprehensive improvements can be made to the real-time performance, stability, energy efficiency, and driving comfort of the electric tricycle suspension system. This enhances the robustness and adaptability of the control system while reducing computational complexity and system resource requirements, enabling efficient and stable suspension control with limited hardware resources.

[0115] Table 3 shows some data of the electric tricycle in round 2352 of the simulation phase, reflecting that the control system provided by the present invention brings good stability and maneuverability to the electric tricycle.

[0116] Table 3. Partial System Response Data

[0117] Simulated steps Lateral tilt angle longitudinal tilt angle Center of gravity offset Response time Step 21 2.21° 1.34° 20.3cm 12.1ms Step 22 1.98° 1.25° 16.4cm 11.8ms Step 23 1.63° 0.99° 14.2cm 12.0ms Step 24 1.39° 0.82° 9.1cm 11.5ms Step 25 0.91° 0.71° 4.5cm 11.3ms

[0118] This invention provides an adaptive telescopic control system for the drive wheels of an electric tricycle. First, it collects comprehensive data about the tricycle from all directions using the static and dynamic data acquisition units of the tricycle data acquisition module, providing data support for control. Then, it extracts the tricycle's turning posture state matrix using the tricycle posture extraction module, comprehensively reflecting the tricycle's state during turning. Finally, the suspension parameter generation module and suspension parameter adjustment module work together as a whole to control the suspension, ultimately achieving efficient and stable adaptive telescopic control of the drive wheels under limited hardware resources.

[0119] Example 2

[0120] When tuning the suspension of a tricycle, a company introduced an adaptive telescopic control system for the drive wheels of an electric tricycle, as provided by this invention, to achieve adaptive telescopic control of the drive wheels when the electric tricycle is turning. The specific implementation method is as follows:

[0121] The tricycle data acquisition module is used to collect data from tricycles, including a static data acquisition unit and a dynamic data acquisition unit;

[0122] Furthermore, static data sensors are installed on the electric tricycle to collect static data, including the vehicle's length, width, and weight. This static data is acquired, preprocessed, and integrated to obtain a static data vector F1. Table 4 shows the data for some tricycle models.

[0123] Table 4. Static Data

[0124] tricycle number Vehicle body length Vehicle width Vehicle weight A001 2.4m 1.4m 290kg A002 2.4m 1.4m 300kg A003 2.4m 1.4m 310kg A004 2.4m 1.4m 320kg A005 2.4m 1.4m 330kg

[0125] Furthermore, a dynamic data sensor is installed on the electric tricycle to collect dynamic data, which includes data collected from the front wheel, the left wheel, and the right wheel. Table 5 shows some of the data collected from the right wheel.

[0126] Table 5. Partial Data Collected from the Right Wheel

[0127] Number of collections Right wheel speed Right wheel load Right wheel suspension travel Right wheel traction Right tire pressure 39th 77.31 rpm 125.4kg 11.52mm 230.4N 298.3 kPa 40th 76.82 rpm 127.8kg 10.48mm 229.8N 300.1 kPa 41st 79.38 rpm 126.3kg 9.24mm 229.4N 299.4 kPa 42nd time 76.40 rpm 127.2kg 8.29mm 230.3N 300.2 kPa 43rd 77.49 rpm 129.5kg 8.78mm 230.6N 299.5 kPa

[0128] The tricycle posture extraction module is used to acquire and process the data from the dynamic data acquisition unit to obtain a turning posture state matrix. The turning posture state matrix includes a first state vector, a second state vector, a lateral state vector, and a third state vector. The first state vector is obtained based on the front wheel data vector and the left wheel data vector. The second state vector is obtained based on the front wheel data vector and the right wheel data vector. The lateral state vector is obtained based on the left wheel data vector and the right wheel data vector. The third state vector is obtained based on the power data of the electric tricycle.

[0129] Furthermore, the specific steps for constructing the turning posture state matrix of the electric tricycle based on the front wheel data vector, left wheel data vector, and right wheel data vector include:

[0130] The first state vector f1 is constructed based on the front wheel data vector and the left wheel data vector. The first state vector includes first state data and second state data. The first state data includes data of different properties, and the second state data includes data of the same properties.

[0131] The second state vector f2 is constructed based on the front wheel data vector and the right wheel data vector. The second state vector includes third state data and fourth state data. The third state data includes data of different properties, and the fourth state data includes data of the same property.

[0132] The lateral state vector f3 is constructed based on the difference between the left wheel data vector and the right wheel data vector;

[0133] The power data of the electric tricycle is acquired and processed to obtain the third state vector f4, wherein the power data includes the output power of the electric motor;

[0134] The first state vector, the second state vector, the lateral state vector, and the third state vector are obtained and integrated to obtain the turning posture state matrix of the electric tricycle.

[0135] The suspension parameter generation module is used to generate a suspension parameter adjustment value vector in real time based on the observation state space. It includes an observation state space unit and an RNN-TD3 model unit. The observation state space unit acquires data from the static data acquisition unit and the tricycle attitude extraction module, and integrates them to obtain the observation state space. The RNN-TD3 model unit acquires the observation state space and processes it using the RNN-TD3 model to obtain the suspension parameter adjustment value vector, which includes suspension stiffness and suspension travel.

[0136] Furthermore, the observation state space unit includes:

[0137] The static data vector and the turning attitude state matrix are acquired and integrated to obtain the first observation state space. The data of the suspension parameter adjustment value vector of the electric tricycle at the previous moment are acquired to obtain the second observation state space. The data of the first observation state space and the second observation state space are acquired and integrated to obtain the observation state space.

[0138] Furthermore, the RNN-TD3 model includes:

[0139] A first action network and a first target action network, the first action network and the first target action network having the same action network structure, the action network structure including an input layer, an RNN layer, a fully connected layer, an action normalization layer and an output layer;

[0140] The evaluation network comprises a first evaluation network, a first target evaluation network, a second evaluation network, and a second target evaluation network. The first evaluation network, the first target evaluation network, the second evaluation network, and the second target evaluation network have the same evaluation network structure, which includes an input layer, an RNN layer, a fully connected layer, and an output layer.

[0141] Furthermore, the training process of the RNN-TD3 model includes:

[0142] Step S1: Initialize the entire network;

[0143] Step S2: Input the observed state space and the historical observed state space into the first action network, output the suspension parameter adjustment value vector, and obtain the suspension parameter reward and the next state of the electric tricycle corresponding to the suspension parameter adjustment value vector; construct training samples based on the observed state space, the suspension parameter adjustment value vector, the reward and the next state, and store them in the experience replay pool;

[0144] Step S3: Sample training samples from the experience replay pool, and update the parameters of the first evaluation network and the second evaluation network based on the training samples;

[0145] Step S4: After updating the parameters of the first evaluation network and the second evaluation network at intervals of D, update the parameters of the first action network;

[0146] Step S5: Update the parameters of the first target action network, the first target evaluation network, and the second target evaluation network according to the soft update rules;

[0147] Step S6: Repeat steps S3 to S5 until the training round is reached.

[0148] Furthermore, the suspension parameter bonus includes:

[0149] R=ω1·(-(|θ′|+|θ″|))+ω2·(-|ΔG|)+ω3·(-E)+ω4·(-T);

[0150] Wherein, R represents the suspension parameter bonus, ω1 represents the stability weighting coefficient, θ′ represents the lateral tilt angle of the electric tricycle, θ″ represents the longitudinal tilt angle of the electric tricycle, ω2 represents the weighting coefficient of the center of gravity offset ΔG, ω3 represents the weighting coefficient of the suspension system adjustment energy consumption E, and ω4 represents the weighting coefficient of the suspension system response time T.

[0151] Furthermore, the formula for the soft update rule is:

[0152] θ t ′ ar =τ·θ cur +(1-τ)·θ tar ;

[0153] Where, θ t ′ ar θ represents the updated parameters of the target network. cur θ represents the parameters of the current network, namely the evaluation network and the action network. tar This represents the parameters of the target network before the update, and τ represents the soft update coefficient, which ranges from 0 to 1.

[0154] The suspension parameter adjustment module is used to obtain the suspension parameter adjustment value vector and control it using an incremental PID algorithm.

[0155] Table 6. Partial System Response Data

[0156] Simulated steps Lateral tilt angle longitudinal tilt angle Center of gravity offset Response time Step 19 2.01° 1.23° 10.32cm 11.2ms Step 20 1.92° 1.21° 10.98cm 10.9ms Step 21 2.02° 1.19° 9.38cm 10.5ms Step 22 1.97° 1.15° 9.57cm 10.3ms Step 23 1.98° 1.18° 9.19cm 10.6ms

[0157] Furthermore, the specific steps for control using the incremental PID algorithm include:

[0158] Obtain the suspension parameter adjustment value vector ΔE output by the RNN-TD3 model unit and adjust the parameters according to the incremental PID algorithm. The calculation formula is as follows:

[0159]

[0160] in, K represents the control variable. p Let K represent the proportional gain, ΔE(t) represent the suspension parameter adjustment vector at time t, ΔE(t-1) represent the suspension parameter adjustment vector at time t-1, ΔE(t-2) represent the suspension parameter adjustment vector at time t-2, and K... i K represents the integral gain. d This represents the differential gain.

[0161] Table 6 shows some of the system response data in round 4828 of the simulation phase. It can be seen that the adaptive telescopic control system for the drive wheels of electric tricycles provided by this invention can realize efficient operation and telescopic movement of the drive wheels, shorten the response time, and improve the driving stability of electric tricycles when turning.

[0162] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. An adaptive telescopic control system for the drive wheels of an electric tricycle, characterized in that, include: The tricycle data acquisition module is used to collect data from tricycles, including a static data acquisition unit and a dynamic data acquisition unit; The tricycle posture extraction module is used to acquire and process the data from the dynamic data acquisition unit to obtain a turning posture state matrix. The turning posture state matrix includes a first state vector, a second state vector, a lateral state vector, and a third state vector. The first state vector is obtained based on the front wheel data vector and the left wheel data vector. The second state vector is obtained based on the front wheel data vector and the right wheel data vector. The lateral state vector is obtained based on the left wheel data vector and the right wheel data vector. The third state vector is obtained based on the power data of the electric tricycle. The suspension parameter generation module is used to generate a suspension parameter adjustment value vector in real time based on the observation state space. It includes an observation state space unit and an RNN-TD3 model unit. The observation state space unit acquires data from the static data acquisition unit and the tricycle attitude extraction module, and integrates them to obtain the observation state space. The RNN-TD3 model unit acquires the observation state space and processes it using the RNN-TD3 model to obtain the suspension parameter adjustment value vector, which includes suspension stiffness and suspension travel. The RNN-TD3 model includes: A first action network and a first target action network, the first action network and the first target action network having the same action network structure, the action network structure including an input layer, an RNN layer, a fully connected layer, an action normalization layer and an output layer; A first evaluation network, a first target evaluation network, a second evaluation network, and a second target evaluation network, wherein the first evaluation network, the first target evaluation network, the second evaluation network, and the second target evaluation network have the same evaluation network structure, wherein the evaluation network structure includes an input layer, an RNN layer, a fully connected layer, and an output layer; The suspension parameter adjustment module is used to obtain the suspension parameter adjustment value vector and control it using an incremental PID algorithm; The specific steps for control using the incremental PID algorithm include: Obtain the suspension parameter adjustment value vector output by the RNN-TD3 model unit. The parameters are then adjusted using an incremental PID algorithm, and the calculation formula is as follows: ； in, Indicates the control quantity. Indicates proportional gain. express The suspension parameter adjustment value vector at time t, express The suspension parameter adjustment value vector at time t, express The suspension parameter adjustment value vector at time t, Indicates integral gain. This represents the differential gain.

2. The adaptive telescopic control system for the drive wheels of an electric tricycle according to claim 1, characterized in that, The static data acquisition unit includes: A static data sensor is installed on the electric tricycle to collect static data, including the vehicle's length, width, and weight. The static data is acquired, preprocessed, and integrated to obtain a static data vector. .

3. The adaptive telescopic control system for the drive wheels of an electric tricycle according to claim 1, characterized in that, The dynamic data acquisition unit includes: A dynamic data sensor is installed on the electric tricycle to collect dynamic data, which includes data collected from the front wheel, the left wheel, and the right wheel. The front wheel data collected includes front wheel angle, front wheel load, front wheel speed, and front wheel friction. The front wheel angle is the rotation angle of the front wheel monitored in real time by an angle sensor. The front wheel load is the load change of the front wheel monitored in real time by a load sensor. The front wheel friction is the friction data between the front wheel and the ground obtained by a traction sensor. The data collected from the front wheels is preprocessed and integrated to obtain the front wheel data vector. .

4. The adaptive telescopic control system for the drive wheels of an electric tricycle according to claim 3, characterized in that, The data collected by the left wheel and the right wheel include: The left-wheel data collected includes left-wheel speed, left-wheel load, left-wheel suspension travel, left-wheel vibration data, left-wheel traction force, left-wheel braking force, and left-wheel tire pressure. The left-wheel data is acquired, preprocessed, and integrated to obtain a left-wheel data vector. ; The data collected from the right wheel includes right wheel speed, right wheel load, right wheel suspension travel, right wheel vibration data, right wheel traction force, right wheel braking force, and right tire pressure. The collected right wheel data is acquired, preprocessed, and integrated to obtain a right wheel data vector. ; The left wheel data vector and the right wheel data vector have the same data format.

5. The adaptive telescopic control system for the drive wheels of an electric tricycle according to claim 1, characterized in that, The specific steps for constructing the turning posture state matrix of the electric tricycle based on the front wheel data vector, left wheel data vector, and right wheel data vector include: The first state vector is constructed based on the front wheel data vector and the left wheel data vector. The first state vector includes first state data and second state data; The second state vector is constructed based on the front wheel data vector and the right wheel data vector. The second state vector includes third state data and fourth state data; The lateral state vector is constructed based on the difference between the left wheel data vector and the right wheel data vector. ; The power data of the electric tricycle is acquired and processed to obtain the third state vector. The power data includes the output power of the electric motor; The first state vector, the second state vector, the lateral state vector, and the third state vector are obtained and integrated to obtain the turning posture state matrix of the electric tricycle. .

6. The adaptive telescopic control system for the drive wheels of an electric tricycle according to claim 1, characterized in that, The observation state space unit includes: The static data vector and the turning attitude state matrix are acquired and integrated to obtain the first observation state space. The data of the suspension parameter adjustment value vector of the electric tricycle at the previous moment are acquired to obtain the second observation state space. The data of the first observation state space and the second observation state space are acquired and integrated to obtain the observation state space.

7. The adaptive telescopic control system for the drive wheels of an electric tricycle according to claim 1, characterized in that, The training process of the RNN-TD3 model includes: Step S1: Initialize the entire network; Step S2: Input the observed state space and the historical observed state space into the first action network, output the suspension parameter adjustment value vector, and obtain the suspension parameter reward and the next state of the electric tricycle corresponding to the suspension parameter adjustment value vector; construct training samples based on the observed state space, the suspension parameter adjustment value vector, the reward and the next state, and store them in the experience replay pool; Step S3: Sample training samples from the experience replay pool, and update the parameters of the first evaluation network and the second evaluation network based on the training samples; Step S4: Interval After the parameters of the first evaluation network and the second evaluation network are updated, the parameters of the first action network are updated. Step S5: Update the parameters of the first target action network, the first target evaluation network, and the second target evaluation network according to the soft update rules; Step S6: Repeat steps S3 to S5 until the training round is reached.

8. The adaptive telescopic control system for the drive wheels of an electric tricycle according to claim 7, characterized in that, The suspension parameter bonus includes: ； in, This indicates the reward for the suspension parameters. Weighting coefficients representing stability This indicates the lateral tilt angle of the electric tricycle. Indicates the longitudinal tilt angle of the electric tricycle. Indicates the offset of the center of gravity The weighting coefficients, Indicates that the suspension system adjusts energy consumption. The weighting coefficients, Indicates the suspension system response time The weighting coefficients.

Images