Motor driving technology device based on combination of BP neural network adaptive algorithm and PID control algorithm

Abstract

The invention discloses a motor driving technology based on the combination of a BP neural network and a PID control algorithm, and relates to the technical field of motor driving control, in particular to a motor driving technology device based on the combination of a BP neural network adaptive algorithm and a PID control algorithm. The invention provides a motor driving technology device based on the combination of a BP neural network and a PID control algorithm and applied to a middle motor of an electric power-assisted bicycle. The technology device comprises the BP neural network, a TMR and HALL sensing observer and a PID controller. The TMR and HALL sensing observers are used for obtaining operation data of the electric power-assisted bicycle in real time; according to the operation data, the expected speed is obtained through the self-adaptive driving algorithm of the BP neural network, then the control current is obtained through calculation, and the PID controller can control the target motor according to the control current so as to achieve assistance. And the feedback circuit also compares the actual operation data with the expected operation data in real time to continuously adjust the PWM signal, so that stable motor assistance is realized, and the assistance performance and the driving stability of the electric moped are improved.

Description

Technical Field

[0001] The present application relates to the field of motor drive control technology, and in particular to a motor drive technology device based on the combination of a BP neural network adaptive algorithm and a PID control algorithm. Technical Background

[0002] As the world's attention to reducing carbon emissions and air pollution continues to increase, electric-assisted bicycles have begun to rise in order to meet the global demand for new environmentally friendly transportation tools. Compared with cars and motorcycles, electric-assisted bicycles have low carbon emissions and do not produce air pollution. At the same time, electric-assisted bicycles can be used as an efficient and flexible means of transportation to alleviate urban traffic pressure. It also provides the fun of riding, while reducing physical exertion during riding, allowing more people to participate in cycling sports, thereby better meeting the public's fitness and exercise needs. However, the electric-assisted bicycles in front of us are still facing the problem of low motor efficiency, which also makes electric-assisted bicycles have a lot of room for improvement in terms of speed control accuracy, energy consumption optimization and riding stability. Therefore, improving the efficiency and power density of the motor is a challenge, which directly affects the riding experience and endurance. And the ideal intelligent control system needs to provide smooth assistance under different road conditions and riding conditions, and the optimization of traditional control algorithms still needs further research. Summary of the invention

[0003] The purpose of the embodiment of the present application is to provide a motor drive technology device based on the combination of BP neural network adaptive algorithm and PID control algorithm. The desired speed is accurately predicted by the adaptive algorithm of the BP neural network, and then the speed is controlled by the PID speed controller, thereby improving the motor power performance, energy efficiency and power-assisted riding stability of the target vehicle.

[0004] The first aspect of the present application provides a motor drive technology based on the combination of BP neural network adaptive algorithm and PID control algorithm, including: BP neural network based on CNN-LSTM expected speed prediction adaptive algorithm; HALL and TMR sensor observers; Constructing a PID speed controller according to the electric-assisted bicycle riding dynamics model; Deploy the BP neural network and the PID controller into the motor of the target electric-assisted bicycle; The target vehicle's running data is collected in real time by using HALL and TMR sensors; wherein the data includes at least one of vehicle speed, cadence and torque; According to the above operating data and the expected speed obtained by the BP neural network, the PID speed controller adjusts the PWM signal according to the expected speed and the dynamic model to determine the control current; the motor of the target is controlled according to the control current to achieve power assistance.

[0005] In the above implementation process, the method can collect the operation data of the target vehicle in real time through the HALL and TMR sensors, and then analyze the force state of the target vehicle based on the constructed dynamic model of the electric power-assisted bicycle during riding. At the same time, the BP neural network based on the CNN-LSTM expected speed prediction adaptive algorithm that has been trained on a large amount of rider data accurately calculates the accurate expected speed, and transmits it to the control current determined according to the PID speed controller, so as to drive the motor to achieve power assistance. The feedback circuit compares the measured output power and the expected output power in real time, and continuously adjusts the PWM signal of the PID speed controller, so that the power assistance of the electric power-assisted bicycle is more stable and the control accuracy is improved.

[0006] The BP neural network based on the CNN-LST adaptive algorithm for predicting the expected speed according to the collected running data of the target vehicle includes: A CNN layer is used to capture the spatial relationship between input features from the input time series while retaining the temporal information, and send the captured structural information to the LSTM layer; Multiple LSTM layers are responsible for extracting temporal information from input features; An attention layer that gives higher weights to important time steps to capture more valuable information; A fully connected layer that outputs the predicted expected speed of the rider under the current state.

[0007] Further, according to the HALL and TMR sensor observer, it includes: The HALL sub-module and the TMR sub-module both use wireless energy transmission technology to generate an alternating magnetic field in the transmitting coil through high-frequency alternating current, and then use the induced current obtained by electromagnetic induction of the receiving coil to measure the target vehicle's operating data after rectification, filtering and voltage stabilization, thereby solving the problem of signal cable entanglement and wear during the rotation of the pedal shaft.

[0008] The second aspect of the present application provides a motor drive device based on a combination of a BP neural network adaptive algorithm and a PID control algorithm, comprising: The first construction part is used for the BP neural network based on the CNN-LSTM expected speed prediction adaptive algorithm; The second construction part is used for HALL and TMR sensor observers; A third construction part, a PID speed controller according to the construction; A deployment part, used for deploying the BP neural network based on the CNN-LSTM desired speed prediction adaptive algorithm, the HALL and TMR sensor observers and the PID speed controller in the motor of the target vehicle; A data acquisition part, used for using HALL and TMR sensors to collect the running data of the target vehicle in real time; wherein the data includes at least one of vehicle speed, torque and pedaling frequency; The data determination part obtains the expected speed based on the above-mentioned operation data and the CNN-LSTM adaptive algorithm in the BP neural network, and the PID speed controller adjusts the PWM signal to determine the control current according to the expected speed; The driving control part controls the motor of the target according to the control current to achieve power assistance.

[0009] Furthermore, the first construction part includes: A CNN layer is used to capture the spatial relationship between input features from the input time series while retaining the temporal information, and send the captured structural information to the LSTM layer; Multiple LSTM layers are responsible for extracting the timing information from the input features. Due to the characteristics of LSTM, both the long-term and short-term characteristics of the input sequence can be learned, which can better model the riding status and motor assistance. Constructing a conversion equation between the riding operation data including at least one of torque, cadence and vehicle speed and the expected speed corresponding to the pre-learned data under the actual operation data according to the LSTM layer; An attention layer that gives higher weights to important time steps to capture more valuable information; A fully connected layer that outputs the predicted expected speed of the rider in the current state;

[0010] Furthermore, the third construction part comprises: Constructing a conversion relationship between the expected speed and the motor control current according to the conversion equation of the expected speed corresponding to the actual running data based on the riding running data including at least one of the torque, the cadence and the vehicle speed and the pre-learned data; According to the PID speed controller; The feedback circuit compares the actual operating data with the operating data under the expected state, continuously feeds back to the PID speed controller, and continuously adjusts the PWM signal to achieve smooth power assistance.

[0011] Furthermore, the data acquisition unit comprises: The sensor system consists of an electric power control board, a HALL sensor, a TMR sensor and a signal receiving board. The system adopts a wireless energy transmission method to solve the power supply and signal acquisition, and is used to better and more accurately receive the real-time operation data of the electric power vehicle.

[0012] The third aspect of the present application provides an electronic device, including a sensor, a memory and a processor, wherein the sensor is used to collect relevant operating data generated by the device, the memory is used to store a computer program, and the processor runs the computer program to enable the electronic device to execute the motor drive technology based on the combination of BP neural network adaptive algorithm and PID control algorithm as described in any one of the first aspect of the present application.

[0013] The fourth aspect of the present application provides a computer-readable storage medium storing computer program instructions. When the computer program instructions are read and executed by a processor, the motor drive technology based on the combination of BP neural network adaptive algorithm and PID control algorithm described in any one of the first aspect of the present application is executed.

[0014] Beneficial effects of this application: (1) The motor drive technology based on the combination of BP neural network adaptive algorithm and PID control algorithm can enable the electric-assisted bicycle to achieve accurate and stable electric-assisted effect, thereby improving driving stability and the riding experience of the driver; (2) Through the BP neural network based on the CNN-LSTM expected speed prediction adaptive algorithm, it provides the rider with exclusive and accurate expected speed prediction to achieve smooth assistance; (3) Through the HALL and TMR sensor observers, wireless energy transmission method is adopted to solve the problems of power supply and signal acquisition. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the embodiments of the present application will be briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present application and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0018] Figure 1 A flow chart of a motor drive technology based on the combination of a BP neural network adaptive algorithm and a PID control algorithm provided in an embodiment of the present application. Figure 2 A schematic diagram of a motor drive technology device based on a combination of a BP neural network adaptive algorithm and a PID control algorithm provided in an embodiment of the present application.

Claims

1. A motor drive technology device based on the combination of BP neural network adaptive algorithm and PID control algorithm, characterized in that: include: BP neural network based on CNN-LSTM expected speed prediction adaptive algorithm; HALL and TMR sensor observers; PID speed controller; Deploy the BP neural network and the PID controller into the motor of the target electric-assisted bicycle; The target vehicle's running data is collected in real time by using HALL and TMR sensors; wherein the data includes at least one of vehicle speed, cadence and torque; According to the above-mentioned operating data and the expected speed obtained by the electric BP neural network, the PID speed controller will adjust the PWM signal according to the expected speed to determine the control current; according to the control current, the motor of the target is controlled to achieve power assistance.

2. The motor drive technology device based on the combination of BP neural network adaptive algorithm and PID control algorithm according to claim 1 is characterized in that: The BP neural network based on the CNN-LST adaptive algorithm for predicting the expected speed according to the collected running data of the target vehicle includes: A CNN layer is used to capture the spatial relationship between input features from the input time series while retaining the temporal information, and send the captured structural information to the LSTM layer; Multiple LSTM layers are responsible for extracting temporal information from input features; An attention layer that gives higher weights to important time steps to capture more valuable information; A fully connected layer that outputs the predicted expected speed of the rider under the current state.

3. The motor drive technology device based on the combination of BP neural network adaptive algorithm and PID control algorithm according to claim 1 is characterized in that: The HALL and TMR sensor observer comprises: The HALL sub-module and the TMR sub-module both use wireless energy transmission technology to generate an alternating magnetic field in the transmitting coil through high-frequency alternating current, and then use the induced current obtained by electromagnetic induction of the receiving coil to measure the target vehicle's operating data after rectification, filtering and voltage stabilization, thereby solving the problem of signal cable entanglement and wear during the rotation of the pedal shaft.

4. A motor drive technology device based on the combination of BP neural network adaptive algorithm and PID control algorithm, characterized in that the motor drive device based on the combination of BP neural network adaptive algorithm and PID control algorithm includes: The first construction part is used to construct a dynamic model of an electric-assisted bicycle; The second construction part is used for the BP neural network based on the CNN-LSTM expected speed prediction adaptive algorithm; The third construction part is used for HALL and TMR sensor observers; The fourth construction part is to construct a PID speed controller according to the riding dynamics model of the electric-assisted bicycle; A deployment part is used to deploy the electric-assisted bicycle dynamics model, the BP neural network based on the CNN-LSTM desired speed prediction adaptive algorithm, the HALL and TMR sensor observers, and the PID speed controller constructed according to the electric-assisted bicycle riding dynamics model in the motor of the target vehicle; A data acquisition part, used for using HALL and TMR sensors to collect the running data of the target vehicle in real time; wherein the data includes at least one of vehicle speed, torque and pedaling frequency; The data determination part obtains the expected speed according to the above-mentioned operation data, the riding dynamics model of the electric-assisted bicycle and the CNN-LSTM adaptive algorithm in the BP neural network, and the PID speed controller adjusts the PWM signal according to the expected speed and the dynamics model to determine the control current; The driving control part controls the motor of the target according to the control current to achieve power assistance.

5. An electronic device, characterized in that: The electronic device includes a memory and a processor, wherein the memory can be used to store a computer program, and the processor runs the computer program to enable the electronic device to execute the motor drive technology based on the combination of a BP neural network adaptive algorithm and a PID control algorithm as described in any one of claims 1 to 3.

6. A readable storage medium, characterized in that: The readable storage medium stores computer program instructions, and when the computer program instructions are read and executed by a processor, the motor drive technology based on the combination of BP neural network adaptive algorithm and PID control algorithm as described in any one of claims 1 to 3 is executed.

Images