Regenerative braking energy feedback control method based on machine learning

Abstract

The invention relates to the technical field of braking energy feedback, in particular to a regenerative braking energy feedback control method based on machine learning, and the method comprises the steps: respectively collecting vehicle driving style characteristic data, road condition characteristic data and energy consumption level data of n continuous time nodes in a historical vehicle driving process; calculating a mean value of the vehicle driving style feature data, respectively importing the mean value into a relevancy evaluation strategy to output relevancy with the energy consumption level data, carrying out descending sort, and outputting vehicle driving style key feature data; inputting the vehicle driving style key feature data into an energy consumption level prediction model construction strategy to construct an energy consumption level prediction model taking the driving style key feature data and the road condition data as input and taking the energy consumption level of m time nodes in the future as output; and when the energy consumption level predicted by the energy consumption level prediction model is higher than the energy consumption safety range, adjusting the feature weight of the neural network model so as to adjust the vehicle driving style and control the feature data influencing the energy consumption level.

Description

Background Art

[0002] The development background of regenerative braking energy feedback control technology is deeply influenced by the global energy crisis and environmental protection requirements. With the rapid expansion of the electric vehicle (EV) and hybrid electric vehicle (HEV) market, improving the energy efficiency and range of these vehicles has become a research focus. Regenerative braking technology, by converting the kinetic energy of the vehicle during braking into electrical energy and storing it in the battery, can effectively improve energy utilization, reduce dependence on fossil fuels, and reduce vehicle operating costs and environmental impact.

[0003] In the development and optimization of regenerative braking systems, the introduction of machine learning technology is the key to improving the intelligence and efficiency of the system. The machine learning model can predict the vehicle's operating status and energy recovery potential, and dynamically adjust the regenerative braking strategy to maximize energy recovery efficiency based on multiple factors such as real-time road conditions, vehicle speed, and battery status. Use machine learning for predictive maintenance, by analyzing the data of the braking system, identifying potential faults and performance degradation, and performing preventive maintenance to reduce unexpected parking and repair costs.

[0004] Problems in the prior art: Currently, there is no solution that uses the driver's driving habits and style data to obtain the timing and direction of regenerative braking intervention to make the braking process smoother and energy recovery more efficient. Summary of the invention

[0005] The main purpose of the present invention is to provide a regenerative braking energy feedback control method based on machine learning to effectively solve the above-mentioned problems mentioned in the background technology.

[0006] The technical solution of the present invention is as follows:

[0007] A regenerative braking energy feedback control method based on machine learning is proposed, which includes the following steps:

[0008] S1. Collecting vehicle driving style characteristic data, road condition characteristic data and energy consumption level data at n consecutive time nodes during the historical vehicle driving process;

[0009] S2. Calculate the average value and change rate of the brake pedal opening, the average value and change rate of the accelerator pedal opening, the average value and change rate of the steering wheel angle, the average speed of the acceleration-to-brake process, and the average speed of the braking-to-acceleration process at n consecutive time nodes in the historical vehicle driving process, import the correlation between the output and the energy consumption level data into the correlation evaluation strategy respectively, and arrange them in descending order, and output the key feature data of the vehicle driving style;

[0010] S3, inputting the key feature data of vehicle driving style into the energy consumption level prediction model construction strategy to construct an energy consumption level prediction model with the key feature data of driving style and road condition data as input and the energy consumption level of m future time nodes as output;

[0011] S4. When the energy consumption level predicted by the energy consumption level prediction model is higher than the energy consumption safety range, the feature weights of the neural network model are adjusted, and then the vehicle driving style is adjusted to control the feature data that affects the energy consumption level.

[0012] A further improvement of the present invention is that the vehicle driving style characteristic data in S1 includes brake pedal opening, accelerator pedal opening, steering wheel angle, speed of acceleration-to-braking process, standard speed of acceleration-to-braking process, speed of braking-to-acceleration process and standard speed of braking-to-acceleration process; the road condition characteristic data includes suburban road condition data, urban road condition data and highway road condition data.

[0013] A further improvement of the present invention is that the relevance strategy in S2 includes the following specific contents:

[0014] S21, extracting features from the average value and change rate of the brake pedal opening, the average value and change rate of the accelerator pedal opening, the average value and change rate of the steering wheel angle, the average speed of the acceleration-to-braking process, and the average speed of the braking-to-acceleration process at n consecutive time nodes in the historical vehicle driving process;

[0015] S22, constructing a correlation analysis model, wherein the correlation analysis model is a neural network model with three layers, wherein the neural network model comprises an input layer, a hidden layer, and an output layer, wherein each node corresponds to a vehicle driving style feature data, and the correlation is used as an output, and the output value is between 0 and 1;

[0016] S23. Arrange the correlations output by the correlation analysis model in descending order, and output the vehicle driving style data with a correlation with the energy consumption level greater than 0.4 as the vehicle driving style key feature data.

[0017] A further improvement of the present invention is that the energy consumption level prediction model construction strategy in S3 includes the following specific contents: 70% of the feature data is used as a training set to input the neural network model for training to obtain an initial neural network model, 30% of the feature data is used as a test set to test the initial neural network model, and the optimal initial neural network model that meets the accuracy of the historical vehicle energy consumption level change curve data is output as the neural network model.

[0018] A further improvement of the present invention is that the output formula of a specific neuron in the neural network model is: in is the output of the p-term neuron in the i+1 layer, is the connection weight between the i-th layer neuron j and the i+1 layer p-item neuron, is the output of neuron j in the i-th layer, Represents the bias of the linear relationship, and σ is the sigmoid activation function.

[0019] The technical effects of the present invention are as follows:

[0020] A regenerative braking energy feedback control method based on machine learning is constructed. The present invention collects vehicle driving style feature data, road condition feature data and energy consumption level data at n consecutive time nodes in the historical vehicle driving process; calculates the average value and change rate of the brake pedal opening, the average value and change rate of the accelerator pedal opening, the average value and change rate of the steering wheel angle, the average speed of the acceleration-to-braking process, and the average speed of the braking-to-acceleration process at n consecutive time nodes in the historical vehicle driving process, and imports the correlation between the output and the energy consumption level data into the correlation evaluation strategy and arranges them in descending order, and outputs the key feature data of the vehicle driving style; inputs the key feature data of the vehicle driving style into the energy consumption level prediction model construction strategy to construct an energy consumption level prediction model with the key feature data of the driving style and the road condition data as input and the energy consumption level of the future m time nodes as output; when the energy consumption level predicted by the energy consumption level prediction model is higher than the energy consumption safety range, adjusts the feature weight of the neural network model, and then adjusts the vehicle driving style to control the feature data affecting the energy consumption level. The timing and direction of regenerative braking intervention are obtained through the driver's driving habits and style data, making the braking process smoother and energy recovery more efficient. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Other features, objects and advantages of the present invention will become more apparent from the detailed description of non-limiting embodiments made with reference to the following drawings:

[0022] Figure 1 It is a flow chart of the regenerative braking energy feedback control method based on machine learning of the present invention. DETAILED DESCRIPTION

[0023] The present invention aims to propose a regenerative braking energy feedback control method based on machine learning, by collecting vehicle driving style feature data, road condition feature data and energy consumption level data at n consecutive time nodes in the historical vehicle driving process; calculating the average value and change rate of the brake pedal opening, the average value and change rate of the accelerator pedal opening, the average value and change rate of the steering wheel angle, the speed average of the acceleration-to-braking process, and the speed average of the braking-to-acceleration process at n consecutive time nodes in the historical vehicle driving process, respectively importing the correlation between the output and the energy consumption level data into a correlation evaluation strategy and arranging them in descending order, and outputting the key feature data of the vehicle driving style; inputting the key feature data of the vehicle driving style into an energy consumption level prediction model construction strategy to construct an energy consumption level prediction model with the key feature data of the driving style and the road condition data as input and the energy consumption level of the future m time nodes as output; when the energy consumption level predicted by the energy consumption level prediction model is higher than the energy consumption safety range, adjusting the feature weight of the neural network model, and then adjusting the vehicle driving style, and controlling the feature data affecting the energy consumption level. The timing and direction of regenerative braking intervention are obtained through the driver's driving habits and style data, making the braking process smoother and energy recovery more efficient.

[0024] Embodiment 1:

[0025] This embodiment proposes a regenerative braking energy feedback control method based on machine learning, such as Figure 1 As shown, the specific steps include:

[0026] S1. Collecting vehicle driving style characteristic data, road condition characteristic data and energy consumption level data at n consecutive time nodes during the historical vehicle driving process;

[0027] S2. Calculate the average value and change rate of the brake pedal opening, the average value and change rate of the accelerator pedal opening, the average value and change rate of the steering wheel angle, the average speed of the acceleration-to-brake process, and the average speed of the braking-to-acceleration process at n consecutive time nodes in the historical vehicle driving process, import the correlation between the output and the energy consumption level data into the correlation evaluation strategy respectively, and arrange them in descending order, and output the key feature data of the vehicle driving style;

[0028] S3, inputting the key feature data of vehicle driving style into the energy consumption level prediction model construction strategy to construct an energy consumption level prediction model with the key feature data of driving style and road condition data as input and the energy consumption level of m future time nodes as output;

[0029] S4. When the energy consumption level predicted by the energy consumption level prediction model is higher than the energy consumption safety range, the feature weights of the neural network model are adjusted, and then the vehicle driving style is adjusted to control the feature data that affects the energy consumption level.

[0030] In this embodiment, the vehicle driving style characteristic data in S1 include brake pedal opening, accelerator pedal opening, steering wheel angle, speed of acceleration-to-braking process, standard speed of acceleration-to-braking process, speed of braking-to-acceleration process and standard speed of braking-to-acceleration process; the road condition characteristic data include suburban road condition data, urban road condition data and highway road condition data. These data reflect the driver's driving habits.

[0031] In this embodiment, data collected based on different road environments (suburban, urban, and highway) in which the vehicle is located is crucial to understanding changes in vehicle energy consumption, because different road conditions have a significant impact on energy consumption. Real-time monitoring and recording of vehicle energy consumption is a key indicator for evaluating the effect of regenerative braking and optimizing strategies.

[0032] In this embodiment, the relevance strategy in S2 includes the following specific contents:

[0033] S21, extracting features from the average value and change rate of the brake pedal opening, the average value and change rate of the accelerator pedal opening, the average value and change rate of the steering wheel angle, the average speed of the acceleration-to-braking process, and the average speed of the braking-to-acceleration process at n consecutive time nodes in the historical vehicle driving process;

[0034] S22, constructing a correlation analysis model, wherein the correlation analysis model is a neural network model with three layers, wherein the neural network model comprises an input layer, a hidden layer, and an output layer, wherein each node corresponds to a vehicle driving style feature data, and the correlation is used as an output, and the output value is between 0 and 1;

[0035] S23. Arrange the correlations output by the correlation analysis model in descending order, and output the vehicle driving style data with a correlation with the energy consumption level greater than 0.4 as the vehicle driving style key feature data.

[0036] In this embodiment, the energy consumption level prediction model construction strategy in S3 includes the following specific contents: 70% of the feature data is input into the neural network model as a training set for training to obtain an initial neural network model, and 30% of the feature data is used as a test set to test the initial neural network model, and the optimal initial neural network model that meets the accuracy of the historical vehicle energy consumption level change curve data is output as the neural network model.

[0037] In this embodiment, the output formula of a specific neuron in the neural network model is: in is the output of the p-term neuron in the i+1 layer, is the connection weight between the i-th layer neuron j and the i+1 layer p-item neuron, is the output of neuron j in the i-th layer, Represents the bias of the linear relationship, and σ is the sigmoid activation function.

[0038] In this embodiment, when the prediction model indicates that the energy consumption level may exceed the safe range, the system will adjust the driving style by adjusting the feature weights in the neural network model. This means adjusting the frequency and method of using the brake pedal and accelerator pedal, as well as the steering wheel angle, etc. through machine learning to optimize energy consumption.

[0039] This approach combines traditional driving data analysis with modern machine learning technology, aiming to reduce energy consumption and improve energy recovery efficiency by optimizing driving style.

[0040] Embodiment 2:

[0041] This embodiment provides an electronic device, including: a processor and a memory, wherein the memory stores a computer program that can be called by the processor; the processor executes the above-mentioned regenerative braking energy feedback control method based on machine learning by calling the computer program stored in the memory.

[0042] The electronic device may have relatively large differences due to different configurations or performances, and may include one or more processors (Central Processing Units, CPU) and one or more memories, wherein the memory stores at least one computer program, which is loaded and executed by the processor to implement the regenerative braking energy feedback control method based on machine learning provided by the above method embodiment. The electronic device may also include other components for implementing the functions of the device, for example, the electronic device may also have components such as a wired or wireless network interface and an input and output interface to input and output data. This embodiment will not be described in detail here.

[0043] Those skilled in the art know that the present invention can be implemented as a system, method or computer program product. Therefore, the present disclosure can be specifically implemented in the following forms, namely: it can be complete hardware, it can be complete software (including firmware, resident software, microcode, etc.), and it can also be a combination of hardware and software, which is generally referred to as "circuit", "module" or "system" herein. In addition, in some embodiments, the present invention can also be implemented in the form of a computer program product in one or more computer-readable media, and the computer-readable medium contains computer-readable program code.

[0044] Any combination of one or more computer-readable media may be used. A computer-readable medium may be a computer-readable signal medium or a computer-readable storage medium. More specific examples of computer-readable storage media (a non-exhaustive list) include: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In this document, a computer-readable storage medium may be any tangible medium containing or storing a program that may be used by or in conjunction with an instruction execution system, apparatus, or device.

[0045] The present invention is described with reference to flowcharts and block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process or block in the flowchart and block diagram, as well as the combination of processes and blocks in the flowchart or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowcharts. Figure 1 Process or multiple processes and boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0046] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 Process or multiple processes and boxes Figure 1 A step that specifies a function in one or more boxes.

[0047] The embodiments of the present invention are described above in conjunction with the accompanying drawings, but the present invention is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are merely illustrative and not restrictive. Under the enlightenment of the present invention, ordinary technicians in this field can also make many forms without departing from the scope of protection of the purpose of the present invention and the claims, which all fall within the protection of the present invention.

Claims

1. A regenerative braking energy feedback control method based on machine learning, characterized in that: The specific steps include: S1. Collecting vehicle driving style characteristic data, road condition characteristic data and energy consumption level data at n consecutive time nodes during the historical vehicle driving process; S2. Calculate the average value and change rate of the brake pedal opening, the average value and change rate of the accelerator pedal opening, the average value and change rate of the steering wheel angle, the average speed of the acceleration-to-brake process, and the average speed of the braking-to-acceleration process at n consecutive time nodes in the historical vehicle driving process, import the correlation between the output and the energy consumption level data into the correlation evaluation strategy respectively, and arrange them in descending order, and output the key feature data of the vehicle driving style; S3, inputting the key feature data of vehicle driving style into the energy consumption level prediction model construction strategy to construct an energy consumption level prediction model with the key feature data of driving style and road condition data as input and the energy consumption level of m future time nodes as output; S4. When the energy consumption level predicted by the energy consumption level prediction model is higher than the energy consumption safety range, the feature weights of the neural network model are adjusted, and then the vehicle driving style is adjusted to control the feature data that affects the energy consumption level.

2. The regenerative braking energy feedback control method based on machine learning according to claim 1, characterized in that: The vehicle driving style characteristic data in S1 include brake pedal opening, accelerator pedal opening, steering wheel angle, speed of acceleration-to-braking process, standard speed of acceleration-to-braking process, speed of braking-to-acceleration process and standard speed of braking-to-acceleration process; The road condition characteristic data includes suburban road condition data, urban road condition data, and highway road condition data.

3. The regenerative braking energy feedback control method based on machine learning according to claim 2 is characterized in that: The relevance strategy in S2 includes the following specific contents: S21, extracting features from the average value and change rate of the brake pedal opening, the average value and change rate of the accelerator pedal opening, the average value and change rate of the steering wheel angle, the average speed of the acceleration-to-braking process, and the average speed of the braking-to-acceleration process at n consecutive time nodes in the historical vehicle driving process; S22, constructing a correlation analysis model, wherein the correlation analysis model is a neural network model with three layers, wherein the neural network model comprises an input layer, a hidden layer, and an output layer, wherein each node corresponds to a vehicle driving style feature data, and the correlation is used as an output, and the output value is between 0 and 1; S23. Arrange the correlations output by the correlation analysis model in descending order, and output the vehicle driving style data with a correlation with the energy consumption level greater than 0.4 as the vehicle driving style key feature data.

4. The regenerative braking energy feedback control method based on machine learning according to claim 3 is characterized in that: The energy consumption level prediction model construction strategy in S3 includes the following specific contents: 70% of the feature data is used as a training set to input the neural network model for training to obtain an initial neural network model, 30% of the feature data is used as a test set to test the initial neural network model, and the optimal initial neural network model that meets the accuracy of the historical vehicle energy consumption level change curve data is output as the neural network model.

5. The regenerative braking energy feedback control method based on machine learning according to claim 4 is characterized in that: The output formula of a specific neuron in the neural network model is: in is the output of the p-term neuron in the i+1 layer, is the connection weight between the i-th layer neuron j and the i+1 layer p-item neuron, is the output of neuron j in the i-th layer, Represents the bias of the linear relationship, and σ is the sigmoid activation function.

Images