Electric vehicle dual-motor cooperative driving energy-saving optimization method and system

Abstract

The invention discloses an electric vehicle dual-motor cooperative driving energy-saving optimization method and system, and relates to the technical field of energy-saving optimization, and the method comprises the steps: collecting the data of a vehicle operation state; the torque distribution ratio alpha of the double motors is optimized according to a set objective function, the moving direction and speed are adjusted according to the optimal position of the double motors and the optimal position of a group, the optimal point is found, and the optimal torque distribution ratio is obtained; and generating a corresponding torque distribution instruction according to the optimal torque distribution ratio, sending the torque distribution instruction to a motor controller, and executing corresponding torque control. Parameters of the ES-PSO algorithm can be adjusted according to specific problems, and the flexibility helps to better adapt to different drive distribution strategy optimization problems.

Description

Technical Field

[0001] The present invention relates to the technical field of energy-saving optimization, and particularly relates to an energy-saving optimization method and system for collaborative driving of dual motors of an electric vehicle. Background Art

[0002] The driving process of a vehicle can usually be broken down into a combination of a series of typical driving events. Due to multiple factors such as differences in driving experience, driver mood, and external road environment, in a series of driving events, a driver often exhibits his own unique driving behaviors and habits. Driving Style is a human-defined setting based on the comparison of these differences in driving behaviors / habits.

[0003] Only by reasonably distributing the torque of multiple power sources can the working efficiency of the power sources be maintained in the high-efficiency range in different driving styles of the vehicle. Relevant domestic and foreign enterprises and institutions have done a lot of work in the research of dual-motor drive and regenerative braking control strategies. For different dual-motor drive configurations, relevant enterprises and institutions have developed different drive control strategies. For dual-motor coupling, torque coupling, speed coupling, and multi-mode control strategies of a single motor are mostly adopted.

[0004] For dual-motor distributed drive, a control strategy of driving torque distribution is mostly adopted. Wang Da of Jilin University introduced a "human-vehicle-environment" closed-loop decision-making system for the energy consumption mode of a multi-power-source drive system, used the method of fuzzy inference for driving intention identification, and selected and switched between single and dual-motor drive modes according to the driving conditions, driving intention, and road resistance. At the same time, the torque of the dual motors was dynamically distributed with the dual-motor efficiency as the evaluation index.

[0005] The rule-based control method refers to directly controlling operations such as torque distribution of dual motors according to information such as the operating state of the vehicle and environmental conditions by presetting a set of control rules and thresholds. These rules are usually based on engineering experience, experimental data, or simulation analysis, have a clear logical structure, and can quickly implement vehicle control decisions. However, this type of method has the following disadvantages:

[0006] Due to the relatively fixed preset rules, it is difficult to adapt to complex or changing working conditions. The system cannot adaptively adjust the control strategy according to the actual driving scenario, resulting in difficulty in achieving the best energy efficiency management. The control rules rely on empirical settings and lack global optimization ability, which may lead to sub-optimal control strategies, especially in non-linear and multi-condition cases. When the driving environment or working conditions change, the system needs to re-adjust or expand the rule set and it is difficult to automatically learn new driving modes or working conditions. Summary of the Invention

[0007] The object of the present invention is to provide a method and system for energy-saving optimization of dual-motor collaborative drive for electric vehicles in view of the deficiencies of the above-mentioned prior art, so as to solve the problems in the prior art.

[0008] The present invention specifically provides the following technical solutions:

[0009] A method for energy-saving optimization of dual-motor collaborative drive for electric vehicles, comprising:

[0010] Collecting data on the running state of the vehicle;

[0011] Based on the data, with the maximum energy utilization rate of the dual motors as the objective function, the particle swarm optimization algorithm ES-PSO adjusts the moving direction and speed of the particles according to the best position of the particles themselves and the best position of the group to find the optimal point, and obtains the optimal torque distribution ratio α of the dual motors through the optimal point; the particles represent potential torque distribution ratios;

[0012] Generating corresponding torque distribution instructions according to the optimal torque distribution ratio, and sending the torque distribution instructions to the motor controller to execute corresponding torque control.

[0013] Preferably, before collecting the data on the running state of the vehicle, it further includes:

[0014] Constructing a vehicle driving environment;

[0015] Constructing a driving style recognition system by using an improved supervised learning algorithm RE-BPNN, and identifying the vehicle driving environment through the driving style recognition system to obtain data on the running state of the vehicle.

[0016] Preferably, identifying the vehicle driving environment through the driving style recognition system to obtain data on the running state of the vehicle includes:

[0017] Obtaining the vehicle speed, longitudinal acceleration, steering wheel angle and angular velocity in the driving environment, and taking the vehicle speed, longitudinal acceleration, steering wheel angle and angular velocity as the input values of the driving style recognition system, and setting the expected output of the driving style recognition system;

[0018] Inputting the input values into the driving style recognition system to obtain driving style parameters, comparing the driving style parameters with the expected output for error, and when the error threshold requirement is not met, recalculating the hidden layer unit error and error gradient, and performing network weight learning to obtain an updated driving style recognition system;

[0019] Iteratively recognizing the input values again through the updated driving style recognition system until the driving style parameters and the expected output meet the error threshold, outputting the current driving style parameters, and taking the current driving style parameters as the data on the running state of the vehicle.

[0020] Preferably, the data of the vehicle operating state includes: vehicle speed signal, accelerator pedal signal, motor speed signal, and driving style parameter.

[0021] Preferably, before sending the torque distribution instruction to the motor controller, it further includes smoothing the torque shock to obtain the final torque processing result, specifically:

[0022] Under the condition of meeting the vehicle demand torque, find the overall vehicle motor efficiency ηf with the optimal energy utilization efficiency of the vehicle drive system in the motor efficiency map, and obtain the torque distribution coefficient m through the overall vehicle motor efficiency ηf.

[0023] The present invention provides an energy-saving optimization system for cooperative driving of dual motors of an electric vehicle, including:

[0024] An acquisition module for acquiring data of the vehicle operating state;

[0025] A distribution module for, based on the data, taking the maximum energy utilization rate of the dual motors as the objective function, using the particle swarm algorithm ES-PSO to adjust the moving direction and speed of the particles according to the best positions of the particles themselves and the best position of the group, find the optimal point, and obtain the optimal torque distribution ratio α of the dual motors through the optimal point; the particles represent potential torque distribution ratios;

[0026] A control module for generating a corresponding torque distribution instruction according to the optimal torque distribution ratio and sending the torque distribution instruction to the motor controller to perform corresponding torque control.

[0027] The present invention provides a computer device, including a memory and a processor. When a program stored in the memory is executed by the processor, the processor executes the steps of the above-mentioned energy-saving optimization method for cooperative driving of dual motors of an electric vehicle.

[0028] The present invention provides a storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the above-mentioned energy-saving optimization method for cooperative driving of dual motors of an electric vehicle are realized.

[0029] Compared with the prior art, the present invention has the following remarkable advantages:

[0030] Based on the collected vehicle operation state data, with the maximized energy utilization rate of the dual motors as the objective function, the particle swarm optimization algorithm ES-PSO adjusts the moving direction and speed of the particles according to the best positions of the particles themselves and the best position of the group to find the optimal point. Through the optimal point, the optimal torque distribution ratio α of the dual motors is obtained. In this process, the global optimal solution of the dual-motor torque distribution is obtained through the ES-PSO algorithm to balance various performance indicators. When dealing with complex or variable working conditions and driving environments, the parameters of the ES-PSO algorithm can be specifically adjusted according to specific problems. This flexibility helps to better adapt to different drive distribution strategy optimization problems and better learn new driving modes or working conditions. Description of the Drawings

[0031] Figure 1 It is a BP neural network diagram;

[0032] Figure 2 It is the flow chart of the RE-BPNN algorithm of the present invention;

[0033] Figure 3 It is the model diagram of the torque distribution control strategy of the present invention;

[0034] Figure 4 It is the flow chart of an energy-saving optimization method for the coordinated drive of dual motors of an electric vehicle according to the present invention. Detailed Embodiment

[0035] The following combines the drawings in the present invention to clearly and completely describe the technical solutions of the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the scope of protection of the present invention.

[0036] At present, the research on driving styles is diverse and complex, but there is no clear definition of driving styles. The current mainstream view believes that driving styles are mainly affected by subjective factors such as driving experience and driving emotions and external factors such as road environment. At present, the main classification criteria for driving styles are: classification based on fuel consumption and classification based on the aggressive behavior of drivers.

[0037] The foreign research and application of driving styles mainly focus on aspects such as driving style classification and recognition algorithm research, vehicle longitudinal control, and dangerous driving behavior recognition.

[0038] In terms of driving style classification and recognition, the style labels of drivers are often unknown in advance. Therefore, clustering methods in unsupervised machine learning are often used to identify the driving styles reflected in the data set. The unsupervised learning clustering methods currently used in driving style research mainly include K-means, K-medoids, and Fuzzy-c means (FCM), etc. Kalsoom et al. proposed an innovative idea for driver style classification in 2013. They collected features such as the number of driver braking times, average gear, maximum gear, average speed, and maximum speed, and used K-means and hierarchical clustering to distinguish slow, normal, and fast driving styles. It was found that for the collected data set, the K-means method was better than hierarchical clustering. Henrik et al. used K-means clustering and t-SNE dimensionality reduction method to classify the driving styles of GPS data on road network hotspots to improve the effectiveness of processing large driving data sets. Chen et al. used the Strategic Highway Research Program 2 (SHRP2) database, selected features such as vehicle speed, lateral and longitudinal acceleration, and used the K-Medoids method to cluster driving styles, and explored the correlation between driving styles and driver attributes. It was found that clustering could reflect the individual differences in driving styles to a certain extent. Kwo k Tai Chui used electrocardiogram signals as input and aimed to minimize the number of clusters, and developed a driver stress recognition algorithm based on multi-objective genetic algorithm and fuzzy C-means clustering. The results showed that the number of clusters was reduced by 2, and the time complexity and model complexity were reduced by 68% respectively.

[0039] In China, the research and application of driving styles mainly focus on the research of driving behavior characteristics, the longitudinal dynamics control of vehicles, the trajectory prediction of autonomous vehicles, etc.

[0040] Due to the complexity of the real road environment and the differences between different drivers, driving data often contains multi-source information such as vehicles, drivers, and road environments. Therefore, the current definition and classification of driving behavior characteristics are not clear. He Xin et al. developed a driving behavior standardization algorithm based on a gated sparse autoencoder network, realized driving feature extraction through self-supervised training, and mined and represented the hidden features of driving data through phase space reconstruction method, and achieved the composite task of driving style classification and abnormal driving behavior recognition. Li Lun based on the high-frequency operation data of shared electric vehicles, took driving characteristics and dangerous driving behaviors as the main research objects, selected vehicle speed, longitudinal acceleration, steering wheel angle and angular velocity as the main characteristic parameters to describe driving behaviors, and explored the statistical characteristics and correlations of the parameters to establish an algorithm for identifying dangerous driving behaviors, and at the same time classified and rated driving styles.

[0041] In terms of the longitudinal dynamics control of vehicles, Gao Bingzhao et al. proposed a personalized Adaptive Cruise Control (ACC) system based on driving style recognition and Model Predictive Control (MPC) to adapt to different driving styles and take into account the comprehensive performance requirements of vehicles. A driving data set was collected from 66 randomly recruited drivers, and then clustered into three categories: conservative, moderate, and aggressive through a Self-organizing map (SOM) neural network. On this basis, the control problem of the personalized ACC system was described as a multi-objective optimization problem solved by the MPC method.

[0042] The simulation results show that the proposed personalized ACC system can meet the requirements of different driving styles and ensure the longitudinal performance of the vehicle; Yang Caixia collected real vehicle data, based on the driving style classification results of the improved K-means method, clarified the optimization objectives corresponding to each driving style, established the relationship between the parameters corresponding to each optimization objective and the parameters describing the driving style, to guide the solution of the MPC optimization controller, and finally realized the longitudinal multi-objective optimization control of the vehicle that can meet different driving style preferences.

[0043] Zheng Qi selected the US-101 data set in NGSIM as the research object, and proposed a feature extraction and dimensionality reduction method of LSTM denoising autoencoder suitable for driving high-dimensional time series to extract low-dimensional features for style clustering and analysis, focusing on defining driving style through the dynamic changes of drivers; on the basis of considering driving style and vehicle information around, proposed a method for identifying vehicle lane-changing intention based on Bi-GRU; and introduced driver style and vehicle information into trajectory prediction, and proposed a multi-information fusion vehicle trajectory prediction algorithm based on the attention mechanism. Deng Zejian reduced the dimensionality of driving features based on the principal component analysis method, completed driving style clustering through the K-means clustering method, and quantitatively described the trajectory selection characteristics of drivers with different styles to decide the reference value of the steering wheel angle at the bottom layer of the vehicle. The MPC algorithm was selected to design the trajectory following controller. The co-simulation results show that the vehicle trajectory tracking controller based on MPC can effectively follow the reference trajectory while ensuring the stability of vehicle driving.

[0044] Only by reasonably allocating the torques of multiple power sources can the working efficiency of the power sources be maintained in the high-efficiency range. Relevant domestic and foreign enterprises and universities have done a lot of work in the research of dual-motor drive and regenerative braking control strategies. For different dual-motor drive configurations, relevant enterprises and universities have developed different drive control strategies. For dual-motor coupling, torque coupling, speed coupling, and multi-mode control strategies of single motors are mostly used.

[0045] For dual-motor distributed drive, a control strategy of driving torque distribution is mostly adopted. Wang Da from Jilin University introduced a closed-loop decision-making system of "human-vehicle-environment" for the energy consumption mode of multi-power-source drive systems, used the method of fuzzy inference to identify driving intentions, and selected and switched between single- and dual-motor drive modes according to driving conditions, driving intentions, and road resistance. At the same time, the torque of the dual motors was dynamically distributed with the dual-motor efficiency as the evaluation index.

[0046] Wang Xiaowei et al. proposed a torque distribution strategy with a two-layer architecture. The upper layer is the driving torque and yaw moment decision-making layer, and the lower layer is the motor torque distribution layer. A control strategy was designed with economy, stability, and motor failure as multi-objective optimization functions.

[0047] Yan Chunhui et al. proposed a distributed electric vehicle torque optimization distribution control strategy for four-wheel in-wheel motors based on different working conditions, which ensured the real-time requirement of motor torque control when the vehicle was driving at high speed and the stability of vehicle steering at the same time.

[0048] Wang Jia et al. divided the driving states of the vehicle into economic, power, and normal modes. Among them, the power mode is applicable to vehicle acceleration, relatively harsh road conditions, overtaking, etc. The economic mode is applicable to low-battery conditions, and the normal mode is applicable to normal commuting road conditions. Using the method of fuzzy control, the working modes of the dual-motor drive system were allocated to realize the recognition and switching of the three control modes, improving the anti-interference ability of vehicle control

[30] .

[0049] Jiang Tao et al. took an electric vehicle with independent front and rear axle drives as the research object, focused on the mode selection problem between single-motor drive and dual-motor drive, and the sudden change of the vehicle's total torque caused by the switching between different drive modes, resulting in the problem of vehicle power output smoothness. A vehicle control strategy model was built and hardware-in-the-loop testing was carried out for verification.

[0050] Han Zongzhi et al. proposed a fault-tolerant control method for a four-wheel drive intelligent agent system based on sliding mode control to improve the fault tolerance and robustness of the drive slip control system of four-wheel drive distributed electric vehicles and reduce the computational complexity of the system.

[0051] Yuan Jiajia et al. analyzed the relationship between the optimal control parameters of active rear-wheel steering and the pros and cons of control performance, and to overcome the limitations of single steering control under special working conditions, proposed a coordinated control strategy for active rear-wheel steering and torque distribution based on the determination principle of the phase plane method.

[0052] Holdstock proposed a variable-speed system with a dual-motor configuration, no clutch, and uninterrupted power, introduced the layout method, modeling method of the dual-motor configuration, and a mode switching strategy based on efficiency optimization. Finally, the energy consumption and performance of traditional single-motor drive systems and dual-motor drive systems were compared.

[0053] Xu Shuoshuo et al. proposed a torque distribution control strategy based on driver intention for electric vehicles with in-wheel motors and four-wheel independent drive. They adopted a simulated annealing particle swarm optimization particle filter algorithm to establish a state estimation model during vehicle driving. According to the model predictive control algorithm, different torque distribution strategies were formulated for different target states to improve the mobility and stability of the vehicle.

[0054] Guo Shuai et al. took the dual-axis independent drive configuration as the research object based on the analysis and research of the dual-motor drive configuration, developed and designed a vehicle controller for distributed dual-motor electric vehicles, and verified the practicability and reliability of the upper control strategy of the dual-motor electric vehicle control system.

[0055] Michael A. Miller designed different working modes and control strategies for the Chevrolet Volt based on the output torque and efficiency of the vehicle system, and made a regional map of vehicle speed - output torque, which detailedly divided the operation regions of different modes.

[0056] Xu Dan et al. from Xi'an Jiaotong University proposed a torque distribution strategy for four-wheel drive electric vehicles with multiple power sources, which can improve vehicle operation stability and driving economy, and has certain reference significance for formulating the torque distribution strategy of dual motors.

[0057] Chu Hong et al. analyzed the characteristics of energy loss caused by motor energy loss and tire slip for the structure and energy consumption characteristics of four-wheel drive electric vehicles. Taking the motor energy loss and tire slip energy loss as the performance index function, they established an optimization control problem of torque distribution considering the overall vehicle energy consumption in the master's thesis of China University of Petroleum (East China).

[0058] Wu Zhe from Shandong University of Technology proposed a control strategy for the timely drive of dual motors for the drive configuration of dual motors and dual axes. The front and rear motors enter the working or "sleep" state timely according to the change of demand torque to provide power for the whole vehicle.

[0059] Xiong Huiyuan et al. from Sun Yat-sen University established an optimization model with the maximum utilization efficiency of dual motors as the goal based on the Map efficiency characteristics of motors for the dual-motor drive configuration, and proposed a torque distribution control strategy for the dual-motor drive system under different driving conditions.

[0060] Cui Huamei from Anhui University took the front and rear axle dual-motor drive system as the research object, proposed a distributed dual-motor torque distribution control strategy based on SMC for the safety and stability requirements during vehicle driving, modeled and simulated the control strategy, and obtained the conclusion that the torque distribution strategy has a certain effect by analyzing the results of co-simulation.

[0061] The purpose of the present invention is to realize the recognition of driving styles in real life. Therefore, the selection criteria for the driving environment are listed as follows:

[0062] 1. The driving environment should be as close to the actual situation as possible, and the information around the road should be rich enough to serve as an important reference for the driver to judge information such as the current vehicle speed, acceleration, yaw rate, and vehicle position.

[0063] 2. The selected driving environment should be compatible with drivers of different styles. On the one hand, the driving environment should not be too complex to prevent misoperations by conservative-style drivers; on the other hand, the driving environment should enable the most aggressive drivers to fully exert their own levels and vehicle performance.

[0064] 3. Considering the variability of driving styles, a driving environment that can reflect the transient characteristics rather than the steady-state characteristics of the human-vehicle system should be selected.

[0065] Construct an intelligent driving style recognition system (RE-BPNN algorithm):

[0066] The BPNN algorithm is a classic supervised learning method. As Figure 1 shown, BPNN includes an input layer, a hidden layer, and an output layer, and is used to train a feedforward neural network. This algorithm adjusts the network weights and biases by using the gradient descent method by comparing the error between the network output and the expected output to achieve the training and learning of the network.

[0067] Innovatively proposed the RE-BPNN (Restart Enhance-Back Propagation Neural Network) algorithm. The relationship between driving style features and pattern recognition may be very complex. RE-BPNN can capture these non-linear relationships through multiple layers of neurons. At the same time, RE-BPNN can adaptively adjust the weights and biases to cope with different types of input data, which is very beneficial for processing highly variable data such as driving behaviors. And the implementation of the RE-BPNN algorithm is relatively simple, and there are many mature neural network libraries to choose from, such as scikit-learn and TensorFlow of Python.

[0068] Establishment of a dual-motor model and parameter verification:

[0069] A coaxial dual-motor drive system is composed of two electric motors with different specifications, and a dual-motor drive control strategy is built with Simulink. According to the parameter matching results, a simulation model is established using Cruise, as shown in Figure 3 . Set vehicle calculation tasks such as acceleration performance, climbing performance, and NEDC cycle conditions.

[0070] Construction of a torque distribution strategy:

[0071] The PSO algorithm is a stochastic search optimization algorithm based on swarm intelligence, proposed by American scholars James Kennedy and Russell Eberhart in 1995. Inspired by the flocking behavior of birds, this algorithm solves complex optimization problems by simulating the process of a flock of birds searching for food. The PSO algorithm refers to each solution to be optimized as a "particle". Each particle flies in the solution space and continuously adjusts its speed and position based on its individual best position and the global best position. The innovative technology of this invention proposes a brand-new ES-PSO (Enhance Search-Particle Swarm Optimization) algorithm. The ES-PSO algorithm has good global search ability and can jump out of local optimal solutions, which is crucial for finding the optimal drive allocation strategy. In the design of vehicle power systems, it is necessary to find the global optimal solution to balance various performance indicators. At the same time, the parameters of the ES-PSO algorithm (such as inertia weight, acceleration coefficient, etc.) can be adjusted according to specific problems, and this flexibility helps to better adapt to different drive allocation strategy optimization problems. In one scenario, with vehicle speed, load signal, motor speed, and a constant as inputs, operations of multiple comparison constants, logical operators, gains, partitions, quantifications, and 2-D look-up tables are performed, and two desired torques are output through two switches.

[0072] As Figure 2 and Figure 4 shown, based on this, an embodiment of the present invention provides an energy-saving optimization method for collaborative driving of dual motors in an electric vehicle, including:

[0073] Step S1: Collect data on the vehicle operating state.

[0074] The system needs to collect data on the vehicle operating state in real time, including but not limited to vehicle speed signals, accelerator pedal signals, motor speed signals, etc. These information are the basis for calculating the required torque command in the current state.

[0075] The data obtained through sensors or other means, as well as driving style parameters, will be input into the control system for preliminary processing to ensure the validity and accuracy of the data. Before collecting data on the vehicle operating state, it also includes:

[0076] Construct the vehicle driving environment.

[0077] Use the improved supervised learning algorithm RE-BPNN to construct a driving style recognition system, and identify the vehicle driving environment through the driving style recognition system to obtain data on the vehicle operating state. Specifically including:

[0078] Obtain the vehicle speed, longitudinal acceleration, steering wheel angle, and angular velocity in the driving environment, and use the vehicle speed, longitudinal acceleration, steering wheel angle, and angular velocity as input values for the driving style recognition system. Set the expected output of the driving style recognition system.

[0079] Input the input values into the driving style recognition system, obtain the driving style parameters, and compare the error between the driving style parameters and the expected output. When the error threshold requirement is not met, recalculate the hidden layer unit error and error gradient, perform network weight learning, and obtain an updated driving style recognition system. When all errors meet the requirements, end and output.

[0080] Iteratively recognize the input values again through the updated driving style recognition system until the driving style parameters and the expected output meet the error threshold. Then, output the current driving style parameters and use the current driving style parameters as part of the vehicle operating state data.

[0081] Step S2: Based on the data, with the maximization of the energy utilization rate of the dual motors as the objective function, the particle swarm optimization algorithm ES-PSO adjusts the moving direction and speed of the particles according to the best position of the particles themselves and the best position of the group, finds the optimal point, and obtains the optimal torque distribution ratio α of the dual motors through the optimal point; the particles represent potential torque distribution ratios.

[0082] Particle Swarm Optimization Algorithm (ES-PSO) Module: Based on the data obtained in the previous step, the ES-PSO module optimizes the torque distribution ratio α of the dual motors according to the set objective function (such as maximizing the energy utilization rate ηZ).

[0083] The ES-PSO algorithm searches for the optimal solution by simulating the foraging behavior of a bird flock. In this process, each "particle" represents a potential solution, and they adjust their moving direction and speed according to their own best position and the best position of the group, and finally find the optimal point. After the algorithm iteration ends, a set of torque distribution ratios that can optimize the objective function will be obtained.

[0084] Step S3: Generate corresponding torque distribution instructions according to the optimal torque distribution ratio, and send the torque distribution instructions to the motor controller to execute the corresponding torque control.

[0085] Generate corresponding torque distribution instructions according to the optimal torque distribution ratio obtained by the ES-PSO algorithm, and send them to the motor controller (MCU) to execute the corresponding torque control.

[0086] Before sending the torque distribution instructions to the motor controller, this stage may also involve smoothing the torque shock to reduce mechanical vibration or an unsmooth driving experience caused by rapidly changing torque demands.

[0087] Apply the particle swarm optimization algorithm to the torque distribution coefficient optimization of the dual-motor drive system. Under the condition of meeting the vehicle's required torque, find the optimal overall vehicle motor efficiency ηf of the vehicle drive system's energy utilization efficiency in the motor efficiency map, and obtain the torque distribution coefficient m through the overall vehicle motor efficiency ηf.

[0088] The particle position in the particle swarm optimization algorithm represents the torque distribution coefficient m, denoted as x i ; the particle velocity represents the change value of the torque distribution coefficient, denoted as v i ; during the iteration process, its individual extreme value is denoted as P i , and the global extreme value is denoted as P g . The particle position and velocity are iteratively updated as follows:

[0089]

[0090] where ω is the inertia weight; t is the current iteration number; c 1 is the self-learning factor, taking 0.5; c 2 is the swarm learning factor, taking 0.5; r 1 , r 2 are random numbers in the range of 0 to 1.

[0091] Use the PreScan software and the PanoSim platform to conduct simulations and analyze data:

[0092] The PreScan professional simulation software can quickly and conveniently build a visual driving scenario model, effectively supporting the research of driving style recognition. PanoSim is an automotive intelligent driving simulation software developed by Panasonic Automotive Systems Europe GmbH. It is an integrated tool and platform that focuses on realizing the virtual R & D of automobiles through simulation technology. PanoSim supports offline simulation and can conduct comprehensive simulation tests on the autonomous driving system before real-world testing. Its advantage lies in being able to simulate complex traffic scenarios, including the behaviors of vehicles, pedestrians, bicycles, and other transportation tools.

[0093] The driving environment should be as close to the actual situation as possible, and the information around the road should be rich enough to serve as an important reference for the driver to judge information such as the current vehicle speed, acceleration, yaw rate, and vehicle position. The selected driving environment should be compatible with different styles of drivers. On the one hand, the driving environment should not be too complex to prevent misoperations by conservative-style drivers; on the other hand, the driving environment should enable the most aggressive-style drivers to fully exert their own levels and vehicle performance.

[0094] Based on the above method, the present invention provides an energy-saving optimization system for cooperative driving of dual motors of an electric vehicle, including: a collection module, a distribution module, and a control module.

[0095] Among them, the collection module is used to collect data on the running state of the vehicle; the distribution module is used to, based on the data, take the maximum energy utilization rate of the dual motors as the objective function, and the particle swarm algorithm ES-PSO adjusts the moving direction and speed of the particles according to the best position of the particles themselves and the best position of the group to find the optimal point, and obtains the optimal torque distribution ratio α of the dual motors through the optimal point; the particles represent potential torque distribution ratios; the control module is used to generate corresponding torque distribution instructions according to the optimal torque distribution ratio and send the torque distribution instructions to the motor controller to execute the corresponding torque control.

[0096] The present invention also provides a computer device, including a memory and a processor. When a program stored in the memory is executed by the processor, the processor executes the steps of an energy-saving optimization method for cooperative driving of dual motors of an electric vehicle.

[0097] According to the disclosed embodiments, the computer device can communicate with one or more external devices (such as a keyboard, a pointing device, Bluetooth communication, etc.), or communicate with any device (such as a router, a demodulator, etc.) that enables the computing device to communicate with one or more other computing devices.

[0098] The present invention also provides a storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of an energy-saving optimization method for cooperative driving of dual motors of an electric vehicle are implemented.

[0099] According to the disclosed embodiments, the storage medium can be a non-volatile computer-readable storage medium, for example, it can include but is not limited to: 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), 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 the present invention, the storage medium can be any tangible medium that contains or stores a program, and the program can be used by or in combination with an instruction execution system, apparatus, or device.

[0100] The above content is a further detailed description of the present invention in combination with specific preferred embodiments. For those skilled in the technical field of the present invention, without departing from the concept of the present invention, several simple deductions or substitutions can still be made, which should all be regarded as belonging to the protection scope of the present invention.

Claims

1. A method for energy-saving optimization of electric vehicle dual-motor cooperative driving, characterized in that: include: Collect data on vehicle operating status; Based on the data, with the maximum energy utilization of the dual motors as the objective function, the particle swarm algorithm ES-PSO adjusts the moving direction and speed of the particles according to the optimal position of the particles themselves and the optimal position of the group, finds the optimal point, and obtains the optimal torque distribution ratio α of the dual motors through the optimal point; the particles represent the potential torque distribution ratio; The corresponding torque distribution instruction is generated according to the optimal torque distribution ratio, and the torque distribution instruction is sent to the motor controller to execute the corresponding torque control.

2. The energy-saving optimization method for dual-motor cooperative driving of an electric vehicle according to claim 1, characterized in that: Before collecting the data of the vehicle running status, the method further includes: Constructing vehicle driving environment; The improved supervised learning algorithm RE-BPNN is used to construct a driving style recognition system, through which the vehicle driving environment is recognized to obtain vehicle operating status data.

3. The electric vehicle dual-motor cooperative drive energy-saving optimization method according to claim 2, characterized in that: The driving environment of the vehicle is identified by the driving style identification system to obtain data on the vehicle operating status, including: Acquiring vehicle speed, longitudinal acceleration, steering wheel angle and angular velocity in a driving environment, and using the vehicle speed, longitudinal acceleration, steering wheel angle and angular velocity as input values ​​of a driving style recognition system, and setting a desired output of the driving style recognition system; Inputting the input value into the driving style recognition system, obtaining the driving style parameter, and performing an error comparison between the driving style parameter and the expected output, and when the error threshold requirement is not met, recalculating the hidden layer unit error and the error gradient, performing network weight learning, and obtaining an updated driving style recognition system; The input value is iteratively identified again through the updated driving style recognition system until the driving style parameter and the expected output meet the error threshold, the current driving style parameter is output, and the current driving style parameter is used as the data of the vehicle operation status.

4. The electric vehicle dual-motor cooperative drive energy-saving optimization method according to claim 3, characterized in that: The vehicle running status data includes a vehicle speed signal, an accelerator pedal signal, a motor speed signal and a driving style parameter.

5. The electric vehicle dual-motor cooperative drive energy-saving optimization method according to claim 1, characterized in that: Before sending the torque distribution instruction to the motor controller, the torque impact is smoothed to obtain the final torque processing result, specifically: Under the condition that the required torque of the vehicle is met, the whole vehicle motor efficiency ηf with the best energy utilization efficiency of the vehicle drive system is found in the motor efficiency diagram, and the torque distribution coefficient m is obtained through the whole vehicle motor efficiency ηf.

6. An electric vehicle dual-motor cooperative drive energy-saving optimization system, characterized in that: include: A collection module, used to collect data on vehicle operation status; The allocation module is used to adjust the moving direction and speed of the particles according to the optimal position of the particles themselves and the optimal position of the group based on the data, and obtain the optimal torque allocation ratio α of the dual motors through the optimal point; the particles represent the potential torque allocation ratio; The control module is used to generate a corresponding torque distribution instruction according to the optimal torque distribution ratio, and send the torque distribution instruction to the motor controller to perform corresponding torque control.

7. A computer device, characterized in that: It includes a memory and a processor, wherein a program is stored in the memory, and when the program is executed by the processor, the processor executes the steps of an electric vehicle dual-motor cooperative drive energy-saving optimization method as described in any one of claims 1 to 5.

8. A storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the electric vehicle dual-motor cooperative drive energy-saving optimization method described in any one of claims 1 to 5 are implemented.

Images