Online self-adaptive gear shifting control method and system for heavy electric commercial vehicle

Abstract

The invention relates to the field, in particular to an online self-adaptive gear shifting control method and system for a heavy electric commercial vehicle. Constructing an off-line optimization gear shifting control strategy based on a dynamic programming algorithm; generating an optimal gear selection rule table of the electric commercial vehicle; inputting the optimal gear selection rule table into a pre-constructed gear selection model for training to obtain a trained gear selection model; and deploying the trained gear selection model to a vehicle-mounted controller, and enabling the vehicle-mounted controller to adapt to a dynamic working condition based on real-time output of a precise gear instruction. By fusing the real-time road gradient sensing technology and the historical gear shifting frequency optimization algorithm, the problem of cyclic gear shifting existing in a ramp change scene in traditional gear shifting control is effectively solved, the adaptability to complex working conditions such as continuous ramps and load fluctuation is enhanced, stable and reliable gear selection can be kept, and the gear shifting efficiency is improved. And the driving safety and comfort are obviously improved.

Description

Technical Field

[0001] This invention relates to the field of online adaptive gear shifting control method and system for heavy-duty electric commercial vehicles. Background Technology

[0002] Driven by both the "dual carbon" strategy and the green upgrade of the logistics industry, the electrification transformation of heavy-duty commercial vehicles has become an inevitable trend. However, the shift control of their multi-speed automatic mechanical transmissions (AMT) still faces severe challenges under complex slope conditions.

[0003] The high torque output of the electric drive system in heavy-duty electric commercial vehicles exacerbates the reliability risks of multi-speed AMT shift control: traditional rule-based solutions rely on fixed thresholds such as vehicle speed and accelerator pedal opening, failing to dynamically perceive changes in road gradient. This leads to insufficient torque in higher gears or a sharp drop in efficiency in lower gears when climbing hills, significantly increasing the risk of drive system overload damage. In scenarios involving rapid gradient changes and start-stop operations, the lack of a transient load prediction mechanism causes cyclic switching between adjacent gears, accelerating clutch wear and generating additional energy consumption. However, existing shift strategies face the dual challenges of increased energy consumption due to cyclic shifting caused by traditional rule-based methods in dynamic gradient environments, and insufficient real-time performance due to the exponential growth of computational complexity of global optimization methods such as dynamic programming with state variables.

[0004] This contradiction of "offline optimization is unavailable and online strategy is underoptimized" has severely restricted technological breakthroughs in shift control of heavy-duty electric commercial vehicles. Efficient control of its multi-speed electric drive system has become a key technology for improving the economy and driving quality of the entire vehicle. Summary of the Invention

[0005] To address the problems mentioned in the prior art, this invention proposes an online adaptive gear shifting control method and system for heavy-duty electric commercial vehicles, which solves the shortcomings of the prior art in the lack of effective strategies for energy consumption surges caused by dynamic slope changes and frequent gear shifts, and the problem that dynamic programming global optimization cannot be applied in real time because it depends on known operating conditions.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: This invention discloses an online adaptive gear shifting control method for heavy-duty electric commercial vehicles, comprising the following steps: S1. Construct an offline optimized shift control strategy based on dynamic programming algorithm; the offline optimized shift control strategy aims to minimize the energy consumption of the whole vehicle, introduces a gear shifting penalty term, takes the gear shifting action as the control variable, and takes the battery SOC and the current gear as the state variables to solve for the optimal gear, and combines the optimal gear with the vehicle operating conditions to generate the optimal gear selection rule table for electric commercial vehicles. S2. Input the optimal gear selection rule table into the pre-built gear selection model for training to obtain the trained gear selection model; S3. Deploy the trained gear selection model to the vehicle controller, which then outputs precise gear selection commands in real time to adapt to dynamic operating conditions.

[0007] As a further improvement to the present invention, the process of S1 is as follows: Obtain vehicle mass, speed, acceleration, and road gradient as required by driving conditions; The driving conditions are divided into several parts based on time. The battery state of charge value and the current gear are selected at any time k as the state vector, and the gear shifting operation is set as the control variable. Based on the state variables and control variables, a target function is constructed, as shown in the following equation:

[0008] In the formula: The change in SOC; The equivalent energy consumption for shift penalty; The instantaneous cost function; Let be the decision variable for the k-th stage, denoted as: downshift, no shift, upshift, represented by 1, 0, and -1 respectively; Set constraints on the objective function; Based on the vehicle's operating conditions, the objective inner function is solved, and the optimal gear selection rule table is generated with the goal of minimizing the overall vehicle energy consumption.

[0009] As a further improvement to the present invention, the constraint condition is shown in the following equation:

[0010] In the formula: T m min and T m max These represent the minimum and maximum torques of the motor at the current moment, respectively; n m min and n m max These represent the minimum and maximum speeds of the motor at the current moment; SOC min and SOC max These are the minimum and maximum values ​​that SOC can reach, respectively; P bat max and P bat max These represent the lower and upper limits of battery charging and discharging power, respectively.

[0011] As a further improvement of the present invention, the shift penalty is shown in the following formula:

[0012] In the formula: λ is the weighting coefficient. The larger the weighting coefficient, the greater the cost of shifting gears.

[0013] As a further improvement of the present invention, the gear selection model is constructed based on CNN-BiLSTM, including convolutional layers, bidirectional long short-term memory layers and fully connected layers; The convolutional layer is used to extract local nonlinear features of the input parameters; the bidirectional long short-term memory layer learns the long-term dependencies of the time series through forward and backward LSTM layers; the fully connected layer is used to integrate features and output gear prediction results.

[0014] As a further improvement to the present invention, the process of S3 is as follows: The vehicle controller acquires vehicle speed, acceleration, road gradient, and wheel-end torque in real time, inputs them into the gear selection model, and outputs real-time gear control commands.

[0015] This invention proposes an online adaptive shifting control system for heavy-duty electric commercial vehicles, comprising: The offline optimization module is used to construct an offline optimized shift control strategy based on a dynamic programming algorithm. The offline optimized shift control strategy aims to minimize the energy consumption of the entire vehicle, introduces a gear shifting penalty term, uses the gear shifting action as a control variable, and uses the battery SOC and the current gear as state variables to solve for the optimal gear. The optimal gear is then combined with the vehicle operating conditions to generate an optimal gear selection rule table for electric commercial vehicles. The training module is used to input the optimal gear selection rule table into the pre-built gear selection model for training, so as to obtain the trained gear selection model. The online control module is used to deploy the trained gear selection model to the vehicle controller, which then outputs precise gear commands in real time to adapt to dynamic operating conditions.

[0016] As a further improvement of the present invention, it also includes a sensing module, which includes a speed sensor, a gyroscope, a GPS and a lidar and a binocular camera, for collecting vehicle speed, acceleration and road slope signals and transmitting them to the online control module.

[0017] This invention proposes an online adaptive gear shifting control device for heavy-duty electric commercial vehicles, comprising a processor and a memory, wherein the processor executes a computer program stored in the memory to implement the online adaptive gear shifting control method for heavy-duty electric commercial vehicles as described above.

[0018] A computer-readable storage medium for storing a computer program, wherein the computer program, when executed by a processor, implements the online adaptive shift control method for heavy-duty electric commercial vehicles as described above.

[0019] Compared with the prior art, the present invention achieves the following technical effects: This invention effectively solves the problem of cyclic shifting in traditional shift control systems under varying slope scenarios by integrating real-time road gradient sensing technology with historical shift frequency optimization algorithms. The system can dynamically sense continuous changes in road gradient and intelligently adjust the shifting strategy based on the vehicle's current operating status, significantly enhancing its adaptability to complex conditions such as continuous slopes and load fluctuations. Especially in road conditions with frequent gradient changes, such as mountain roads and urban overpasses, it can maintain stable and reliable gear selection, significantly improving driving safety and comfort.

[0020] This invention employs a collaborative control architecture combining dynamic programming and neural networks. While ensuring optimal gear selection, it achieves a dual improvement in energy economy and driving smoothness. By introducing a shift frequency penalty mechanism, unnecessary gear switching is effectively suppressed, reducing energy loss. At the same time, the deep learning-based gear prediction model can accurately predict the optimal shifting time, making the shifting process smoother and more natural. This not only improves energy efficiency but also significantly enhances the driving experience.

[0021] This invention adopts an offline optimization + online application approach. The global optimization calculation is completed in the offline stage, and a lightweight neural network model is deployed in the online stage. This not only retains the scientific nature of the global optimization algorithm, but also perfectly solves the problems of high computational complexity and poor real-time performance of traditional optimization algorithms. It enables the system to achieve the best shift control effect while meeting the real-time requirements of the vehicle controller, providing a reliable technical guarantee for the intelligent control of heavy-duty electric commercial vehicles. Attached Figure Description

[0022] Figure 1 This is a schematic diagram of the offline shift control strategy of the present invention.

[0023] Figure 2 This is a schematic diagram of the online gear selection of the present invention.

[0024] Figure 3 This is a schematic diagram of the overall process of the present invention.

[0025] Figure 4 This is a schematic diagram of the motor under actual operating conditions according to the present invention.

[0026] Figure 5 This is a schematic diagram of the vehicle's energy consumption under actual operating conditions according to the present invention.

[0027] Figure 6 This is a schematic diagram illustrating the SOC variation under actual operating conditions of the present invention.

[0028] Figure 7 This is a schematic diagram of adaptive gear changes under actual working conditions.

[0029] Figure 8This is a schematic diagram showing the gear changes based on motor efficiency under actual operating conditions. Detailed Implementation

[0030] The present invention will now be described in further detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and not intended to limit it. Furthermore, it should be noted that, for ease of description, the accompanying drawings show only the parts relevant to the present invention, and not all of the structures.

[0031] See Figure 3 This invention proposes an online adaptive shifting control method for heavy-duty electric commercial vehicles, comprising the following steps: S1. Construct an offline optimized shift control strategy based on dynamic programming algorithm; the offline optimized shift control strategy aims to minimize the energy consumption of the whole vehicle, introduces a gear shifting penalty term, takes the gear shifting action as the control variable, and takes the battery SOC and the current gear as the state variables to solve for the optimal gear, and combines the optimal gear with the vehicle operating conditions to generate the optimal gear selection rule table for electric commercial vehicles. S2. Input the optimal gear selection rule table into the pre-built gear selection model for training to obtain the trained gear selection model; S3. Deploy the trained gear selection model to the vehicle controller, which then outputs precise gear selection commands in real time to adapt to dynamic operating conditions.

[0032] The present invention will be further explained below with reference to the accompanying drawings and specific embodiments: First, to meet the control accuracy requirements in energy management, the driving condition is divided into several parts based on time, with each part having a step size of 1 second. The battery state of charge value and the current gear are selected at any time k (k=0, 1, 2, 3, ..., N) as the state vector, and the gear shifting operation is set as the control variable.

[0033] The discrete form of the state equation can be expressed as:

[0034]

[0035] In the formula: x(k+1) and x(k) are the values ​​of the state vector at time k+1 and time k, respectively; u(k) is the value of the control vector at time k; This indicates the current gear position.

[0036] The gear shifting of the AMT is selected as the decision variable in the dynamic programming, as shown in the following equation:

[0037] In the formula: u(k) is the decision variable for the k-th stage, set as downshift, no shift, upshift, represented by 1, 0, and -1 respectively.

[0038] This decision variable can be linked to the gear position, that is:

[0039] Construct the objective function for dynamic programming, whose expression is as follows:

[0040]

[0041]

[0042]

[0043] Where: Where: The change in SOC; The equivalent energy consumption for shift penalty; The instantaneous cost function; I k It is the instantaneous current; Q b Battery capacity; λ This is the weighting coefficient; the larger the weighting coefficient, the greater the cost of shifting gears.

[0044] Set the constraints as follows: To improve optimization efficiency, constraints were applied to the shift speeds during the optimization process, as shown in the following expressions:

[0045] In the formula: T m min and T m max These represent the minimum and maximum torques of the motor at the current moment, respectively; n m min and n m max These represent the minimum and maximum speeds of the motor at the current moment; SOC min and SOC max These are the minimum and maximum values ​​that SOC can reach, respectively; P bat max and P bat max These represent the lower and upper limits of battery charging and discharging power, respectively.

[0046] Calculating the vehicle's actual energy consumption: Since the shift penalty is an additional judgment and does not actually exist, this energy loss needs to be subtracted. Therefore, the vehicle's actual energy consumption can be expressed as:

[0047] In the formula: E allThis represents the vehicle's total energy consumption.

[0048] Based on the WTVC (Wheel Threatened Variable Valve) operating condition, this study investigates shift control for heavy-duty electric commercial vehicles using both dynamic programming and motor efficiency-based shift strategies. With the goal of minimizing overall vehicle energy consumption, and incorporating a slope-adaptive penalty term, a global optimization solution is performed to generate an optimal gear mapping rule table covering all operating conditions. This table serves as the dataset for training the next neural network model. The specific process is as follows: Figure 1 As shown.

[0049] Secondly, the specific process of S2 is as follows: Model Construction: The calculation formulas for the input gate, forget gate, and output gate of the Long Short-Term Memory (LSTM) network are as follows:

[0050] In the formula: σ It is a sigmoid activation function; W f , b f Here are the weight matrix and bias vector for the forget gate; f t for t The output vector of the forget gate at each time step; i t for t The output vector of the input gate at any given time; W i This is the weight matrix of the input gate; b i The bias vector for the input gate; c t -1 and c t for t- 1. Time and t The information stored in the state unit at any given time; O t为 Output value of the output gate; h t for t The output value of the state unit at that moment; w o and b o These are the weight matrix and bias vector of the output gate.

[0051] The computational model of the Bidirectional Long Short-Term Memory (BiLSTM) neural network is as follows:

[0052] In the formula: The hidden layer state of the forward LSTM at time t; The hidden layer state of the inverse LSTM at time t; Forward LSTM unit; It is an inverse LSTM unit; x t This is the input at time t; The hidden layer state of the forward LSTM at time t-1; The hidden layer state of the inverse LSTM at time t+1; H t Let t represent the hidden layer state of the BiLSTM.

[0053] CNN network feature extraction output:

[0054] In the formula: Y represents the extracted features; is the activation function; W is the weight matrix; X is the time series; b is the bias vector.

[0055] In this invention, the accuracy of the prediction is calculated by the degree of similarity between the predicted gear and the original gear at each moment. The calculation formula is as follows:

[0056] In the formula: E is the accuracy of gear prediction; m is the number of predicted gears that are equal to the original gears; N is the total number of gears.

[0057] Training of the gear selection model: This invention addresses the complex nonlinear mapping relationship between vehicle speed, acceleration, gradient, wheel-end torque, and SOC by constructing a CNN-BiLSTM-based gear selection model. The optimal gear selection rule table obtained in the above steps is used as the training set and input into the model for training. The model integrates vehicle dynamic characteristics (torque, vehicle speed, acceleration) and environmental characteristics (gradient). It extracts local spatial features (such as sudden torque changes and steep gradient changes) through CNN layers and uses BiLSTM layers to learn time dependencies bidirectionally: the forward LSTM captures historical states, and the backward LSTM learns future influences, jointly modeling long-term dynamics (such as energy consumption during continuous uphill climbing). The fully connected layer integrates spatiotemporal features and outputs target parameters (such as gear / SOC).

[0058] Experimental results show that, compared to a single LSTM model, this model significantly improves prediction accuracy under sudden changes in operating conditions (rapid acceleration, steep slopes) by coupling local patterns and long-term dependencies, overcoming the bottleneck of traditional models in expressing strongly nonlinear relationships. The specific process is as follows: Figure 2 As shown.

[0059] The specific optimization process of S3 in this embodiment is as follows: An online adaptive shift control method is established, and the specific process is as follows: Figure 3As shown, the online adaptive shift control method for heavy-duty electric commercial vehicles that considers road gradient and shift frequency is based on the "offline optimization + online application" approach, including dynamic programming shift strategy and CNN-BiLSTM gear selection strategy.

[0060] In the offline optimization phase, the required torque is calculated based on vehicle operating conditions (mass, speed, acceleration, and gradient). With the goal of minimizing overall vehicle energy consumption, a gradient-adaptive penalty function is introduced. A dynamic programming algorithm (control variables: gear shifting action; state variables: battery SOC and current gear) generates an optimal gear rule table covering all operating conditions. The results show that the number of gear shifts based on dynamic programming is significantly less than that based on motor efficiency optimization, with 109 and 298 shifts respectively. The dynamic programming control strategy reduces the gear shifting frequency by 60.2%, significantly lowering the overall frequency. Under both control strategies, the SOC decreases along the same trend, exhibiting fluctuations during the decrease. The operating regions of the motor MAP are also largely the same, mostly operating in the high-efficiency region (efficiency > 80%). The final SOC value for the dynamic programming strategy is 0.725, with an energy consumption of 324 MJ; the final SOC value for the motor efficiency optimization strategy is 0.715, with an energy consumption of 342 MJ. Under the dynamic programming shifting strategy, energy consumption was reduced by 18 MJ, and energy efficiency was reduced by 5.26%. The dynamic programming shifting strategy reduces energy consumption while decreasing the shifting frequency.

[0061] In the online decision-making stage, vehicle speed, acceleration, gradient, and wheel-end torque are used as inputs. A CNN-BiLSTM hybrid model (convolutional layers extracting local features + bidirectional LSTM capturing temporal dependencies) is used to train the gear selection model. In the real-time control stage, the model is embedded in the vehicle controller, and gear commands are output online to the AMT actuator. In the actual road conditions used, the number of gears predicted by the CNN-BiLSTM model that are not equal to the original gears is 102, and the gear prediction accuracy reaches 97.9%. The number of gears predicted by the LSTM model that are not equal to the original gears is 102, and the gear prediction accuracy is 94.1%. The gear prediction accuracy of the CNN-BiLSTM model is 3.8% higher than that of the LSTM model. To further analyze the gear selection process, the number of predicted gears for each gear is counted. As can be seen from the figure, the number of predicted gears for 1st and 4th gears is relatively high, while the number of predicted gears for 2nd and 3rd gears is relatively low. This is because lower gears prioritize power, while higher gears prioritize economy.

[0062] Online Control Result Analysis: The trained CNN-BiLSTM neural network model was loaded into the vehicle controller for online application, and a shift control strategy based on motor efficiency optimization was used as the benchmark to verify the performance of the control method. A specific measured operating condition was used as the operating condition for the online control application, which differs from the operating condition for the offline control application. Based on the measured driving conditions, the online adaptive shift control method and the shift control strategy based on motor efficiency optimization proposed in this study were simulated and verified. The motor MAP diagrams under the two control strategies are shown below. Figure 4 As shown; the change in the SOC value of the power battery is as follows Figure 6 As shown; the gear changes are as follows Figure 7 , 8 As shown.

[0063] Depend on Figure 7 , 8 It can be seen that the number of gear shifts in the online adaptive shift control method is less than that in the optimal motor efficiency-based method, with 64 and 156 gear shifts respectively. The gear shift frequency is reduced by 58.97% under the online adaptive shift control method. Figure 4 , 6 It can be seen that under both control strategies, the downward trend of SOC is the same, with fluctuations and increases occurring during the decline. The operating regions of the motor MAP are also basically the same, mostly operating in the high-efficiency region. The final SOC value under the dynamic programming strategy is 0.758, with an energy consumption of 168 MJ, while the final SOC value under the motor efficiency optimization strategy is 0.756, with an energy consumption of 172 MJ, representing a 2.3% reduction in energy consumption. When applied online, the online adaptive shift control method effectively suppresses the problem of cyclic shifting in heavy-duty electric commercial vehicles and reduces the overall vehicle energy consumption. The energy losses of major components are as follows: Figure 5 As shown. Component energy loss refers to the energy consumption caused by the component's own efficiency or transmission efficiency not reaching 100%. The energy loss of the online adaptive shift control method is ultimately 14.75 MJ, while the energy loss of the shift strategy based on motor efficiency optimization is ultimately 18.73 MJ. Compared with the latter, the energy loss rate of the former is reduced by 21.25%.

[0064] Based on the same inventive concept, this invention also provides an online adaptive shifting control system for heavy-duty electric commercial vehicles. Since the principle of solving the problem by this online adaptive shifting control system for heavy-duty electric commercial vehicles is similar to that of the aforementioned online adaptive shifting control method for heavy-duty electric commercial vehicles, the implementation of this online adaptive shifting control system for heavy-duty electric commercial vehicles can refer to the implementation of the online adaptive shifting control method for heavy-duty electric commercial vehicles, and the repeated parts will not be described again.

[0065] In specific implementation, the online adaptive shifting control system for heavy-duty electric commercial vehicles provided in this embodiment of the invention specifically includes: The offline optimization module is used to construct an offline optimized shift control strategy based on a dynamic programming algorithm. The offline optimized shift control strategy aims to minimize the energy consumption of the entire vehicle, introduces a gear shifting penalty term, uses the gear shifting action as a control variable, and uses the battery SOC and the current gear as state variables to solve for the optimal gear. The optimal gear is then combined with the vehicle operating conditions to generate an optimal gear selection rule table for electric commercial vehicles. The training module is used to input the optimal gear selection rule table into the pre-built gear selection model for training, so as to obtain the trained gear selection model. The online control module is used to deploy the trained gear selection model to the vehicle controller, which then outputs precise gear commands in real time to adapt to dynamic operating conditions.

[0066] The system also includes a perception module, which includes a speed sensor, gyroscope, GPS, lidar, and binocular camera, used to collect vehicle speed, acceleration, and road slope signals and transmit them to the online control module.

[0067] Accordingly, this embodiment of the invention also provides an online adaptive shifting control device for heavy-duty electric commercial vehicles, including a processor and a memory, wherein the processor executes a computer program stored in the memory to implement the online adaptive shifting control method for heavy-duty electric commercial vehicles as provided in this embodiment of the invention.

[0068] For more detailed information on the above methods, please refer to the relevant content disclosed in the foregoing embodiments, which will not be repeated here.

[0069] Accordingly, embodiments of the present invention also provide a computer-readable storage medium for storing a computer program, wherein the computer program, when executed by a processor, implements the above-described online adaptive gear shifting control method for heavy-duty electric commercial vehicles provided in embodiments of the present invention.

[0070] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the systems, devices, and storage media disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the descriptions are relatively simple; relevant parts can be referred to the method section.

[0071] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0072] The steps of the methods or algorithms described in conjunction with the embodiments disclosed herein can be implemented directly by hardware, a software module executed by a processor, or a combination of both. The software module can be located in random access memory (RAM), main memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, removable disk, CD-ROM, or any other form of storage medium known in the art.

[0073] Finally, it should be noted that in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0074] The above provides a detailed description of the online adaptive shifting control method, system, device, and storage medium for heavy-duty electric commercial vehicles provided by this invention. Specific examples have been used to illustrate the principles and implementation methods of this invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of this invention. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this invention. Therefore, the content of this specification should not be construed as a limitation of this invention.

Claims

1. A method for online adaptive gear shifting control of a heavy-duty electric commercial vehicle, characterized in that, Includes the following steps: S1. Construct an offline optimized shift control strategy based on dynamic programming algorithm; the offline optimized shift control strategy aims to minimize the energy consumption of the whole vehicle, introduces a gear shifting penalty term, takes the gear shifting action as the control variable, and takes the battery SOC and the current gear as the state variables to solve for the optimal gear, and combines the optimal gear with the vehicle operating conditions to generate the optimal gear selection rule table for electric commercial vehicles. S2. Input the optimal gear selection rule table into the pre-built gear selection model for training to obtain the trained gear selection model; S3. Deploy the trained gear selection model to the vehicle controller, which then outputs precise gear selection commands in real time to adapt to dynamic operating conditions.

2. The online adaptive shifting control method for heavy-duty electric commercial vehicles according to claim 1, characterized in that, The process of S1 is as follows: Obtain vehicle mass, speed, acceleration, and road gradient as required by driving conditions; The driving conditions are divided into several parts based on time. The battery state of charge value and the current gear are selected at any time k as the state vector, and the gear shifting operation is set as the control variable. Based on the state variables and control variables, a target function is constructed, as shown in the following equation: In the formula: The change in SOC; The equivalent energy consumption for shift penalty; The instantaneous cost function; Let be the decision variable for the k-th stage, denoted as: downshift, no shift, upshift, represented by 1, 0, and -1 respectively; Set constraints on the objective function; Based on the vehicle's operating conditions, the objective inner function is solved, and the optimal gear selection rule table is generated with the goal of minimizing the overall vehicle energy consumption.

3. The online adaptive shifting control method for heavy-duty electric commercial vehicles according to claim 1, characterized in that, The constraint conditions are shown in the following equation: In the formula: T m min and T m max These represent the minimum and maximum torques of the motor at the current moment, respectively; n m min and n m max These are the minimum and maximum speeds of the motor at the current moment, respectively. SOC min and SOC max These are the minimum and maximum values ​​that SOC can reach, respectively; P bat max and P bat max These represent the lower and upper limits of battery charging and discharging power, respectively.

4. The online adaptive shifting control method for heavy-duty electric commercial vehicles according to claim 2, characterized in that, The shift penalty is shown in the following formula: In the formula: λ is the weighting coefficient. The larger the weighting coefficient, the greater the cost of shifting gears.

5. The online adaptive shifting control method for heavy-duty electric commercial vehicles according to claim 1, characterized in that, The gear selection model is built on CNN-BiLSTM and includes convolutional layers, bidirectional long short-term memory layers, and fully connected layers. The convolutional layer is used to extract local nonlinear features of the input parameters; The bidirectional long short-term memory layer learns the long-term dependencies of time series through forward and backward LSTM layers; the fully connected layer is used to integrate features and output gear prediction results.

6. The online adaptive shifting control method for heavy-duty electric commercial vehicles according to claim 1, characterized in that, The process in S3 is as follows: The vehicle controller acquires vehicle speed, acceleration, road gradient, and wheel-end torque in real time, inputs them into the gear selection model, and outputs real-time gear control commands.

7. An online adaptive shifting control system for heavy-duty electric commercial vehicles, characterized in that, include: The offline optimization module is used to construct offline optimized shift control strategies based on dynamic programming algorithms. The offline optimized shift control strategy aims to minimize the energy consumption of the entire vehicle. It introduces a shift penalty term, uses the shift action as a control variable, and uses the battery SOC and the current gear as state variables to solve for the optimal gear. The optimal gear is then combined with the vehicle operating conditions to generate an optimal gear selection rule table for electric commercial vehicles. The training module is used to input the optimal gear selection rule table into the pre-built gear selection model for training, so as to obtain the trained gear selection model. The online control module is used to deploy the trained gear selection model to the vehicle controller, which then outputs precise gear commands in real time to adapt to dynamic operating conditions.

8. The online adaptive shifting control system for heavy-duty electric commercial vehicles according to claim 7, characterized in that, It also includes a perception module, which includes a speed sensor, gyroscope, GPS, lidar and binocular camera, used to collect vehicle speed, acceleration and road slope signals and transmit them to the online control module.

9. A heavy-duty electric commercial vehicle online adaptive shifting control device, characterized in that, It includes a processor and a memory, wherein the processor executes a computer program stored in the memory to implement the online adaptive shift control method for heavy-duty electric commercial vehicles as described in any one of claims 1 to 6.

10. A computer-readable storage medium, characterized in that, Used to store a computer program, wherein the computer program, when executed by a processor, implements the online adaptive shift control method for heavy-duty electric commercial vehicles as described in any one of claims 1 to 6.

Images