Intelligent control method for automatic transmission of commercial vehicle based on VCU architecture

Abstract

The invention relates to an intelligent control method for an automatic transmission of a commercial vehicle based on a VCU (vehicle control unit) architecture, which comprises the following steps of: based on a rule-based control mode, determining an operation mode of an AMT (automated mechanical transmission) of the vehicle according to an optimal principle according to a vehicle operation condition, a road condition and an intention of a driver, and performing intelligent switching control on economical and dynamic gear shifting rules of the transmission by the VCU, and the economic mode is selected under the condition that the accelerator opening degree is smaller than or equal to a preset threshold value under the light load or the small road slope or the uphill working condition. And selecting a power mode under the conditions of heavy load, relatively large gradient and greater accelerator opening degree than a preset value under the uphill working condition. Intelligent switching control is carried out on economical efficiency and dynamic property gear shifting rules of the transmission, it is guaranteed that an engine works in a proper rotating speed interval, and the climbing dynamic property is improved on the premise that economical efficiency is guaranteed. And meanwhile, the VCU improves the feeling of a driver on the fuel economy and the dynamic property of the vehicle through the multi-stage accelerator Map switching control method.

Description

Technical Field

[0001] The present invention belongs to the technical field of automatic transmissions and relates to an intelligent control method for a commercial vehicle automatic transmission based on a VCU architecture. Background Art

[0002] With the rapid development of automotive electronic control technology, traditional manual transmissions can no longer meet users' demand for advanced automotive technology products. In view of the complex driving conditions and harsh environments of commercial vehicles, most OEMs have chosen to develop cost-effective, energy-saving and fuel-efficient AMT transmissions to adapt to different road conditions and meet customers' economic and power needs. According to the real-time status of the vehicle, road conditions and the driver's intention (throttle, engine speed, vehicle load, etc.), the throttle map with different characteristics is intelligently switched, the vehicle transmission operating mode is judged, and the target operating mode of the automatic transmission (AMT) is determined by the vehicle control unit (VCU). The transmission controller (TCU) receives and executes the E / P mode switching request (economic mode E, power mode P) issued by the VCU to meet the driver's needs for vehicle fuel economy, power, etc. The existing intelligent control solutions for commercial vehicle automatic transmissions are as follows: 1. Commercial vehicle multi-level throttle switching control solution: 1) Commercial vehicles use a power-saving switch to control the switching of different throttle maps. Drivers manually switch between these modes, but this has issues such as low utilization of manual control. 2) Commercial vehicles, based on the VCU architecture, only have one throttle map, which does not distinguish between vehicle operating status and driving intent, failing to achieve optimal fuel economy and power for the vehicle. 2. Automatic transmission operating mode switching control scheme: 1) The E / P switching of commercial vehicle AMT automatic transmissions is manually operated by the driver. There are cases where the mode is not switched, resulting in user complaints such as insufficient climbing power and low manual control utilization; 2) Commercial vehicle AMT automatic transmissions use the vehicle's onboard electronic map module (e-Horizon) to identify upcoming slopes in advance. The controller uses an algorithm to automatically switch the transmission from economy mode to power mode to meet the climbing dynamics requirements. However, this method requires an electronic map module, which is relatively expensive. To date, there is no low-cost control technology that can automatically switch control based on vehicle status, driver intentions, and road conditions. Summary of the Invention

[0003] The present invention provides a method for intelligent switching control of the operating mode of an automatic transmission for commercial vehicles. Based on the current vehicle state, road conditions, and driver intention, the method intelligently switches and controls the vehicle's throttle map and the transmission's shifting schedule for economy and power, ensuring that the engine operates within an appropriate speed range and improving power while ensuring economy. The control method of the present invention is integrated into the VCU without increasing the cost of the entire vehicle.

[0004] The present invention is achieved by adopting the following technical solutions: A method for intelligent control of an automatic transmission for a commercial vehicle includes a rule-based control mode. The rule-based control mode performs throttle map switching and transmission operating mode decision-making based on the current vehicle state, current vehicle load, road slope, and driver intention. The VCU intelligently switches and controls the transmission's economic and power mode shifting patterns. The VCU controls and switches the throttle map based on the current gear position. The VCU makes operating mode decisions and controls switching based on the current road slope classification and accelerator pedal opening. The method ensures that the engine operates in an economical speed range as much as possible and always operates in an efficient engine operating condition, thereby improving power while ensuring economy.

[0005] As a further illustration of the present invention: When the load is light, the road slope is small, or the throttle opening is less than or equal to the preset threshold in uphill conditions, the economy mode is selected and sent to the transmission TCU; when the load is heavy, the slope is large, and the throttle opening is greater than the preset threshold in uphill conditions, the power mode is selected and sent to the transmission TCU; the transmission TCU selects the economy shift schedule according to the E-mode status signal sent by the VCU; the transmission TCU selects the power shift schedule according to the P-mode status signal sent by the VCU; Preferably, when the load is light, or the road is flat, or when going up a gentle slope and the throttle opening is less than or equal to the throttle opening set for the gentle slope, or when going up a steep slope and the throttle opening is less than or equal to the throttle opening set for the steep slope, the economic mode is selected and sent to the transmission TCU; when the load is heavy and the slope is gentle and the throttle opening is greater than the throttle opening set for the gentle slope, or when the load is heavy and the slope is steep and the throttle opening is greater than the throttle opening set for the steep slope, and all three conditions are met at the same time, the power mode is selected and sent to the transmission TCU.

[0006] As a further illustration of the present invention: Set different throttle thresholds APP for steep slopes and gentle slopes, where the throttle threshold for steep slopes is APP1 and the throttle threshold for gentle slopes is APP2; Because climbing a gentle slope requires less power, when the vehicle is heavily loaded and on a gentle slope and the throttle opening is greater than APP2, it switches to power mode. When the vehicle is heavily loaded and on a gentle slope and the throttle opening is less than or equal to APP2, it switches to economy mode. Because climbing steep slopes requires a lot of power, when the vehicle is heavily loaded and the slope is steep and the throttle opening is greater than APP1, it switches to power mode. When the vehicle is heavily loaded and the slope is steep and the throttle opening is less than or equal to APP1, it switches to economy mode. Set the throttle threshold APP1 for steep slopes to 45% and the throttle threshold APP2 for gentle slopes to 65%.

[0007] As a further explanation of the present invention: the VCU controls the throttle map according to the current gear position, including: Based on the VCU's assessment of driving intent, different throttle maps are selected based on the operating conditions of different gears. Gear state 1 switches to throttle Map 1, gear state 2 switches to throttle Map 2, and gear state 3 switches to throttle Map 3, giving consideration to both power and economy, thus improving vehicle driving performance. Through big data statistical analysis of the gear distribution of currently running vehicles, the gear ranges are divided according to different gearbox models: gear state 1 is the low gear range, gear state 2 is the middle gear range, and gear state 3 is the high gear range; Throttle Map status division: Map1 is the low-gear area for starting and mountain climbing; Map2 is the medium-gear area for high-speed climbing, taking into account both power and economy; Map3 is the high-gear area for high-speed cruising conditions to meet economy.

[0008] As a further illustration of the present invention: When the actual gear position of the transmission is higher than the low-range threshold, the throttle MAP state 1 in the low-range is switched to the throttle MAP state 2 in the middle-range, and the torque control is smoothed; When the actual gear position of the transmission is higher than the threshold of the middle range, the throttle MAP state 2 in the middle range is switched to the throttle MAP state 3 in the high range, and the torque control is smoothed; When the actual gear position of the transmission is lower than the middle gear threshold, the throttle MAP state 2 in the middle gear is switched to the throttle MAP state 1 in the low gear, and the torque control is smoothed; When the actual gear position of the transmission is lower than the high gear threshold, the throttle MAP state 3 in the high gear is switched to the throttle MAP state 2 in the middle gear, and the torque control is smoothed.

[0009] As a further illustration of the present invention: The VCU makes decisions and controls switching of operating modes based on the current road gradient classification state and the accelerator pedal opening state, including: making decisions and controlling switching of transmission operating modes based on the current vehicle load, the current road gradient classification state, and the accelerator pedal opening state; Road slope classification status: Road slope state 1: flat road; Road slope state 2: gentle slope; Road slope state 3: steep slope; Road slope classification status judgment jump principle: When the slope is greater than S1 and remains greater than the time threshold, the road slope state 1 jumps to the road slope state 2; When the slope is less than or equal to S1 and remains greater than the time threshold, the road slope state 2 jumps to the road slope state 1; When the slope is greater than S2 and remains greater than the time threshold, the road slope state 2 jumps to the road slope state 3; When the slope is less than or equal to S2 and remains greater than the time threshold, the road slope state 3 is changed to the road slope state 2; When the slope is greater than S2 and remains greater than the time threshold, the road slope state 1 jumps to the road slope state 3; Throttle opening state setting: The throttle opening state 1 is greater than the throttle opening APP2 on a gentle slope; Throttle opening state 2 is less than or equal to throttle opening APP2 on a gentle slope; The throttle opening state 3 is greater than the throttle opening APP1 on a steep slope; Throttle opening state 4 is less than or equal to the throttle opening APP1 on a steep slope; Among them, S1 is 3%, S2 is 7%, and the slope unit is percentage; less than S1 is a flat road, greater than S2 is a steep slope, and between S1-S2 is a gentle slope.

[0010] As a further explanation of the present invention: the intelligent control method is characterized in that it specifically includes the following steps: After the transmission E / P intelligent switching is activated, the default mode is economic mode E, which determines the vehicle load threshold, road slope threshold, and accelerator pedal opening threshold in real time; The conditions for switching the transmission operating mode from economy mode E to power mode P are as follows: 1) When the vehicle load is greater than the set threshold m, the time maintained in road slope state 2 is greater than the set time threshold, and the time maintained in throttle opening state 1 is greater than the set time threshold, the transmission operating mode switches from economy mode E to power mode P; 2) When the vehicle load is greater than the set threshold m, the time maintained in road slope state 3 is greater than the set time threshold, and the time maintained in throttle opening state 3 is greater than the set time threshold, the transmission operating mode switches from economy mode E to power mode P; The conditions for switching the transmission operating mode from power mode P to economy mode E are as follows: 1) When the vehicle load is less than or equal to the set threshold m, the power mode P is switched to the economic mode E; 2) When the vehicle load is greater than the set threshold m and the time it maintains in road slope state 1 is greater than the set time threshold, the power mode P is switched to the economy mode E; 3) When the vehicle load is greater than the set threshold m, the time maintained in road slope state 2 is greater than the set time threshold, and the time maintained in throttle opening state 2 is greater than the set time threshold, the power mode P is switched to the economy mode E; 4) When the vehicle load is greater than the set threshold m, the time maintained in road slope state 3 is greater than the set time threshold, and the time maintained in throttle opening state 4 is greater than the set time threshold, the power mode P is switched to the economy mode E.

[0011] As a further illustration of the present invention: The intelligent control method also includes a control mode based on an artificial intelligence algorithm. Each time the vehicle is powered on and started, it first enters a rule-based control mode. After the vehicle has accumulated 200 kilometers of driving after starting, it automatically switches to a control mode based on an artificial intelligence algorithm. The control mode based on the artificial intelligence algorithm identifies driving style based on past vehicle operation data, performs data cluster analysis based on vehicle operation data, clusters and samples driving conditions and formulates driving cycles, and classifies the road type and driving behavior of the vehicle. The road type is identified through the distribution of vehicle speed, slope, and gear operation. The driving behavior extracts feature vectors including maximum vehicle speed, average vehicle speed, maximum lateral acceleration, and maximum acceleration pedal change rate information, which are standardized. The classification tree classification algorithm is trained to derive three driving style models: aggressive, general, and moderate. The aggressive driving style model correspondingly switches to the power mode, and the general or moderate driving style models correspondingly switch to the economy mode.

[0012] As a further illustration of the present invention: The control mode based on artificial intelligence algorithm specifically includes the following steps: S1, sample preparation: First, a data collection route covering different road levels, including urban, highway, national highway, mountainous, and hilly, was developed as a driving cycle sample. The idling, acceleration, deceleration, and uniform speed distribution and ratios of the driving cycle sample were set, and the road type was identified based on the distribution of vehicle speed, slope, and gear operation. Next is driver selection. Three types of drivers, aggressive, moderate, and average, are selected. Driving data from these drivers on the specified driving cycle samples is collected. The driving data is cleaned and sliced. After data processing, a comprehensive score is given to the drivers' driving behaviors within the driving cycle samples. S2, feature engineering: First, the mean, variance, and maximum mathematical features of samples in different scenarios are extracted as basic features and standardized. In terms of feature selection, low-variance features are first removed. Then, a two-sample t-test is performed to select features with significant differences in different comprehensive driving styles. Finally, redundant features are removed through correlation analysis. The collected driving data is divided into multiple separate small samples to facilitate cluster analysis in step S3. S3, cluster analysis: The driving style recognition problem is considered as an unsupervised learning problem. First, the collected driving data is analyzed offline. The sample data is clustered using the K-means clustering method without PCA. The clustering result is used as a sample label, and the corresponding driving style label is added based on the specific result. S4, classification algorithm development: After labeling the samples through cluster analysis, the labeled samples were used to train the classification algorithm. A generalized linear classifier (GLCL) was selected to classify the data using a supervised learning approach using a support vector machine (SVM). Features related to vehicle speed, lateral and longitudinal acceleration, acceleration, and accelerator pedal change rate were input into the classification algorithm. Algorithm training was implemented using an m-file script. The distance from a sample point to the near-overclocked operation itself reflects the aggressiveness of the sample point. Driving style is no longer classified into three isolated categories: aggressive, moderate, and normal. Instead, it is linearly scored for aggressiveness. Aggressive samples have a higher distance from the aggressive hyperplane than normal samples, and normal samples have a higher distance from the aggressive hyperplane than moderate samples. After algorithm training, the trained algorithm was integrated into Simulink using a MATLAB function module to implement algorithm modeling. After the model was built, C++ code was generated from the model using Embedded Coder to create a driving style recognition app. The app ultimately calculates and outputs the three driving style labels: aggressive, moderate, and normal.

[0013] As a further explanation of the present invention: Preferably, in step S1, Data cleaning includes noise smoothing, outlier detection, and missing value processing to ensure data quality. Noise smoothing removes high-frequency noise through simple sliding window averaging and low-pass filtering. Outlier detection sets a threshold based on vehicle physics knowledge and combines multiple sensor data to determine rationality. Missing value processing is divided into small missing segments and large missing segments. Small missing segments are processed using linear interpolation or forward filling, while large missing segments are directly removed from the affected time period. Different slicing rules are designed for different scenarios. There are three methods for data slicing. The first method uses a fixed time window as a segment, which is suitable for steady-state analysis. The second method is event-driven, which defines rules for sudden acceleration, sudden braking, and sharp turns, and extracts 5 seconds of data before and after the event. The speed of high-speed driving is higher and changes less, while when driving on national highways, there are frequent starts and stops, and sudden acceleration and sudden braking are more common. The third method is driving state segmentation, which is based on the idle, acceleration, deceleration, and constant speed states set in the driving cycle, and is segmented based on state change points.

[0014] Compared with the prior art, the present invention has the following beneficial effects: 1) The present invention's intelligent switching control method for commercial vehicle automatic transmission operating modes includes a rule-based control mode that identifies and predicts driver intent based on the current vehicle state, road conditions, and past vehicle data. It then intelligently switches and controls the vehicle's throttle map and the transmission's shifting schedule for economy and power, ensuring that the engine operates within an economical speed range as much as possible and always operates in an efficient engine operating condition, thereby improving power while ensuring economy.

[0015] 2) The intelligent switching control method for the automatic transmission operating mode of a commercial vehicle of the present invention also includes a control mode based on an artificial intelligence algorithm. Each time the vehicle is powered on and started, it first enters a rule-based control mode. After the vehicle has accumulated 200 kilometers of driving after starting, it automatically switches to a control mode based on an artificial intelligence algorithm. The control mode based on the artificial intelligence algorithm is that after the driving style recognition APP completes learning using vehicle operation data, the APP issues a driving style label. An aggressive driving style label corresponds to switching to power mode P, and a general or moderate driving style label corresponds to switching to economy mode E, thereby realizing adaptive control of the transmission operating mode. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 This is a diagram of the VCU-based transmission intelligent switching control architecture of the present invention (rule-based control mode); Figure 2 This is a schematic diagram of the multi-stage throttle switching control based on the gear range; Figure 3 This is a schematic diagram of the jump principle for judging the road slope classification state of the present invention; Figure 4 This is a flowchart of the intelligent E / P mode switching process of the present invention; Figure 5 A flow chart is developed for the driving style recognition algorithm in the control mode based on the artificial intelligence algorithm of the present invention. DETAILED DESCRIPTION

[0017] In order to enable those skilled in the art to better understand the technical solutions of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts should fall within the scope of protection of the present invention.

[0018] Definitions of Abbreviations and Key Terms: VCU: Vehicle Control System; TCU: Transmission Control System; ECU: Engine Control System; Throttle MAP, also known as pedal map, is also called accelerator pedal characteristics. It is a corresponding curve graph between pedal depth, engine speed and engine power.

[0019] Example 1 The present invention provides an intelligent control method for a commercial vehicle automatic transmission, wherein the intelligent control method includes a rule-based control mode, such as Figure 1 As shown in the VCU-based transmission intelligent switching control architecture diagram (rule-based control mode), the vehicle control system VCU selects different throttle maps in real time according to the vehicle operating conditions (real-time collection of engine speed, throttle opening, automatic / manual transmission mode status, vehicle load, transmission gear status, etc.), road conditions and driver's intentions, and determines the vehicle AMT transmission operating mode according to the optimal principle.

[0020] The VCU switches the throttle map and determines the transmission operating mode based on the current vehicle status, road grade, and driver intent. The VCU intelligently controls the automatic transmission's shift schedule for economic and power modes, and switches between different throttle maps based on the current gear. The VCU makes operating mode decisions and switches based on the current road grade classification and accelerator pedal position, ensuring the engine operates within the most economical speed range possible, maintaining high efficiency and improving power without sacrificing economy. Engine speed is a reference value used to determine vehicle start, and the economical engine speed range is controllable between 1000-1700 rpm.

[0021] Based on vehicle weight, road slope, and throttle opening, the VCU intelligently switches between economical and dynamic transmission shift patterns. In light-load situations, on a low road slope, or with a throttle opening less than a preset value, the VCU selects the economic mode and sends it to the TCU. In heavy-load situations, with a steep slope and a large throttle opening, the TCU selects the dynamic mode and sends it to the TCU, ensuring vehicle dynamics. The transmission TCU selects the economical shift pattern based on the E-mode status signal sent by the VCU; it selects the dynamic shift pattern based on the P-mode status signal sent by the VCU.

[0022] The throttle map switching and transmission operating mode decisions are made based on the current vehicle status and driver intention. Parameters such as the road slope threshold S for E / P mode switching, the vehicle load threshold m, and the throttle opening threshold APP need to be finely calibrated according to the vehicle to meet the vehicle performance requirements of different market segments.

[0023] like Figure 2 The principle diagram of multi-stage throttle switching control based on gear range is shown as follows: Through big data statistical analysis of the gear distribution of currently running vehicles, the gear ranges are divided according to different gearbox models, and the following settings are made: gear state 1 is the low gear area, gear state 2 is the middle gear area, and gear state 3 is the high gear area.

[0024] Throttle Map status setting: Throttle Map 1 is the low-gear area to cope with starting and mountain climbing; Throttle Map 2 is the medium-gear area to cope with high-speed climbing, taking into account power on the premise of economy; Throttle Map 3 is the high-gear area to cope with high-speed cruising conditions to meet economy.

[0025] When the actual gear position of the transmission is higher than the low-range threshold, the throttle MAP state 1 in the low-range jumps to the throttle MAP state 2 in the middle-range; When the actual gear position of the transmission is higher than the threshold of the middle range, the throttle MAP state 2 in the middle range jumps to the throttle MAP state 3 in the high range; When the actual gear position of the transmission is lower than the middle gear threshold, the throttle MAP state 2 in the middle gear area jumps to the throttle MAP state 1 in the low gear area; When the actual gear position of the transmission is lower than the high gear threshold, the throttle MAP state 3 in the high gear area jumps to the throttle MAP state 2 in the middle gear area.

[0026] Under the premise of judging driving intention through VCU, the present invention selects different throttle MAPs according to the operating conditions of different gears. Gear state 1 switches to throttle Map1, gear state 2 switches to throttle Map2, and gear state 3 switches to throttle Map3, taking into account both power and economy and improving the driving performance of the entire vehicle.

[0027] like Figure 3The road gradient classification state judgment jump principle diagram of the present invention is shown as follows: Road slope state 1: flat road; Road slope state 2: gentle slope; Road slope state 3: steep slope.

[0028] Road slope classification status judgment jump principle: When the slope is greater than S1 and remains greater than the time threshold, the road slope state 1 jumps to the road slope state 2; When the slope is less than or equal to S1 and remains greater than the time threshold, the road slope state 2 jumps to the road slope state 1; When the slope is greater than S2 and remains greater than the time threshold, the road slope state 2 jumps to the road slope state 3; When the slope is less than or equal to S2 and remains greater than the time threshold, the road slope state 3 is changed to the road slope state 2; When the slope is greater than S2 and remains greater than the time threshold, the road slope state jumps from 1 to 3.

[0029] Throttle opening state setting: The throttle opening state 1 is greater than the throttle opening APP2 on a gentle slope; Throttle opening state 2 is less than or equal to throttle opening APP2 on a gentle slope; The throttle opening state 3 is greater than the throttle opening APP1 on a steep slope; The throttle opening state 4 is less than or equal to the throttle opening APP1 on a steep slope.

[0030] The default rule-based control method is economic mode E, which determines the vehicle load threshold, road slope threshold, and accelerator pedal opening threshold in real time.

[0031] The conditions for switching the transmission operating mode from economy mode E to power mode P are as follows: 1) When the vehicle load is greater than the set threshold m, the time maintained in road slope state 2 is greater than the set time threshold, and the time maintained in throttle opening state 1 is greater than the set time threshold, the transmission operating mode switches from economy mode E to power mode P; 2) When the vehicle load is greater than the set threshold m, the time maintained in road slope state 3 is greater than the set time threshold, and the time maintained in throttle opening state 3 is greater than the set time threshold, the transmission operating mode switches from economy mode E to power mode P.

[0032] The throttle opening state 1 is greater than the throttle opening APP2 on a gentle slope; The throttle opening state 3 is greater than the throttle opening APP1 on a steep slope; The throttle opening state 4 is less than or equal to the throttle opening APP1 on a steep slope.

[0033] The conditions for switching the transmission operating mode from power mode P to economy mode E are as follows: 1) When the vehicle load is less than or equal to the set threshold m, the power mode P is switched to the economic mode E; 2) When the vehicle load is greater than the set threshold m and the time it maintains in road slope state 1 is greater than the set time threshold, the power mode P is switched to the economy mode E; 3) When the vehicle load is greater than the set threshold m, the time maintained in road slope state 2 is greater than the set time threshold, and the time maintained in throttle opening state 2 is greater than the set time threshold, the power mode P is switched to the economy mode E; 4) When the vehicle load is greater than the set threshold m, the time maintained in road slope state 3 is greater than the set time threshold, and the time maintained in throttle opening state 4 is greater than the set time threshold, the power mode P is switched to the economy mode E.

[0034] Example 2 The present invention provides an intelligent control method for a commercial vehicle automatic transmission, including a rule-based control mode and an artificial intelligence algorithm-based control mode. Each time the vehicle is powered on and started, it first enters the rule-based control mode. After the vehicle has accumulated 200 kilometers of driving after starting, it automatically switches to the artificial intelligence algorithm-based control mode. The flow chart of the intelligent switching control of the operating mode of the commercial vehicle automatic transmission of the present invention is shown in FIG. Figure 4 As shown, state machine A is a rule-based control mode, and state machine B is a control mode based on an artificial intelligence algorithm. After the transmission EP intelligent switching is activated, the control mode based on the artificial intelligence algorithm intervenes after the vehicle has accumulated 200 kilometers after starting, and the control of state machine A is switched to state machine B, which can achieve automatic control.

[0035] The control method based on artificial intelligence algorithm is that after the driving style recognition APP completes learning using vehicle operation data, the APP issues a driving style label. The aggressive driving style label switches to power mode P, and the general and mild driving style label switches to economy mode E, thereby realizing adaptive control of the transmission operation mode.

[0036] The present invention provides an intelligent control method for a commercial vehicle automatic transmission. Figure 5 The present invention is based on the artificial intelligence algorithm control mode driving style recognition algorithm development process, based on the vehicle operation data of the past driving style recognition, the development of a real-time driving style recognition algorithm based on machine learning, based on the vehicle operation data of the data cluster analysis, the main implementation methods are as follows: S1, data acquisition: First, a data collection route covering different road levels such as cities, highways, national highways, mountainous areas, and hills is developed as a driving cycle (acceleration ratio, deceleration ratio, constant speed ratio, and idle speed ratio), and the road type is identified through the distribution of vehicle speed, slope, and gear operation.

[0037] The second is the driver's choice. We select three types of drivers: aggressive, moderate, and general. We collect driving data from these drivers in the specified driving cycles, and perform data cleaning and data slicing. The data differences in different driving scenarios will adjust the data cleaning and slicing accordingly.

[0038] Data cleaning processes noise, anomalies, and missing values ​​to ensure data quality. Noise smoothing removes high-frequency noise through simple sliding window averaging and low-pass filtering. Outlier detection sets thresholds based on vehicle physics and combines data from multiple sensors to determine rationality. Missing value processing is divided into small and large missing values. Small missing values ​​are handled using linear interpolation or forward filling, while large missing values ​​are simply removed from the affected time period.

[0039] Different slicing rules are designed for different scenarios. Data slicing uses three methods. The first uses a fixed time window as a segment, such as a 30-second segment for highway driving. This is suitable for steady-state analysis and can be shortened to 20 seconds for national highway driving. The second is event-driven, defining rules for sudden acceleration, sudden braking, and sharp turns, extracting 5 seconds of data before and after each event. Highway driving has higher speeds and less variability, while national highway driving has frequent starts and stops, with more sudden acceleration and braking. The third method is driving state segmentation. Based on the idle, acceleration, deceleration, and constant speed states defined in the driving cycle, segmentation is based on state change points, such as from acceleration to constant speed.

[0040] After data processing is complete, a comprehensive score is assigned to the driver's driving behavior within the driving cycle. A base score is set based on the distribution of driving data such as acceleration, braking, turning, lane changing, steady-state driving, idling, and unreasonable use of air conditioning. The frequency of occurrence and other factors are then multiplied by corresponding weight coefficients to obtain a comprehensive score. This comprehensive score serves as an important basis for selecting the clustering algorithm later. S2, feature engineering: First, we extract mathematical features such as the mean, variance, and maximum of samples from different scenarios as basic features. To minimize the influence of different values, we standardize the features. For feature selection, we first remove features with low variance. Then, we use a two-sample t-test to select features with significant differences in driving styles. Finally, we use correlation analysis to remove redundant features. Through these operations, the collected data is divided into multiple, separate, small samples for cluster analysis.

[0041] S3, cluster analysis: Because the aggressiveness of driving style is difficult to assess objectively, labeling relies primarily on subjective experience. However, labeling large amounts of data is labor-intensive, and labeling accuracy is difficult to assess. Therefore, we first approach driving style recognition as an unsupervised learning problem. We collect a large amount of driving data, analyze it offline, and use a clustering algorithm to cluster the sample data. We then assign labels based on the results. This can significantly reduce the workload of algorithm development. Here, we use the K-means clustering method without PCA for clustering, and the clustering result serves as a sample label.

[0042] Currently, the common clustering algorithms include K-means clustering, Gaussian mixture clustering, hierarchical clustering, density clustering, and spectral clustering. The patent mainly focuses on K-means clustering and Gaussian mixture clustering. Through clustering result analysis, the results obtained by the two methods are similar. Due to the large amount of data, the K-means clustering method without PCA converges faster and is more timely.

[0043] S4. Classification algorithm development: Due to the influence of external conditions, it is difficult for drivers to maintain a fixed driving style throughout the entire driving cycle. Therefore, the driver's driving style throughout the entire driving cycle cannot be 100% classified as aggressive, normal, or moderate. After adding labels to the samples through cluster analysis, the labeled samples are used to train the classification algorithm. Here, the support vector machine, a generalized linear classifier that classifies data using a supervised learning method, is selected. Relevant features such as vehicle speed, lateral and longitudinal acceleration, acceleration, and accelerator pedal change rate are input into the classification algorithm. Algorithm training can be implemented using m-file scripts. The distance from a sample point to the near-overclocked operation can itself reflect the aggressiveness of the sample point. The driving style is no longer divided into three isolated categories (aggressive, moderate, and normal), but a linear score of the aggressiveness. The distance from the aggressive sample to the aggressive hyperplane is higher than that of the normal sample, and the distance from the normal sample to the aggressive hyperplane is higher than that of the moderate sample. After algorithm training is complete, the trained algorithm is integrated into Simulink via the MATLAB function module to implement algorithm modeling development. After the model is built, C++ code is generated from the model through Embedded Coder to create a driving style recognition app. The driving style app ultimately calculates and outputs a driving style label (aggressive, moderate, or normal).

[0044] The above description is merely an example of the embodiments of the present invention and does not constitute any form of limitation to the present invention. The scope of protection of the present invention shall be based on the claims and is not limited by the above specific embodiments. Any simple modification or equivalent changes and modifications made to the above embodiments based on the technical essence of the present invention shall fall within the scope of protection of the present invention.

Claims

1. A commercial vehicle automatic transmission intelligent control method based on VCU architecture, characterized by: The intelligent control method includes a rule-based control mode, which performs throttle map switching and transmission operating mode decision-making according to the current vehicle state, the current vehicle load, the road slope and the driver's intention. The VCU intelligently switches the transmission economy and power mode shifting rules, and at the same time, the VCU controls the throttle map according to the current gear position; the VCU makes operating mode decisions and controls switching according to the current road slope classification status and the accelerator pedal opening status; ensures that the engine operates in the economic speed range as much as possible, always operates in the engine's high-efficiency working condition, and improves power while ensuring economy.

2. The intelligent control method for commercial vehicle automatic transmission according to claim 1, characterized in that: When the load is light, or the road slope is small, or the throttle opening is less than or equal to the preset threshold value in uphill conditions, the economic mode is selected and sent to the transmission TCU; when the three conditions are met at the same time, such as heavy load, large slope and throttle opening greater than the preset threshold value in uphill conditions, the power mode is selected and sent to the transmission TCU; the transmission TCU selects the economical shifting schedule to shift gears according to the E-mode status signal sent by the VCU; the transmission TCU selects the power-type shifting schedule to shift gears according to the P-mode status signal sent by the VCU; Alternatively, when the load is light, or on a flat road, or when going up a gentle slope and the throttle opening is less than or equal to the throttle opening set for the gentle slope, or when going up a steep slope and the throttle opening is less than or equal to the throttle opening set for the steep slope, the economy mode is selected and sent to the transmission TCU; when the load is heavy and the slope is gentle and the throttle opening is greater than the throttle opening set for the gentle slope, or when the load is heavy and the slope is steep and the throttle opening is greater than the throttle opening set for the steep slope, and all three conditions are met at the same time, the power mode is selected and sent to the transmission TCU.

3. The intelligent control method for commercial vehicle automatic transmission according to claim 2, characterized in that: Set different throttle thresholds APP for steep slopes and gentle slopes, where the throttle threshold for steep slopes is APP1 and the throttle threshold for gentle slopes is APP2; Since the power demand for climbing a gentle slope is small, when the vehicle is heavily loaded and the slope is gentle and the throttle opening is greater than APP2, the vehicle switches to the power mode; when the vehicle is heavily loaded and the slope is gentle and the throttle opening is less than or equal to APP2, the vehicle switches to the economy mode; Because climbing steep slopes requires a lot of power, when the vehicle is heavily loaded and the slope is steep and the throttle opening is greater than APP1, it switches to power mode; when the vehicle is heavily loaded and the slope is steep and the throttle opening is less than or equal to APP1, it switches to economy mode; The throttle threshold APP1 for the steep slope is set to 45%, and the throttle threshold APP2 for the gentle slope is set to 65%.

4. The intelligent control method for commercial vehicle automatic transmission according to claim 1, characterized in that: The VCU controls the throttle map according to the current gear position, including: Based on the premise of judging the driving intention through VCU, different throttle MAPs are selected according to the operating conditions of different gears. Gear state 1 switches to throttle Map1, gear state 2 switches to throttle Map2, and gear state 3 switches to throttle Map3, taking into account both power and economy, and improving the driving performance of the whole vehicle; Through big data statistics and analysis of the gear distribution of currently running vehicles, the gear ranges are divided according to different gearbox models: gear state 1 is the low gear area, gear state 2 is the middle gear area, and gear state 3 is the high gear area; Throttle Map status division: Map1 is the low gear area to cope with starting and mountain climbing; Map2 is the middle gear area to cope with high-speed climbing, taking into account both power and economy; Map3 is the high gear area to cope with high-speed cruising conditions to meet economy.

5. The intelligent control method for commercial vehicle automatic transmission according to claim 4, characterized in that: When the actual gear position of the transmission is higher than the low gear threshold, the throttle MAP state 1 in the low gear is switched to the throttle MAP state 2 in the middle gear, and the torque control is smoothed; When the actual gear position of the transmission is higher than the threshold of the middle gear area, the throttle MAP state 2 in the middle gear area is switched to the throttle MAP state 3 in the high gear area, and the torque control is smoothed; When the actual gear position of the transmission is lower than the threshold of the middle gear area, the throttle MAP state 2 in the middle gear area is switched to the throttle MAP state 1 in the low gear area, and the torque control is smoothed; When the actual gear position of the transmission is lower than the high gear threshold, the throttle MAP state 3 in the high gear is switched to the throttle MAP state 2 in the middle gear, and the torque control is smoothed.

6. The intelligent control method for commercial vehicle automatic transmission according to claim 1, characterized in that: The VCU makes a decision and controls the switching of the operation mode according to the current road slope classification state and the accelerator pedal opening state, including: making a decision and controlling the switching of the transmission operation mode according to the current vehicle load, the current road slope classification state and the accelerator pedal opening state; Road slope classification status: Road slope state 1: flat road; Road slope state 2: gentle slope; Road slope state 3: steep slope; Road slope classification status judgment jump principle: When the slope is greater than S1 and remains greater than the time threshold, the road slope state 1 jumps to the road slope state 2; When the slope is less than or equal to S1 and remains greater than the time threshold, the road slope state 2 is changed to the road slope state 1; When the slope is greater than S2 and remains greater than the time threshold, the road slope state 2 is changed to the road slope state 3; When the slope is less than or equal to S2 and remains greater than the time threshold, the road slope state 3 is changed to the road slope state 2; When the slope is greater than S2 and remains greater than the time threshold, the road slope state 1 jumps to the road slope state 3; Throttle opening state setting: The throttle opening state 1 is greater than the throttle opening APP2 on a gentle slope; Throttle opening state 2 is less than or equal to throttle opening APP2 on a gentle slope; The throttle opening state 3 is greater than the throttle opening APP1 under the steep slope; The throttle opening state 4 is less than or equal to the throttle opening APP1 on a steep slope; Among them, S1 is 3%, S2 is 7%, and the slope unit is percentage; less than S1 is a flat road, greater than S2 is a steep slope, and between S1-S2 is a gentle slope.

7. The intelligent control method for commercial vehicle automatic transmission according to claim 1, characterized in that: The specific steps include: After the transmission E / P intelligent switching is activated, the default is economic mode E, which determines the vehicle load threshold, road slope threshold, and accelerator pedal opening threshold in real time; The conditions for switching the transmission operating mode from the economic mode E to the power mode P are as follows: 1) When the vehicle load is greater than the set threshold m, the time maintained in the road slope state 2 is greater than the set time threshold, and the time maintained in the throttle opening state 1 is greater than the set time threshold, the transmission operation mode is switched from the economic mode E to the power mode P when the three conditions are met at the same time; 2) When the vehicle load is greater than the set threshold m, the time maintained in the road slope state 3 is greater than the set time threshold, and the time maintained in the throttle opening state 3 is greater than the set time threshold, the transmission operation mode is switched from the economic mode E to the power mode P; The conditions for switching the transmission operating mode from power mode P to economy mode E are as follows: 1) When the vehicle load is less than or equal to the set threshold m, the power mode P is switched to the economic mode E; 2) When the vehicle load is greater than the set threshold m and the time maintained in the road slope state 1 is greater than the set time threshold, the power mode P is switched to the economic mode E; 3) When the vehicle load is greater than the set threshold m, the time maintained in the road slope state 2 is greater than the set time threshold, and the time maintained in the throttle opening state 2 is greater than the set time threshold, the power mode P is switched to the economic mode E; 4) When the vehicle load is greater than the set threshold m, the time maintained in the road slope state 3 is greater than the set time threshold, and the time maintained in the throttle opening state 4 is greater than the set time threshold, the power mode P is switched to the economic mode E.

8. The intelligent control method for commercial vehicle automatic transmission according to claim 1, characterized in that: The intelligent control method also includes a control mode based on an artificial intelligence algorithm, where each time the vehicle is powered on and started, it first enters a rule-based control mode, and automatically switches to a control mode based on an artificial intelligence algorithm after the vehicle has accumulated 200 kilometers of driving after starting; The control mode based on the artificial intelligence algorithm identifies the driving style according to the past vehicle operation data, performs data cluster analysis based on the vehicle operation data, clusters and samples the driving conditions and formulates the driving cycle, classifies the road type and driving behavior of the vehicle, identifies the road type through the distribution of vehicle speed, slope, and gear operation, extracts the characteristic vector of the driving behavior including the maximum vehicle speed, average vehicle speed, maximum lateral acceleration, and maximum acceleration pedal change rate information, performs standardization processing, and obtains three driving style models of aggressive, general, and mild through training of the classification tree classification algorithm. The aggressive driving style model is switched to the power mode, and the general or mild driving style model is switched to the economy mode.

9. The intelligent control method for commercial vehicle automatic transmission according to claim 8, characterized in that: The control mode based on artificial intelligence algorithm specifically includes the following steps: S1, sample preparation: First, a data collection route covering different road levels such as cities, highways, national roads, mountainous areas, and hills is formulated as a driving cycle sample, and the idling, acceleration, deceleration, and uniform speed distribution and proportion of the driving cycle sample are set, and the road type is identified through the distribution of vehicle speed, slope, and gear operation; The second step is driver selection. Drivers with three driving styles, namely, aggressive, moderate and general, are selected. The driving data of these drivers on the specified driving cycle samples are collected. The driving data are cleaned and sliced. After the data processing is completed, the driving behaviors of the drivers in the driving cycle samples are comprehensively scored. S2, feature engineering: First, the mean, variance, and maximum mathematical features of samples in different scenarios are extracted as basic features, and the basic features are standardized. In terms of feature selection, the extracted low-variance features are first deleted. Secondly, the two-sample t-test is used to select features with large differences in different comprehensive driving styles. Finally, redundant features are deleted through correlation analysis. The collected driving data is made into multiple separate small samples to facilitate cluster analysis in step S3. S3, cluster analysis: The driving style recognition problem is regarded as an unsupervised learning problem. First, the collected driving data is analyzed offline. The sample data is clustered using the K-means clustering method without PCA. The clustering result is used as a sample label, and the corresponding driving style label is added according to the specific result. S4, classification algorithm development: After adding labels to the samples through cluster analysis, the labeled samples are used to train the classification algorithm. The generalized linear classifier that supports vector machines to classify data in a supervised learning manner is selected. The vehicle speed, lateral and longitudinal acceleration, acceleration, and accelerator pedal change rate related features are input into the classification algorithm. The algorithm training is implemented by using m-file scripts. The distance from the sample point to the near-overclocking operation can itself reflect the aggressiveness of the sample point. The driving style is no longer divided into three isolated categories: aggressive, mild, and general, but a linear score for the degree of aggressiveness. The distance from the aggressive sample to the aggressive hyperplane is higher than that of the general sample, and the distance from the general sample to the aggressive hyperplane is higher than that of the mild sample. After the algorithm training is completed, the trained algorithm is integrated into Simulink through the matlab function module to realize the modeling development of the algorithm. After the model is built, the model is generated into C++ code through Embedded Coder to make a driving style recognition APP. The driving style APP finally calculates and outputs three labels of driving style: aggressive, mild, and general.

10. The intelligent control method for commercial vehicle automatic transmission according to claim 9, characterized in that: In step S1, The data cleaning includes noise smoothing, outlier detection and missing value processing to ensure data quality; noise smoothing removes high-frequency noise through simple sliding window averaging and low-pass filtering, outlier detection sets thresholds based on vehicle physics knowledge and judges rationality based on multiple sensor data; missing value processing is divided into small segment missing and large segment missing, small segment missing uses linear interpolation or forward filling, large segment missing directly removes the affected time period; Different slicing rules are designed for different scenarios. There are three methods for data slicing. The first method takes a fixed time window as a segment, which is suitable for steady-state analysis. The second method is event-driven, which defines rules for sudden acceleration, sudden braking, and sharp turns, and extracts 5 seconds of data before and after the event. The speed of high-speed driving is higher and changes less, while when driving on national highways, the vehicle starts and stops frequently, and there are more sudden accelerations and sudden braking. The third method is driving state segmentation, which is based on the idle, acceleration, deceleration, and constant speed states set in the driving cycle, and is segmented based on state change points.

Images