Search

CN-121973782-A - AMT (automated mechanical transmission) and EPB (electronic pressure brake) and AutoHold combined starting control method for commercial vehicle

CN121973782ACN 121973782 ACN121973782 ACN 121973782ACN-121973782-A

Abstract

The invention provides a combined starting control method of a commercial vehicle AMT, an EPB and an AutoHold, which belongs to the technical field of vehicle starting control, the invention constructs a cooperative control model of an AMT controller, an EPB controller and a AutoHold controller by adopting a graph convolution neural network, estimates the gradient and the vehicle mass based on an extended Kalman filtering algorithm, the continuous switching of braking force and driving force is realized by utilizing a wheel end torque preloading mechanism, the friction coefficient is corrected by a digital twin model, the torsional vibration of the transmission system is restrained by combining an active damping control strategy, the learning rate and the prediction time domain are dynamically adjusted, and the technical problem of poor starting smoothness caused by insufficient coordination control precision of a braking system and a power transmission system when the commercial vehicle starts on a hill is solved.

Inventors

  • Wang Lvjun
  • YU SONGLIN

Assignees

  • 东鼎重科传动科技(青岛)有限公司

Dates

Publication Date
20260505
Application Date
20260310

Claims (10)

  1. 1. A combined starting control method for a commercial vehicle AMT, an EPB and an AutoHold is characterized by comprising the following steps of acquiring starting parameters in real time through a CAN bus by an AMT controller in a vehicle starting preparation stage, measuring a longitudinal inclination angle and a transverse inclination angle of the vehicle through a triaxial acceleration sensor, utilizing an extended Kalman filtering algorithm to fuse torque of a power source of the whole vehicle, a vehicle speed, the longitudinal inclination angle and the transverse inclination angle of the vehicle to estimate the current gradient and the vehicle quality, triggering parking judgment according to the fact that the vehicle speed is smaller than a preset vehicle speed threshold value and the depth of a brake pedal is larger than a preset pedal depth threshold value, activating AutoHold a parking mode when the current gradient is smaller than a first gradient threshold value, switching to the EPB parking mode when the current gradient is larger than a second gradient threshold value or the duration of the AutoHold parking mode exceeds the preset duration threshold value, adopting a wheel end torque preloading mechanism to maintain braking force continuity in the switching process, calculating a clutch target position and an engine limiting torque based on a cooperative control model, inputting the current transmission torque of the clutch, the ramp resistance and the tire attaching limit torque into the cooperative control model, outputting an allowable signal confidence level and a preset demand track in the current moment, transmitting a future demand signal, triggering a signal to a trigger signal in a preset time window, when the current gradient is larger than the second gradient threshold value or the EPT controller is larger than the second gradient threshold value or AutoHold, transmitting the signal to the EPT signal is smaller than the preset threshold value and the EPT controller through the EPT threshold value, and the EPT controller, the EPT controller is smaller than the threshold value, and the EPT controller is set to be 3, and the threshold value is set to be smaller than 24, and the threshold value is set to be smaller, and the threshold, and the engine speed signal is 1, and the threshold value is controlled and the threshold is controlled and is simultaneously is controlled and smaller by the threshold and the speed is 1 and the threshold is controlled and the speed is 25 and the threshold is controlled and the threshold is simultaneously is controlled. And in the clutch sliding stage, the AMT controller monitors the temperature of the friction plate in real time, inputs the temperature of the friction plate into a digital twin model to correct the friction coefficient, dynamically adjusts the learning rate and the prediction time domain of the cooperative control model by utilizing the self-adaptive adjusting function, and ensures that the torque handover is completed when the difference value between the rotating speed of the engine and the rotating speed of a first shaft of the gearbox is smaller than a preset rotating speed difference threshold value.
  2. 2. The method of claim 1, wherein the specific structure of the cooperative control model abstracts the AMT controller, the EPB controller, the AutoHold controller, the engine management system, and the brake system into graph structure nodes, wherein CAN bus communication links, mechanical connection relationships, and torque transmission paths among the nodes form directed edges, a graph convolution layer is used to extract state feature vectors of the graph structure nodes, and information aggregation among the graph structure nodes is realized through a message transfer layer.
  3. 3. The method of claim 2, wherein the message transfer layer distributes weights of the neighborhood nodes by adopting an attention mechanism, each graph structure node updates its own control instruction according to state feature vectors and edge weights of the neighborhood nodes, and the global loss function is formed by weighting four items of starting time, sliding friction energy loss, rotational speed synchronization error and sliding distance.
  4. 4. The method of claim 3, wherein a training data set of the cooperative control model is established, a plurality of standard ramps with different gradients are arranged on a test site, three vehicle weight working conditions of no-load, half-load and full-load are respectively tested under each standard ramp, starting operation is repeated for a plurality of times under each vehicle weight working condition, and time series data of engine rotation speed, clutch target position, brake pressure, vehicle speed, longitudinal acceleration and friction plate temperature are acquired.
  5. 5. The method of claim 4, wherein the collaborative control model is trained, the weights of the initialized graph convolution layer are uniformly distributed by using Xavier, an initial learning rate is set, gradient descent is performed by using Adam optimizer, and a total loss function is a weighted sum of three terms of mean square error loss, starting smoothness constraint loss and safety constraint loss.
  6. 6. The method of claim 5, wherein the state vector of the extended kalman filter algorithm comprises three state quantities of a longitudinal speed, a gradient angle and a vehicle mass of the vehicle, the observation vector comprises a longitudinal acceleration of a triaxial acceleration sensor, a lateral acceleration and a linear speed of a wheel rotation speed sensor, and the state transfer equation is established based on a longitudinal dynamics model of the vehicle.
  7. 7. The method of claim 6, wherein the wheel end torque preloading is performed by the AMT controller gradually increasing the current torque transmitted by the clutch to a predetermined proportion of the rated torque for a predetermined time before switching AutoHold the parking mode to the EPB parking mode, so that the wheel end torque reaches a predetermined percentage of the ramp resistance torque, and the EPB actuator pre-tightens the brake caliper in advance.
  8. 8. The method of claim 7, wherein the calculation method of the confidence level of the start permission signal normalizes the current transmission torque of the clutch, the resistance torque of the ramp and the tire adhesion limit torque to construct a confidence evaluation vector, and the confidence evaluation vector comprises three components of a torque margin ratio, a torque change rate and sensor data consistency.
  9. 9. The method of claim 8, wherein the implementation of the active damping control strategy is implemented by creating a two-degree-of-freedom torsional vibration model of the drive train, generating a feedforward damping torque by multiplying a differential signal of torque of a power source of the whole vehicle by a compensation coefficient, setting a notch filter, and processing a difference between the rotational speed of the engine and the rotational speed of a first shaft of the gearbox by the notch filter to obtain a feedback control quantity.
  10. 10. The method of claim 9, wherein the launch parameters include vehicle power source torque, vehicle speed, brake pedal depth, EPB on-off state, engine speed, transmission one-shaft speed.

Description

AMT (automated mechanical transmission) and EPB (electronic pressure brake) and AutoHold combined starting control method for commercial vehicle Technical Field The invention belongs to the technical field of vehicle starting control, and particularly relates to a combined starting control method for an AMT, an EPB and an AutoHold of a commercial vehicle. Background The commercial vehicle automatic transmission system is required to realize stable switching of braking force and driving force in a hill start scene. The traditional technology adopts a distributed architecture of independent control of an AMT controller, an EPB controller and a AutoHold controller, a brake release and clutch engagement action is triggered through preset sequential logic, a brake release moment is judged by a fixed threshold value, and an open loop feedforward compensation strategy is adopted for clutch torque control. In the current hill start control method, due to the fact that a real-time sharing and collaborative optimization mechanism of state information is lacked among all controllers, a time window mismatch phenomenon exists between braking force release and driving force establishment, and accordingly sliding or impact occurs in the starting process. The traditional method is difficult to dynamically adjust the control strategy according to multidimensional state parameters such as real-time gradient, vehicle mass, clutch temperature and the like, and clutch torque control depends on a calibrated MAP table and cannot adapt to working condition changes. That is, in the prior art, there is a technical problem that poor starting smoothness is caused by insufficient coordination control precision of a braking system and a power transmission system when a commercial vehicle starts on a slope. Disclosure of Invention In view of the above, the invention provides a method for controlling AMT, EPB and AutoHold combined starting of a commercial vehicle, which can solve the technical problem of poor starting smoothness caused by insufficient coordination control precision of a braking system and a power transmission system of the commercial vehicle during hill start in the prior art. In the vehicle starting preparation stage, an AMT controller acquires starting parameters in real time through a CAN bus, measures the longitudinal inclination angle and the transverse inclination angle of the vehicle through a three-axis acceleration sensor, and utilizes an extended Kalman filtering algorithm to fuse the torque, the vehicle speed, the longitudinal inclination angle and the transverse inclination angle of the power source of the whole vehicle to estimate the current gradient and the vehicle quality; the method comprises the steps of triggering parking judgment according to the condition that the vehicle speed is smaller than a preset vehicle speed threshold value and the depth of a brake pedal is larger than a preset pedal depth threshold value, activating AutoHold a parking mode when the current gradient is smaller than a first gradient threshold value, switching to an EPB parking mode when the current gradient is larger than a second gradient threshold value or the duration of the AutoHold parking mode exceeds a preset duration threshold value, adopting a wheel end torque preloading mechanism to keep braking force continuity in the switching process, inputting the current transmission torque of the clutch, the ramp resistance torque and the tire attachment limit torque into a cooperative control model based on a cooperative control model by an AMT controller, outputting a start permission signal confidence level at the current moment and a torque demand track in a future preset time window, when the start permission signal confidence level is larger than the preset confidence level threshold value, sending TransmissionReadyforBrakeRelease signals in ETC7 messages to an EPB controller and a AutoHold controller through a time-triggered Ethernet TSN bus, enabling signal transmission time jitter of the ETC7 messages to be 1, gradually reducing the EPB controller and the AutoHold controller to be smaller than the preset threshold value after receiving TransmissionReadyforBrakeRelease signals, controlling the rotation speed of the engine to be gradually reduced to a zero through an active speed control speed of an active speed control axle according to a differential value of the differential of the clutch and the dynamic speed control of the pneumatic speed of the engine, and in the clutch sliding stage, the AMT controller monitors the temperature of the friction plate in real time, inputs the temperature of the friction plate into a digital twin model to correct the friction coefficient, dynamically adjusts the learning rate and the prediction time domain of the cooperative control model by utilizing the self-adaptive adjusting function, and ensures that the torque handover is completed when the difference value between the rotating s