A torque smoothing control method for dual-redundant EPS fault cutting edge
By dynamically adjusting the filtering coefficient of the first-order low-pass filter algorithm, the torque mutation problem during fault switching in the EPS dual-redundancy architecture is solved, achieving smooth control under different operating conditions and improving driving comfort and steering response.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- BOSCH HUAYU STEERING SYST CO LTD
- Filing Date
- 2026-04-02
- Publication Date
- 2026-06-26
AI Technical Summary
In the existing EPS dual-redundancy architecture, during fault switching, there is a sudden difference in the power assist torque command between the main and slave control units, which leads to steering system sluggishness and discomfort. The existing solution cannot meet the smooth requirements under different operating conditions.
By dynamically adjusting the filter coefficient of the first-order low-pass filter algorithm, and dynamically adjusting the filter parameters based on the torque difference and vehicle speed, smooth torque control is achieved during EPS fault switching. Real-time filtering processing is performed using a data acquisition module, torque sensor, MasterECU, and SlaveECU.
It achieves smooth torque transition under different operating conditions, balancing switching smoothness and steering response, avoiding changes in driver feel and steering lag, and meeting the requirements of driving comfort and safety.
Smart Images
Figure CN122276004A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of steering system technology, specifically a torque smoothing control method for dual-redundant EPS fault shearing. Background Technology
[0002] With the development of autonomous driving technology, the EPS system, as a core component of vehicle lateral control, needs to have "failure-safe operation" capability, and dual-redundancy architecture has become standard in high-level autonomous vehicles. Existing EPS dual-redundancy architectures typically include two independent control units, a master and a slave. When the master control unit fails, it quickly switches to the slave control unit to take over the power steering control, ensuring the continuous operation of the steering system.
[0003] However, existing redundancy switching technologies have significant drawbacks: during fault switching, the power assist torque commands from the master and slave control units are prone to abrupt differences, causing a sudden change in the steering motor's output torque. Drivers will experience abrupt steering wheel sensations such as "jerkiness" and "rebound," affecting driving comfort and handling stability. To address this issue, some solutions employ a fixed-rate torque increment / decrement strategy or a fixed-coefficient low-pass filtering algorithm, but neither can adapt to the smoothing requirements under different torque differences and vehicle speeds.
[0004] Existing solutions only achieve fault switching by adjusting the assist distribution ratio between the master and slave units, without designing a smooth control mechanism for torque surges; or they focus on calculating the assist output value of redundant control units, without addressing torque fluctuation suppression during the switching process. Another solution uses a fixed-coefficient low-pass filter, but if the coefficient is set too small, it will cause steering response lag; if the coefficient is set too large, it cannot effectively filter out torque surge spikes, making it difficult to balance smoothness and responsiveness.
[0005] Therefore, a torque smoothing control scheme that can dynamically adjust the filtering parameters according to the switching operating conditions is needed to solve the problems of poor adaptability and poor smoothing effect of existing technologies. Summary of the Invention
[0006] To overcome the shortcomings of the prior art, this invention provides a torque smoothing control method for dual redundant EPS fault cutting edge. By dynamically adjusting the first-order low-pass filter coefficient, it achieves precise smooth control under different torque differences and different vehicle speed conditions, taking into account both switching smoothness and steering response.
[0007] To achieve the above objectives, a torque smoothing control method for dual-redundant EPS fault edge cutting is designed, including a data acquisition module, a torque sensor, a MasterECU, a SlaveECU, and an assist motor. The specific control method flow is as follows: S1, the data acquisition module collects in real time the first assist torque command output by the MasterECU, the second assist torque command output by the SlaveECU, the current vehicle speed, the fault level signal of the MasterECU, and the fault level signal of the SlaveECU. S2, When the fault detection module inside the MasterECU detects a fault in the MasterECU and triggers redundancy switching, it calculates the torque difference ΔT between the first assist torque and the second torque at the moment of switching. S3, based on the torque difference ΔT and the current vehicle speed collected from the CAN-BUS, dynamically adjusts the filtering coefficient α of the first-order low-pass filter algorithm; S4, during the redundancy switching cycle, uses an adjusted first-order low-pass filtering algorithm to perform real-time filtering on the second assist torque output by the SlaveECU, and outputs a smooth assist torque command to the assist motor. S5, after the switching is completed, the filter coefficient α is linearly restored to the preset initial value, the smoothing filter process is stopped, and the SlaveECU directly outputs the assist torque command.
[0008] The discretization calculation formula of the first-order low-pass filtering algorithm is: Y(n)=α·X(n)+(1-α)·Y(n-1); where Y(n) is the smoothed assist torque of the current output, X(n) is the second assist torque of the current SlaveECU output, and Y(n-1) is the smoothed assist torque command of the previous sampling period.
[0009] The range of values for the filtering coefficient α follows the rule that the larger the torque difference ΔT and the lower the vehicle speed, the smaller the filtering coefficient α; conversely, the larger the filtering coefficient α is.
[0010] The MasterECU is the master electronic control unit; the SlaveECU is the slave electronic control unit; and CAN-BUS is the controller local area network bus.
[0011] The outputs of the data acquisition module and the torque sensor are connected to the inputs of the MasterECU and SlaveECU, respectively. The MasterECU has a fault detection module, and the SlaveECU has a fault detection module, a switching control module, a filter coefficient adjustment module, and a smoothing filter module. The fault detection modules in the MasterECU and SlaveECU are bidirectionally connected. The output of the fault detection module in the SlaveECU is connected to the inputs of the switching control module and the filter coefficient adjustment module, respectively. The output of the filter coefficient adjustment module is connected to the smoothing filter module. The outputs of the MasterECU and the smoothing filter module are combined and connected to the power assist motor.
[0012] Compared with the prior art, the present invention provides a torque smoothing control method when a dual redundant EPS fault cuts off. By dynamically adjusting the first-order low-pass filter coefficient, it achieves precise smooth control under different torque differences and different vehicle speed conditions, taking into account both switching smoothness and steering response. Attached Figure Description
[0013] Figure 1 This is a block diagram of the torque smoothing control system of the present invention.
[0014] Figure 2 This is a schematic diagram of the torque smoothing control method of the present invention. Detailed Implementation
[0015] The present invention will now be further described with reference to the accompanying drawings.
[0016] like Figure 1 , Figure 2 As shown, a torque smoothing control method for fault switching in an EPS dual-redundancy architecture is presented. The core of this method lies in dynamically adjusting the filtering coefficient α of a first-order low-pass filter algorithm based on the torque difference at the moment of switching and the current vehicle speed, thereby achieving targeted smoothing control. The core principle is as follows: When a fault is detected in the MasterECU and redundancy switching is triggered, the torque difference ΔT between the first assist torque and the second torque at the moment of switching is calculated. The first-order low-pass filter algorithm filters out high-frequency torque abrupt changes by weighted superposition of the current torque command and the smoothing torque command from the previous cycle. Dynamically adjusting the filtering coefficient α enhances the filtering effect when the torque difference is large and the vehicle speed is low (high smoothness requirements), and weakens the filtering effect when the torque difference is small and the vehicle speed is high (high responsiveness requirements), thus balancing the two core requirements.
[0017] The torque smoothing control system of this invention includes a data acquisition module, a torque sensor, a Master ECU, a Slave ECU, and an assist motor. The outputs of the data acquisition module and the torque sensor are connected to the inputs of the Master ECU and Slave ECU, respectively. The Master ECU has a fault detection module, and the Slave ECU has a fault detection module, a switching control module, a filter coefficient adjustment module, and a smoothing filter module. The fault detection modules in the Master ECU and Slave ECU are bidirectionally connected. The output of the fault detection module in the Slave ECU is connected to the inputs of the switching control module and the filter coefficient adjustment module, respectively. The output of the filter coefficient adjustment module is connected to the smoothing filter module. The outputs of the Master ECU and the smoothing filter module are combined and connected to the assist motor. The Master ECU is the master electronic control unit; the Slave ECU is the slave electronic control unit; and CAN-BUS is the controller local area network bus.
[0018] The torque smoothing control method for dual-redundant EPS fault edge trimming according to the present invention is as follows: S1, the data acquisition module collects in real time the first assist torque command output by the MasterECU, the second assist torque command output by the SlaveECU, the current vehicle speed, the fault level signal of the MasterECU, and the fault level signal of the SlaveECU. S2, When the fault detection module inside the MasterECU detects a fault in the MasterECU and triggers redundancy switching, it calculates the torque difference ΔT between the first assist torque and the second torque at the moment of switching. S3, based on the torque difference ΔT and the current vehicle speed collected from the CAN-BUS, dynamically adjusts the filtering coefficient α of the first-order low-pass filter algorithm; S4, during the redundancy switching cycle, uses an adjusted first-order low-pass filtering algorithm to perform real-time filtering on the second assist torque output by the SlaveECU, and outputs a smooth assist torque command to the assist motor. S5, after the switching is completed, the filter coefficient α is linearly restored to the preset initial value, the smoothing filter process is stopped, and the SlaveECU directly outputs the assist torque command.
[0019] The range of values for the filter coefficient α follows the rule that the larger the torque difference ΔT and the lower the vehicle speed, the smaller the filter coefficient α; conversely, the larger the filter coefficient α is.
[0020] The discretization calculation formula of the first-order low-pass filter algorithm is: Y(n)=α·X(n)+(1-α)·Y(n-1), where Y(n) is the smoothed boost torque of the current output, X(n) is the second boost torque of the current SlaveECU output, and Y(n-1) is the smoothed boost torque command of the previous sampling period.
[0021] The specific steps of the control method in this embodiment are as follows: S1: The data acquisition module collects the first assist torque command T1 output by the MasterECU, the second assist torque command T2 output by the SlaveECU, and the current vehicle speed V collected through the CANFD bus in real time. The fault detection module identifies the MasterECU fault level as a serious fault. S2: The fault detection module triggers redundancy switching and calculates the torque interpolation ΔT=|T2-T1|=6N·m at the moment of switching. The current vehicle speed V=20km / h (low speed condition). S3: The filter coefficient adjustment module, based on ΔT=6N·m and V=20km / h, queries the built-in parameter mapping table and uses an interpolation algorithm to determine the filter coefficient α=0.15 (approaching the lower limit value to enhance the filtering effect). S4: The smoothing filter module executes a first-order low-pass filter algorithm, with the calculation formula Y(n)=0.15×T2(n)+0.85×Y(n-1). It outputs a smoothing torque command to the power assist motor every 10ms to suppress torque changes in real time. Using this method, the torque fluctuation during the switching process will not cause the driver to have a significant change in the feel. S5: After the redundancy switch is completed, the switching control module triggers the linear recovery of the filter coefficient α, with the recovery rate set to 0.003 / ms. After 283ms, it recovers to the initial value of 0.8 (the weak filtering state during normal operation). The smoothing filter process stops, and the SlaveECU directly outputs the power assist torque command, and the steering response returns to normal.
[0022] The beneficial effects of this invention are: 1. Strong adaptability: Breaking through the limitations of existing fixed parameter filtering, the filter coefficient is dynamically adjusted based on torque difference and vehicle speed, which can cover different fault scenarios and different driving conditions, ensuring smooth torque transition in all scenarios without the driver's perception. 2. Excellent responsiveness: Through dynamic adjustment and switching of the filter coefficient and the linear recovery mechanism, it avoids sudden torque changes and does not cause steering response lag, ensuring steering accuracy and meeting the safety requirements of high-speed driving.
Claims
1. A torque smoothing control method for dual-redundant EPS fault edge cutting, comprising a data acquisition module, a torque sensor, a MasterECU, a SlaveECU, and an assist motor, characterized in that: The specific process of the control method is as follows: S1, the data acquisition module collects in real time the first assist torque command output by the MasterECU, the second assist torque command output by the SlaveECU, the current vehicle speed, the fault level signal of the MasterECU, and the fault level signal of the SlaveECU. S2, When the fault detection module inside the MasterECU detects a fault in the MasterECU and triggers redundancy switching, it calculates the torque difference ΔT between the first assist torque and the second torque at the moment of switching. S3, based on the torque difference ΔT and the current vehicle speed collected from the CAN-BUS, dynamically adjusts the filtering coefficient α of the first-order low-pass filter algorithm; S4, during the redundancy switching cycle, uses an adjusted first-order low-pass filtering algorithm to perform real-time filtering on the second assist torque output by the SlaveECU, and outputs a smooth assist torque command to the assist motor. S5, after the switching is completed, the filter coefficient α is linearly restored to the preset initial value, the smoothing filter process is stopped, and the SlaveECU directly outputs the assist torque command.
2. The torque smoothing control method for dual-redundant EPS fault trimming according to claim 1, characterized in that: The discretization calculation formula of the first-order low-pass filtering algorithm is: Y(n)=α·X(n)+(1-α)·Y(n-1); where Y(n) is the smoothed assist torque of the current output, X(n) is the second assist torque of the current SlaveECU output, and Y(n-1) is the smoothed assist torque command of the previous sampling period.
3. The torque smoothing control method for dual-redundant EPS fault trimming according to claim 1 or 2, characterized in that: The range of values for the filtering coefficient α follows the rule that the larger the torque difference ΔT and the lower the vehicle speed, the smaller the filtering coefficient α; conversely, the larger the filtering coefficient α is.
4. The torque smoothing control method for fault-cutting edge of dual-redundant EPS according to claim 1, characterized in that: The MasterECU is the master electronic control unit; the SlaveECU is the slave electronic control unit; and CAN-BUS is the controller local area network bus.
5. The torque smoothing control method for fault-cutting edge of dual-redundant EPS according to claim 1, characterized in that: The outputs of the data acquisition module and the torque sensor are connected to the inputs of the MasterECU and SlaveECU, respectively. The MasterECU has a fault detection module, and the SlaveECU has a fault detection module, a switching control module, a filter coefficient adjustment module, and a smoothing filter module. The fault detection modules in the MasterECU and SlaveECU are bidirectionally connected. The output of the fault detection module in the SlaveECU is connected to the inputs of the switching control module and the filter coefficient adjustment module, respectively. The output of the filter coefficient adjustment module is connected to the smoothing filter module. The outputs of the MasterECU and the smoothing filter module are combined and connected to the power assist motor.