An electric control brake method and system based on axle load adaptation

Abstract

The application discloses an electric control brake method and system based on axle load self-adaption, comprising: acquiring the equivalent mass of a vehicle; judging whether the electric actuator has brake force output capacity according to the equivalent mass; entering the comfortable braking section, querying the deceleration calibration point table according to v_filtered, calculating the target deceleration a_target(v) by using the Smoothstep cubic polynomial interpolation, and reading the road slope G_road; calculating the feedforward brake force instruction; outputting the feedforward brake force instruction to the electric actuator after processing by the asymmetric slope limiter to execute braking; switching to the peristalsis suppression section when v_filtered <= 1 km / h; and switching to the parking keeping section when v_filtered <= 0.3 km / h and lasting for no less than M control periods. The application can realize smooth braking and accurate point parking of the full load spectrum.

Description

Technical Field

[0001] This invention relates to an electric braking method and system, and more particularly to an electric braking method and system based on shaft load adaptation. Background Technology

[0002] Truck cranes are special vehicles that use a dedicated chassis as a platform and are equipped with a boom lifting mechanism. Their equivalent driving mass can vary significantly, ranging from approximately 42 tons when the chassis is unloaded to approximately 100 tons when fully loaded, a variation of up to 138%. In the braking control design for autonomous driving systems, truck cranes face three prominent and differentiated challenges:

[0003] (1) Extremely large range of mass variation. Traditional braking control strategies for passenger cars or commercial vehicles are calibrated based on the premise that the vehicle mass parameters are basically fixed (passenger car mass fluctuation <10%, truck mass fluctuation <30%). If directly applied to the crane scenario, the braking force is seriously excessive under light load, resulting in sudden braking impact, and the braking force is insufficient under heavy load, leading to the risk of rear-end collision.

[0004] (2) Vibration interference from hydraulic suspension. The natural vibration frequency of the crane's hydraulic suspension is about 0.8 to 2 Hz. In the low-speed range (<1km / h), it highly overlaps with the frequency of the wheel speed pulse signal, resulting in a deterioration of the signal-to-noise ratio. At the same time, the longitudinal acceleration channel of the IMU is affected by the suspension vibration and generates pseudo-acceleration signals (±0.05 to 0.10 m / s²). If these signals are used directly in closed-loop control without processing, they will cause high-frequency oscillations in the braking force command.

[0005] (3) Dual precision constraints. The longitudinal positioning error of the outrigger deployment point must be controlled within ±15 cm, and the elastic displacement of the boom end can reach ±30 cm when the braking impact (Jerk) exceeds 1.5 g / s, which may trigger a structural safety warning.

[0006] Low-speed end-of-range braking (5 km / h to 0) is a critical range determining parking accuracy and comfort. Traditional pneumatic braking systems suffer from pressure build-up delay (200–300 ms) and abrupt braking force fluctuations, making precise position control difficult. Electronic parking brake (EPB) systems, driven by permanent magnet synchronous motors, offer advantages such as short response delay (approximately 50 ms) and continuously adjustable clamping force (0–100%), making them an ideal choice for low-speed end-of-range braking actuators.

[0007] Existing research includes CN114123456A, which discloses a method for intelligent vehicle comfort parking control based on the Smoothstep function. This method constructs a low-speed deceleration mapping curve using the Smoothstep function to control Jerk (rate of deceleration). CN113879267A discloses a braking comfort optimization method based on model predictive control (MPC), which constructs an optimization problem using Jerk as a constraint. Furthermore, the braking control system for unmanned container trucks in ports (CN115384470A) achieves a positioning accuracy of ±30 cm.

[0008] The aforementioned existing technologies share the following common limitations:

[0009] (1) Assumption of fixed mass parameters. The above methods all take vehicles with basically fixed mass parameters as the control object, and the calibration of control parameters is only carried out near a single mass point. For truck cranes with equivalent mass changes of more than twice, the fixed parameter strategy will inevitably lead to excessive braking force under light load or insufficient braking force under heavy load.

[0010] (2) No explicit decoupling between mass and deceleration curves has been established. In existing methods, changes to mass parameters require recalibrating the entire deceleration curve or control gain. There is a lack of a structured framework that separates mass from comfort design, resulting in a large amount of repetitive calibration work required when switching between multiple operating conditions.

[0011] (3) Lack of differentiated processing for signal quality degradation in the low-speed domain. Existing methods do not fully recognize the physical fact that the wheel speed signal signal-to-noise ratio deteriorates sharply when the vehicle speed is below 1 km / h. The same sensor signal source and control algorithm are used throughout the braking process, resulting in deterioration of end-control accuracy.

[0012] (4) Lack of actuator capability prediction before braking starts. Existing methods do not assess whether the actuator has sufficient force output capability before braking begins, which may lead to insufficient force during braking and thus safety risks. Summary of the Invention

[0013] Purpose of the invention: The purpose of this invention is to propose an electric braking method and system based on axle load adaptation, which can achieve smooth braking and precise stopping across the entire load spectrum without recalibrating control parameters for different load conditions.

[0014] Technical solution: This invention includes the following steps:

[0015] S1. Obtain the vehicle's equivalent mass m_eff before braking;

[0016] S2. Based on the equivalent mass m_eff, predict the actuator's capability and determine whether the electric actuator has the ability to output braking force. If the capability is insufficient, refuse to enter the braking state and issue an alarm.

[0017] S3. Entering the comfort braking phase, the wheel speed signal is low-pass filtered to obtain v_filtered. Based on v_filtered, the deceleration calibration point table is consulted, and the target deceleration a_target(v) is calculated using Smoothstep cubic polynomial interpolation. The road gradient G_road is read. The feedforward braking force command is calculated according to the formula u_ff = [m_eff × (a_target + g × G_road / 100)] / F_max × 100%, where F_max is the rated clamping force of the electric actuator. The feedforward braking force command is processed by the asymmetric slope limiter and then output to the electric actuator to perform braking. When v_filtered ≤ 1 km / h, the system switches to the creep suppression phase.

[0018] S4. Enter the creep suppression phase, abandon the wheel speed signal, use the longitudinal acceleration of the inertial measurement unit as the feedback variable for closed-loop control, and output braking force command to the electric actuator; when v_filtered ≤ 0.3 km / h and lasts for no less than M control cycles, switch to the stop holding phase;

[0019] S5. Enter the parking holding phase, increase the braking force command to the preset holding value, and activate the mechanical locking mechanism of the electric actuator.

[0020] The flow of the control process in this invention follows the principle of unidirectional irreversibility: idle → comfort braking → creep suppression → stop holding, and cannot be reversed.

[0021] The Smoothstep cubic polynomial interpolation is specifically defined as follows: the normalized velocity parameter t = (v - v_{i+1}) / (v_i - v_{i+1}), t ∈ [0, 1]; the deceleration interpolation a_target(v) = a_{i+1} + (a_i - a_{i+1}) × (3t² - 2t³), where (v_i, a_i) and (v_{i+1}, a_{i+1}) are adjacent deceleration calibration points.

[0022] The asymmetric ramp limiter imposes a direction-independent upper limit constraint on the change in the feedforward braking force command in each control cycle: the upper limit of the rate of change in the increasing direction, Δu_up, is different from the upper limit of the rate of change in the decreasing direction, Δu_dn.

[0023] The closed-loop control is a pure proportional control, and the control law is: u(k) = u(k-1) + K_p × [a_target -a_measured(k)] × T_s; where K_p is the proportional gain and T_s is the control period.

[0024] The output command limit of the creep suppression section is dynamically calculated based on the equivalent mass:

[0025] u_min = (m_eff × α_low) / F_max × 100 %

[0026] u_max = (m_eff × α_high) / F_max × 100 %

[0027] Here, α_low and α_high are preset acceleration boundary parameters in m / s². When the equivalent mass changes, the upper and lower limits are scaled proportionally to maintain the same force / mass ratio.

[0028] The equivalent mass m_eff is obtained by measuring the axle load at the four corners using axle load sensors installed at the piston rod ends of the hydraulic cylinders of each outrigger. The sum of the four points is the current total weight of the vehicle, which is then divided by the gravitational acceleration g.

[0029] The actuator capability prediction is specifically as follows: calculate the prediction value u_check = (m_eff × a_target, max) / F_max × 100%; where a_target is the preset target deceleration; if u_check > preset safety threshold, then the electric actuator capability margin is determined to be insufficient.

[0030] The present invention also provides an electric braking system based on shaft load adaptation, comprising:

[0031] The perception layer is used to collect vehicle motion status and environmental parameters, including wheel speed sensors for measuring wheel speed, inertial measurement units for measuring longitudinal acceleration, axle load sensors for measuring axle load to obtain equivalent mass, and slope sensors for measuring road gradient.

[0032] The decision-making layer includes a host computer and a chassis controller, the chassis controller of which runs the aforementioned axle load adaptive electric control braking method;

[0033] The execution layer includes electric actuators for receiving braking force commands from the chassis controller and performing braking actions.

[0034] The execution layer also includes a pneumatic braking system as a backup actuator. The pneumatic braking system is physically isolated from the electric actuator and has an independent controller, actuator and power circuit. When the electric actuator fails, the chassis controller switches to the pneumatic braking system to take over the braking.

[0035] Beneficial effects: The present invention has the following advantages:

[0036] (1) Full-load adaptive operation. Through an architecture that completely decouples the deceleration curve from the mass parameter, only an equivalent mass parameter m_eff needs to be updated before operation to cover all operating conditions from light load (42 tons) to heavy load (100 tons), without the need to repeatedly calibrate the control curve for different tonnages or operating conditions. Simulation verification: Automatic adaptation is achieved for three typical load conditions (42t / 65t / 100t), and the performance meets all standards.

[0037] (2) Effective constraint of braking impact. Smoothstep cubic interpolation eliminates the Jerk spike at the calibration point from the mathematical source, and the asymmetric ramp limiter eliminates the sudden change in braking force command, so that the Jerk peak value is stabilized at 0.61 g / s - lower than the passenger car comfort limit (1.0 g / s) and the crane boom protection threshold (1.5 g / s).

[0038] (3) Intelligent switching driven by signal quality. Based on the frequency domain overlap analysis of wheel speed pulse frequency and suspension vibration frequency, the sensor signal source and control algorithm (feedforward → closed loop, wheel speed → IMU) are automatically switched at 1 km / h, effectively addressing the problem of signal quality degradation in the low-speed domain.

[0039] (4) Preventive safety design. The actuator capacity prediction eliminates the risk of "insufficient force" before braking begins, the asymmetric ramp limiter avoids jerking impact, and the irreversible state transition avoids state machine jitter caused by signal jumps.

[0040] (5) Low engineering implementation cost. The core conversion formula involves only one floating-point multiplication and division operation; Smoothstep interpolation is a rational operation of cubic polynomials; all control algorithms do not require solving optimization problems or online system identification within a 10 ms control cycle. Attached Figure Description

[0041] Figure 1 This is a system block diagram of the present invention;

[0042] Figure 2 This is the state transition diagram of the three-stage finite state machine of the present invention;

[0043] Figure 3 This is a single-cycle flowchart of the control algorithm of the present invention. Detailed Implementation

[0044] The invention will now be further described with reference to the accompanying drawings.

[0045] Example 1

[0046] like Figure 1 As shown, the axle load adaptive electric braking system in this embodiment estimates or measures the vehicle's current mass (including load) in real time through the control system and dynamically adjusts the braking force command to ensure that the deceleration curve does not drift with changes in mass. It includes: a perception layer, a decision layer, an execution layer, and a communication bus.

[0047] The perception layer is responsible for collecting vehicle motion status and environmental parameters, including the following sensors:

[0048] Four-wheel speed sensor: A magnetoresistive wheel speed sensor is used, which is installed on the wheel hub of each wheel and outputs four-channel wheel speed pulse signals. After differential calculation by the ECU, the longitudinal speed of the vehicle, v_raw, is obtained.

[0049] Inertial Measurement Unit (IMU): A six-axis IMU (three-axis acceleration + three-axis angular velocity) is installed on the chassis crossbeam near the vehicle's center of gravity. This invention primarily uses the longitudinal acceleration a_x, processed by moving average, output from the IMU's longitudinal channel, for closed-loop feedback control in the creep suppression phase.

[0050] Four-corner axle load cells: These are resistance strain gauge load cells installed at the ends of the piston rods of the hydraulic cylinders of each outrigger. The sum of the four points is the total current weight of the vehicle. Dividing this weight by the gravitational acceleration g yields the equivalent mass m_eff, which is the equivalent mass actually involved in the longitudinal dynamics.

[0051] Slope sensor: An electronic tilt sensor based on a MEMS accelerometer is installed in the middle of the vehicle frame longitudinal beam, outputting the longitudinal slope percentage G_road for slope compensation calculation. G_road = 100 × tan(θ) ≈ 100 × sin(θ) (small angle approximation). The gravity component along the slope direction is generated as F_grade = m_eff·g·G_road / 100.

[0052] The decision-making layer consists of a host computer and a chassis domain controller (ECU).

[0053] The host computer is a high-performance in-vehicle computing platform that runs the autonomous driving operating system and is responsible for global path planning, behavior decision-making, and task scheduling. In this invention, the host computer's function is to send a "comfort parking enable" signal and the current equivalent mass parameter m_eff to the ECU via the CAN bus when it is a certain distance away from the parking point.

[0054] Chassis Controller (ECU): A highly reliable real-time controller running AUTOSAR CP or a similar real-time operating system, with a fixed control task cycle of 10 ms. The ECU internally runs the complete control algorithm of this invention, including: a capability prediction module, a five-state finite state machine (FSM), a comfort braking segment control module, a creep suppression segment control module, a parking hold segment control module, and non-volatile RAM (NVRAM) for state persistence.

[0055] The execution layer is responsible for receiving braking force commands from the ECU and executing braking actions.

[0056] Integrated electric EPB assembly: Composed of a permanent magnet synchronous motor (PMSM), planetary gearbox, ball screw / wedge force amplification mechanism, clamping force sensor, and rear-mounted mechanical locking mechanism. Rated clamping force F_max = 150,000 N. Receives braking force percentage command u_out (0~100%) from the ECU and outputs the corresponding clamping force.

[0057] Air Braking System (Backup): Employs a traditional pneumatic braking system, including an air compressor, air tank, brake valve, and brake chamber. It is physically isolated from the EPB and has its own independent controller, actuator, and power circuit. In the event of an EPB failure (communication timeout, motor overcurrent, sensor failure, etc.), the ECU switches to air braking control.

[0058] Communication bus

[0059] The system components communicate with each other via CAN. Key signals include: EBC_BrakeForceCmd (braking force command), EBC_VehicleMass (equivalent mass), EBC_CtrlState (control status), EBC_LongAccel (IMU acceleration), and WheelSpeed_FrontLeft / Right (wheel speed).

[0060] Example 2

[0061] like Figure 2 and Figure 3 As shown, the axle load-adaptive electric braking method in this embodiment uses an electronic control unit (ECU) as its core. Based on sensor signals (vehicle speed, gradient, mass, acceleration), it calculates and outputs braking force commands to the electric actuator (EPB) online, achieving a controlled deceleration-stopping process from a low-speed range to complete standstill. The method includes the following steps:

[0062] S1. Pre-operation mass acquisition: Before braking, the equivalent mass m_eff is acquired through the axle load sensor and sent to the chassis domain controller (ECU); m_eff = m_vehicle + m_cargo + m_rotational, which includes the vehicle body mass, load mass and the moment of inertia of rotating parts (wheels, drive shafts, etc.), unit: kg.

[0063] S2, Actuator Capability Prediction: Predicting the actuator capability before braking provides preventative safety protection.

[0064] Before transitioning from the IDLE state to the COMFORT_BRAKE state, the ECU performs a one-time pre-judgment check: u_check = (m_eff × a_target,max) / F_max × 100%; if u_check > 95%, it indicates that the EPB's capacity margin is insufficient, and the ECU refuses to enter the braking state and issues an alarm. This pre-judgment mechanism advances the risk of "insufficient force detected midway through braking" to "detection before braking even begins," which is a preventative safety design.

[0065] S3, Comfort Braking Section (5–1 km / h): Speed ​​feedforward control, low-pass filtering of wheel speed signal to obtain v_filtered; based on v_filtered, query the deceleration calibration point table, and use Smoothstep cubic polynomial interpolation to calculate target deceleration a_target(v); read road gradient G_road; calculate feedforward braking force command according to the core conversion formula u_ff = [m_eff ×(a_target + g × G_road / 100)] / F_max × 100%; process u_ff through an asymmetric slope limiter (increase rate Δu_up ≠ decrease rate Δu_dn) to obtain actual output u_out and send it to electric actuator EPB; when v_filtered ≤ 1 km / h, switch to creep suppression section; in this stage, control algorithm and mass parameters are decoupled.

[0066] The deceleration target curve is completely decoupled from the mass parameters—the deceleration curve is determined only by comfort requirements, and mass changes are fully reflected through the core conversion formula, without the need for recalibration.

[0067] The deceleration target curve is interpolated using Smoothstep cubic polynomials between adjacent calibration points (v_i, a_i) and (v_{i+1}, a_{i+1}):

[0068] Define the normalized velocity parameter t = (v - v_{i+1}) / (v_i - v_{i+1}), t ∈ [0, 1];

[0069] The Smoothstep function is S(t) = 3t² − 2t³, which satisfies that the first derivative at the endpoints is zero. It is used for deceleration mapping interpolation between calibration points, eliminating Jerk spikes from their mathematical source.

[0070] The deceleration interpolation a_target(v) = a_{i+1} + (a_i - a_{i+1}) × (3t² - 2t³).

[0071] Mathematical basis: S'(t) = 6t(1-t), at the endpoints S'(0) = 0 and S'(1) = 0. By Jerk chain decomposition J = da / dt = (da / dv) × (dv / dt) = (da / dv) × (-a), at the calibration point da / dv = 0 (because S' = 0), J = 0 is automatically guaranteed, thus eliminating the Jerk spike at the calibration point from a mathematical perspective.

[0072] Mass decoupling: Substituting the simplified longitudinal dynamic equation m_eff × (dv / dt) = -F_epb + m_eff × g × (G_road / 100) into the target deceleration a_target (taking dv / dt = -a_target), and combining the linear characteristic of EPB F_epb = (u / 100) × F_max, the core conversion formula is derived:

[0073] u_ff = [m_eff × (a_target + g × G_road / 100)] / F_max × 100 (%)

[0074] When the crane switches from light load to heavy load (mass change of 161%), or switches to a model with different baseline parameters, only one parameter, m_eff, needs to be updated. The deceleration curve itself and all control gains do not need to be recalibrated.

[0075] Asymmetric ramp limiter: applies an independent upper limit constraint on the change of braking force command u_out in each control cycle: the values ​​of increasing direction Δu_up and decreasing direction Δu_dn are different (Δu_up ≠ Δu_dn) to balance braking response speed and smoothness.

[0076] Regardless of how the target command u_ff calculated by the feedforward changes, the change in the actual output u_out within each control cycle is rigidly limited to Δu_up (increase direction) and Δu_dn (decrease direction), and Δu_up ≠ Δu_dn.

[0077] Δu_up = 2.0% / cycle, Δu_dn = 1.5% / cycle (control cycle 10 ms), meaning the rate of increase in the direction is greater than the rate of decrease in the direction. The engineering basis for the asymmetric design is that the establishment of braking force requires positive drive of the motor, and a larger rate of increase allows for a moderately fast response; while the reduction of braking force is more sensitive to comfort (the human body perceives a decrease in force more strongly than an increase in force), and a slightly slower rate of reduction helps to avoid the jerky feeling of "releasing and re-clamping".

[0078] The ramp limiter and Smoothstep interpolation work together: Smoothstep eliminates Jerk spikes caused by derivative discontinuities at calibration points, while the ramp limiter eliminates abrupt changes in braking force commands. Together, they protect the upper limit of Jerk from two aspects: "curve design" and "command limiting".

[0079] Jerk upper limit quantitative verification: The maximum increase in braking force per cycle ΔF_max = Δu_up × F_max, the equivalent deceleration increase Δa = ΔF_max / m_eff, corresponding to Jerk J_max = Δa / T_s. Calculated with typical parameters, J_max ≈ 0.61 g / s, which is lower than the boom protection threshold of 1.5 g / s and the comfort limit of 1.0 g / s in GB / T 7258.

[0080] S4. Creep Suppression Phase (1–0.3 km / h): Wheel speed signals are abandoned, and the longitudinal acceleration of the inertial measurement unit (IMU) is used as the feedback variable. A pure proportional (P) controller is employed for closed-loop control: u(k) = u(k-1) + K_p × [a_target -a_measured(k)] × T_s; T_s is the control cycle, the execution cycle of the control task, which is 10 ms in this embodiment. The output is limited to [u_min, u_max] dynamically scaled according to m_eff. When v_filtered ≤ 0.3 km / h and continues for at least M control cycles, the system switches to the stop-hold phase. During this phase, the sensor source and control algorithm are switched based on signal quality.

[0081] The output command limit of the creep suppression section is not a fixed value, but is dynamically calculated based on the equivalent mass:

[0082] u_min = (m_eff × α_low) / F_max × 100 %

[0083] u_max = (m_eff × α_high) / F_max × 100 %

[0084] α_low and α_high are preset acceleration boundary parameters (in m / s²). When the equivalent mass changes, the upper and lower limits scale proportionally to maintain the same force / mass ratio. The lower limit u_min ensures that the braking force will not excessively retract and cause the vehicle to slip when the IMU value is abnormally low, while the upper limit u_max protects the EPB motor from overheating due to continuous high load.

[0085] S5, Parking Holding Phase (<0.3 km / h and after continuous de-jittering): The braking force command is linearly increased from the end value of the creep phase to 100% at a rate of Δu_lock; a mechanical locking command is sent to activate the EPB mechanical locking mechanism; the locked state is maintained after waiting for the locking confirmation signal. During this phase, the mechanical locking provides power-off holding capability.

[0086] This embodiment adopts a differentiated signal processing strategy to address the differences in signal quality characteristics across different speed ranges:

[0087] Comfort braking mode: A first-order low-pass digital filter (cutoff frequency f_c = 5 Hz) is applied to the wheel speed signal, with a filter coefficient α = 1 - e^(-2πf_cT_s), providing an attenuation of approximately -12 dB in the suspension vibration frequency band of 0.8 to 2 Hz;

[0088] Creep suppression section: Apply N-frame moving average processing (window 50 ms) to the longitudinal acceleration of the IMU, providing approximately -6 dB attenuation for suspension vibration pseudo-signals, and has a stronger effect on suppressing spike pulses than the first-order low-pass filter.

[0089] like Figure 2 As shown, the five-state finite state machine (FSM) of this embodiment includes five states: IDLE, COMFORT_BRAKE, CREEP_SUPPRESS, STANDSTILL, and FAULT.

[0090] Normal flow follows the "one-way irreversible" principle: IDLE → COMFORT_BRAKE → CREEP_SUPPRESS → STANDSTILL. Once a state is entered, it cannot be reversed. This design avoids repeated state transitions caused by instantaneous signal fluctuations. The FAULT state can be entered from any state, providing a global safety net.

[0091] The physical basis for the phase switching is as follows: when the vehicle speed is below 1 km / h, the wheel speed pulse frequency drops to about 5-10 Hz, which severely overlaps with the suspension vibration frequency band (0.8-2 Hz), and the signal-to-noise ratio deteriorates to the 10:1 level. Therefore, the creep suppression phase abandons the wheel speed signal and switches to IMU longitudinal acceleration closed-loop control to achieve "signal quality-driven controller switching".

Claims

1. An electric control braking method based on shaft load adaptation, characterized in that, Includes the following steps: S1. Obtain the vehicle's equivalent mass m_eff before braking; S2. Based on the equivalent mass m_eff, predict the actuator's capability and determine whether the electric actuator has the ability to output braking force. If the capability is insufficient, refuse to enter the braking state and issue an alarm. S3. Entering the comfort braking phase, the wheel speed signal is low-pass filtered to obtain v_filtered. Based on v_filtered, the deceleration calibration point table is consulted, and the target deceleration a_target(v) is calculated using Smoothstep cubic polynomial interpolation. The road gradient G_road is read. The feedforward braking force command is calculated according to the formula u_ff = [m_eff × (a_target + g × G_road / 100)] / F_max ×100%, where F_max is the rated clamping force of the electric actuator. The feedforward braking force command is processed by the asymmetric slope limiter and then output to the electric actuator to perform braking. When v_filtered ≤ 1 km / h, the system switches to the creep suppression phase. S4. Enter the creep suppression phase, abandon the wheel speed signal, use the longitudinal acceleration of the inertial measurement unit as the feedback variable for closed-loop control, and output braking force command to the electric actuator; when v_filtered ≤ 0.3 km / h and lasts for no less than M control cycles, switch to the stop holding phase; S5. Enter the parking holding phase, increase the braking force command to the preset holding value, and activate the mechanical locking mechanism of the electric actuator.

2. The electric control braking method based on shaft load adaptation according to claim 1, characterized in that, The flow of the control process follows the principle of unidirectional irreversibility: idle → comfort braking → creep inhibition → stop holding, and cannot be reversed.

3. The electric control braking method based on shaft load adaptation according to claim 1, characterized in that, The Smoothstep cubic polynomial interpolation is specifically defined as follows: the normalized velocity parameter t = (v - v_{i+1}) / (v_i - v_{i+1}), t ∈ [0, 1]; the deceleration interpolation a_target(v) = a_{i+1} + (a_i - a_{i+1}) ×(3t² - 2t³), where (v_i, a_i) and (v_{i+1}, a_{i+1}) are adjacent deceleration calibration points.

4. The electric control braking method based on shaft load adaptation according to claim 1, characterized in that, The asymmetric ramp limiter imposes a direction-independent upper limit constraint on the change in the feedforward braking force command in each control cycle: the upper limit of the rate of change in the increasing direction, Δu_up, is different from the upper limit of the rate of change in the decreasing direction, Δu_dn.

5. The electric control braking method based on shaft load adaptation according to claim 1, characterized in that, The closed-loop control is a pure proportional control, and the control law is: u(k) = u(k-1) + K_p × [a_target - a_measured(k)] × T_s; where K_p is the proportional gain and T_s is the control period.

6. The electric control braking method based on shaft load adaptation according to claim 5, characterized in that, The output command limit of the creep suppression section is dynamically calculated based on the equivalent mass: u_min = (m_eff × α_low) / F_max × 100 % u_max = (m_eff × α_high) / F_max × 100 % Here, α_low and α_high are preset acceleration boundary parameters in m / s². When the equivalent mass changes, the upper and lower limits are scaled proportionally to maintain the same force / mass ratio.

7. The electric control braking method based on shaft load adaptation according to claim 1, characterized in that, The equivalent mass m_eff is obtained by measuring the axle load at the four corners using axle load sensors installed at the piston rod ends of the hydraulic cylinders of each outrigger. The sum of the four points is the current total weight of the vehicle, which is then divided by the gravitational acceleration g.

8. The electric control braking method based on shaft load adaptation according to claim 1, characterized in that, The actuator capability prediction is specifically as follows: calculate the prediction value u_check = (m_eff × a_target, max) / F_max × 100%; where a_target is the preset target deceleration; if u_check > preset safety threshold, then the electric actuator is determined to have insufficient capability margin.

9. An electrically controlled braking system based on shaft load adaptation, characterized in that, include: The perception layer is used to collect vehicle motion status and environmental parameters, including wheel speed sensors for measuring wheel speed, inertial measurement units for measuring longitudinal acceleration, axle load sensors for measuring axle load to obtain equivalent mass, and slope sensors for measuring road gradient. The decision-making layer includes a host computer and a chassis controller, wherein the chassis controller runs the electric control braking method based on axle load adaptation as described in any one of claims 1 to 8; The execution layer includes electric actuators for receiving braking force commands from the chassis controller and performing braking actions.

10. The electric braking system based on shaft load adaptation according to claim 9, characterized in that, The execution layer also includes a pneumatic braking system as a backup actuator. The pneumatic braking system is physically isolated from the electric actuator and has an independent controller, actuator and power circuit. When the electric actuator fails, the chassis controller switches to the pneumatic braking system to take over the braking.

Images