A vehicle longitudinal motion intention recognition and torque pre-control method and a control system thereof

Abstract

The application discloses a vehicle longitudinal motion intention recognition and torque pre-control method and a control system thereof. The control method comprises the following steps: collecting brake pedal stroke signals, brake pedal release rates, accelerator pedal opening degree signals and accelerator pedal response time delays in real time, and calculating brake-accelerator interaction timing characteristic parameters; according to the timing characteristic parameters, classifying driver longitudinal motion intentions into several grades; for different intention grades, calling pre-stored torque output curve templates to generate target torque instructions; sending the target torque instructions to a torque execution layer after low-pass filtering, and performing closed-loop correction according to actual vehicle speed feedback. The control method is a vehicle longitudinal motion intention recognition and torque pre-control method based on accelerator-brake interaction timing characteristics. By analyzing the timing characteristics of brake release and accelerator response, the driver intention can be accurately recognized, the adaptive output of torque is realized, and the vehicle can stably run in various working conditions.

Description

Technical Field

[0001] This invention relates to the field of vehicle power control technology, and in particular to a method and control system for recognizing the longitudinal motion intention of a vehicle and pre-controlling torque. Background Technology

[0002] During vehicle start-up and low-speed crawling, the power domain controller needs to output appropriate drive torque according to the driver's intention; in existing technologies, the following two solutions are typically used:

[0003] Option 1: Fixed Torque Creep. After the vehicle is engaged in Drive (D), the controller outputs a fixed torque (e.g., 10-15 Nm) to make the vehicle move slowly. The problem with this option is that: when starting on a flat road, the torque is too high, causing the vehicle to lurch forward (acceleration can reach 1.0-1.5 m / s²); when starting on an incline, the torque is insufficient, causing the vehicle to roll backward.

[0004] Option 2: Slope-based torque compensation. The slope angle is obtained via IMU, and the compensation torque is calculated using the formula T = mg·sinθ·r / i. The problem with this option is that the slope calculation is delayed and noisy, and it completely ignores the driver's actual intentions—the driver may only want to move slowly or may want to start quickly, and a fixed compensation strategy cannot distinguish between these.

[0005] The existing technology has the following shortcomings: the existing solutions are all "one-way" - they only look at the slope or only look at the accelerator opening, and no one combines the timing relationship between the two actions of "how to release the brake" and "how to press the accelerator" for analysis; for example, the closed-loop control method based on the driver's longitudinal acceleration intention disclosed in patent CN113147734A; in fact, experienced drivers and novice drivers release the brake and press the accelerator in completely different ways, and this difference contains the most direct information about the driver's intention. Summary of the Invention

[0006] To address the shortcomings of existing technologies, this invention provides a method and control system for recognizing vehicle longitudinal motion intention and pre-controlling torque, in order to accurately recognize the driver's intention and achieve graded adaptive torque output.

[0007] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is as follows:

[0008] This application provides a method for recognizing the longitudinal motion intention of a vehicle and pre-controlling torque, including the following steps:

[0009] S1: Real-time acquisition of brake pedal travel signal, brake pedal release rate, accelerator pedal opening signal and accelerator pedal response delay, and calculation of brake-accelerator interaction timing characteristic parameters;

[0010] S2: Based on the aforementioned temporal characteristic parameters, the driver's longitudinal movement intention is divided into several levels;

[0011] S3: For different intent levels, call the pre-stored torque output curve template to generate the target torque command;

[0012] S4: The target torque command is sent to the torque execution layer after being low-pass filtered, and closed-loop correction is performed based on the actual vehicle speed feedback.

[0013] Further:

[0014] In step S2, the driver's longitudinal movement intention is divided into four levels: slow start intention, normal start intention, rapid start intention, and risk of rolling back.

[0015] In step S1, the brake-throttle interaction timing characteristic parameters include:

[0016] Brake release rate V_brk is defined as the average release speed during the process of the brake pedal travel decreasing from 90% to 10%, and the unit is % / s;

[0017] Throttle response delay T_delay is defined as the time interval from when the brake pedal travel drops to 10% to when the accelerator pedal opening first exceeds 3%, in milliseconds.

[0018] The timing overlap O_lap is defined as the duration during which the brake pedal travel is less than 10% and the accelerator pedal opening is greater than 3%, in milliseconds.

[0019] In step S2, the logic for determining the intent level is as follows:

[0020] When V_brk < 20% / s, T_delay > 300ms, and O_lap = 0, it is determined to be a slow start intention;

[0021] When 20% / s ≤ V_brk ≤ 60% / s, 100ms ≤ T_delay ≤ 300ms, and O_lap < 100ms, it is determined to be a normal start-up intention;

[0022] When V_brk > 60% / s, T_delay < 100ms, and O_lap ≥ 100ms, it is determined to be an intention to start abruptly.

[0023] When V_brk > 30% / s, T_delay > 500ms, and the throttle opening is consistently below 3%, it is determined to be an intention to cause the vehicle to roll away.

[0024] In step S3, the torque output strategy corresponding to different intent levels is as follows:

[0025] Slow start intention: The target torque increases slowly in an S-shaped curve, with an increase gradient not exceeding 15 Nm / s, and the final torque is 80% of the torque requested by the driver;

[0026] Normal start-up intention: The target torque increases along a linear curve with an increase gradient of 30-50 Nm / s, and the final torque is equal to the torque requested by the driver;

[0027] Intended rapid start: The target torque rises rapidly according to an exponential curve, with an increase gradient of 80-120 Nm / s, and the final torque is 110% of the torque requested by the driver. The duration does not exceed 0.5 seconds before returning to the requested torque.

[0028] Risk of runaway vehicle: Actively apply anti-runaway torque, with a torque value of T_hold=m·g·sinθ·r / i·k, where θ is the real-time slope angle and k is the safety factor of 1.2-1.5.

[0029] In step S4, the closed-loop correction uses an incremental PID controller.

[0030] The incremental PID controller strategy is as follows:

[0031] ΔT(k)=Kp·e(k)+Ki·∑e(j)+Kd·[e(k)-e(k-1)]

[0032] Where e(k) is the deviation between the target vehicle speed and the actual vehicle speed, Kp = 1.5, Ki = 0.3, Kd = 0.05, the PID output is superimposed on the target torque command and then filtered by a first-order low-pass filter with a filtering time constant τ = 50ms.

[0033] This application provides a vehicle longitudinal motion intention recognition and torque pre-control system for implementing the method, comprising:

[0034] The signal acquisition module is used to acquire brake pedal travel, accelerator pedal opening and vehicle speed signals in real time;

[0035] The timing feature calculation module is used to calculate the brake release rate, throttle response delay, and timing overlap.

[0036] The intent recognition module is used to determine the driver's intent level based on temporal feature parameters;

[0037] The torque template matching module is used to call the corresponding torque output curve template according to the intent level;

[0038] The torque execution module is used to send the target torque to the motor controller after filtering and correction.

[0039] The sampling period of the signal acquisition module is 10ms, the determination period of the intent recognition module is 50ms, and the output period of the torque execution module is 20ms.

[0040] Compared with the prior art, the present invention has the following advantages:

[0041] The vehicle longitudinal motion intention recognition and torque pre-control method is based on the timing characteristics of throttle-brake interaction. By analyzing the timing characteristics of brake release and throttle response, it accurately identifies the driver's intention and achieves graded adaptive torque output, ensuring smooth vehicle operation under various working conditions. Attached Figure Description

[0042] The following is a brief explanation of the contents of each of the accompanying drawings and the markings in the drawings:

[0043] Figure 1 This is a schematic diagram illustrating the decision-making logic for determining the intent of this invention.

[0044] Figure 2 This is a schematic diagram of the normal starting torque output of the present invention. Detailed Implementation

[0045] The specific embodiments of the present invention will be further described in detail below with reference to the accompanying drawings and through the description of the examples.

[0046] Although the invention has been shown and described herein with reference to specific embodiments, it is not intended to be limited to the details shown. Rather, various modifications in detail may be made within the equivalent scope and scope of the claims without departing from the invention. In the drawings, the same item numbers refer to the same elements.

[0047] Existing solutions are all "one-way"—either only looking at the slope or only looking at the accelerator opening, and no one combines the timing relationship between the actions of "how to release the brake" and "how to press the accelerator" for analysis; in fact, experienced drivers and novices release the brake and press the accelerator in completely different ways, and this difference contains the most direct information about the driver's intention.

[0048] like Figure 1 As shown, this application provides a method for vehicle longitudinal motion intention recognition and torque pre-control based on throttle-brake interaction timing characteristics. By analyzing the timing characteristics of brake release and throttle response, the method accurately identifies the driver's intention and achieves hierarchical adaptive torque output.

[0049] A driver's starting intention is reflected not only in "how much the accelerator is pressed," but also in "how the brake is released and when the accelerator is pressed again." This invention quantifies this temporal characteristic, establishes an intention hierarchy model, and achieves "personalized" torque output.

[0050] The specific method for vehicle longitudinal motion intention recognition and torque pre-control based on throttle-brake interaction timing characteristics includes the following steps:

[0051] S1: Real-time acquisition of brake pedal travel signal, brake pedal release rate, accelerator pedal opening signal and accelerator pedal response delay, and calculation of brake-accelerator interaction timing characteristic parameters;

[0052] S2: Based on the aforementioned timing characteristic parameters, the driver's longitudinal movement intention is divided into four levels: slow start intention, normal start intention, rapid start intention, and intention to avoid rolling back.

[0053] S3: For different intent levels, call the pre-stored torque output curve template to generate the target torque command;

[0054] S4: The target torque command is sent to the torque execution layer after being low-pass filtered, and closed-loop correction is performed based on the actual vehicle speed feedback.

[0055] Its braking-throttle interaction timing characteristic parameters include:

[0056] Brake release rate V_brk is defined as the average release speed during the process of the brake pedal travel decreasing from 90% to 10%, and the unit is % / s;

[0057] Throttle response delay T_delay is defined as the time interval from when the brake pedal travel drops to 10% to when the accelerator pedal opening first exceeds 3%, in milliseconds.

[0058] The timing overlap O_lap is defined as the duration during which the brake pedal travel is less than 10% and the accelerator pedal opening is greater than 3%, in milliseconds.

[0059] The logic for determining the level of intent is as follows:

[0060] When V_brk < 20% / s, T_delay > 300ms, and O_lap = 0, it is determined to be a slow start intention;

[0061] When 20% / s ≤ V_brk ≤ 60% / s, 100ms ≤ T_delay ≤ 300ms, and O_lap < 100ms, it is determined to be a normal start-up intention;

[0062] When V_brk > 60% / s, T_delay < 100ms, and O_lap ≥ 100ms, it is determined to be an intention to start abruptly.

[0063] When V_brk > 30% / s, T_delay > 500ms, and the throttle opening is consistently below 3%, it is determined to be an intention to cause the vehicle to roll away.

[0064] The torque output strategies corresponding to its different intent levels are as follows:

[0065] Slow start intention: The target torque increases slowly in an S-shaped curve, with an increase gradient not exceeding 15 Nm / s, and the final torque is 80% of the torque requested by the driver;

[0066] Normal start-up intention: The target torque increases along a linear curve with an increase gradient of 30-50 Nm / s, and the final torque is equal to the torque requested by the driver;

[0067] Intended rapid start: The target torque rises rapidly according to an exponential curve, with an increase gradient of 80-120 Nm / s, and the final torque is 110% of the torque requested by the driver. The duration does not exceed 0.5 seconds before returning to the requested torque.

[0068] Risk of runaway vehicle: Actively apply anti-runaway torque, with a torque value of T_hold=m·g·sinθ·r / i·k, where θ is the real-time slope angle and k is the safety factor of 1.2-1.5.

[0069] The closed-loop correction uses an incremental PID controller; its incremental PID controller strategy is as follows:

[0070] ΔT(k)=Kp·e(k)+Ki·∑e(j)+Kd·[e(k)-e(k-1)]

[0071] Where e(k) is the deviation between the target vehicle speed and the actual vehicle speed, Kp = 1.5, Ki = 0.3, Kd = 0.05, the PID output is superimposed on the target torque command and then filtered by a first-order low-pass filter with a filtering time constant τ = 50ms.

[0072] This application provides a vehicle longitudinal motion intention recognition and torque pre-control system for implementing the method, comprising:

[0073] The signal acquisition module is used to acquire brake pedal travel, accelerator pedal opening and vehicle speed signals in real time;

[0074] The timing feature calculation module is used to calculate the brake release rate, throttle response delay, and timing overlap.

[0075] The intent recognition module is used to determine the driver's intent level based on temporal feature parameters;

[0076] The torque template matching module is used to call the corresponding torque output curve template according to the intent level;

[0077] The torque execution module is used to send the target torque to the motor controller after filtering and correction.

[0078] The sampling period of the signal acquisition module is 10ms, the determination period of the intent recognition module is 50ms, and the output period of the torque execution module is 20ms.

[0079] The vehicle longitudinal motion intention recognition and torque pre-control method of this application is based on the timing characteristics of throttle-brake interaction. By analyzing the timing characteristics of brake release and throttle response, it accurately identifies the driver's intention and realizes graded adaptive torque output, so that the vehicle can drive smoothly under various operating conditions; as shown in Table 1, which compares the improvement of this application with traditional control methods.

[0080]

[0081] Table 1 compares the improvements of this application compared to traditional control methods.

[0082] This solution breaks away from the traditional method of judging the driver's starting intention solely based on the accelerator pedal opening. Instead, it uses the interactive timing characteristics of the brake pedal release and accelerator pedal response as the basis for judgment. It quantifies three key parameters: the speed of brake release, the accelerator lag time, and the pedal operation overlap time. This allows for graded recognition of the driver's longitudinal starting intention, matching differentiated torque output curves to achieve personalized torque pre-control tailored to each individual driver, while also ensuring safety on slopes and smooth driving.

[0083] The vehicle longitudinal motion intent recognition and torque pre-control method of this application specifically involves four execution steps:

[0084] Feature parameter acquisition and calculation: Braking stroke, brake release rate, throttle opening, and throttle response delay are acquired at 10ms intervals, and three major time-series feature parameters, namely brake release rate, throttle response delay, and time-series overlap, are calculated.

[0085] Intent classification and determination: Based on parameter threshold boundaries, the intent is divided into four categories: slow start, normal start, rapid start, and risk of rollback. Each category is set with clear determination conditions for rate, delay, and overlap duration.

[0086] Graded torque template output: Match exclusive torque rise curves, gradients and amplitudes for different intentions: Slow start adopts S-shaped gradual rise and fall power output, normal start linear full output, rapid start index short-term overpower burst followed by drop, and the risk of rolling back is directly substituted into the slope formula to calculate the parking anti-rollback torque.

[0087] Closed-loop filter correction execution: The target torque is sent out after a 50ms first-order low-pass filter. It is then compensated by a fixed-parameter incremental PID controller based on the vehicle speed deviation in a closed-loop manner. Finally, the torque execution layer outputs the torque to the motor controller at a 20ms cycle.

[0088] Example 1: Detailed explanation of intent recognition logic, such as Figure 1 As shown;

[0089] This invention defines three key timing parameters:

[0090] Parameter 1: Brake release rate V_brk:

[0091] Brake release rate reflects the "quickness" of the driver in releasing the brake; calculation method:

[0092] V_brk = (S_brk_90% - S_brk_10%) / (t_90% - t_10%);

[0093] Where S_brk_90% is the moment corresponding to 90% of the brake pedal travel, and S_brk_10% is the moment corresponding to 10% of the travel;

[0094] For novice drivers releasing the brake: V_brk ≈ 10-20% / s (very slow, to avoid the car moving).

[0095] Normal brake release: V_brk ≈ 30-50% / s;

[0096] Sudden release of the brake: V_brk ≈ 70-100% / s (Experienced drivers release the brake completely in one go).

[0097] Parameter 2: Throttle response delay T_delay:

[0098] T_delay = t_acc_3% - t_brk_10%;

[0099] That is, the time difference between when the brake is released to 10% and when the accelerator first exceeds 3%;

[0100] Slowly follow the oil flow: T_delay > 300ms;

[0101] Normal oil flow: T_delay ≈ 100-300ms;

[0102] Start-up with clutch engaged: T_delay < 100ms (the accelerator is pressed before the brake is fully released).

[0103] Parameter 3: Timing overlap O_lap:

[0104] O_lap = ∫[S_brk(t)<10% ∩ S_acc(t)>3%]dt;

[0105] The overlap time between when the brake has been fully released but the accelerator has been pressed;

[0106] Normal start: O_lap ≈ 0 (release first, then press, without overlap);

[0107] Launch start: O_lap > 100ms (releasing the pedal while pressing it, or even pressing it first and then releasing it).

[0108] Example 2: Torque output curve template, such as Figure 2 As shown;

[0109] This invention pre-stores four torque output curve templates, taking normal start-up as an example:

[0110] Slow start: slope 15 Nm / s, peak torque 80% of requested torque, lasting 2 seconds;

[0111] Normal start: slope 40 Nm / s, peak torque = requested torque, lasting 0.5 s;

[0112] Rapid start: slope 100 Nm / s, peak torque = requested torque × 110%, linear retraction after 0.3 s;

[0113] Risk of rollback: Constant torque T_hold until throttle opening >5% or vehicle speed >0.5m / s.

[0114] Example 3: Proactive Intervention for Car Runaway Risk:

[0115] When a risk of rollback is detected, the system does not simply output a fixed torque, but instead:

[0116] T_hold = m·g·sinθ·(r / i)·k_safety;

[0117] in:

[0118] m: Vehicle weight (obtained from CAN bus);

[0119] θ: Real-time slope angle (after IMU filtering, sampling period 50ms);

[0120] r / i: The ratio of the final drive ratio to the tire radius;

[0121] k_safety: Safety factor, dynamically adjusted based on the road surface adhesion coefficient;

[0122] Methods for estimating the road surface adhesion coefficient μ:

[0123] μ = min(1.0, T_wheel / (m·g·cosθ·r));

[0124] When μ < 0.3 (on icy or snowy roads), k_safety is increased from 1.2 to 1.5 to prevent excessive torque from causing wheel slippage.

[0125] Example 4: Closed-loop correction strategy:

[0126] Even if the intent is correctly recognized, the actual vehicle speed may still deviate from the expected speed (e.g., suddenly going uphill). Therefore, this invention superimposes an incremental PID closed-loop on top of the open-loop torque command:

[0127] The target vehicle speed v_target is determined by the intent level:

[0128] Slow start: v_target = 0.3 m / s;

[0129] Normal start-up: v_target = 0.8 m / s;

[0130] Rapid start: v_target = 1.5 m / s;

[0131] The PID parameters were calibrated on a real vehicle: Kp=1.5, Ki=0.3, Kd=0.05, with an output limit of ±30Nm to prevent excessive correction from causing secondary impact.

[0132] The PID output is filtered by a first-order low-pass filter (τ=50ms) and then superimposed onto the open-loop torque, resulting in a final output cycle of 20ms.

[0133] This invention presents a method for recognizing vehicle longitudinal motion intent and pre-controlling torque based on the timing characteristics of throttle-brake interaction. By real-time acquisition of brake pedal release rate, accelerator pedal response delay, and the timing overlap characteristics of the two, the driver's starting intent is classified into four levels: slow start, normal start, rapid start, and risk of rollback. A mapping relationship between the intent level and the torque output curve is established to achieve graded pre-control of torque. This invention requires no additional hardware and is implemented through software algorithms based solely on existing sensor signals in the power domain. It effectively solves the problems of single starting torque response and easy lurching or rollback in traditional creep control, as shown in the comparison table 2 below. Real vehicle tests show that this method reduces starting acceleration fluctuation by 62%, reduces the rollback distance on a 5° slope to zero, and improves the driver's subjective comfort score by 38%.

[0134]

[0135] The present invention has been described above by way of example with reference to the accompanying drawings. Obviously, the specific implementation of the present invention is not limited to the above-described manner. Any non-substantial improvements made using the concept and technical solution of the present invention, or the direct application of the concept and technical solution of the present invention to other occasions without modification, are all within the protection scope of the present invention.

Claims

1. A method for recognizing the longitudinal motion intention of a vehicle and pre-controlling torque, characterized in that: Includes the following steps: S1: Real-time acquisition of brake pedal travel signal, brake pedal release rate, accelerator pedal opening signal and accelerator pedal response delay, and calculation of brake-accelerator interaction timing characteristic parameters; S2: Based on the aforementioned temporal characteristic parameters, the driver's longitudinal movement intention is divided into several levels; S3: For different intent levels, call the pre-stored torque output curve template to generate the target torque command; S4: The target torque command is sent to the torque execution layer after being low-pass filtered, and closed-loop correction is performed based on the actual vehicle speed feedback.

2. The vehicle longitudinal motion intention recognition and torque pre-control method as described in claim 1, characterized in that: In step S2, the driver's longitudinal movement intention is divided into four levels: slow start intention, normal start intention, rapid start intention, and risk of rolling back.

3. The vehicle longitudinal motion intention recognition and torque pre-control method as described in claim 2, characterized in that: In step S1, the brake-throttle interaction timing characteristic parameters include: Brake release rate V_brk is defined as the average release speed during the process of the brake pedal travel decreasing from 90% to 10%, and the unit is % / s; Throttle response delay T_delay is defined as the time interval from when the brake pedal travel drops to 10% to when the accelerator pedal opening first exceeds 3%, in milliseconds. The timing overlap O_lap is defined as the duration during which the brake pedal travel is less than 10% and the accelerator pedal opening is greater than 3%, in milliseconds.

4. The vehicle longitudinal motion intention recognition and torque pre-control method as described in claim 3, characterized in that: In step S2, the logic for determining the intent level is as follows: When V_brk < 20% / s, T_delay > 300ms, and O_lap = 0, it is determined to be a slow start intention; When 20% / s ≤ V_brk ≤ 60% / s, 100ms ≤ T_delay ≤ 300ms, and O_lap < 100ms, it is determined to be a normal start-up intention; When V_brk > 60% / s, T_delay < 100ms, and O_lap ≥ 100ms, it is determined to be an intention to start abruptly. When V_brk > 30% / s, T_delay > 500ms, and the throttle opening is consistently below 3%, it is determined to be an intention to cause the vehicle to roll away.

5. The vehicle longitudinal motion intention recognition and torque pre-control method as described in claim 1, characterized in that: In step S3, the torque output strategy corresponding to different intent levels is as follows: Slow start intention: The target torque increases slowly in an S-shaped curve, with an increase gradient not exceeding 15 Nm / s, and the final torque is 80% of the torque requested by the driver; Normal start-up intention: The target torque increases along a linear curve with an increase gradient of 30-50 Nm / s, and the final torque is equal to the torque requested by the driver; Intended rapid start: The target torque rises rapidly according to an exponential curve, with an increase gradient of 80-120 Nm / s, and the final torque is 110% of the torque requested by the driver. The duration does not exceed 0.5 seconds before returning to the requested torque. Risk of runaway vehicle: Actively apply anti-runaway torque, with the torque value being T_hold=m·g·sinθ·r / i·k, where θ is the real-time slope angle and k is the safety factor of 1.2-1.

5.

6. The vehicle longitudinal motion intention recognition and torque pre-control method as described in claim 1, characterized in that: In step S4, the closed-loop correction uses an incremental PID controller.

7. The vehicle longitudinal motion intention recognition and torque pre-control method as described in claim 6, characterized in that: The incremental PID controller strategy is as follows: ΔT(k)=Kp·e(k)+Ki·∑e(j)+Kd·[e(k)-e(k-1)] Where e(k) is the deviation between the target vehicle speed and the actual vehicle speed, Kp = 1.5, Ki = 0.3, Kd = 0.05, the PID output is superimposed on the target torque command and then filtered by a first-order low-pass filter with a filtering time constant τ = 50ms.

8. A vehicle longitudinal motion intention recognition and torque pre-control system implementing the method as described in any one of claims 1 to 7, characterized in that: include: The signal acquisition module is used to acquire brake pedal travel, accelerator pedal opening, and vehicle speed signals in real time. The timing feature calculation module is used to calculate the brake release rate, throttle response delay, and timing overlap. The intent recognition module is used to determine the driver's intent level based on temporal feature parameters; The torque template matching module is used to call the corresponding torque output curve template according to the intent level; The torque execution module is used to send the target torque to the motor controller after filtering and correction.

9. The control system as described in claim 8, characterized in that: The sampling period of the signal acquisition module is 10ms, the determination period of the intent recognition module is 50ms, and the output period of the torque execution module is 20ms.

Images