Muscle stiffness prediction and haptic guidance based automatic takeover control method for autonomous driving

Abstract

The application is suitable for the technical field of automatic driving human-computer interaction, and provides an automatic driving takeover control method based on muscle stiffness prediction and tactile guidance, a takeover state recognition, a takeover request triggering and a takeover preparation time window determination, a target stiffness prediction, a target arm muscle stiffness and a target damping required by a driver at a predicted takeover time are calculated, a stiffness loading and a tactile guidance, a steering impedance torque is gradually loaded in the takeover preparation time window, and a high-frequency low-amplitude vibration signal is superimposed for tactile guidance, and a control handover, a driver state is judged at the predicted takeover time, and the control handover is completed after success. The method helps the driver to prepare for takeover through muscle stiffness prediction, and simultaneously introduces tactile guidance to actively help the driver to establish a correct control direction, shortens the takeover reaction time of the driver, and solves the problem of human-machine confrontation caused by the inconsistency between the control intention of the driver and the planning of the automatic driving system.

Description

Technical Field

[0001] This invention belongs to the field of autonomous driving human-computer interaction technology, and particularly relates to an autonomous driving takeover control method based on muscle stiffness prediction and tactile guidance. Background Technology

[0002] With the rapid development of autonomous driving technology, when encountering conditions exceeding the designed operating range, the driver needs to respond to a takeover request to regain control of the vehicle. However, in autonomous driving mode, the driver's hand muscles are in a relaxed state with low stiffness. When control is transferred instantaneously, the driver's relaxed muscle resistance cannot resist the steering reaction force, easily leading to "muscle mismatch" that causes steering wheel overshoot, vehicle deviation, or even rollover, seriously threatening driving safety. Although existing technical solutions have achieved basic takeover prompts, they still have the following limitations in practical applications:

[0003] Chinese invention patent CN120735798B, entitled "An Autonomous Driving Takeover Method and System Based on Cognitive-Execution Two-Level Decision-Making," introduces a multimodal large model for risk assessment to determine the takeover mode and guides the driver to prepare for takeover through multi-level warning signals of vision, hearing, and touch. It calculates the driver's readiness level based on driver state information such as eye movement trajectory and grip strength, and dynamically adjusts control weights according to the driver's readiness state to achieve a smooth transfer of human-machine control. However, this method does not consider the impact of changes in driver muscle stiffness on maneuverability in emergency situations, and the tactile warning remains a passive alert, lacking an active tactile guidance mechanism to guide driver operation.

[0004] The Chinese invention patent CN116975671B, entitled "Method and System for Assessing Driver Trust in Level 3 Autonomous Vehicles," assesses driver readiness by monitoring the driver's eye, hand, and foot states. It combines this with road traffic information risk assessment to calculate the compatibility between the two, providing visual cues through different dashboard light colors. However, this method relies on visual signals for interaction. The driver's interpretation of different signal levels of danger can increase reaction time during takeover attempts, making it difficult to quickly establish correct control intentions in emergency obstacle avoidance situations. Summary of the Invention

[0005] The purpose of this invention is to provide an autonomous driving takeover control method based on muscle stiffness prediction and tactile guidance, which aims to solve the problems mentioned in the background art.

[0006] The present invention is implemented as follows: an autonomous driving takeover control method based on muscle stiffness prediction and tactile guidance includes the following steps:

[0007] Step 1: Takeover status identification;

[0008] By monitoring the vehicle's autonomous driving status and surrounding environment in real time through onboard sensors, a takeover request is triggered when the current operating condition is determined to exceed the autonomous driving safety control capabilities, and the expected takeover time is determined. Simultaneously record the time when the takeover request was issued. Forming a takeover preparation time window ;

[0009] Step 2: Target stiffness prediction;

[0010] Obtain current vehicle driving condition data and calculate the driver's expected takeover time. Target arm muscle stiffness required for safe vehicle control and target damping ;

[0011] Step 3: Stiffness loading and haptic guidance;

[0012] During the takeover preparation window Internally, the steer-by-wire motor is controlled according to the target arm muscle stiffness. Gradual application of steering resistance torque Until the expected takeover time is reached. Simultaneously superimposed high-frequency low-amplitude vibration signals Tactile guidance is provided to the driver;

[0013] Step 4: Control Transfer;

[0014] At the expected takeover time It determines whether the driver has applied sufficient steering torque and gripped the steering wheel firmly. If the determination is successful, it completes the transfer of control, stops the tactile guidance vibration, and smoothly decays the steering system impedance parameters to manual driving mode.

[0015] A further technical solution, the specific steps of step 1 are as follows:

[0016] Risk distance for collecting obstacles or risk points ahead Current vehicle speed obstacle movement speed Static obstacles ; Calculate relative velocity and collision time :

[0017]

[0018]

[0019] Set collision time threshold When the collision time When the system determines that the autonomous driving safety control capabilities have been exceeded, a takeover request is triggered; the current moment is recorded. And calculate the expected takeover time. :

[0020]

[0021] in, This is the preset time for the handover of autonomous driving control.

[0022] According to the time the takeover request was issued and expected takeover time To form a preparation time window .

[0023] A further technical solution involves setting a collision time threshold in step 1. It is 4 seconds. Take 3.5 seconds.

[0024] A further technical solution, the specific steps of step 2 are as follows:

[0025] Based on current vehicle speed and road curvature Calculate the current operating condition risk index :

[0026]

[0027] in, These are the weighting coefficients;

[0028] Calculate the target arm muscle stiffness required for the driver to stably control the vehicle. ,as follows:

[0029]

[0030] in, Muscle stiffness in a relaxed state; Muscle stiffness under tension; This represents the current operating condition risk index. The slope coefficient; Risk center value;

[0031] Synchronous calculation of target damping :

[0032]

[0033] in, The damping ratio; This is the equivalent rotational inertia of the steering system.

[0034] A further technical solution is provided in step 2. Set the value to 1.0; Take 2.5 N·m / rad; Take 12 N·m / rad; Take 0.5; Take 3.0; Set the value to 1.0; Take 0.05 kg·m 2 .

[0035] A further technical solution, the specific steps of step 3 are as follows:

[0036] Step 3.1: Stiffness loading;

[0037] During the takeover preparation window Internally, set stiffness From the current basic stiffness linearly increases over time to Damping It changes accordingly;

[0038] Current time is ( The formulas for calculating virtual stiffness and damping are as follows:

[0039]

[0040]

[0041] Calculated stiffness With damping The input is fed into the impedance control model of the steer-by-wire motor to calculate the steering impedance torque. :

[0042]

[0043] in, The target steering wheel angle for planning a path for the autonomous driving system; The current steering wheel angle; This is the current steering wheel angular velocity;

[0044] Step 3.2: Tactile guidance;

[0045] Determining direction by road curvature, if the road curvature This means that a left turn is required, and a high-frequency, low-amplitude vibration signal is superimposed on the left side of the steering wheel. If the road curvature This means that a right turn is required, and a high-frequency, low-amplitude vibration signal is superimposed on the right side of the steering wheel. High-frequency low-amplitude vibration signal The expression is:

[0046]

[0047] in, The amplitude; For frequency.

[0048] A further technical solution is provided in step 3. Take 1.5 N·m; Select 30 Hz.

[0049] A further technical solution, the specific steps of step 4 are as follows:

[0050] Hand grip force is read via steering wheel pressure sensor. And calculate the rate of change of the torque applied by the driver to the steering wheel. :

[0051]

[0052] When hand grip strength Reaching the grip strength threshold Torque change rate Reaching the threshold of torque change rate and duration Reaching the collision time threshold At that time, it is determined to be a valid takeover, and the transfer of driving control is complete;

[0053] The steering system impedance parameters are smoothly attenuated to the stiffness characteristics of the conventional EPS assist mode, and tactile-guided vibrations are stopped.

[0054] The autonomous driving takeover control method based on muscle stiffness prediction and tactile guidance provided in this invention has the following beneficial effects:

[0055] (1) By predicting the target stiffness and loading the stiffness, the dynamic matching between the driver's muscle state and the steering resistance is realized, which effectively solves the problem of instability caused by "muscle force mismatch" at the moment of takeover.

[0056] (2) By superimposing directional tactile guidance signals, the driver's takeover reaction time is shortened and the conflict between human and machine control intentions is eliminated;

[0057] (3) By using impedance parameter smooth transition and dual determination of takeover status, the smooth and safe transfer of control is achieved;

[0058] (4) A complete closed-loop framework of “state recognition - stiffness prediction - tactile guidance - control handover” was constructed, which improved the overall performance of the autonomous driving takeover process. Attached Figure Description

[0059] Figure 1 A flowchart of an autonomous driving takeover control method based on muscle stiffness prediction and tactile guidance provided in an embodiment of the present invention;

[0060] Figure 2 A schematic diagram comparing virtual stiffness and driver muscle stiffness within the preparation time window for takeover;

[0061] Figure 3 A schematic diagram of the steering impedance torque loading during the pre-takeover preparation time window. Detailed Implementation

[0062] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0063] The specific implementation of the present invention will be described in detail below with reference to specific embodiments.

[0064] like Figure 1 As shown, an embodiment of the present invention provides an autonomous driving takeover control method based on muscle stiffness prediction and haptic guidance, which includes the following steps:

[0065] Step 1: Takeover status identification;

[0066] By monitoring the vehicle's autonomous driving status and surrounding environment in real time through onboard sensors, a takeover request is triggered when the current operating condition is determined to exceed the autonomous driving safety control capabilities, and the expected takeover time is determined. Simultaneously record the time when the takeover request was issued. Forming a takeover preparation time window The specific steps are as follows:

[0067] Risk distance for collecting obstacles or risk points ahead Current vehicle speed obstacle movement speed Static obstacles Calculate relative velocity and collision time :

[0068]

[0069]

[0070] Set collision time threshold When the collision time If the system is deemed to have exceeded the autonomous driving safety control capabilities, a takeover request is triggered. Record the current moment. And calculate the expected takeover time. :

[0071]

[0072] in, This is the preset handover time for autonomous driving control, set to 3.5 seconds.

[0073] According to the time the takeover request was issued and expected takeover time To form a preparation time window .

[0074] Step 2: Target stiffness prediction;

[0075] Acquire current vehicle operating condition data and calculate the driver's expected takeover time. Target arm muscle stiffness required for safe vehicle control and target damping The specific steps are as follows:

[0076] Based on current vehicle speed and road curvature Calculate the current operating condition risk index :

[0077]

[0078] in, This is the weighting coefficient, set to 1.0.

[0079] Calculate the target arm muscle stiffness required for the driver to stably control the vehicle. ,as follows:

[0080]

[0081] in, This is the muscle stiffness in a relaxed state, taken as 2.5 N·m / rad; This is the muscle stiffness under tension, taken as 12 N·m / rad; It is the current working condition risk index; This is the slope coefficient, taken as 0.5; This is the risk center value, set to 3.0.

[0082] Meanwhile, to prevent system oscillations caused by stiffness changes and to ensure the smoothness of the connection, the target damping needs to be calculated simultaneously. :

[0083]

[0084] in, This is the damping ratio, taken as 1.0; It is the equivalent rotational inertia of the steering system, a pre-calibrated inherent parameter of the system, taken as 0.05 kg·m. 2 .

[0085] Step 3: Stiffness loading and haptic guidance;

[0086] During the takeover preparation time window, control the steer-by-wire motor according to the target arm muscle stiffness. Gradual application of steering resistance torque Until the expected takeover time is reached. Simultaneously superimposed high-frequency, low-amplitude vibration signals. Tactile guidance is then provided to the driver. The specific steps are as follows:

[0087] Step 3.1: Stiffness loading;

[0088] During the takeover preparation window Internally, set stiffness From the current basic stiffness linearly increases over time to Damping The steering wheel will gradually become heavier, forcing the driver to subconsciously tighten their grip on the wheel, creating a state of muscle tension and preparing to take over.

[0089] Current time is ( The formulas for calculating virtual stiffness and damping are as follows:

[0090]

[0091]

[0092] Calculated stiffness With damping The input is fed into the impedance control model of the steer-by-wire motor to calculate the steering impedance torque. :

[0093]

[0094] in, It is the target steering wheel angle for the path planned by the autonomous driving system; It is the current steering wheel angle; It is the current steering wheel angular velocity.

[0095] Step 3.2: Tactile guidance;

[0096] While applying stiffness, a directional pulse vibration signal is superimposed through the steering wheel's built-in tactile feedback device. This superimposed tactile guiding torque on top of the impedance torque provides steering tactile guidance. The superimposed high-frequency, low-amplitude pulse signal assists the driver in quickly establishing the correct handling intention. Direction is determined by road curvature; if the road curvature... This means that a left turn is required, and a high-frequency, low-amplitude vibration signal is superimposed on the left side of the steering wheel. If the road curvature This means that a right turn is required, and a high-frequency, low-amplitude vibration signal is superimposed on the right side of the steering wheel. , The expression is:

[0097]

[0098] in, The amplitude is taken as 1.5 N·m; The frequency is set to 30 Hz.

[0099] Step 4: Control handover;

[0100] At the expected takeover time The control handover module determines whether the driver has applied sufficient steering torque and is gripping the steering wheel firmly. Upon successful determination, control is transferred, tactile guidance vibrations cease, and the steering system impedance parameters are smoothly attenuated to manual driving mode. The specific steps are as follows:

[0101] Hand grip force is read via steering wheel pressure sensor. And calculate the rate of change of the torque applied by the driver to the steering wheel. :

[0102]

[0103] When hand grip strength Reaching the grip strength threshold Torque change rate Reaching the threshold of torque change rate and duration Reaching the collision time threshold When the system is deemed to have taken over effectively and the transfer of driving control is complete, the steering system impedance parameters are gradually and smoothly reduced to the stiffness characteristics of the normal EPS power assist mode, and the tactile guidance vibration is stopped.

[0104] To verify the effectiveness of the autonomous driving takeover control method based on muscle stiffness prediction and tactile guidance proposed in this invention, this embodiment selects a wet and slippery road surface with limited curve stability as the verification scenario. The following test conditions are constructed in a driving simulator environment: The vehicle is traveling at 72 km / h on a wet and slippery road surface in autonomous driving mode, about to enter a left-turn ramp with a radius of R=100m. Since the onboard perception system detects the slippery road surface, the road adhesion coefficient is estimated to be 0.4. The lateral acceleration required by the vehicle exceeds the maximum lateral adhesion capacity provided by the tires. The system determines that the current condition poses a risk of sideslip, exceeding the autonomous driving safety control capability, and immediately triggers a takeover request. A schematic diagram comparing the virtual stiffness within the takeover preparation time window calculated by this method with the driver's muscle stiffness is shown below. Figure 2 As shown in the diagram, the steering impedance torque loading within the preparation time window is as follows: Figure 3 As shown.

[0105] like Figure 2 As shown, within the takeover preparation time window, the virtual stiffness increases linearly and smoothly, while the driver's muscle stiffness gradually increases from the initial relaxed state. The figure shows that in the initial stage of takeover (approximately t=0~0.5s), the virtual stiffness increases linearly according to the preset strategy, slightly exceeding the driver's initial muscle stiffness, which helps the driver establish grip force earlier. The driver's muscle stiffness rapidly increases from the relaxed state (approximately 2.5 N·m / rad), exhibiting a non-linear surge. In the middle and later stages of takeover, the driver's muscle stiffness continues to increase with the increase of virtual stiffness, and the trends of the two are basically consistent, indicating that the stiffness loading strategy designed in this invention can effectively match the driver's muscle adjustment process, achieving dynamic matching between the driver's muscle state and steering resistance.

[0106] like Figure 3 As shown, within the takeover preparation time window, the steering system resistance torque gradually increases over time, exhibiting a smooth upward trend, guiding the driver to gradually apply steering wheel torque and establish control intention. Simultaneously, the resistance torque change process is continuous without abrupt changes, effectively avoiding torque shocks and oscillations that may occur during traditional takeover processes, achieving a smooth and safe transfer of control.

[0107] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. An autonomous driving takeover control method based on muscle stiffness prediction and haptic guidance, characterized in that, Includes the following steps: Step 1: Takeover status identification; By monitoring the vehicle's autonomous driving status and surrounding environment in real time through onboard sensors, a takeover request is triggered when the current operating condition is determined to exceed the autonomous driving safety control capabilities, and the expected takeover time is determined. Simultaneously record the time when the takeover request was issued. Forming a takeover preparation time window ; Step 2: Target stiffness prediction; Obtain current vehicle driving condition data and calculate the driver's expected takeover time. Target arm muscle stiffness required for safe vehicle control and target damping ; Step 3: Stiffness loading and haptic guidance; During the takeover preparation window Internally, the steer-by-wire motor is controlled according to the target arm muscle stiffness. Gradual application of steering resistance torque Until the expected takeover time is reached. Simultaneously superimposed high-frequency low-amplitude vibration signals Tactile guidance is provided to the driver; Step 4: Control handover; At the expected takeover time It determines whether the driver has applied sufficient steering torque and gripped the steering wheel firmly. If the determination is successful, it completes the transfer of control, stops the tactile guidance vibration, and smoothly decays the steering system impedance parameters to manual driving mode. The specific steps of step 1 are as follows: Risk distance for collecting obstacles or risk points ahead Current vehicle speed obstacle movement speed Static obstacles ; Calculate relative velocity and collision time : Set collision time threshold When the collision time When the system determines that the autonomous driving safety control capabilities have been exceeded, a takeover request is triggered; the current moment is recorded. And calculate the expected takeover time. : in, This is the preset time for the handover of autonomous driving control. According to the time the takeover request was issued and expected takeover time To form a preparation time window ; The specific steps of step 2 are as follows: Based on current vehicle speed and road curvature Calculate the current operating condition risk index : in, These are the weighting coefficients; Calculate the target arm muscle stiffness required for the driver to stably control the vehicle. ,as follows: in, Muscle stiffness in a relaxed state; Muscle stiffness under tension; This represents the current operating condition risk index. The slope coefficient; Risk center value; Synchronous calculation of target damping : in, The damping ratio; This is the equivalent rotational inertia of the steering system.

2. The autonomous driving takeover control method based on muscle stiffness prediction and tactile guidance according to claim 1, characterized in that, In step 1, a collision time threshold is set. It is 4 seconds. Take 3.5 seconds.

3. The autonomous driving takeover control method based on muscle stiffness prediction and tactile guidance according to claim 1, characterized in that, In step 2, Set the value to 1.0; Take 2.5 N·m / rad; Take 12 N·m / rad; Take 0.5; Take 3.0; Set the value to 1.0; Take 0.05 kg·m 2 .

4. The autonomous driving takeover control method based on muscle stiffness prediction and tactile guidance according to claim 1, characterized in that, The specific steps of step 3 are as follows: Step 3.1: Stiffness loading; During the takeover preparation window Internally, set stiffness From the current basic stiffness linearly increases over time to Damping It changes accordingly; Current time is The formulas for calculating virtual stiffness and damping are as follows: Calculated stiffness With damping The input is fed into the impedance control model of the steer-by-wire motor to calculate the steering impedance torque. : in, The target steering wheel angle for planning a path for the autonomous driving system; The current steering wheel angle; This is the current steering wheel angular velocity; Step 3.2: Tactile guidance; Determining direction by road curvature, if the road curvature This means that a left turn is required, and a high-frequency, low-amplitude vibration signal is superimposed on the left side of the steering wheel. If the road curvature This means that a right turn is required, and a high-frequency, low-amplitude vibration signal is superimposed on the right side of the steering wheel. High-frequency low-amplitude vibration signal The expression is: in, The amplitude; For frequency.

5. The autonomous driving takeover control method based on muscle stiffness prediction and tactile guidance according to claim 4, characterized in that, In step 3, Take 1.5 N·m; Select 30 Hz.

6. The autonomous driving takeover control method based on muscle stiffness prediction and tactile guidance according to claim 4, characterized in that, The specific steps of step 4 are as follows: Hand grip force is read via steering wheel pressure sensor. And calculate the rate of change of the torque applied by the driver to the steering wheel. : When hand grip strength Reaching the grip strength threshold Torque change rate Reaching the threshold of torque change rate and duration Reaching the collision time threshold At that time, it is determined to be a valid takeover, and the transfer of driving control is complete; The steering system impedance parameters are smoothly attenuated to the stiffness characteristics of the conventional EPS assist mode, and tactile-guided vibrations are stopped.

Images