Autopilot Vehicle Control Device and Method Based on Safety Officer Training

Abstract

The present invention belongs to the technical field of autonomous driving vehicles, and particularly relates to a control device and method for autonomous driving vehicles based on safety officer training. Aiming at the problems that the existing autonomous driving safety officer training relies on the operation of a driving robot, the scenarios are single and complex working conditions cannot be dynamically simulated, the present invention provides a training device based on a drive-by-wire vehicle, a low-torque motor, an integrated navigation system and an electronic map. Its technical solution includes: simulating the driver's gripping operation through a low-torque motor, real-time positioning compensation by an INS / GPS integrated navigation system, a bus tool executing drive-by-wire instructions, and a control system generating a path plan in combination with an electronic map and a training plan. This device can simulate a mileage of more than 1000 kilometers, complex weather and emergency scenarios, and is used to train the emergency response ability and vehicle control skills of safety officers.

Description

Technical Field

[0001] The present invention relates to the technical field of vehicle driving training. More specifically, the present invention relates to a control device and method for an autonomous driving vehicle based on safety officer training. Background Art

[0002] Currently, the training of autonomous driving safety officers mainly relies on installing a full set of driving robots, and the route is replayed through "driving memory", that is, simulating an autonomous driving vehicle. Then, in the program of the driving robot, the steering wheel is suddenly turned on a certain section to train the safety officer. Robot training has problems such as high cost, limited venue, difficulty in reproducing extreme scenarios, etc., and it is impossible to accurately quantify the operation behavior of safety officers. Although using traditional simulators can reduce costs, they generally have defects such as static scenarios, poor interactive real-time performance, and lack of physical feedback, resulting in the disconnection between the training effect and the actual driving scenario. The deficiencies of the above existing technologies are specifically reflected in: (1) The simulation of the steering wheel holding state is not realistic, and it cannot meet the requirements of alarm thresholds and driving intention recognition at the same time; (2) The path tracking accuracy is limited by a single navigation system and is easily affected by GPS signal occlusion; (3) The training plan lacks a dynamic adjustment mechanism and it is difficult to optimize the training content according to the performance of safety officers; (4) The safety protection strategy is simple and hierarchical braking and risk adaptive control are not realized.

[0003] The root cause of the above problems lies in the lack of a closed-loop control architecture for multi-system collaboration and in-depth modeling of vehicle dynamics characteristics in complex scenarios. In addition, traditional training methods do not fully utilize the flexibility of drive-by-wire technology, resulting in difficulties in balancing training efficiency and safety. Summary of the Invention

[0004] An object of the present invention is to provide a control device and method for an autonomous driving vehicle based on safety officer training, which uses existing vehicles to construct a low-cost training system and ensures accurate path control.

[0005] To achieve these objects and other advantages of the present invention, according to one aspect of the present invention, there is provided a control device for an autonomous driving vehicle based on safety officer training, including:

[0006] An autonomous driving vehicle equipped with an automatic assisted navigation driving system and a drive-by-wire system;

[0007] A low-torque motor installed on the steering wheel for periodically outputting torque to simulate the driver's holding operation;

[0008] An INS / GPS integrated navigation system for real-time output of vehicle position fusion data;

[0009] A bus tool for sending steering, throttle, and brake commands to the drive-by-wire system based on the drive-by-wire protocol;

[0010] The control system includes:

[0011] A site map digitization system that divides the roads in the training site into multiple Links. Each Link includes an ID, latitude and longitude information, a direction attribute, and nodes (Nodes), and establishes a road network topology relationship;

[0012] Training plan data, which is used to generate vehicle control instructions;

[0013] The control system controls the vehicle to travel along a predetermined path according to the site map data and the training plan through a bus tool.

[0014] Preferably, the output period of the low-torque motor is greater than the reminder period of the vehicle steering wheel grip alarm, and the output torque is greater than the grip sensing threshold of the steering wheel but less than the torque threshold for exiting the autonomous driving function.

[0015] Preferably, the position data of the INS / GPS integrated navigation system is matched with the Links in the site map, and dynamic compensation is performed by calculating the perpendicular distance between the vehicle position and the Link to ensure path tracking accuracy.

[0016] Preferably, the calculation method of the perpendicular distance is as follows:

[0017] Convert the Link start point, end point, and vehicle position into a plane rectangular coordinate system;

[0018] Calculate the perpendicular distance from the vehicle to the Link line segment through vector operation formulas;

[0019] When the perpendicular distance exceeds the threshold, generate a compensation amount according to the deviation direction and adjust the vehicle direction to make it return to the Link center line.

[0020] Preferably, the INS / GPS integrated navigation system includes an error correction module, which is used for:

[0021] When the GPS signal is available, correct the INS integration error through Kalman filtering to make the position error ≤ ±0.1 meter;

[0022] When the GPS signal is lost, switch to the INS standalone mode and suppress the error through zero velocity correction to make the position error ≤ ±0.5 meter.

[0023] Preferably, the error correction module outputs position data at intervals of 0.01 second in the INS standalone mode and fuses it with the GPS data after the GPS is restored to avoid position jumps.

[0024] Preferably, the training plan data includes:

[0025] The preset training mileage ≥ 1000 kilometers;

[0026] The number threshold of steering operations at different angles;

[0027] A complex scenario simulation module that supports bad weather, night driving, and sudden traffic scenarios;

[0028] Monitor the reactions of safety officers and the vehicle status in real time, and dynamically adjust the training content.

[0029] Preferably, it further includes a safety strategy module for triggering a hierarchical braking command according to the risk level when the vehicle is predicted to touch the site boundary.

[0030] Preferably, the safety strategy module:

[0031] Divide into three levels of risks, corresponding to sound and light warnings, progressive braking, and emergency braking respectively;

[0032] Generate an adaptive braking curve based on the vehicle dynamics model, restricting the lateral acceleration ≤ 0.3g;

[0033] Adopt redundant communication of CAN bus and Ethernet, and trigger local autonomous braking when the command is lost;

[0034] Dynamically adjust the risk determination threshold according to the historical reaction time of the safety officer.

[0035] The present invention also provides a control method for an automatic driving vehicle control device based on safety officer training, including the following steps:

[0036] Digitize the training site road into an electronic map containing Links and topological relationships;

[0037] Generate vehicle control parameters according to the training plan;

[0038] Apply a periodic torque to the steering wheel through a low-torque motor;

[0039] Obtain INS / GPS pose data in real time and match it with the electronic map;

[0040] Send a by-wire command through a bus tool to control the vehicle to execute the training path.

[0041] The present invention has at least the following beneficial effects: The present invention provides a control device and method for an autonomous driving vehicle based on safety officer training. The device is directly docked with a bus tool through a by-wire system, without the need to additionally modify the vehicle actuator; a low-torque motor is used to replace manual steering wheel holding, avoiding triggering the vehicle alarm function and reducing the cost of manual intervention; the site map electronic system abstracts the physical road into a Link topology structure, providing a digital basis for path planning; combining the fusion positioning data of INS / GPS, the real-time matching of the vehicle position and the electronic map is realized. The corresponding method combines the electronic map and the by-wire protocol to realize the closed-loop control of path planning → positioning matching → instruction execution; it reduces the cumbersome operations of manually setting paths and manually intervening the steering wheel; the parameters of the low-torque motor adopted are adapted to the existing vehicle, avoiding physical modification of the steering wheel structure.

[0042] Other advantages, objectives and features of the present invention will be partially reflected by the following description and partially understood by those skilled in the art through the research and practice of the present invention. Brief Description of the Drawings

[0043] Figure 1 It is a schematic structural diagram of the control device for an autonomous driving vehicle based on safety officer training described in the present invention. Detailed Embodiments

[0044] The following further detailed description of the present invention is made in conjunction with the drawings and specific embodiments, so that those skilled in the art can implement it according to the description in the specification.

[0045] It should be understood that the terms such as "having", "comprising" and "including" used herein do not exclude the existence or addition of one or more other elements or their combinations.

[0046] It should be noted that the experimental methods described in the following embodiments are all conventional methods unless otherwise specified, and the reagents and materials can be obtained from commercial channels unless otherwise specified.

[0047] As Figure 1 shown, the present invention provides a control device for an autonomous driving vehicle based on safety officer training, including:

[0048] An autonomous driving vehicle equipped with an automatic assisted navigation driving system and a by-wire system;

[0049] A low-torque motor installed on the steering wheel for periodically outputting torque to simulate the holding operation of a driver;

[0050] An INS / GPS integrated navigation system for real-time output of vehicle position fusion data;

[0051] A bus tool that sends steering, throttle, and brake commands to a by-wire system based on a by-wire protocol;

[0052] A control system, comprising:

[0053] A site map digitization system that divides the roads in a training site into multiple Links. Each Link contains an ID, latitude and longitude information, a direction attribute, and nodes (Nodes), and establishes a road network topology relationship;

[0054] Training scenario data for generating vehicle control commands;

[0055] The control system controls the vehicle to travel along a predetermined path through the bus tool according to the site map data and the training scenario.

[0056] In the above technical solution, the automatic driving vehicle control device based on safety officer training includes: an automatic driving vehicle equipped with an automatic assisted navigation driving system and a by-wire system; a low-torque motor installed on the steering wheel of the automatic driving vehicle for outputting low torque within a fixed period to simulate the operation of a driver holding the steering wheel tightly; an INS / GPS integrated navigation system installed on the automatic driving vehicle for real-time output of the position data fusion information of the vehicle's INS and GPS, and a bus tool for sending steering, throttle, and brake commands to the by-wire system according to the by-wire protocol; and further includes a control system in which a site map digitization system is provided for digitizing the roads in a training site, dividing them into multiple Links, each Link containing an ID, latitude and longitude information, one-way / bi-directional direction, and nodes (Nodes), and establishing the topological relationship of the road network; and training scenario data; the control system combines the data in the site map digitization system and the training scenario data to control the bus tool to operate the vehicle. The above device utilizes existing vehicles to build a low-cost training system and ensures precise path control. It directly docks the by-wire system with the bus tool without additional modification of the vehicle actuator; uses a low-torque motor to replace manual holding of the steering wheel, avoiding triggering of the vehicle alarm function and reducing the cost of manual intervention; the site map digitization system abstracts the physical road into a Link topological structure, providing a digital basis for path planning; and combines the fusion positioning data of INS / GPS to achieve real-time matching of the vehicle position and the electronic map.

[0057] In another technical solution, the output period of the low-torque motor is greater than the reminder period of the vehicle's steering wheel grip alarm, and the output torque is greater than the grip sensing threshold of the steering wheel but less than the torque threshold for exiting the autonomous driving function. In this technical solution, the output period of the low-torque motor is greater than the reminder period of the vehicle's "hold the steering wheel tightly" alarm, and the output torque is greater than the vehicle's sensing torque but less than the torque threshold for exiting the NOA function. This avoids the motor triggering the vehicle safety alarm or causing the autonomous driving function to exit. Through the precise setting of the period and torque threshold, it not only simulates the tactile feedback of the driver's grip but also ensures that the vehicle maintains the autonomous driving state (such as the NOA function does not exit); it prevents the training from being frequently paused due to vehicle alarms or function interruptions.

[0058] In another technical solution, the position data of the INS / GPS integrated navigation system is matched with the Link in the site map, and dynamic compensation is performed by calculating the vertical distance between the vehicle position and the Link to ensure path tracking accuracy. In this technical solution, the position data fusion information output by the INS / GPS integrated navigation system is matched with the Link in the site map, the vertical distance between the GPS point and the Link is calculated, and compensation is performed based on this distance to ensure precise control of the vehicle during training.

[0059] In another technical solution, the calculation method of the vertical distance is as follows:

[0060] Convert the starting point, ending point of the Link and the vehicle position into a plane rectangular coordinate system;

[0061] Calculate the vertical distance from the vehicle to the Link line segment through the vector operation formula;

[0062] When the vertical distance exceeds the threshold, generate a compensation amount according to the deviation direction and adjust the vehicle direction to make it return to the center line of the Link.

[0063] In this technical solution, the specific calculation method of the vertical distance is as follows:

[0064] Convert the longitude and latitude coordinates of the starting point and ending point of the Link into a plane rectangular coordinate system;

[0065] Convert the GPS point of the vehicle into the same plane rectangular coordinate system;

[0066] Calculate the vertical distance d from the vehicle GPS point to the Link line segment, and the formula is:

[0067] ;

[0068] Where, is the vector of the Link, is the vector from the vehicle position data fusion information to the starting point of the Link;

[0069] The compensation method includes:

[0070] Setting a threshold value d for the vertical distance threshold ;

[0071] When d > d thresholdd trigger the compensation mechanism;

[0072] According to the vertical distance d and the current moving direction of the vehicle, calculate the compensation amount Δx:

[0073] If the vehicle deviates to the left of the Link, the compensation amount is ;

[0074] If the vehicle deviates to the right of the Link, the compensation amount is Δx = k × d; where k is the compensation coefficient; control the bus tool through the control system to manipulate the vehicle and adjust the direction of the vehicle to make it return to the center line of the Link. Avoid the vehicle deviating from the predetermined route due to positioning errors or control deviations. Quantify the degree of vehicle deviation through plane coordinate system conversion and vector operations (formulaically calculate the vertical distance); force the vehicle to return to the Link center line through dynamic compensation (such as Δx = k × d), and control the path tracking error within the centimeter level; use mathematical formulas to clarify the compensation logic for easy debugging and optimization.

[0075] In another technical solution, the INS / GPS integrated navigation system includes an error correction module for:

[0076] When the GPS signal is available, correct the INS integration error through Kalman filtering to make the position error ≤ ±0.1 m;

[0077] When the GPS signal is lost, switch to the INS standalone mode and suppress the error through zero velocity correction to make the position error ≤ ±0.5 m.

[0078] In this technical solution, the INS / GPS integrated navigation system further includes an error correction module for dynamically adjusting the INS integration error according to the availability of the GPS signal, and when the GPS signal is lost, providing continuous position data through the INS alone to ensure precise positioning of the vehicle during the training process;

[0079] Among them, the working process of the error correction module includes:

[0080] When the GPS signal is available, receive the GPS position data once every 0.1 s and use the Kalman filtering algorithm to correct the INS integration error to ensure that the position error is within the range of ±0.1 m;

[0081] When the GPS signal is lost, switch to the INS standalone working mode. Utilize the accelerometer and gyroscope data of the INS to calculate the vehicle's position, speed, and attitude through integration. And through the zero-speed correction technology, correct the errors of the INS when the vehicle is stationary to ensure that the error does not exceed ±0.5 meters.

[0082] Among them, dynamically adjusting the integration error of the INS includes:

[0083] When the GPS signal strength is lower than -130 dBm, it is determined that the GPS signal is lost, and automatically switch to the INS standalone working mode;

[0084] When the GPS signal is available, the error correction module corrects the integration error of the INS every 1 second to ensure that the corrected position error is within the range of ±0.1 meters.

[0085] In another technical solution, the error correction module outputs position data at an interval of 0.01 seconds in the INS standalone mode and fuses it with the GPS data after the GPS is restored to avoid position jumps. In this technical solution, in the INS standalone working mode, position data is output once every 0.01 seconds to ensure the continuity and real-time nature of the data;

[0086] After the GPS signal is restored, the error correction module fuses the position data of the INS with the position data of the GPS to ensure smooth transition and avoid position jumps;

[0087] During the period when the GPS signal is lost, through the position data provided by the INS alone, combined with the Link information in the site map, the vertical distance between the vehicle and the Link is calculated in real time and compensated to ensure that the vehicle always stays on the predetermined route. It realizes continuous high-precision positioning when the GPS signal is unstable or lost. When the GPS is available, the position error is limited to ±0.1 meters through Kalman filtering; when the GPS is lost, the error of the INS working alone ≤ ±0.5 meters, meeting the accuracy requirements of the training scenario; through data fusion for smooth transition, avoiding control instruction jitter caused by position jumps; ensuring continuous training in GPS-denied environments such as tunnels and underground garages.

[0088] In another technical solution, the training program data includes:

[0089] The preset training mileage ≥ 1000 kilometers;

[0090] The number threshold of steering operations at different angles;

[0091] A complex scenario simulation module that supports bad weather, night driving, and sudden traffic scenarios;

[0092] Real-time monitor the reactions of safety officers and the vehicle status, and dynamically adjust the training content.

[0093] In this technical solution, the training program data includes the following:

[0094] Training mileage: more than 1000 kilometers;

[0095] Number of emergency steering trainings: 200 times of 80 - 90° steering and 100 times of 60° - 80° steering;

[0096] After the vehicle stabilizes on a specific section, a steering instruction is sent through the bus tool to complete the steering operation;

[0097] After the vehicle completes the steering operation, the safety officer takes over the vehicle and restarts the vehicle's automatic driving function again after driving normally for 5 seconds;

[0098] During the training process, the vehicle's status and the safety officer's reactions are monitored in real time, and the training program is dynamically adjusted;

[0099] Support the simulation of various complex scenarios, including bad weather scenarios, sudden traffic scenarios, and night driving scenarios;

[0100] The driving behavior of the safety officer is recorded in real time, a training evaluation report is generated, and it is displayed in real time through the in - vehicle display screen or mobile terminal;

[0101] The vehicle's position and speed are monitored in real time, and the safety strategy is dynamically adjusted to ensure vehicle safety.

[0102] Use scientific design of training content to cover diverse driving scenarios, achieve quantitative indicators (mileage, number of steering times) to ensure that the safety officer proficiently masters routine operations; improve the ability to respond to emergencies (such as tire skidding in rainy days, low visibility at night) through complex scenario simulation; adopt real - time monitoring and dynamic adjustment functions to optimize training efficiency and avoid repeated or ineffective training.

[0103] In another technical solution, it also includes a safety strategy module, which is used to trigger a hierarchical braking instruction according to the risk level when the vehicle predicts to touch the boundary of the site. In this technical solution, the safety strategy module is used to send a by - wire chassis instruction to the bus tool for braking when the safety officer fails to take over the vehicle in time and the vehicle has a risk of touching the boundary of the site.

[0104] In another technical solution, the safety strategy module:

[0105] Divides into three levels of risks, corresponding to sound and light warnings, progressive braking, and emergency braking respectively;

[0106] Generates an adaptive braking curve based on the vehicle dynamics model, restricting the lateral acceleration ≤ 0.3g;

[0107] Adopt redundant communication via CAN bus and Ethernet, and trigger local autonomous braking when instructions are lost;

[0108] Dynamically adjust the risk judgment threshold according to the historical response time of the safety officer.

[0109] In this technical solution, the safety policy module specifically includes:

[0110] Risk level classification:

[0111] Level 1 risk: The predicted distance between the vehicle and the boundary of the site is 2 - 5 meters, and the speed is higher than 30 km / h, triggering an audible and visual warning and prompting the safety officer to take over;

[0112] Level 2 risk: The predicted distance between the vehicle and the boundary of the site is 1 - 2 meters, or the speed is higher than 50 km / h, triggering progressive braking, and the braking force increases at a rate of 0.5g / s;

[0113] Level 3 risk: The predicted distance between the vehicle and the boundary of the site is less than 1 meter, or the speed is higher than 70 km / h, triggering emergency braking, and the braking force instantaneously reaches 1.2g;

[0114] Dynamic trajectory prediction:

[0115] Based on the vehicle's real-time position, attitude, and kinematic model, calculate the trajectory within the next 3 seconds, and combine with the boundary information in the site electronic map to judge the risk level in real time;

[0116] Adaptive brake curve generation:

[0117] According to the vehicle mass, tire friction coefficient, and current speed in the control system, calculate the optimal brake curve in real time through the vehicle dynamics model to ensure vehicle stability during braking, and the lateral acceleration does not exceed 0.3g;

[0118] If the vehicle is in a turning state, an additional steering compensation instruction is introduced to adjust the front wheel steering angle by ±2° to prevent tail-swing caused by braking;

[0119] Redundant communication and fault-tolerant calculation:

[0120] Main communication channel: Send the wire control braking instruction via CAN bus, with a response delay less than 50ms;

[0121] Standby channel: When the CAN bus communication fails, switch to the redundant communication protocol based on Ethernet to ensure the reliability of instruction transmission;

[0122] Local actuator autonomous judgment: If the bus instruction is lost, the braking system triggers emergency braking autonomously according to the local inertial navigation data, and the braking force is fixed at 0.8g;

[0123] Safety officer behavior prediction and dynamic threshold adjustment:

[0124] Analyze the historical takeover response time of the safety officer and dynamically adjust the risk trigger threshold:

[0125] If the safety officer's recent response time is shortened by 20%, the first-level risk determination distance is relaxed to 3 meters;

[0126] If the response time is extended by 30%, the second-level risk determination distance is tightened to 1.5 meters;

[0127] The training data includes the safety officer's eye movement tracking and hand movement sensor data to improve the prediction accuracy.

[0128] Through the above safety strategy module, a rapid response in an emergency is achieved to avoid vehicle out of control or collision. Through hierarchical safety intervention: the first-level risk (acoustic and light warning) gives priority to prompting the safety officer to take over, leaving room for manual handling; the second-level risk (progressive braking) balances the braking efficiency and comfort by increasing the braking force at 0.5g / s; the third-level risk (1.2g emergency braking) minimizes the collision risk. By restricting the lateral acceleration ≤ 0.3g, it prevents the vehicle from skidding due to braking; steering compensation (±2° front wheel adjustment) is used to offset the centrifugal force effect during braking on a curve; redundant communication and local autonomous braking ensure the foolproof transmission of instructions.

[0129] The present invention also provides a control method for an autonomous driving vehicle control device based on safety officer training, including the following steps:

[0130] Digitize the training site road into an electronic map containing Links and topological relationships;

[0131] Generate vehicle control parameters according to the training plan;

[0132] Apply a periodic torque to the steering wheel through a low-torque motor;

[0133] Obtain INS / GPS pose data in real time and match it with the electronic map;

[0134] Send by-wire instructions through the bus tool to control the vehicle to execute the training path.

[0135] In this technical solution, the training method specifically includes the following steps:

[0136] Digitally process the roads in the training site to generate electronic map data containing multiple Links. Each Link records the ID, longitude and latitude information, one-way / bidirectional attribute, and associated node Node, and establish a road network topological relationship;

[0137] Generate vehicle control parameters according to a preset training plan, where the control parameters include steering angle, throttle opening, and braking force; apply a periodic torque to the steering wheel of the autonomous vehicle through a low-torque motor, and the torque value is set to a value below the steering wheel grip detection threshold;

[0138] Obtain the vehicle pose fusion data output by the INS / GPS integrated navigation system in real time;

[0139] Based on the topological relationship of the electronic map and the real-time pose data, send steering, throttle, and braking commands to the vehicle by-wire system through a bus tool according to the by-wire protocol, so that the vehicle executes the training plan along the predetermined path.

[0140] In this way, the systematic implementation of the training process is realized and the hardware dependence is reduced. By combining the electronic map and the by-wire protocol, a closed-loop control of path planning → positioning matching → command execution is realized; the cumbersome operations of manually setting paths and manually intervening the steering wheel are reduced; the parameters of the low-torque motor are adapted to existing vehicles, avoiding physical modification of the steering wheel structure.

[0141] In another technical solution, the low-torque motor further includes a dynamic torque adjustment module for dynamically adjusting the output torque and period according to the real-time state of the vehicle, including the following steps:

[0142] 1. Obtain the real-time motion parameters of the vehicle: collect the vehicle speed v and the steering angle θ of the steering wheel through the CAN bus;

[0143] 2. Calculate the dynamic torque coefficient:

[0144] Base torque T base = Vehicle grip sensing threshold × 1.2;

[0145] Speed correction factor α = 1 + (v / 100 km / h) × 0.5, with α ∈ [1, 2];

[0146] Steering angle correction factor β = 1 + |θ| / 180° × 0.8, with β ∈ [1, 1.8];

[0147] Actual output torque T = T base × α × β;

[0148] 3. Adjust the output period:

[0149] When v > 60 km / h, the period is adjusted to 1.5 times the alarm period;

[0150] When θ > 45°, the period is shortened to 0.7 times the alarm period;

[0151] 4. Drive the low-torque motor through the motor controller and output according to the adjusted torque T and period.

[0152] As above, it solves the technical problem that the fixed torque output cannot adapt to the change of the steering wheel resistance in different driving states (such as high speed and sharp turning), resulting in false alarms or interfering with the execution of the steer-by-wire command. Calculate the torque in real time based on the vehicle speed and steering angle, increase the torque at high speed to prevent false alarms (α factor), and increase the torque during sharp turning to simulate the driver's firm grip (β factor); for example: if v = 80 km / h and θ = 90°, then T = T base × 1.4 × 1.4 = 1.96 T base , ensuring that the torque always covers the resistance change. In addition, by extending the output period at high speed, the interference with the steer-by-wire is reduced; during sharp turning, the period is shortened to enhance the authenticity of the grip simulation. In the straight-line driving scenario at 100 km / h, the torque of the above device is increased to 1.5 T base , and the alarm trigger rate drops by 90%; in the 60° steering scenario, the period is shortened to 0.7 times, and the steering wheel control error drops from ±3° to ±0.5°.

[0153] In another technical solution, the dynamic torque adjustment module further includes a steering wheel torque sensor and a closed-loop control module, where:

[0154] The steering wheel torque sensor is installed on the steering column of the steering wheel and is used to detect the actual torque T applied by the driver's hand in real time real ;

[0155] The closed-loop control module adjusts the output torque through the following steps:

[0156] 1. Calculate the target torque T target and the difference ΔT = |T real - T target - T real | between the actual torque T

[0157] 2. If ΔT > 0.2 N•m, adjust the motor output torque according to the proportional-integral (PI) algorithm:

[0158] Proportional term: P = Kp × ΔT, where Kp = 0.8;

[0159] Integral term: I = Σ(Ki × ΔT × Δt), where Ki = 0.05 and Δt is the control period (0.1 second);

[0160] Corrected torque T adjust = T target + (P + I);

[0161] 3. Drive the low-torque motor to output T through the motor controller adjust 。

[0162] As described above, the open-loop control of the dynamic torque adjustment module cannot respond to the actual resistance change of the steering wheel in real time, resulting in the accumulation of torque output errors (such as when the frictional resistance suddenly changes, T real deviates from T target ). By using a torque sensor to monitor the actual torque in real time and combining the PI algorithm to dynamically correct the output, the torque error ΔT is controlled within ±0.1 N•m. For example, if the frictional resistance of the steering wheel increases by 20% due to temperature rise, the system automatically increases the output torque compensation to avoid triggering an alarm. In bumpy road or emergency steering scenarios, the interference of external vibration on torque output is suppressed, and the simulated stability of steering wheel holding is improved by 40%.

[0163] Example: Emergency steering and braking coordination training in heavy rain

[0164] Scene parameter settings

[0165] Road environment:

[0166] Simulate a two-way two-lane road in a closed test site, including a left-turn curve with a radius of 50 meters;

[0167] Road surface adhesion coefficient μ = 0.3 (wet and slippery state);

[0168] Environment simulation: heavy rain (visibility ≤ 50 meters), night lighting (only low beams are turned on);

[0169] Training objectives:

[0170] Evaluate the emergency takeover ability of the safety officer in extreme weather;

[0171] Verify the vehicle's lateral control accuracy (requirement: deviation from the center line ≤ ±0.3 meters);

[0172] Test the hierarchical braking response of the safety strategy module;

[0173] System configuration:

[0174] Vehicle: A mass-produced vehicle equipped with the NOA system, and the by-wire system supports 100Hz command updates;

[0175] Low-torque motor: Output cycle 1.2 seconds, torque 0.8N・m;

[0176] Integrated navigation: INS update rate 100Hz, GPS update rate 10Hz;

[0177] System response process

[0178] Initial stage:

[0179] The control system loads the electronic map and divides the curve into 3 Link segments (Link1 - Link3);

[0180] The low - torque motor starts to output periodically to simulate the normal gripping state of the driver;

[0181] The integrated navigation system is initialized, and the initial error of the INS is corrected through the GPS signal;

[0182] Training trigger:

[0183] The vehicle enters Link1 (straight - line segment) at a speed of 40 km / h, and the system calculates the distance between the vehicle and the curb in real - time (required to maintain 40 - 60 cm);

[0184] When the vehicle travels to the curve entrance (30 meters away from the intersection), the control system sends instructions through the bus tool:

[0185] Steering angular velocity 360° / s;

[0186] The throttle maintains the current opening to prepare for the acceleration demand after steering;

[0187] Sudden - scenario simulation:

[0188] After the vehicle enters the curve, the system simulates a stationary vehicle ahead (20 meters away);

[0189] The integrated navigation system detects a short - term loss of the GPS signal (simulating being blocked by heavy rain) and switches to the INS - only mode:

[0190] The zero - speed correction module suppresses the error through the wheel - speed sensor signal (position error controlled within 0.3 meters);

[0191] The vertical - distance compensation algorithm continuously calculates the deviation between the vehicle and the center line of the Link;

[0192] Safety strategy activation:

[0193] The predicted time to collision (TTC) is calculated as 1.8 seconds (triggering a secondary risk):

[0194] The audible and visual alarm is activated (buzzer + warning light);

[0195] Progressive braking starts, and the deceleration linearly increases to - 0.2g;

[0196] If the safety officer does not take over in time and the TTC drops to 1.2 seconds (triggering a tertiary risk):

[0197] Emergency braking is activated, with a maximum deceleration of -0.5g (lateral acceleration of 0.28g, meeting the requirement of ≤0.3g);

[0198] The system sends braking instructions through dual-redundant channels of CAN bus + Ethernet;

[0199] Takeover and recovery:

[0200] After the safety officer takes over manually, the system records the takeover time (0.6 seconds) and the lateral offset of the vehicle (0.25 meters);

[0201] After the vehicle drives stably for 5 seconds, the control system automatically resumes the NOA function;

[0202] Training effect evaluation

[0203] Quantitative indicators:

[0204] Path tracking accuracy: The maximum lateral deviation is 0.28 meters (meeting the requirement of ≤0.3 meters);

[0205] Braking response time: It takes 0.8 seconds from risk identification to full braking;

[0206] Safety officer takeover efficiency: The average response time is 0.6 seconds (better than the industry benchmark of 0.8 seconds);

[0207] System optimization:

[0208] According to the training data of this time, dynamically adjust the TTC threshold for subsequent curve training (from 1.8 seconds to 1.5 seconds);

[0209] Optimize the combined navigation fusion algorithm and reduce the position jump error after GPS recovery from 0.15 meters to 0.1 meter;

[0210] Scenario expansion:

[0211] Subsequently, compound scenarios such as sudden lane changes of oncoming vehicles and pedestrians crossing the road can be superimposed;

[0212] Add night high-beam switching training to assess the visual adaptation ability of safety officers.

[0213] Although the embodiments of the present invention have been disclosed as above, they are not limited to only the applications listed in the specification and embodiments. It can be fully applied to various fields suitable for the present invention. For those familiar with the field, additional modifications can be easily made. Therefore, without departing from the general concept defined by the claims and the equivalent scope, the present invention is not limited to specific details and the illustrated and described examples here.

Claims

1. An automatic driving vehicle control device based on safety officer training, characterized in that: include: Autonomous driving vehicles, equipped with autopilot and drive-by-wire systems; A low-torque motor is installed on the steering wheel and is used to periodically output torque to simulate the driver's grip operation; The low-torque motor further includes a dynamic torque adjustment module for dynamically adjusting the output torque and cycle according to the real-time state of the vehicle, comprising the following steps: Step 1: Obtain the real-time motion parameters of the vehicle: collect the vehicle speed v and steering wheel angle θ through the CAN bus; Step 2: Calculate the dynamic torque coefficient: Basic torque T base = Vehicle grip sensing threshold × 1.2; Speed ​​correction factor α = 1 + (v / 100 km / h) × 0.5, limit α ∈ [1, 2]; Steering angle correction factor β = 1 + |θ| / 180°×0.8, constrained by β∈ [1, 1.8]; Actual output torque T = T base ×α×β; Step 3: Adjust the output cycle: When v > 60 km / h, the period is adjusted to 1.5 times the alarm period; When θ> 45°, the period is shortened to 0.7 times the alarm period; Step 4: Drive the low torque motor through the motor controller and output according to the adjusted torque T and cycle; INS / GPS combined navigation system, real-time output of vehicle posture fusion data; Bus tool, which sends steering, throttle and brake commands to the wire control system based on the wire control protocol; Control system, including: The electronic site map system divides the training site roads into multiple links. Each link contains ID, longitude and latitude information, direction attributes and nodes, and establishes the road network topology relationship. Training program data, used to generate vehicle control instructions; the training program data includes preset training mileage and the number of steering operations at different angles, and the control instructions include steering angle, throttle opening and braking force; The control system acquires the vehicle posture fusion data output by the INS / GPS combined navigation system in real time; based on the electronic map topological relationship and the real-time posture data, the control system sends steering, throttle and brake commands to the vehicle's wire control system through the bus tool in accordance with the wire control protocol, so that the vehicle executes the training plan along the predetermined path; the safety officer takes over the vehicle after the vehicle completes the steering operation, and restarts the vehicle's automatic driving function after 5 seconds of normal driving; during the training process, the vehicle status and the safety officer's response are monitored in real time, and the training plan is dynamically adjusted.

2. The automatic driving vehicle control device based on safety officer training according to claim 1, characterized in that: The output torque of the low-torque motor is greater than the steering wheel grip sensing threshold, but less than the torque threshold for exiting the autonomous driving function.

3. The automatic driving vehicle control device based on safety officer training according to claim 1, characterized in that: The position data of the INS / GPS integrated navigation system is matched with the Link in the site map, and dynamic compensation is performed by calculating the vertical distance between the vehicle position and the Link to ensure the path tracking accuracy.

4. The automatic driving vehicle control device based on safety officer training according to claim 3 is characterized in that: The vertical distance is calculated as follows: Convert the starting point, end point and vehicle position of the Link into a plane rectangular coordinate system; Calculate the vertical distance from the vehicle to the Link segment using the vector operation formula; When the vertical distance exceeds the threshold, a compensation amount is generated according to the deviation direction, and the vehicle direction is adjusted to return to the Link center line.

5. The automatic driving vehicle control device based on safety officer training according to claim 3, characterized in that: The INS / GPS integrated navigation system includes an error correction module for: When GPS signals are available, the INS integral error is corrected by Kalman filtering to make the position error ≤ ±0.1 m; When the GPS signal is lost, it switches to INS standalone mode and suppresses the error through zero-speed correction to make the position error ≤±0.5 meters.

6. The automatic driving vehicle control device based on safety officer training according to claim 5, characterized in that: The error correction module outputs position data at 0.01 second intervals in INS standalone mode and merges with GPS data after GPS recovery to avoid position jumps.

7. The automatic driving vehicle control device based on safety officer training according to claim 1, characterized in that: It also includes a safety strategy module for triggering graded braking instructions based on the risk level when the vehicle is predicted to touch the boundary of the site.

8. The automatic driving vehicle control device based on safety officer training according to claim 7, characterized in that: The security policy module: There are three levels of risk, corresponding to sound and light warning, progressive braking and emergency braking; Generates adaptive braking curve based on vehicle dynamics model, limiting lateral acceleration to ≤ 0.3g; Uses CAN bus and Ethernet redundant communication and triggers local autonomous braking when command is lost; Dynamically adjust the risk assessment threshold based on the safety officer’s historical response time.

9. A control method for an automatic driving vehicle control device based on safety officer training according to any one of claims 1 to 8, characterized in that: The following steps are involved: Digitize the training site roads into an electronic map containing links and topological relationships; Generate vehicle control parameters according to the training plan; Applying periodic torque to the steering wheel via a low torque motor; Acquire INS / GPS posture fusion data in real time and match it with electronic maps; Send control-by-wire commands through the bus tool to control the vehicle to execute the training path.

Images