A control method and system for automatic driving fixed-point parking

Abstract

The present invention discloses a control method and system for autonomous driving fixed-point parking, belonging to the technical field of automotive intelligent driving. The control method includes the following steps: determining a target parking position, automatically generating target trajectory points, an estimated braking distance, and an estimated coasting distance; calculating the longitudinal distance between the current position and the target end point according to the target trajectory points and the current control deviation; collecting vehicle driving information in real time, and calculating the current vehicle speed, slope, and vehicle braking response time information; updating the longitudinal distance at a fixed frequency according to the longitudinal distance from the target end point and the real-time vehicle speed; and based on the control state decision logic, estimating the distance from the target parking point in real time to determine the vehicle control state. The control system includes a vehicle information collection module, a position estimation module, and a control state decision module. By adopting this control method, high-precision fixed-point parking can be achieved during the process of autonomous driving or assisted driving, and at the same time, driving comfort can be taken into account.

Description

Technical Field

[0001] The present invention belongs to the technical field of intelligent driving for vehicles, and particularly relates to a control method and system for autonomous driving fixed-point parking. Background Art

[0002] In autonomous vehicles, in many cases, it is required that the vehicle can accurately stop at the target position, such as bus stops, electric bus charging, tractor-trailer docking, terminal cargo transportation, etc.

[0003] Conventional autonomous driving trajectory tracking and speed real-time control methods are limited by the response time and accuracy of the actuators. The following three actual situations make the parking position accuracy of conventional autonomous driving control insufficient in heavy trucks with AMT and pneumatic braking systems: (1) Usually, the virtual driver updates the vehicle position every 100 milliseconds, which means that even when the speed is 1 km / h, the vehicle has moved 3.6 centimeters between two position information updates, and the positioning update frequency is low; (2) For most commercial vehicles, the clutch actuator and the brake control system are both pneumatic actuator systems, and the response lag is very long; for example, from applying the brake to the brake cylinder establishing the required pressure, the tractor needs more than 600 milliseconds, and the trailer needs more than 1000 milliseconds; on the other hand, both the clutch control and the braking force control are very unstable and inaccurate within a low amplitude range; (3) To achieve a parking accuracy of 10 cm level, it is necessary to use a very low (1-2 km / h) and stable (even more important) vehicle speed. For most commercial vehicles with AMT, the system needs to control the clutch to work in a slipping state within this speed range, which poses high requirements on the control accuracy of the clutch and the brake.

[0004] In view of the above situation, it is necessary to develop and design a specific control function to achieve high-precision fixed-point parking by controlling the clutch and the brake system, while taking into account driving comfort. Summary of the Invention

[0005] In order to solve the above problems, the present invention proposes a control method and system for autonomous driving fixed-point parking, which realizes the high-precision fixed-point parking function and maximum comfort.

[0006] The technical solution of the present invention is as follows:

[0007] A control method for autonomous driving fixed-point parking includes the following steps:

[0008] S1. Determine the target parking position, and automatically generate target trajectory points, estimated braking distance, and estimated coasting distance;

[0009] S2. Calculate the longitudinal distance between the current position and the target end point according to the target trajectory points and the current control deviation;

[0010] S3. Collect vehicle driving information in real time, and calculate the current vehicle speed, slope, and vehicle braking response time information;

[0011] S4. Update the longitudinal distance at a fixed frequency according to the longitudinal distance from the target end point and the real-time vehicle speed;

[0012] S5. Based on the control state decision logic, estimate the distance to the target stop point in real time to determine the vehicle control state.

[0013] Furthermore, the target trajectory points are a list of points (p0, p1, …, pn-1, pn), where p0 is the current position point, Pn is the target stop point, and (p1…pn-1) are the discrete predicted positions between the current position point and the target stop point.

[0014] Furthermore, the calculation and update of the vehicle longitudinal distance adopt a fusion algorithm based on Kalman filter; the fixed frequency is preset to 10 ms.

[0015] Furthermore, the content of the control state decision logic is as follows:

[0016] If the longitudinal distance from the target end point ≤ the estimated braking distance, enter the braking control stage;

[0017] If the longitudinal distance from the target end point ≤ the estimated coasting distance, enter the decelerating coasting control stage;

[0018] If the current vehicle speed ≥ the target vehicle speed at steady state + the compensation vehicle speed, enter the decelerating control stage;

[0019] Otherwise, the vehicle remains in the steady-state speed control stage.

[0020] Furthermore, in the decelerating stage, control the vehicle to decelerate slowly to the set stable driving vehicle speed; in the steady-state speed stage, control the clutch to be in a semi-slip friction state so that the vehicle speed can be stabilized in a fixed vehicle speed range; in the decelerating coasting stage, control to open the clutch and decelerate creep; in the braking stage, control the vehicle to stop slowly to the target position point according to the distance from the target end point.

[0021] A control system for autonomous driving fixed-point parking uses the control method for autonomous driving fixed-point parking as described above to achieve fixed-point parking of vehicle autonomous driving;

[0022] The control system includes a vehicle information acquisition module, a position estimation module, and a control state decision module; among them,

[0023] The vehicle information acquisition module uses a vehicle active speed sensor that can achieve millimeter-level accuracy of the wheel rolling distance to collect vehicle driving information in real time, calculate the vehicle speed, slope, and vehicle braking response time information, and send this information to the position estimation module;

[0024] The position estimation module uses a fusion algorithm based on Kalman filtering. Based on the vehicle information transmitted by the vehicle information acquisition module, and together with the yaw rate with a 10ms update rate in the vehicle's own ESC module, it performs real-time calculation of the vehicle position.

[0025] The control state decision module integrates a control state decision logic control program, and controls the vehicle state according to the vehicle distance estimated in real time by the position estimation module and the current vehicle speed.

[0026] Furthermore, a control system with three cascades is used. The outer cascade controls the vehicle position, the middle cascade controls the vehicle speed, and the inner cascade is used to control the braking force applied by the braking system and the drive train force applied by the clutch control.

[0027] Furthermore, the control logic for completing fixed-point parking is as follows:

[0028] 1) After entering the fixed-point parking control mode, the vehicle slowly decelerates to a preset constant maneuvering speed;

[0029] 2) Use the motion control mode to smoothly maintain the vehicle's maneuvering speed until the distance of the vehicle from the target parking point reaches the pre-opening distance of the clutch calculated from the driving resistance and the clutch opening lag time;

[0030] 3) Keep the clutch open and let the vehicle roll freely until the distance of the vehicle from the target parking point reaches the pre-application distance of the brake calculated from the driving resistance and the brake application lag time;

[0031] 4) Once the vehicle stops, apply full braking to keep the vehicle statically safe.

[0032] Furthermore, the motion control mode completes the smooth control of the vehicle speed through the joint cooperation of traction control, braking control, and clutch actuator control.

[0033] The beneficial technical effects brought by the present invention:

[0034] By adopting vehicle sensors and a fusion algorithm based on Kalman filtering, the received positioning information is predicted in a shorter cycle, and more real-time and higher-precision positioning is obtained; through the control method of the present invention, the clutch can be controlled to remain in the slip state, which is beneficial to quickly releasing the power system drive chain during parking, making the clutch control and braking force control very stable and accurate within a low amplitude range, taking into account driving comfort; by planning the distance to the target parking point at a stable and controllable vehicle speed, before parking, the clutch and the braking system are controlled simultaneously to reach the best parking position. Description of the Drawings

[0035] Figure 1 It is a structural diagram of the automatic driving fixed-point parking control method of the present invention;

[0036] Figure 2 This is the vehicle speed diagram for the vehicle positioning and parking control stage of the present invention. Detailed implementation manners

[0037] The present invention will be further described in detail below in conjunction with the accompanying drawings and specific implementation manners:

[0038] In order to achieve the high-precision vehicle positioning and parking function, a control system for autonomous vehicle positioning and parking uses a control system with three cascades. The outer cascade controls the vehicle position, the middle cascade controls the vehicle speed, and the inner cascade is used to control the braking force applied by the braking system and the transmission force applied by the clutch control.

[0039] The entire control system mainly includes a vehicle information acquisition module, a position estimation module, and a control state decision-making module. Among them, the vehicle information acquisition module uses a vehicle active speed sensor that can achieve millimeter-level accuracy of the wheel rolling distance to collect vehicle driving information in real time, calculate vehicle speed, slope, vehicle braking response time information, and send this information to the position estimation module; the position estimation module uses a fusion algorithm based on Kalman filtering, based on the vehicle information transmitted by the vehicle information acquisition module, and at the same time, with the yaw rate with a 10ms update rate in the vehicle's own ESC module, to perform real-time calculation of the vehicle position; the control state decision-making module integrates a control state decision-making logic control program to control the vehicle state according to the vehicle distance estimated in real time by the position estimation module and the current vehicle speed.

[0040] As Figure 1 shown, a control method for autonomous vehicle positioning and parking includes the following steps:

[0041] Step 1: First, based on the control system, determine the target parking position, automatically generate target trajectory points (Pt0…Ptn), estimated braking distance, and estimated coasting distance;

[0042] Step 2: According to the target trajectory points (target trajectory points automatically generated by the virtual driver in autonomous driving) and the current control deviation dx (the deviation between the actual vehicle position point and the target trajectory point, provided by the virtual driver), calculate the longitudinal distance between the current position and the target end point. The update frequency of this distance depends on the frequency of sending the target trajectory points, and the preset value of the frequency can be set to 10ms or less than 10ms;

[0043] Step 3: Real-time collect vehicle information through the sensors installed on the vehicle and calculate information such as the current vehicle speed v_speed, slope, and vehicle braking response time;

[0044] Step 4: Update the longitudinal distance dx_end according to the longitudinal distance from the target end point calculated in Step 2 and the real-time vehicle speed. The update frequency of this distance is the calculation frequency of the system controller, which is 10 ms or less than 10 ms.

[0045] Step 5: Based on information such as the current vehicle speed, slope, and vehicle braking response time, estimate the distance to the target stop point in real time based on the control state decision logic to determine when the vehicle enters the control states of the deceleration stage, steady-state speed stage, coasting stage, or braking stage. The specific judgment logics for each stage are as follows:

[0046] (1) If dx_end ≤ the estimated braking distance, enter the braking control stage. In this stage, control the braking system according to the distance to the target end point to slowly stop the vehicle at the target position point.

[0047] (2) If dx_end ≤ the estimated coasting distance, enter the decelerating coasting control stage. In this stage, control the opening of the clutch to decelerate and creep at a certain slope.

[0048] (3) If the current vehicle speed v_speed ≥ the target vehicle speed at steady state (1 km / h) + the compensated vehicle speed (1 km / h), enter the deceleration control stage. In this stage, slowly decelerate to the set vehicle speed (1 km / h) at steady state at a certain slope, and this slope can be determined according to the distance to the end point.

[0049] (4) Otherwise, the vehicle remains in the steady-state speed control stage. The goal of this stage is to control the semi-slip friction state of the clutch so that the vehicle speed can be stabilized in a relatively low vehicle speed range (1 km / h).

[0050] The above 1 km / h is the set value for conventional vehicle models, and this value can be specifically set according to vehicle models, environments, etc., and is not unique.

[0051] The essence of the present invention actually includes several key aspects as follows.

[0052] I. Higher-frequency positioning estimation

[0053] In conventional autonomous driving, due to hardware or computing resource limitations, the virtual driver updates the vehicle position at a rate exceeding 100 ms and fuses the positioning with RTK, GPS, IMU, and lidar SLAM. To achieve more precise parking, a higher positioning rate, such as 10 ms, must be adopted to reach a horizontal stop accuracy of 10 cm. Therefore, the vehicle information acquisition module of the control system of the present invention adopts a low-cost solution, that is, the Kalman position estimation algorithm based on high-precision vehicle sensor information. Since the information from high-precision vehicle sensors is more accurate and direct, such as the wheel rolling distance with millimeter-level accuracy from an active speed sensor, and assisted by the yaw rate from the ESC module (with a 10 ms update rate), the position estimation accuracy and stability will be much better than the fusion result from the IMU.

[0054] II. Position Control

[0055] The distance information from the current position to the desired parking position is provided by the virtual driver through the trajectory interface, which includes continuously updating the distance to the desired target point longitudinally (x) and laterally (y). The trajectory interface is defined as a list of points (p0, p1, …, pn-1, pn), where p0 is the current position point of the position control, Pn is the desired stop position point, and (p1…Pn-1) are the discrete predicted positions between the current position point and the desired stop position point. In the control system of the present invention, the end stop position point is found according to the trajectory point list, and the longitudinal distance (dx_end) between the current position of the vehicle and the found end stop position is calculated in real time at a period of 10 ms or less than 10 ms.

[0056] On the other hand, to overcome the long response lag of the clutch and brake drive systems, a special model prediction method based on vehicle dynamics is adopted to achieve longitudinal control in the target parking mode, and the clutch and brake are controlled in advance according to the estimated distance between the current position and the end stop position.

[0057] The fixed-point parking program logic of the special model prediction method based on vehicle dynamics is as follows:

[0058] 1) When the virtual driver selects to enter the fixed-point parking control mode, the vehicle slowly decelerates to a preset constant maneuvering speed (1 km / h);

[0059] 2) Use a special motion control mode to smoothly maintain the maneuvering speed until dx_end reaches the clutch pre-opening distance calculated from the driving resistance and the clutch opening lag time;

[0060] 3) Keep the clutch open and let the vehicle roll freely until dx_end reaches the brake pre-application distance calculated from the driving resistance and the brake application lag time;

[0061] 4) Once the vehicle stops, full braking will be applied to keep the vehicle statically safe.

[0062] III. Motion Control Modes

[0063] The motion control mode accomplishes the smooth speed control of the vehicle through the coordinated cooperation of traction control, braking control, and clutch actuator control.

[0064] (1) Traction Control

[0065] For longitudinal motion control, the traction force (positive or negative) to reach or maintain the desired speed is used as feedforward. The traction force highly depends on road slope, rolling resistance, and air resistance, which can be estimated through the kinematic values and vehicle dynamic characteristics in other functional modules of the autonomous driving system. In addition, a feedback controller using the actual vehicle speed provided by the motion module is also used to minimize the speed control error.

[0066] In addition to the above common control methods, special control methods are required under certain special working conditions. For example, when the vehicle is in a low driving resistance situation (e.g., a truck driving alone or a truck-trailer on a small downhill), since the dry clutch system cannot accurately control the clutch torque near small torques, it is difficult to keep the vehicle speed stable at the target value. To overcome this situation, a slight braking force will be applied while the clutch is being engaged. This will help balance the clutch torque and driving resistance at a deeper position, making the clutch torque control more stable and controllable.

[0067] (2) Braking Control

[0068] The braking controller splits the negative traction force (braking force) into the braking forces required for each actuator output, including components such as service brakes, retarders, motors / non-internal combustion engines, etc., and controls the air pressure of the service brakes, the torque (output) of the retarders, and the engines accordingly. To better utilize the regenerative power function of the motors / non-internal combustion engines during braking, the interaction between the traditional brakes and them is particularly important. In addition, due to the diversity of the braking system, motors / non-internal combustion engines, their functions are unified into an abstract and general hardware layer.

[0069] To improve driving comfort by reducing the braking shock at the end of the parking operation, the redistribution of braking pressure from the front axle to the rear axle implemented in the soft stop function is used.

[0070] (3) Clutch Actuator Control

[0071] Due to the low-speed control range, the positive traction during target controlled parking is handled by the clutch control within the transmission control function. In this case, a clutch torque control function should be provided to support this special motion control. Since the clutch torque request is not defined in J1939, there is no AMT system on the market that supports this interface. Therefore, the clutch torque is controlled by the virtual throttle pedal position.

[0072] In addition, in addition to precise position control, the driving comfort of passengers should be maximized by requiring the vehicle to move continuously (continuous speed, acceleration, and jerk), thereby reducing braking shock. Since high driving comfort may reduce the accuracy of the parking position, a maximum comfort mode (deceleration of braking > -1 m / s²) and a maximum precision mode (distance from the target position < 10 cm when the vehicle stops) can be defined to meet the main objectives of the driving task.

[0073] Figure 2 For a situation of autonomous driving fixed-point parking, the vehicle automatically travels on the road of the predetermined parking point. The sensors installed on the vehicle collect the vehicle driving information in real time, calculate the current vehicle speed, slope, vehicle braking response time and other information in real time, and estimate the distance from the target parking point in real time. When the vehicle travels to the preset value from the target point and the vehicle speed ≥ the target vehicle speed (1 km / h) at steady state + the compensation vehicle speed (1 km / h), the vehicle enters the deceleration stage, decelerates, and slowly decelerates to the set vehicle speed (1 km / h) at steady state at a certain slope; then, it starts to enter the steady-state speed stage. At this time, the control clutch is in a semi-slip friction state, so that the vehicle speed can be stabilized in a relatively small vehicle speed range; when the distance from the target point ≤ the preset coasting distance, the vehicle enters the deceleration coasting stage. At this time, the vehicle controls the clutch to open and decelerate creep at a certain slope; when the distance from the target point ≤ the preset braking distance, the vehicle enters the braking control stage, and controls the braking system according to the target end distance to make the vehicle slowly stop at the target position point.

[0074] Of course, the above description is not a limitation of the present invention, and the present invention is not limited to the above examples. Changes, modifications, additions or substitutions made by those skilled in the art within the substantial scope of the present invention should also fall within the protection scope of the present invention.

Claims

1. A control method for autonomous driving fixed-point parking, characterized in that It includes the following steps: S1. Determine the target parking position, and automatically generate target trajectory points, estimated braking distance, and estimated coasting distance; S2. Calculate the longitudinal distance between the current position and the target end point according to the target trajectory points and the current control deviation; S3. Collect vehicle driving information in real time, and calculate the current vehicle speed, slope, and vehicle braking response time information; S4. Update the longitudinal distance at a fixed frequency according to the longitudinal distance from the target end point and the real-time vehicle speed; S5. Based on the control state decision logic, estimate the distance to the target parking point in real time to determine the vehicle control state; The target trajectory points are a list of points (p0, p1, …, pn-1, pn), where p0 is the current position point, Pn is the target parking point, and (p1…pn-1) are the discrete predicted positions between the current position point and the target parking point; The calculation and update of the vehicle longitudinal distance adopt a fusion algorithm based on Kalman filter; the fixed frequency is preset to 10 ms; The content of the control state decision logic is as follows: If the longitudinal distance from the target end point ≤ the estimated braking distance, enter the braking control stage; If the longitudinal distance from the target end point ≤ the estimated coasting distance, enter the decelerating coasting control stage; If the current vehicle speed ≥ the target vehicle speed at steady state + compensation vehicle speed, enter the decelerating control stage; Otherwise, the vehicle remains in the steady-state speed control stage; In the decelerating stage, control the vehicle to slowly decelerate to the set stable driving vehicle speed; in the steady-state speed stage, control the clutch to be in a semi-slip state so that the vehicle speed can be stabilized in a fixed vehicle speed range; in the decelerating coasting stage, control to open the clutch for decelerating creep; in the braking stage, control the vehicle to slowly stop at the target position point according to the distance to the target end point.

2. A control system for autonomous parking at a fixed point, characterized in that, Adopt the control method for autonomous driving fixed-point parking as described in Claim 1 to achieve fixed-point parking of vehicle autonomous driving; The control system includes a vehicle information acquisition module, a position estimation module, and a control state decision module; among them, The vehicle information acquisition module uses a vehicle active speed sensor that reaches millimeter-level accuracy of wheel rolling distance to collect vehicle driving information in real time, calculate vehicle speed, slope, and vehicle braking response time information, and send this information to the position estimation module; The position estimation module adopts a fusion algorithm based on Kalman filter, and based on the vehicle information transmitted by the vehicle information acquisition module, and at the same time with the yaw rate with a 10 ms update rate in the vehicle's own ESC module, performs real-time calculation of the vehicle position; The control state decision module integrates a control state decision logic control program to control the vehicle state according to the vehicle distance estimated in real time by the position estimation module and the current vehicle speed.

3. The control system for autonomous fixed-point parking according to claim 2, characterized in that, Use a control system with three cascades. The outer cascade controls the vehicle position, the middle cascade controls the vehicle speed, and the inner cascade is used to control the braking force applied by the braking system and the drive force applied by the clutch control.

4. The control system for autonomous fixed-point parking according to claim 3, characterized in that, The control logic for completing fixed-point parking is as follows: 1) After entering the fixed-point parking control mode, the vehicle slowly decelerates to the preset constant maneuvering speed; 2) Use the motion control mode to smoothly maintain the vehicle maneuvering speed until the distance of the vehicle from the target parking point reaches the clutch pre-opening distance calculated from the driving resistance and the clutch opening lag time; 3) Keep the clutch open and let the vehicle roll freely until the distance of the vehicle from the target stopping point reaches the pre-application braking distance calculated from the driving resistance and the braking application lag time; 4) Once the vehicle stops, apply full braking to keep the vehicle statically safe.

5. The control system for autonomous fixed-point parking according to claim 4, characterized in that, The described motion control mode accomplishes the smooth speed control of the vehicle through the combined cooperation of traction control, braking control, and clutch actuator control.

Images