A lateral control system, method, electronic device, and storage medium

Abstract

The present disclosure relates to a lateral control system, method, electronic device, and storage medium. The system includes: a reference trajectory planning module for providing a current reference trajectory offset; an unmanned vehicle model for determining a current deviation and a composite disturbance according to current driving information, a preset delay, and the current reference trajectory offset; an extended state observer for observing the current deviation and the composite disturbance; a tracking differentiator for filtering a current error, where the current error is obtained based on an observed value of the current reference trajectory offset and the current deviation; a feedback controller for determining a nominal control amount based on the filtered current error; and a first arithmetic unit for calculating a target control amount for lateral control of the unmanned vehicle based on a first compensation gain amount and the nominal control amount. By accurately controlling the lateral direction of the vehicle, the system reduces the impact of lateral offset on path tracking to a certain extent and also reduces the waste of human and material resources.

Description

Technical Field

[0001] The present disclosure relates to the technical field of autonomous driving, and particularly to a lateral control system, method, electronic device, and storage medium. Background Art

[0002] Currently, driverless vehicles are widely used in daily life. Among the key technologies of driverless driving, lateral control is one of the main functions of the driverless control layer, and trajectory tracking is also the focus of realizing driverless technology. Among them, the control algorithm is the core of lateral control, which can control the driverless vehicle to automatically complete the desired tracking task, that is, to achieve trajectory tracking. Therefore, the control algorithm is particularly important for key performance such as the tracking accuracy and riding comfort of driverless vehicles, and also determines the quality of the vehicle to complete the driverless task.

[0003] However, the control algorithm is affected by various factors such as the vehicle itself and the external road environment, resulting in great challenges to its stability, accuracy, and robustness. For example, the control algorithm is easily affected by the following factors: when the number of driverless vehicles increases and it is difficult to perform regular maintenance, the parameters of the vehicle mechanism and the positioning system are more likely to have a zero position drift phenomenon; the road environment in which the vehicle travels is complex; different application scenarios have different requirements for the vehicle, which will further affect the lateral trajectory tracking effect. To sum up, the above factors will cause the driverless vehicle to have a lateral offset relative to the given reference trajectory in terms of the tracking result, and the lateral offset poses a great threat to the tracking performance and even the driving safety of the vehicle. To solve such problems, a large number of engineering parameters are usually used for tuning and trial-and-error work. Although this method reduces the impact of the lateral offset problem on path tracking to a certain extent, it wastes a lot of manpower and material resources and is not suitable for the quantitative production and maintenance of driverless vehicles. Summary of the Invention

[0004] To solve the above technical problems, the present disclosure provides a lateral control system, method, electronic device, and storage medium, which reduce the impact of lateral offset on path tracking to a certain extent by accurately controlling the driverless vehicle, and also reduce the waste of manpower and material resources.

[0005] In a first aspect, an embodiment of the present disclosure provides a lateral control system applied to a driverless vehicle, including a reference trajectory planning module, an extended state observer, a driverless vehicle model, a tracking differentiator, a feedback controller, and a first arithmetic unit, where:

[0006] The reference trajectory planning module is configured to provide the current reference trajectory offset of the driverless vehicle, where the current reference trajectory offset includes a current lateral coordinate offset and a current heading angle offset;

[0007] The unmanned vehicle model is used to determine the current lateral deviation, heading deviation, and composite disturbance of the unmanned vehicle according to the acquired current driving information of the unmanned vehicle, the preset delay, and the current reference trajectory offset;

[0008] The extended state observer is used to observe the current lateral deviation, heading deviation, and composite disturbance of the unmanned vehicle;

[0009] The tracking differentiator is used to filter the lateral error and the heading error, where the lateral error is obtained based on the observed value of the lateral deviation output by the extended state observer and the current lateral coordinate offset, and the heading error is obtained based on the observed value of the heading deviation output by the extended state observer and the current heading angle offset;

[0010] The feedback controller is used to determine the nominal control quantity based on the filtered lateral error and the filtered heading error;

[0011] The first arithmetic unit is used to calculate the target control quantity for lateral control of the unmanned vehicle based on the first compensation gain quantity and the nominal control quantity, where the first compensation gain quantity is obtained according to the observed value of the composite disturbance output by the extended state observer and the compensation gain.

[0012] In a second aspect, an embodiment of the present disclosure provides a lateral control method applied to an unmanned vehicle, and the method includes:

[0013] Obtain the current reference trajectory offset of the unmanned vehicle, where the current reference trajectory offset includes the current lateral coordinate offset and the current heading angle offset;

[0014] Determine the current lateral deviation, heading deviation, and composite disturbance of the unmanned vehicle according to the acquired current driving information of the unmanned vehicle, the preset delay, and the current reference trajectory offset;

[0015] Based on a pre-established extended state observer, observe the current lateral deviation, heading deviation, and composite disturbance of the unmanned vehicle, and determine the observed value of the current lateral deviation, the observed value of the heading deviation, and the observed value of the composite disturbance of the unmanned vehicle;

[0016] Filter the lateral error and the heading error, where the lateral error is determined based on the observed value of the lateral deviation output by the extended state observer and the current lateral coordinate offset, and the heading error is determined based on the observed value of the heading deviation output by the extended state observer and the current heading angle offset;

[0017] Determine a nominal control quantity based on the filtered lateral error and the filtered heading error;

[0018] Calculate a target control quantity for lateral control of the driverless vehicle based on a first compensation gain quantity and the nominal control quantity, where the first compensation gain quantity is calculated based on an observed value of a composite disturbance output by the extended state observer and a compensation quantity gain.

[0019] In a third aspect, an embodiment of the present disclosure provides an electronic device, including:

[0020] A memory;

[0021] A processor; and

[0022] A computer program;

[0023] Wherein, the computer program is stored in the memory and is configured to be executed by the processor to implement the lateral control method as described above.

[0024] In a fourth aspect, an embodiment of the present disclosure provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the lateral control method as described above are implemented.

[0025] An embodiment of the present disclosure provides a lateral control system, method, electronic device, and storage medium. The lateral control system includes a reference trajectory planning module, a driverless vehicle model, a tracking differentiator, a feedback controller, an extended state observer, and a first arithmetic unit, where: the reference trajectory planning module is used to provide the current reference trajectory offset of the driverless vehicle in real time; the driverless vehicle model is used to determine the current lateral deviation, heading deviation, and composite disturbance of the driverless vehicle according to the current driving information, preset delay, and current reference trajectory offset of the driverless vehicle; the extended state observer is used to observe the lateral deviation, heading deviation, and composite disturbance of the driverless vehicle to obtain an observed value of the lateral deviation, an observed value of the heading deviation, and an observed value of the composite disturbance; the tracking differentiator is used to filter the lateral error and the heading error, where the lateral error and the heading error are obtained based on the current reference trajectory offset and the observed values of the respective deviations; the feedback controller is used to determine a nominal control quantity based on the filtered lateral error and the heading error; the first arithmetic unit is used to calculate a target control quantity for lateral control of the driverless vehicle based on a first compensation gain quantity and the nominal control quantity, and the first compensation gain quantity is obtained according to the observed value of the composite disturbance and the compensation quantity gain. The lateral control system provided by the present disclosure reduces the impact of lateral offset on path tracking to a certain extent by accurately controlling the lateral direction of the driverless vehicle, and also reduces the waste of manpower and material resources. Description of the Drawings

[0026] The accompanying drawings herein are incorporated into the specification and constitute a part of this specification, showing embodiments consistent with the present disclosure, and are used together with the specification to explain the principles of the present disclosure.

[0027] To more clearly illustrate the technical solutions in the embodiments of the present disclosure or in the prior art, the following will briefly introduce the accompanying drawings required for use in the description of the embodiments or the prior art. Obviously, for those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0028] Figure 1 Schematic diagram of a lateral control system provided for an embodiment of the present disclosure;

[0029] Figure 2 Geometric schematic diagram of a vehicle in a global coordinate system provided for an embodiment of the present disclosure;

[0030] Figure 3 Schematic flow chart of a lateral control method provided for an embodiment of the present disclosure;

[0031] Figure 4 Schematic flow chart of another lateral control method provided for an embodiment of the present disclosure;

[0032] Figure 5 Schematic plan view provided for an embodiment of the present disclosure;

[0033] Figure 6 Schematic diagram of the structure of an electronic device provided for an embodiment of the present disclosure. Detailed implementation manners

[0034] In order to be able to more clearly understand the above objects, features, and advantages of the present disclosure, the following will further describe the solutions of the present disclosure. It should be noted that, without conflict, the embodiments of the present disclosure and the features in the embodiments can be combined with each other.

[0035] In the following description, many specific details are set forth in order to fully understand the present disclosure, but the present disclosure can also be implemented in other ways different from those described herein; obviously, the embodiments in the specification are only a part of the embodiments of the present disclosure, rather than all the embodiments.

[0036] In the key technologies of driverless driving, lateral control is one of the main functions of the driverless control layer, and trajectory tracking is also the top priority for realizing driverless driving technology. Among them, trajectory tracking refers to calculating the lateral control amount in real time according to the trajectory information output by the lateral planning and the positioning information of the positioning system through a set control algorithm to control the vehicle steering mechanism (steering wheel, power assist motor, vehicle front wheels, etc.) to automatically complete the desired tracking task. In lateral control, the control algorithm is the core, which is particularly important for the key performance of driverless vehicles such as tracking accuracy and ride comfort, and also determines the quality of the driverless vehicle to complete the driverless task. However, affected by various factors such as the driverless vehicle itself and the external road environment, the stability, accuracy, and robustness of the control algorithm have been greatly challenged.

[0037] When the number of applied driverless vehicles increases and it is difficult to maintain them regularly, the parameters of the driverless vehicle mechanism and the positioning system are prone to drift. For example, the steering zero position drift of the vehicle (hereinafter referred to as the vehicle for the driverless vehicle), the drift caused by the uneven left and right distribution of the vehicle tire pressure, and the zero position parameter drift of the Global Positioning System (GPS) receiver. In addition, the road environment where the vehicle travels is complex. The roads faced not only include longitudinal uphill and downhill sections, but also lateral slopes that affect the lateral tracking ability of the system. At the same time, in the face of different application scenarios, especially for application scenarios with passenger-carrying requirements such as driverless taxis, driverless buses, and driverless sightseeing vehicles, the different seat distributions of passengers will also cause uneven mass distribution of the vehicle, which in turn affects the tracking effect of the lateral trajectory. Due to the influence of the above-mentioned many factors, the driverless vehicle will have a lateral deviation relative to the given reference trajectory in terms of the tracking result. If this situation cannot be effectively handled, it poses a great threat to the tracking performance and even the driving safety of the vehicle. To address such problems, a large amount of engineering parameter reconfiguration and trial-and-error work is usually carried out. Although this method can reduce the impact of the lateral deviation problem, it wastes a lot of manpower and material resources and is not suitable for the mass production and maintenance of driverless vehicles.

[0038] To address the above technical problems, embodiments of the present disclosure provide a lateral control system for improving the robust ability of vehicle steering system control and accurately controlling the lateral movement of the vehicle. That is, after determining the target control quantity for lateral control of the vehicle through the lateral control system, the vehicle steering system controls the vehicle to automatically steer based on this target control quantity to correct the lateral deviation of the vehicle caused by reasons such as zero-position drift and road lateral slope of the vehicle steering system, enabling the driverless vehicle to stably, accurately, and robustly track the desired trajectory. It can be understood that during the driving process of the entire vehicle, the lateral movement can be achieved by controlling the front-wheel steering angle of the vehicle. That is, the target control quantity determined by the lateral control system for lateral control of the vehicle can be a control quantity regarding the front-wheel steering angle. Specifically, in the lateral control system, the disturbance of the steering system is defined as a composite disturbance, and a kinematic error model with the composite disturbance is established. Subsequently, an extended state observer is designed to observe the composite disturbance together with the error state of the steering system through the extended state observer and compensate in the control channel. In addition, to suppress the noise influence of the observed value of the extended state observer and the reference trajectory, a tracking differentiator is designed for filtering to reduce the jitter degree of the control quantity. Finally, the extended state observer is discretized based on the Euler method and applied to the embedded computer. The system provided by the present disclosure has generalization ability in different scenarios, can not only solve the adverse effects caused by a single disturbance, but also resist the adverse effects caused by the simultaneous action of multiple disturbances, reducing the influence of lateral deviation on path tracking to a certain extent and also reducing the waste of manpower and material resources. Specifically, it will be described in detail through one or more of the following embodiments.

[0039] Figure 1 FIG. is a schematic structural diagram of a lateral control system provided by an embodiment of the present disclosure. The lateral control system can be implemented in a software and / or hardware manner and is generally integrated in a computer device and configured in a driverless vehicle. Specifically, as Figure 1 shown, the lateral control system 100 includes a reference trajectory planning module 110, an extended state observer 120, a tracking differentiator 130, a feedback controller 140, a first arithmetic unit 150, and a driverless vehicle model 190.

[0040] It can be understood that a steering system is configured inside the vehicle, and the steering system can be communicatively connected to the lateral control system 100 to receive the target control quantity sent by the lateral control system 100. The target control quantity can be understood as a control quantity regarding the front-wheel steering angle for controlling the steering of the steering system. That is, the lateral control system 100 will send the target control quantity to the steering system through the communication network, and then the steering system controls the vehicle to steer according to the target control quantity to complete the actual steering action of the vehicle.

[0041] Among them, the reference trajectory planning module 110 is used to provide the current reference trajectory offset of the driverless vehicle, where the current reference trajectory offset includes the current lateral coordinate offset and the current heading angle offset.

[0042] It can be understood that the reference trajectory planning module 110 is used to provide the real-time reference trajectory offset for the driverless vehicle controller, where the current reference trajectory offset includes the current lateral coordinate offset and the current heading angle offset.

[0043] Among them, the driverless vehicle module 190 is used to determine the current lateral deviation, heading deviation and composite disturbance of the driverless vehicle according to the obtained current driving information of the driverless vehicle, the preset delay and the current reference trajectory offset.

[0044] Among them, the current driving information further includes the current pose, the current front wheel steering angle and the current longitudinal speed.

[0045] Among them, the driverless vehicle model 190 includes a kinematic model and an error model. The kinematic model is used to determine the predicted pose of the driverless vehicle according to the current pose, the current front wheel steering angle, the wheelbase of the driverless vehicle, the preset delay and the current longitudinal speed, where the predicted pose includes the predicted lateral coordinate and the predicted heading angle; the error model is used to determine the current lateral deviation, heading deviation and composite disturbance of the driverless vehicle according to the current reference trajectory offset and the predicted pose.

[0046] It can be understood that the driverless vehicle module 190 is used to obtain the current driving information of the vehicle. Among them, the current driving information of the vehicle includes the current pose, the current front wheel steering angle and the current longitudinal speed, etc. The current pose of the driverless vehicle can be obtained by the perception and positioning system configured in the vehicle, and the perception and positioning system will provide the real-time pose information of the vehicle, and the real-time pose information includes the position and orientation of the vehicle in the global coordinate system.

[0047] It is understandable that the driverless vehicle module 190 includes a kinematic model and an error model. The kinematic model is used to predict the future state of the vehicle based on the actual state of the vehicle. Specifically, the kinematic model determines the predicted pose of the vehicle according to the current pose of the vehicle, the current front-wheel steering angle, the wheelbase of the vehicle, a preset delay, and the current longitudinal speed of the vehicle. The predicted pose can be understood as the position and orientation of the vehicle in the global coordinate system after the preset delay. It is understandable that there is a steering lag or steering delay in the steering system, which will cause the steering system to oscillate. Therefore, the embodiments of the present disclosure predict the future state of the vehicle based on the actual state of the vehicle to reduce the possible oscillation of the steering system. Taking a four-wheel driverless vehicle as an example, the wheelbase of the vehicle is the distance between the centers of the two adjacent front-wheel axles and the rear-wheel axle on the same side. It is understandable that the current pose of the vehicle can be the coordinates and orientation of the center point of the vehicle's rear axle in the global coordinate system, and the predicted pose is also the coordinates and orientation of the center point of the vehicle's rear axle in the global coordinate system. Specifically, the formula for the kinematic model to predict the future state of the vehicle based on the actual state of the vehicle is shown in Formula (1):

[0048]

[0049] In the formula, the predicted pose is (x p , y p , θ p ), the coordinates of the center point of the vehicle's rear axle in the global coordinate system are denoted as (x p , y p ), its heading angle (orientation) is denoted as θ p , the wheelbase of the vehicle is l fr , the preset delay is t s , the preset delay is also the lag time or delay time, δ f is the front-wheel steering angle, the current longitudinal speed of the vehicle is v lon , the current pose of the vehicle is (x, y, θ), (x, y) is the position coordinates of the center point of the vehicle's rear axle in the global coordinate system, and θ is the orientation.

[0050] It is understandable that the kinematic model shown in Formula (1) is obtained by adding a preset delay to the kinematic model shown in Formula (2), and the kinematic model shown in Formula (2) is constructed according to the geometric relationship of the vehicle in the global coordinate system.

[0051] Exemplarily, referring to Figure 2 , Figure 2 is a geometric schematic diagram of a vehicle in the global coordinate system provided by the embodiments of the present disclosure, Figure 2 including a vehicle 210, a center point 220 of the vehicle's rear axle, a vehicle 211 after the preset delay, and a center point 221 of the vehicle's rear axle after the preset delay. The global coordinate system is denoted as XOY. The kinematic model of Formula (2) is as follows:[[]]

[0052]

[0053] In the formula, and are respectively the first derivatives of the position (lateral coordinate) and orientation (heading angle) of the vehicle in the global coordinate system.

[0054] Among them, the extended state observer 120 is used to observe the current lateral deviation, heading deviation and combined disturbance of the driverless vehicle.

[0055] It can be understood that the extended state observer 120 is used to observe and compensate the internal state of the steering system and the combined disturbance. The internal state of the steering system includes the lateral deviation and the heading deviation. The lateral deviation can be denoted as the lateral deviation state variable, and the heading deviation can be denoted as the heading deviation state variable. The combined disturbance of the steering system is the random disturbance that occurs during the driving of the vehicle. In addition, the combined disturbance is the set of factors that cause random deviations (lateral deviation and heading deviation) during the lateral tracking control process. It can be understood that, considering the vehicle actuator and the driving environment, etc., there are situations where the steering system of the vehicle has zero position drift, the GPS has calibration errors, and the driving road surface has a lateral slope, etc., which cause random lateral deviations of the vehicle. Therefore, the embodiments of the present disclosure design a discretized extended state observer 120 to observe and compensate the internal state of the system and the combined disturbance, so as to reduce the possible noise in the process of transmitting vehicle-related information. The discretized extended state observer 120 can be embedded and applied in a computer system, and the application is more convenient, and it is more automated than the traditional operations that require a lot of manpower and material resources, such as map fixed-point marking and regular vehicle maintenance, effectively reducing the waste of manpower and material resources.

[0056] Among them, the extended state observer 120 is specifically used to observe the current lateral deviation, heading deviation and combined disturbance of the driverless vehicle based on the second compensation gain, the current heading deviation, the current combined disturbance and the observation error; wherein, the second compensation gain is generated based on the current target control amount of the system and the compensation amount gain, and the observation error is the error between the current lateral deviation of the system and the observed value of the current lateral deviation of the system.

[0057] Among them, when designing the extended state observer 120 to observe and compensate the internal state of the steering system and the combined disturbance, the construction process of the extended state observer 120 includes:

[0058] Based on the kinematic model shown in Equation (2), a kinematic model with composite disturbances is established. Compared with the traditional vehicle kinematic model, the disturbance effect is injected, making the final control have better robustness. Specifically, according to the variation laws of the lateral deviation and the heading deviation, assuming that the vehicle is traveling at a constant speed, mathematical relationships of variables such as the lateral deviation, the heading deviation, the longitudinal speed of the vehicle, the lateral speed of the vehicle, and the front-wheel steering angle are established, that is, the equations of the nominal system are established. This mathematical relationship is shown in Equation (3):

[0059]

[0060] In the formula, is the first derivative of the heading deviation e θ , is the first derivative of the lateral deviation e y , the longitudinal speed is v lon , and the lateral speed is v lat .

[0061] Based on the above Equation (3), the mathematical formula of the kinematic error model with composite disturbances is derived as shown in the following Equation (4), where the derivative of Equation (3) is bounded.

[0062]

[0063] In the formula, is the first derivative of the heading deviation e θ , is the first derivative of the lateral deviation e y , is the error control quantity, d is the composite disturbance,, δ f is the front-wheel steering angle of the current control quantity, δ ff is the front-wheel steering angle of the current desired value.

[0064] Among them, the kinematic error model with composite disturbances is used to determine the current lateral deviation, heading deviation, and composite disturbance of the driverless vehicle according to the current reference trajectory offset and the predicted pose. Calculate the difference between the predicted lateral coordinate and the current lateral coordinate offset in the current reference trajectory offset as the current lateral deviation of the vehicle, calculate the difference between the predicted heading angle and the current heading deviation in the current reference trajectory offset as the current heading deviation of the vehicle, and then determine the current composite disturbance of the vehicle based on Equation (4).

[0065] After establishing the kinematic error model with compound disturbances, the lateral deviation is taken as the first observation variable of the to-be-established extended state observer, the product of the heading deviation and the longitudinal speed of the driverless vehicle is taken as the second observation variable of the to-be-established extended state observer, and the compound disturbance is taken as the third observation variable of the to-be-established extended state observer. The kinematic error model with compound disturbances is updated according to the first observation variable, the second observation variable and the third observation variable, that is, the above formula (4) is rewritten to obtain the updated kinematic error model with compound disturbances, as specifically shown in formula (5):

[0066]

[0067] In the formula, is the first derivative of the first observation variable, is the first derivative of the second observation variable, is the first derivative of the third observation variable.

[0068] It can be understood that in practical applications, during the driving process of the vehicle, there is certain noise in the vehicle-related information transmitted by the perception and positioning system, etc., resulting in the error state of the steering system not necessarily being completely accurate. Therefore, an extended state observer is designed to observe the internal state of the vehicle during the driving process of the vehicle. On the other hand, due to the influence of the compound disturbance on the steering system, the vehicle cannot complete the desired trajectory (reference trajectory) tracking action. Therefore, the designed extended state observer also takes the compound disturbance as the extended state of the steering system and observes it together with the error state of the vehicle. At the same time, it can also compensate the nominal control quantity according to the observation result of the compound disturbance, realizing the function of online resistance to external disturbances. In addition, this compound disturbance is not limited to any specific form, and only requires the disturbance derivative to be bounded. Specifically, the lateral deviation state variable is taken as the first observation variable of the to-be-established extended state observer, denoted as x1 = e y , the product of the heading deviation state variable and the longitudinal speed of the vehicle is taken as the second observation variable of the to-be-established extended state observer, denoted as x2 = v lon e θ , and the compound disturbance variable is taken as the third observation variable of the to-be-established extended state observer, that is, the compound disturbance d is taken as the extended state x3 of the system.

[0069] After obtaining the updated kinematic model with compound disturbances, based on the updated kinematic model with compound disturbances shown in formula (5), a third-order nonlinear extended state observer is designed, as specifically shown in the following formula (6).

[0070]

[0071] Wherein, z1, z2, and z3 are the observed values of x1, x2, and x3 output by the extended state observer, respectively. are the first derivatives of z1, z2, and z3 respectively, and e1 is the observed error of the lateral deviation. Among them, e1 = z1 - e y , f(e, σ, δ) is a function, where σ1, δ1, σ2, and δ2 are the relevant parameters of the function f(e, σ, δ). β1, β2, and β3 are the gains of the extended state observer.

[0072] After obtaining the extended state observer shown in formula (6), the extended state observer is discretized based on the Euler method. Since the extended state observer needs to be discretized before it can be applied in the embedded system, and considering the computational efficiency of the system after discretization, the extended state observer is discretized by the Euler method x(t1) = x(t0) + f(x0)Δt. The extended state observer can also be discretized by the bilinear discretization method. The specific discretization method is not limited. The discretized extended state observer is shown in the following formula (7):

[0073]

[0074] Wherein, k represents the current cycle, and k + 1 represents the next cycle; z1(k) to z3(k) can be understood as the observed values of the internal error and composite disturbance of the current system output by the extended state observer 120. z1(k) is the observed value of the current lateral deviation of the vehicle, z2(k) is the observed value of the current heading deviation of the vehicle, and z3(k) is the observed value of the current composite disturbance of the vehicle; z1(k) to z3(k) are the current lateral deviation, heading deviation, and composite disturbance of the vehicle output by the above error model. z1(k) is the current lateral deviation e y , z2(k) is the current heading deviation e θ , z3(k) is the composite disturbance d; is the second compensation gain amount. is the target control amount of the current cycle output by the first arithmetic unit 150, and b0 is the compensation gain. Among them, the compensation gain is determined according to the current longitudinal speed in the current driving information and the wheelbase of the driverless vehicle. Specifically, e1(k) is the observed error, which is the error between the current lateral deviation of the system and the observed value of the current lateral deviation of the system, reflecting the correlation between the observed value of the output lateral deviation of the extended state observer 120 and the input lateral deviation of the extended state observer 120. Specifically, e1(k) = z1(k) - e y; h is the sampling period of the lateral control system; β1, β2, and β3 are the gain parameters of the extended state observer.

[0075] Among them, the tracking differentiator 130 is used to filter the lateral error and the heading error. Among them, the lateral error is obtained according to the observed value of the lateral deviation output by the extended state observer and the current lateral coordinate offset, and the heading error is obtained according to the observed value of the heading deviation output by the extended state observer and the current heading angle offset.

[0076] It can be understood that the lateral control system 100 obtains the lateral error according to the observed value of the lateral deviation state variable output by the extended state observer 120 and the current lateral coordinate offset output by the reference trajectory planning module 110. The lateral error is denoted as e'. y , and obtains the heading error according to the observed value of the heading deviation state variable output by the extended state observer 120 and the current heading angle offset output by the reference trajectory planning module 110. The heading error is denoted as e'. θ , where the generation order of the heading error and the lateral error is not limited. Subsequently, the tracking differentiator 130 filters the heading error and the lateral error, and outputs the filtered heading error and the filtered lateral error to the feedback controller 140 to reduce the noise influence of the perception and positioning system in the vehicle, and avoid the tracking jitter or divergence phenomenon caused by noise during the vehicle driving. Compared with the traditional direct feedback control based on the control error, it is smoother and more stable. Among them, the filtering formula of the tracking differentiator 130 is shown in the following formulas (8) to (9).

[0077] Among them, the tracking differentiator 130 filters the lateral error e'. y The filtering formula is shown in formula (8):

[0078]

[0079] In the formula, the change rate of e' y is The change rate The observed value of is The filtered lateral error is r is the fast factor of the tracking differentiator 130, which is an adjustable parameter of the tracking differentiator 130, and fhan(*) is the fastest control synthesis function.

[0080] Among them, the tracking differentiator 130 filters the lateral deviation error e' θ The filtering formula is shown in formula (9):

[0081]

[0082] In the formula, e' θThe rate of change is Rate of change The observed value of The filtered lateral error is

[0083] Among them, the fastest control synthesis function fhan(x1, x2, r, h) is shown in formula (10):

[0084]

[0085] In the formula, fsg(y, c) = (sign(x + c) - sign(x - c)) / 2.

[0086] Among them, the feedback controller 140 is used to determine the nominal control quantity based on the filtered lateral error and the filtered heading error.

[0087] It can be understood that the feedback controller 140 receives the filtered lateral error and the filtered heading error, and determines the nominal control quantity based on the filtered lateral error and the filtered heading error. The calculation formula of the feedback controller 140 is shown in formula (11):

[0088]

[0089] In the formula, u f (k) is the nominal control quantity, f ctrl (*) is any control law that can make the nominal system converge, including but not limited to pure tracking methods, numerical control (Packet Identifier, PID) methods, model predictive control (Model Predictive Control, MPC) methods, etc.

[0090] Among them, the first arithmetic unit 150 is used to calculate the target control quantity for lateral control of the driverless vehicle based on the first compensation gain quantity and the nominal control quantity, where the first compensation gain quantity is obtained according to the observed value and the compensation quantity gain of the composite disturbance output by the extended state observer.

[0091] It can be understood that the lateral control system 100 obtains the first compensation gain quantity according to the observed value and the compensation quantity gain of the composite disturbance output by the extended state observer 120. Subsequently, the first arithmetic unit 150 calculates the target control quantity for lateral control of the vehicle based on the first compensation gain quantity and the nominal control quantity. The formula for the first arithmetic unit 150 to calculate the target control quantity is shown in formula (12):

[0092]

[0093] In the formula, u(k) is the target control quantity, u d(k) is the first compensation gain. Specifically,

[0094] Understandably, after obtaining the target control quantity of the driverless vehicle, the lateral control system 100 can also send the target control quantity to the underlying system of the driverless vehicle through the communication network, and the underlying system completes the actual steering action of the vehicle.

[0095] Wherein, the lateral control system 100 may further include: a second arithmetic unit 160, a third arithmetic unit 170, and a fourth arithmetic unit 180, wherein:

[0096] The second arithmetic unit 160 is configured to determine a lateral error based on the observed value of the lateral deviation and the current lateral coordinate offset output by the reference trajectory planning module 110, and output the lateral error to the tracking differentiator 130.

[0097] Understandably, the second arithmetic unit 160 can be directly arranged inside the lateral control system 100, and is connected to the reference trajectory planning module 110, the extended state observer 120, and the tracking differentiator 130 at the same time. It can also be arranged inside the reference trajectory planning module 110, or arranged inside the tracking differentiator 130, which is not limited herein. The second arithmetic unit 160 receives the current lateral coordinate offset output by the reference trajectory planning module 110 and the observed value of the lateral deviation state variable output by the extended state observer 120, and performs a subtraction operation based on the current lateral coordinate offset and the observed value of the lateral deviation state variable, and outputs the lateral error to the tracking differentiator 130. The formula for the second arithmetic unit 160 to determine the lateral error is shown in formula (13):

[0098] e′ y (k)=offset(x p ,y p ) - z1(k) (13)

[0099] In the formula, e′ y (k) is the lateral error, offset(x p ,y p ) is the current lateral coordinate offset, and z1(k) is the observed value of the current lateral deviation state variable.

[0100] Wherein, the third arithmetic unit 170 is configured to determine the heading error based on the observed value of the heading deviation and the current heading angle offset, and output the heading error to the tracking differentiator 130.

[0101] It is understandable that the third arithmetic unit 170 can also be directly arranged inside the lateral control system 100. The third arithmetic unit 170 is connected to the reference trajectory planning module 110, the extended state observer 120, and the tracking differentiator 130. It can also be arranged inside the reference trajectory planning module 110 or inside the tracking differentiator 130, which is not limited herein. The third arithmetic unit 170 receives the current heading angle offset output by the reference trajectory planning module 110 and the observed value of the heading deviation state variable output by the extended state observer 120, and performs a subtraction operation based on the current heading angle offset and the observed value of the heading deviation state variable, and outputs the heading deviation state variable to the tracking differentiator 130. The formula for the third arithmetic unit 170 to determine the heading deviation state variable is shown in formula (14):

[0102] e′ θ (k)=offset(θ p )-z2(k) (14)

[0103] In the formula, e′ θ (k) is the heading deviation, offset(θ p ) is the current heading angle offset, and z2(k) is the observed value of the current heading deviation.

[0104] Among them, the fourth arithmetic unit 180 is used to generate a second compensation gain based on the current target control amount of the system and the compensation gain amount, where the compensation gain amount is determined according to the current longitudinal speed in the current driving information and the wheelbase of the driverless vehicle.

[0105] It is understandable that the fourth arithmetic unit 180 can also be directly arranged inside the lateral control system 100 and is connected to the first arithmetic unit 150 and the extended state observer 120 at the same time. The fourth arithmetic unit 180 receives the current target control amount output by the first arithmetic unit 150, and performs a multiplication operation based on the current target control amount u(k) and the compensation gain amount b0 to generate a second compensation gain amount, and outputs the second compensation gain amount to the extended state observer 120 as an input of the extended state observer 120.

[0106] The embodiments of the present disclosure provide a lateral control system. The lateral control system includes a reference trajectory planning module, an unmanned vehicle module, a tracking differentiator, a feedback controller, an extended state observer, and a first arithmetic unit. Among them, the reference trajectory planning module is used to provide the current reference trajectory offset of the unmanned vehicle in real time; the unmanned vehicle model is used to determine the predicted pose of the vehicle according to the current driving information of the unmanned vehicle and a preset delay, fully considering the problem of steering system oscillation caused by steering lag or steering delay in the steering system. Based on the predicted pose, the influence of oscillation can be effectively reduced. Subsequently, based on the error model, according to the predicted pose and the current reference trajectory offset, the current lateral deviation, heading deviation, and composite disturbance of the unmanned vehicle are determined; the extended state observer is used to observe the lateral deviation state variable, heading deviation state variable, and composite disturbance, and generate corresponding observed values. Based on the extended state observer, the problem of vehicle lateral deviation caused by adverse factors such as zero position drift of the steering system, uneven left and right tire pressures, and lateral road slopes is effectively handled. The tracking differentiator is used to filter the error state variables (lateral error and heading error) of the steering system. The error state variables are obtained based on the observed values of the lateral deviation state variable, the observed values of the heading deviation state variable, and the current reference trajectory offset. The filtering process can effectively avoid the noise influence introduced after the perception positioning system and the extended state observer are discretized. By filtering and smoothing the error state variables through the tracking differentiator, the tracking jitter or divergence phenomenon caused by noise during the vehicle driving process is also avoided. The feedback controller is used to determine the nominal control quantity based on the filtered lateral error and the filtered heading error. The first arithmetic unit is used to calculate the target control quantity for lateral control of the vehicle based on the first compensation gain quantity and the nominal control quantity. The first compensation control quantity is obtained based on the observed value of the composite disturbance and the compensation quantity gain. By observing the composite disturbance based on the extended state observer, the automatic compensation function for the final control quantity of the vehicle can be realized when a single disturbance acts or multiple disturbances act simultaneously, enabling the vehicle to continuously and stably track the desired trajectory. The system provided by the present disclosure has generalization ability in different scenarios, can not only solve the adverse effects caused by a single disturbance, but also resist the adverse effects caused by the simultaneous action of multiple disturbances, reducing the influence of lateral offset on path tracking to a certain extent, improving the safety and comfort of the vehicle, and at the same time reducing the waste of manpower and material resources.

[0107] Based on the above embodiments, Figure 3 It is a schematic flowchart of a lateral control method provided by an embodiment of the present disclosure, which is applied to an unmanned vehicle and can be applied to the lateral control system 100 configured in the unmanned vehicle. Specifically, it includes the following steps S310 to S360 as shown in Figure 3 follows:

[0108] S310. Obtain the current reference trajectory offset of the driverless vehicle, where the current reference trajectory offset includes a current lateral coordinate offset and a current heading angle offset.

[0109] It can be understood that for the specific implementation steps of S310, refer to the implementation steps of the reference trajectory planning module 110 in the above-mentioned lateral control system 100, which will not be elaborated here.

[0110] S320. Determine the current lateral deviation, heading deviation, and composite disturbance of the driverless vehicle according to the obtained current driving information of the driverless vehicle, a preset delay, and the current reference trajectory offset.

[0111] It can be understood that for the specific implementation steps of S320, refer to the implementation steps of the driverless vehicle model 190 in the above-mentioned lateral control system 100, which will not be elaborated here.

[0112] S330. Based on a pre-established extended state observer, observe the current lateral deviation, heading deviation, and composite disturbance of the driverless vehicle, and determine the observed value of the current lateral deviation, the observed value of the heading deviation, and the observed value of the composite disturbance of the driverless vehicle.

[0113] Among them, the composite disturbance is a random lateral deviation that occurs during the driving process of the driverless vehicle.

[0114] It can be understood that the specific implementation steps of S330 are the same as those of the extended state observer 120 in the above-mentioned lateral control system 100, which will not be elaborated here.

[0115] S340. Filter the lateral error and the heading error.

[0116] Among them, the lateral error is determined based on the observed value of the lateral deviation output by the extended state observer and the current lateral coordinate offset, and the heading error is determined based on the observed value of the heading deviation output by the extended state observer and the current heading angle offset.

[0117] It can be understood that for the specific implementation steps of S340, refer to the implementation steps of the tracking differentiator 130 in the above-mentioned lateral control system 100, which will not be elaborated here.

[0118] S350. Determine the nominal control quantity based on the filtered lateral error and the filtered heading error.

[0119] It can be understood that for the specific implementation steps of S340, refer to the implementation steps of the feedback controller 140 in the above-mentioned lateral control system 100, which will not be elaborated here.

[0120] S360. Calculate the target control quantity for lateral control of the driverless vehicle based on the first compensation gain quantity and the nominal control quantity.

[0121] Wherein, the first compensation gain quantity is calculated based on the observed value of the composite disturbance output by the extended state observer and the compensation quantity gain.

[0122] It can be understood that for the specific implementation steps of S360, refer to the implementation steps of the first arithmetic unit 150 in the above-mentioned lateral control system 100, which will not be elaborated here.

[0123] The embodiment of the present disclosure provides a lateral control method. By using a pre-designed extended state observer to observe the error state and composite disturbance inside the steering system, and at the same time, it can automatically compensate the control quantity, which can effectively resist the adverse effects brought by internal and external factors such as steering zero position drift, uneven left and right tire pressures, and lateral road gradients. Obtain the current reference trajectory offset of the driverless vehicle, and then determine the predicted pose of the driverless vehicle according to the obtained current driving information of the driverless vehicle and the preset delay, which effectively reduces the oscillation caused by steering lag or steering delay of the steering system. Determine the current lateral deviation, heading deviation and composite disturbance of the vehicle according to the predicted pose and the current reference trajectory offset. Subsequently, filter the lateral error and the heading error to avoid the tracking jitter or divergence phenomenon caused by noise during the driving process of the vehicle. Based on the filtered lateral error and the filtered heading error, determine the nominal control quantity, and finally calculate the target control quantity for lateral control of the vehicle based on the first compensation gain quantity and the nominal control quantity, so that the lateral control of the vehicle by the steering system based on the target control quantity has better robustness.

[0124] Based on the above embodiment, Figure 4 It is a schematic flowchart of a lateral control method provided by an embodiment of the present disclosure. For the establishment process of the extended state observer used in the above S330, the extended state observer can be designed and constructed based on a server and a terminal. The server and the terminal can communicate with the vehicle to transmit the designed extended state observer to the vehicle, specifically including the following steps S410 to S450 as Figure 4 shown:

[0125] S410. Based on the kinematic model, determine the predicted pose according to the driving information of the driverless vehicle and the preset delay.

[0126] It is understandable that based on the kinematic model shown in the above formula (1), the predicted pose is determined according to the driving information of the vehicle and a preset delay. The driving information includes the current pose, the current longitudinal speed, the wheelbase, and the front wheel steering angle of the vehicle. The current pose can be the real-time pose information provided by the perception and positioning system configured in the vehicle. Specifically, the detailed implementation steps of formula (1) will not be elaborated here.

[0127] S420. Determine the lateral deviation and the heading deviation of the driverless vehicle according to the predicted pose and the reference trajectory offset of the driverless vehicle.

[0128] It is understandable that on the basis of the above S410, the reference trajectory offset of the vehicle is obtained. The reference trajectory offset can be obtained from the real-time reference trajectory provided by the driverless planning system configured in the vehicle. The predicted pose of the center point of the rear axle of the vehicle, which takes into account the steering delay and is predicted, is used as the pose of the vehicle. Subsequently, according to the predicted pose and the reference trajectory of the vehicle, the lateral state variable and the heading state variable are determined. The reference trajectory also includes the pose information of the vehicle.

[0129] Among them, the determination of the lateral deviation and the heading deviation in the above S420 can be specifically implemented through the following steps:

[0130] Determine the reference pose offset on the reference trajectory offset of the driverless vehicle according to the predicted pose.

[0131] Based on the predicted lateral coordinate in the predicted pose and the reference lateral coordinate offset in the reference pose offset, determine the lateral deviation of the driverless vehicle.

[0132] Based on the predicted heading angle in the predicted pose and the reference heading angle offset in the reference pose offset, determine the heading deviation of the driverless vehicle.

[0133] It is understandable that on the basis of the above S410, a reference point is selected on the reference trajectory offset given by the driverless planning system according to the projection method, and the pose offset corresponding to the reference point is used as the reference pose offset. The predicted pose includes the predicted lateral coordinate and the predicted heading angle, and the reference pose offset also includes the reference lateral coordinate offset and the reference heading angle offset. It is also possible to directly obtain the reference trajectory, select a reference point on the reference trajectory, and use the pose information corresponding to the reference point as the reference pose. The reference pose also includes the reference lateral coordinate and the reference heading angle. Subsequently, the Euclidean distance between the predicted lateral coordinate and the reference lateral coordinate is calculated as the lateral deviation state variable of the vehicle, that is, the Euclidean distance between the reference point and the center point of the rear axle of the vehicle is calculated as the lateral deviation state variable, and the difference between the predicted heading angle and the reference heading angle is calculated and used as the heading deviation state variable.

[0134] Exemplarily, refer to Figure 5 , Figure 5 a schematic plan view provided by an embodiment of the present disclosure, Figure 5 which is a schematic plan view in a global coordinate system (XOY), Figure 5 including a reference trajectory 510, a reference point 520 and the center point 530 of the rear axle of the vehicle. Among them, the coordinates of the center point 530 of the rear axle in the global coordinate system are the predicted lateral coordinates, denoted as (x p , y p ), its longitudinal speed v lon The included angle between the virtual extension line 540 of the heading where it is located and the parallel line 550 of the X-axis is the predicted heading angle θ p . Based on the projection of the center point 530 of the rear axle onto the reference trajectory 510, the reference point 520 is determined on the reference trajectory 510. The reference lateral coordinate of the reference point 520 is denoted as (x r , y r ). The included angle between the tangent line 560 at the reference point 520 and the parallel line 550 of the X-axis is the reference heading angle θ r . The difference between the predicted heading angle θ p and the reference heading angle θ r is the heading deviation state variable e θ , that is, the included angle between the tangent line 560 at the reference point 520 and the straight line 540 where the heading of the vehicle is located is used as the heading deviation state variable, and the Euclidean distance between the reference point 520 and the center point 530 of the rear axle of the vehicle is used as the lateral deviation state variable e y .

[0135] S430. Based on the kinematic model, according to the lateral deviation, the heading deviation and the composite disturbance, establish a kinematic error model with composite disturbance.

[0136] It can be understood that, on the basis of the above S420, the kinematic error model with composite disturbance is the above formula (4), and the specific establishment process will not be elaborated here.

[0137] S440. Take the lateral deviation as the first observation variable of the to-be-established extended state observer, take the product of the heading deviation and the longitudinal speed of the driverless vehicle as the second observation variable of the to-be-established extended state observer, and take the composite disturbance as the third observation variable of the to-be-established extended state observer.

[0138] It can be understood that, on the basis of the above S430, based on the kinematic error model with composite disturbance shown in the above formula (4), design an extended state observer to observe and compensate the internal state of the steering system and the composite disturbance.

[0139] S450. Update the kinematic error model with compound disturbances according to the first observation variable, the second observation variable, and the third observation variable, and establish the extended state observer according to the updated kinematic error model with compound disturbances.

[0140] It can be understood that, based on S440 above, after the kinematic error model with compound disturbances as shown in formula (4) is established, the kinematic model with compound disturbances is updated according to the first observation variable, the second observation variable, and the third observation variable. The updated kinematic error model with compound disturbances is as shown in formula (5) above.

[0141] It can be understood that in practical applications, the embodiments of the present disclosure provide a method for tuning the gain parameters in the extended state observer. Specifically, an engineering formula for obtaining the parameters of a typical third-order extended state observer (ESO) according to the particle swarm optimization algorithm is obtained, where β1, β2, and β3 satisfy the relational expression as shown in formula (15):

[0142]

[0143] In the formula, ω0 is a weight coefficient, which can be used as the parameters of the extended state observer in engineering applications. By adjusting ω0, that is, mapping the three parameters to be tuned to one parameter, the efficiency of parameter tuning is greatly improved, and compared with the traditional disturbance observer, it has more practical application feasibility.

[0144] A lateral control method provided by the embodiments of the present disclosure has generalization ability in different scenarios by designing an extended state observer that simultaneously observes the error state and compound disturbances of the steering system. It can not only solve single disturbances but also resist the adverse effects caused by the simultaneous action of multiple disturbances.

[0145] Figure 6 It is a schematic structural diagram of an electronic device provided by the embodiments of the present disclosure. The electronic device provided by the embodiments of the present disclosure can execute the processing flow provided by the above embodiments. As Figure 6 shown, the electronic device 600 includes: a processor 610, a communication interface 620, and a memory 630; wherein, the computer program is stored in the memory 630 and is configured to be executed by the processor 610 to perform the lateral control method as described above.

[0146] In addition, the embodiments of the present disclosure also provide a computer-readable storage medium, on which a computer program is stored, and the computer program is executed by a processor to implement the lateral control method described in the above embodiments.

[0147] In addition, an embodiment of the present disclosure also provides a computer program product, which includes a computer program or instruction. When the computer program or instruction is executed by a processor, the above-described lateral control method is implemented.

[0148] It should be noted that, in this document, relational terms such as "first" and "second" are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the term "comprising", "including" or any other variant thereof is intended to cover a non-exclusive inclusion, such that a process, method, article or device comprising a series of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article or device. Without further limitation, an element defined by the phrase "comprising a..." does not exclude the presence of additional identical elements in the process, method, article or device comprising the element.

[0149] The above are only specific embodiments of the present disclosure, enabling those skilled in the art to understand or implement the present disclosure. Various modifications to these embodiments will be obvious to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present disclosure. Therefore, the present disclosure will not be limited to these embodiments described herein, but rather to the broadest scope consistent with the principles and novel features disclosed herein.

Claims

1. A lateral control system, applied to a driverless vehicle, includes a reference trajectory planning module, an extended state observer, a tracking differentiator, a feedback controller, a first arithmetic unit, and a driverless vehicle model, wherein: The reference trajectory planning module is used to provide the current reference trajectory offset of the driverless vehicle, where the current reference trajectory offset includes the current lateral coordinate offset and the current heading angle offset; The driverless vehicle model is used to determine the current lateral deviation, heading deviation and composite disturbance of the driverless vehicle according to the acquired current driving information of the driverless vehicle, the preset delay and the current reference trajectory offset; The extended state observer is used to observe the current lateral deviation, heading deviation and composite disturbance of the driverless vehicle; The tracking differentiator is used to filter the lateral error and the heading error, where the lateral error is obtained according to the observed value of the lateral deviation output by the extended state observer and the current lateral coordinate offset, and the heading error is obtained according to the observed value of the heading deviation output by the extended state observer and the current heading angle offset; The feedback controller is used to determine the nominal control quantity based on the filtered lateral error and the filtered heading error; The first arithmetic unit is used to calculate the target control quantity for lateral control of the driverless vehicle based on the first compensation gain quantity and the nominal control quantity, where the first compensation gain quantity is obtained according to the observed value of the composite disturbance output by the extended state observer and the compensation gain, and the compensation gain is determined according to the current longitudinal speed in the current driving information and the wheelbase of the driverless vehicle.

2. The system according to claim 1, wherein The system further includes a second arithmetic unit, which is used to determine the lateral error based on the observed value of the lateral deviation and the current lateral coordinate offset, and output the lateral error to the tracking differentiator.

3. The system according to claim 1, wherein The system further includes a third arithmetic unit, which is used to determine the heading error based on the observed value of the heading deviation and the current heading angle offset, and output the heading error to the tracking differentiator.

4. The system according to claim 1, wherein The extended state observer is specifically used for: observing the current lateral deviation, heading deviation and composite disturbance of the driverless vehicle based on the second compensation gain quantity, the current heading deviation, the current composite disturbance and the observation error; where the second compensation gain quantity is generated based on the current target control quantity of the system and the compensation gain, and the observation error is the error between the current lateral deviation of the system and the observed value of the current lateral deviation of the system.

5. The system according to claim 1, wherein The system further includes a fourth arithmetic unit, which is used to generate the second compensation gain quantity based on the current target control quantity of the system and the compensation gain, where the compensation gain is determined according to the current longitudinal speed in the current driving information and the wheelbase of the driverless vehicle.

6. The system according to claim 1, wherein The current driving information further includes the current pose, the current front wheel steering angle and the current longitudinal speed, The unmanned vehicle model includes a kinematic model and an error model. The kinematic model is used to determine the predicted pose of the unmanned vehicle according to the current pose, the current front wheel steering angle, the wheelbase of the unmanned vehicle, a preset delay, and the current longitudinal speed. Among them, the predicted pose includes a predicted lateral coordinate and a predicted heading angle. The error model is used to determine the current lateral deviation, heading deviation, and composite disturbance of the unmanned vehicle according to the current reference trajectory offset and the predicted pose.

7. A lateral control method, applied to a driverless vehicle, wherein The method includes: Obtain the current reference trajectory offset of the unmanned vehicle, where the current reference trajectory offset includes a current lateral coordinate offset and a current heading angle offset. Determine the current lateral deviation, heading deviation, and composite disturbance of the unmanned vehicle according to the obtained current driving information of the unmanned vehicle, the preset delay, and the current reference trajectory offset. Based on a pre-established extended state observer, observe the current lateral deviation, heading deviation, and composite disturbance of the unmanned vehicle, and determine the observed value of the current lateral deviation, the observed value of the heading deviation, and the observed value of the composite disturbance of the unmanned vehicle. Filter the lateral error and the heading error, where the lateral error is determined based on the observed value of the lateral deviation output by the extended state observer and the current lateral coordinate offset, and the heading error is determined based on the observed value of the heading deviation output by the extended state observer and the current heading angle offset. Determine a nominal control quantity based on the filtered lateral error and the filtered heading error. Calculate a target control quantity for lateral control of the unmanned vehicle based on a first compensation gain quantity and the nominal control quantity, where the first compensation gain quantity is calculated based on the observed value of the composite disturbance output by the extended state observer and a compensation quantity gain, and the compensation quantity gain is determined according to the current longitudinal speed in the current driving information and the wheelbase of the unmanned vehicle.

8. The method according to claim 7, wherein The establishment process of the extended state observer includes: Based on the kinematic model, determine the predicted pose according to the driving information of the unmanned vehicle and the preset delay. Determine the lateral deviation and heading deviation of the unmanned vehicle according to the predicted pose and the reference trajectory offset of the unmanned vehicle. Based on the kinematic model, establish a kinematic error model with composite disturbance according to the lateral deviation, the heading deviation, and the composite disturbance. Take the lateral deviation as the first observation variable of the to-be-established extended state observer, take the product of the heading deviation and the longitudinal speed of the unmanned vehicle as the second observation variable of the to-be-established extended state observer, and take the composite disturbance as the third observation variable of the to-be-established extended state observer. Update the kinematic error model with composite disturbance according to the first observation variable, the second observation variable, and the third observation variable, and establish the extended state observer according to the updated kinematic error model with composite disturbance.

9. The method according to claim 8, wherein Determining the lateral deviation and the heading deviation of the driverless vehicle according to the predicted pose and the reference trajectory offset of the driverless vehicle includes: Determining a reference pose offset on the reference trajectory offset of the driverless vehicle according to the predicted pose; Determining the lateral deviation of the driverless vehicle based on the predicted lateral coordinate in the predicted pose and the reference lateral coordinate offset in the reference pose offset; Determining the heading deviation of the driverless vehicle based on the predicted heading angle in the predicted pose and the reference heading angle offset in the reference pose offset.

10. An electronic device, characterized in that, Comprising: A memory; A processor; And A computer program; Wherein, the computer program is stored in the memory and is configured to be executed by the processor to implement the lateral control method according to any one of claims 7 to 9.

11. A computer-readable storage medium, on which a computer program is stored, characterized in that, When the computer program is executed by the processor, the steps of the lateral control method according to any one of claims 7 to 9 are implemented.

Images