Method and system for lateral control of a vehicle

Abstract

Method and system for lateral control of a vehicle. A computer-implemented method for lateral control of a vehicle includes the following steps performed by computer hardware components: determining a position error of the vehicle; determining an orientation error of the vehicle; determining a cost function based on the position error and the orientation error using a circular transform; and processing the cost function in a model predictive controller to laterally control the vehicle.

Description

Technical Field

[0001] The present disclosure relates to methods and systems for lateral control of a vehicle. Background Art

[0002] Lateral control algorithms in autonomous vehicles are designed to safely and comfortably perform lane centering and merging into the center of a lane.

[0003] Therefore, there is effective and reliable lateral control. Summary of the Invention

[0004] In one aspect, the present invention relates to a computer-implemented method for lateral control of a vehicle, the computer-implemented method comprising the following steps performed (in other words: carried out) by computer hardware components: determining a position error of the vehicle; determining an orientation error of the vehicle; determining a cost function based on the position error and the orientation error using a circular transformation; and processing the cost function in a model predictive controller to laterally control the vehicle.

[0005] In other words, a lateral control method in an autonomously driven vehicle is provided to safely and comfortably perform lane centering and merging into the center of a lane.

[0006] According to an embodiment, the cost function comprises a non-holonomic ellipsoid cost function.

[0007] According to an embodiment, the cost function comprises an integration of an error term, wherein the error term comprises a position error and an orientation error.

[0008] According to an embodiment, the error term comprises a product based on a position error and an orientation error.

[0009] According to one embodiment, the cost function is based on a cosine function.

[0010] According to one embodiment, the cost function is based on a sine function.

[0011] According to one embodiment, the cost function is based on a tangent function.

[0012] According to one embodiment, the cost function is determined according to the following formula:

[0013] e tr =cosθ e 2 *tan -1 (dd ref )+sinθ e 2 *θ e (2)

[0014]

[0015] According to an embodiment, laterally controlling the vehicle comprises determining a lateral offset of the vehicle.

[0016] According to an embodiment, laterally controlling the vehicle comprises determining an orientation error of the vehicle.

[0017] According to an embodiment, laterally controlling the vehicle comprises determining an optimal steering angle value.

[0018] Obtaining the cost function for model predictive control to calculate the steering wheel angle involves determining the vehicle's lateral offset and orientation error relative to the lane center. The result of this compensation is a more human-like cost function for steering and coordination within the controller.

[0019] In another aspect, the present disclosure relates to a computer system comprising a plurality of computer hardware components configured to perform some or all of the steps of the computer-implemented methods described herein.The computer system may be part of a vehicle.

[0020] In another aspect, the invention relates to a vehicle comprising a computer system as described herein and a sensor configured to determine a position error and / or an orientation error.

[0021] According to an embodiment, the sensor comprises at least one of a radar sensor, a lidar sensor, an ultrasonic sensor, a camera or a global navigation satellite system sensor.

[0022] The computer system may include a plurality of computer hardware components (e.g., a processor, such as a processing unit or a processing network; at least one memory, such as a memory unit or a memory network; and at least one non-transitory data storage device). It should be understood that additional computer hardware components may be provided and used to perform the steps of the computer-implemented method in the computer system. The non-transitory data storage device and / or the memory unit may include a computer program for instructing a computer, for example, using the processing unit and the at least one memory unit, to perform some or all steps or aspects of the computer-implemented method described herein.

[0023] In another aspect, the present disclosure is directed to a non-transitory computer-readable medium comprising instructions for performing some or all steps or aspects of the computer-implemented methods described herein. The computer-readable medium can be configured as: an optical medium, such as a compact disc (CD) or a digital versatile disc (DVD); a magnetic medium, such as a hard disk drive (HDD); a solid-state drive (SSD); a read-only memory (ROM), such as a flash memory; and the like. In addition, the computer-readable medium can be configured as a data storage device accessible via a data connection, such as an Internet connection. The computer-readable medium can be, for example, an online data repository or cloud storage.

[0024] The present disclosure is also directed to a computer program that instructs a computer to perform several or all steps or aspects of the computer-implemented method described herein.

[0025] The methods, devices, and systems described herein may be used in advanced driver assistance systems (ADAS). BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Exemplary embodiments and functions of the present disclosure are described herein with reference to the following schematically illustrated drawings:

[0027] Figure 1 is a model predictive controller according to various embodiments;

[0028] Figure 2 is the comparison of steering responses between different cost functions;

[0029] Figure 3 is the cross-tracking error comparison between different cost functions;

[0030] Figure 4 is the comparison of orientation errors between different cost functions;

[0031] Figure 5 is a flowchart illustrating a method for lateral control of a vehicle according to various embodiments; and

[0032] Figure 6 A computer system having a plurality of computer hardware components configured to perform the steps of a computer-implemented method for lateral control of a vehicle according to various embodiments. DETAILED DESCRIPTION

[0033] An autonomous vehicle must not only observe the safe distance to itself and other objects in front and behind it, but also to the sides of the vehicle. This can be done by determining the lateral distance between the center of the lane and the host vehicle.

[0034] According to various embodiments, a mathematical cost function is provided to cause a model predictive controller to minimize the lateral distance and provide an optimal solution. In this regard, the cost function may include an ellipsoid function that incorporates the vehicle's lateral offset and orientation error relative to the lane center. Based on this cost function, system dynamics, and applied constraints, the model predictive controller can provide an optimal steering wheel angle for lateral control of the vehicle. The resulting compensation can be similar to a human steering and coordination cost function within the controller.

[0035] According to various embodiments, an apparatus and method for lateral control of a vehicle using a specific cost function method of a model predictive controller may be provided.

[0036] Figure 1 A diagram 100 of a model predictive controller is shown.

[0037] Model predictive controllers can be powerful controllers. As the name suggests, they can predict the vehicle's state, such as the lateral offset with respect to a reference and the vehicle's orientation error with respect to the reference, over a finite time horizon (T). The vehicle's reference value can be calculated by the planning block and then introduced into the model predictive control (MPC), which then generates a control signal at the current time instant to compensate for the error.

[0038] exist Figure 1 , a time axis 102 (with past time 104 and future time 106), a desired set point 108, a measured state 110, a closed loop input 112, a remeasured state 114, a predicted state 116, an optimal input trajectory 118, a re-predicted state 120, and a re-optimal input trajectory 122 are shown. k ) 124 and the corresponding prediction time domain (T) 126. The backward time domain (t k+1 )128 and the corresponding prediction time domain (T)130.

[0039] The following symbols can be used:

[0040] t k - Current time step;

[0041] t - time axis moving forward in time;

[0042] T-prediction time domain (the duration of the state prediction);

[0043] Δt-sampling period;

[0044] u * (τ k )-optimal control value at the current time step;

[0045] x ref (t k ) – reference trajectory of the state;

[0046] -Prediction status.

[0047] The optimizer in the model predictive controller can use the cost function according to various embodiments to provide the optimal control value at the current time step. In this case, the steering angle can be the optimal control signal sequence (u * ), which is used to actuate the vehicle to minimize the orientation error and lateral offset (the state of the system). Based on the designed system dynamics and the cost function, a trajectory prediction can also be calculated. The obtained control sequence can be a motion control sequence that can be applied to the vehicle.

[0048] One approach for minimizing the cross-tracking error and orientation relative to the reference may be to introduce these terms in the cost function as follows:

[0049]

[0050] In equation (1):

[0051] J represents the cost function that needs to be minimized;

[0052] d represents the initial value of the lateral offset obtained from the vehicle sensor input;

[0053] d ref A reference value that the vehicle should track, such as the center of a lane;

[0054] k d represents the tuning gain to minimize the cross-tracking error;

[0055] θ e represents the orientation error relative to a reference, such as relative to the center of the lane; and

[0056] k θ represents the tuning gain that minimizes the orientation error.

[0057] The disadvantage of this cost is that it cannot compensate for both cross-tracking error and orientation error simultaneously. This can lead to contradictory behavior in the cost evaluation, and thus the steering commands generated from the MPC are a suboptimal set resulting in poor steering performance.

[0058] According to various embodiments, different approaches to the cost function of equation (1) can be provided. According to various embodiments, the orientation error and cross-tracking error can be mapped to a single tracking error signal by using a circular transform. Thus, this tracking signal can be used as a traditional quadratic term in the cost. The axes of the circle can be scaled by the orientation error and cross-tracking error and thus unified based on orientation error units (radians).

[0059] The following equations (2) and (3) show modified cost functions according to various embodiments:

[0060] e tr =cosθ e 2 *tan -1 (dd ref )+sinθ e 2 *θ e (2)

[0061]

[0062] The various variables in equation (2) are the same as those described with reference to equation (1) above.

[0063] Equations (2) and (3) provide incomplete ellipsoid cost functions for model predictive control according to various embodiments.

[0064] In the following, curves and graphs are shown that show a comparison between the steering performance according to the cost function of equation (1) and the cost functions according to equations (2) and (3). The use case of the following graphs is a behavior similar to a lane change, where the required lateral offset from the initial offset is approximately the lane width (in this case, 3.5m). The set speed of the vehicle may be 36m / s.

[0065] Figure 2 Graph 200 shows how, for the same maneuver, the steering response has a smoother rise and significantly lower amplitude. In real-world testing, this results in a more comfortable lane change for the passengers. Horizontal axis 202 represents time, and vertical axis 204 represents steering angle. Curve 206 is obtained using the cost function of equation (1), and curve 208 is obtained using the cost functions of equations (2) and (3).

[0066] Figure 3Graph 300 shows a comparison between the lateral offsets for the two cases, and how the cost (function) according to equations (2) and (3) causes the rate of lateral offset to be less than conventional, again resulting in a smoother maneuver. Furthermore, a difference can be noted in the overshoot between the graphs, indicating that the costs according to equations (2) and (3) have less overshoot for the same maneuver than the cost function according to equation (1). The horizontal axis 302 represents time, and the vertical axis 304 represents cross-tracking error. Curve 306 is obtained using the cost function of equation (1), and curve 308 is obtained using the cost function of equations (2) and (3).

[0067] Figure 4 A graph 400 showing the difference in orientation error is shown. This can have significant differences, where the compensation for the error using the modified cost is much better than the traditional compensation. The horizontal axis 402 represents time, and the vertical axis 404 represents the orientation error. Curve 406 is obtained using the cost function of equation (1), and curve 408 is obtained using the cost functions of equations (2) and (3).

[0068] Figure 5 A flowchart 500 is shown illustrating the flow of a method for lateral control of a vehicle according to various embodiments. At 502, a position error of a vehicle can be determined. At 504, an orientation error of the vehicle can be determined. At 506, a cost function based on the position error and the orientation error using a circular transformation can be determined. At 508, the cost function can be processed in a model predictive controller to laterally control the vehicle.

[0069] The error may be measured about (or relative to) the center of the lane.

[0070] A cost function (which may be an incomplete cost function) may be used in a model predictive controller to calculate an optimal steering angle, which may be used for lateral control of the vehicle.

[0071] According to various embodiments, the cost function may include or may be an incomplete ellipsoid cost function.

[0072] According to various embodiments, the cost function may include an error term or may be an integral of the error terms, wherein the error term includes a position error and an orientation error. The integral may be an integral according to equation (3).

[0073] According to various embodiments, the error term may include or may be based on the product of the position error and the orientation error.

[0074] According to various embodiments, the cost function (or error term) may be based on a cosine function.

[0075] According to various embodiments, the cost function (or error term) may be based on a sine function.

[0076] According to various embodiments, the cost function (or error term) may be based on a tangent function.

[0077] According to various embodiments, the cost function may be determined according to equations (2) and (3). According to various embodiments, the error term may be determined according to equation (2).

[0078] According to various embodiments, laterally controlling the vehicle may include or may be determining a lateral offset of the vehicle.

[0079] According to various embodiments, laterally controlling the vehicle may include or may be determining an orientation error of the vehicle.

[0080] According to various embodiments, laterally controlling the vehicle may include or may be determining an optimal steering angle value.

[0081] According to various embodiments, the cost of MPC may be derived from the above-mentioned error to calculate the optimal steering angle value for lateral control.

[0082] Each of steps 502, 504, 506, 508, and further steps described above, may be performed by computer hardware components.

[0083] Figure 6 A computer system 600 is shown having a plurality of computer hardware components configured to perform the steps of a computer-implemented method for lateral control of a vehicle according to various embodiments. The computer system 600 may include a processor 602, a memory 604, and a non-transitory data storage device 606. A sensor 608 may be provided as part of the computer system 600 (e.g., Figure 6 shown), or may be provided external to the computer system 600.

[0084] Processor 602 may execute instructions provided in memory 604. Non-transitory data storage 606 may store computer programs including instructions that may be transferred to memory 604 and then executed by processor 602. Sensor 608 may be used to determine position error and / or orientation error.

[0085] The processor 602, the memory 604, and the non-transitory data storage device 606 can be coupled to each other for exchanging electrical signals, for example, via an electrical connection 610 (e.g., a cable or a computer bus) or via any other suitable electrical connection. The sensor 608 can be coupled to the computer system 600, for example, via an external interface, or can be provided as part of the computer system (in other words, internal to the computer system, for example, coupled via the electrical connection 610).

[0086] The terms “coupled” or “connected” are intended to include a direct “coupled” (eg, via a physical link) or direct “connection” as well as an indirect “coupled” or indirect “connection” (eg, via a logical link), respectively.

[0087] It should be understood that the content described above for one of the methods can be similarly applied to the computer system 600 .

[0088] Reference Signs List

[0089] 100 Model Predictive Controller According to Various Embodiments

[0090] 102 Timeline

[0091] 104 Past Time

[0092] 106 Future Time

[0093] 108 Desired set point

[0094] 110 Measurement status

[0095] 112 Closed-loop input

[0096] 114 Remeasurement status

[0097] 116 Prediction Status

[0098] 118 Optimal Input Trajectory

[0099] 120 Re-predict status

[0100] 122 Re-optimize input trajectory

[0101] 124 Backward Time Domain

[0102] 126 Prediction Time Domain

[0103] 128 Backward Time Domain

[0104] 130 Prediction Time Domain

[0105] 200 Comparison of Steering Responses Between Different Cost Functions

[0106] 202 horizontal axis

[0107] 204 vertical axis

[0108] 206 Curve

[0109] 208 Curve

[0110] 300 Comparison of Cross-Tracking Error Between Different Cost Functions

[0111] 302 horizontal axis

[0112] 304 vertical axis

[0113] 306 Curve

[0114] 308 Curve

[0115] 400 Comparison of orientation errors between different cost functions

[0116] 402 horizontal axis

[0117] 404 vertical axis

[0118] 406 Curve

[0119] 408 Curve

[0120] 500 is a flowchart showing a method for lateral control of a vehicle according to various embodiments.

[0121] 502 Steps to determine vehicle position error

[0122] 504 Steps for determining the vehicle's orientation error

[0123] 506 Steps to determine the cost function based on position error and orientation error using circular transformation

[0124] 508 Steps to process the cost function in a model predictive controller to control the vehicle laterally

[0125] 600 Computer system according to various embodiments

[0126] 602 processor

[0127] 604 Memory

[0128] 606 Non-transitory data storage devices

[0129] 608 Sensor

[0130] 610 connection

Claims

1. A computer-implemented method for lateral control of a vehicle, the computer-implemented method comprising the following steps performed by computer hardware components: determining (502) a position error of the vehicle; determining (504) an orientation error of the vehicle; determining (506) a cost function based on the position error and the orientation error using a circular transform; as well as processing (508) the cost function in a model predictive controller to laterally control the vehicle, wherein the cost function comprises an integral over an error term, wherein the error term contains both the position error and the orientation error, and The error term includes a product of the position error and the orientation error.

2. The computer-implemented method of claim 1 , wherein: The cost function comprises an incomplete ellipsoid cost function.

3. The computer-implemented method of claim 1 , wherein: The cost function is based on a cosine function.

4. The computer-implemented method of claim 1 , wherein: The cost function is based on a sine function.

5. The computer-implemented method of claim 1 , wherein: The cost function is based on a tangent function.

6. The computer-implemented method of claim 1 , wherein: The cost function is determined according to the following formula: and tr =cosθ e 2 *tan -1 (dd ref )+sinθ e 2 *θ e (2) in, θ e represents the orientation error relative to the reference, d represents the initial value of the lateral offset obtained from the vehicle sensor input, d ref represents the reference value that the vehicle should track, J represents the cost function to be minimized, and T represents the prediction time domain.

7. The computer-implemented method of claim 1 , wherein: Laterally controlling the vehicle includes determining a lateral offset of the vehicle.

8. The computer-implemented method of claim 1 , wherein: Laterally controlling the vehicle includes determining an optimal steering angle value.

9. A computer system (600) comprising a plurality of computer hardware components configured to perform the steps of the computer-implemented method according to any one of claims 1 to 8.

10. A vehicle comprising the computer system (600) according to claim 9 and a sensor (608) configured to determine the position error and / or the orientation error.

11. The vehicle according to claim 10, wherein: The sensor (608) includes at least one of a radar sensor, a lidar sensor, an ultrasonic sensor, a camera, or a global navigation satellite system sensor.

12. A non-transitory computer-readable medium comprising instructions for executing the computer-implemented method according to any one of claims 1 to 8.

Images