A method for determining the longitudinal speed of a vehicle equipped with radar sensors and the mounting orientation of the radar sensors while cornering.

Abstract

The present invention relates to a method for determining the longitudinal speed of a vehicle (1) having at least one radar sensor (2) and the mounting orientation of the at least one radar sensor (2) during cornering, the method comprising the steps of: (a) determining at least one velocity vector (3) of the at least one radar sensor (2) during cornering of the vehicle (1), the at least one velocity vector (3) comprising a longitudinal velocity component and a lateral velocity component of the at least one radar sensor (2); (b) transmitting the at least one velocity vector (3) to a module (14) for estimating the longitudinal speed of the vehicle (1) and the mounting orientation of the at least one radar sensor (2); and (c) estimating the longitudinal speed of the vehicle (1) and the mounting orientation of the at least one radar sensor (2) at least based on the at least one velocity vector (3) transmitted to the module (14) and by means of the module (14).

Description

[0001] The application of radar sensors in the automotive industry is becoming increasingly widespread. To provide driver assistance functions such as distance-adjustable speed controllers, warning systems for identifying vehicles in blind spots while reversing, and lane-changing assist, more and more vehicles are equipped with radar sensors as distance measurement instruments, relative speed measurement instruments, and angle measurement instruments for objects in the vehicle's surrounding environment. From these measurements, an understanding of the complex vehicle's surrounding environment is derived through classification algorithms, which distinguish identified radar targets into different object types with specific behavioral patterns.

[0002] In addition to traditional detection and localization of surrounding objects, future radar sensors should also be used for precise classification of different object types. For this classification, key features must be extracted from the signal shape during signal processing. A key focus here is detecting prominent reflectors in the vehicle's surrounding environment and classifying them as moving or stationary targets. Because the latter may change their relative position to the vehicle or so-called "self-vehicle" (Ego-Fahrzeug), moving targets require more sophisticated localization and control measures than stationary targets.

[0003] In order to accurately perform this classification, the radar system must be geometrically or externally calibrated relative to the vehicle itself, which is guiding the radar. Furthermore, this presupposes a precise understanding of the motion state of the vehicle. Insufficient understanding of the vehicle's motion state directly leads to decreased classification accuracy.

[0004] By Grimm, C., Farhound, R., Fei, T., Warsitz, E., and Breddermann, T., and Umbach, R., in his paper "Detection of moving targets in automotive radar with distorted ego-velocity information," presented at the 2017 Microwave, Radar, and Remote Sensing Symposium (MRRS), describes a method for estimating the longitudinal velocity of a vehicle or its own vehicle, or its own velocity, in the longitudinal direction of vehicle motion. The algorithm used therein considers identified salient reflectors to estimate the vehicle's relative longitudinal self-motion and thus significantly improve target classification. This method is limited to longitudinal vehicle motion.

[0005] The objective of this invention is to improve known methods for determining the longitudinal speed of a vehicle equipped with radar sensors, and in particular to achieve particularly accurate classification of different objects detected by the radar sensors.

[0006] Herein, the features and details described in conjunction with the method according to the invention also apply to the radar system according to the invention and the vehicle according to the invention, and vice versa, so that the disclosures regarding the various aspects of the invention are always mutually referential or can be mutually referenced.

[0007] According to a first aspect of the invention, the task is correspondingly solved by a method for determining the longitudinal speed of a vehicle having at least one radar sensor and the mounting orientation of said at least one radar sensor while turning, wherein the method comprises the following steps:

[0008] (a) Determine at least one velocity vector of the at least one radar sensor during the turning of the vehicle, wherein the at least one velocity vector includes a longitudinal velocity component and a lateral velocity component of the at least one radar sensor.

[0009] (b) Transmitting the at least one velocity vector to a module for estimating the longitudinal velocity of the vehicle and the mounting orientation of the at least one radar sensor; and

[0010] (c) Estimate the longitudinal speed of the vehicle and the mounting orientation of the at least one radar sensor, at least based on the at least one velocity vector transmitted to the module and by means of the module.

[0011] According to the present invention, the task mentioned at the beginning is solved on the one hand by setting the method for determining the longitudinal speed of a vehicle during cornering. On the other hand, the task mentioned at the beginning is solved by estimating, in addition to estimating, the mounting orientation of the at least one radar sensor or the mounting orientation of multiple radar sensors is also estimated and thus determined. This allows for accurate determination of the vehicle's longitudinal speed even during cornering and avoids measurement errors caused by incorrect or inaccurate parameterization or detection of the mounting orientation. Furthermore, it can be specified that the vehicle's longitudinal acceleration is also estimated using the vehicle's longitudinal speed and the mounting orientation of the at least one radar sensor.

[0012] Naturally, in the method according to the invention, multiple velocity vectors can be determined and transmitted to the module so that an estimate can be determined based on the determined multiple velocity vectors. The estimate by means of the module can ultimately oscillate with the increase of the velocity vector or measurement value, i.e., with the increase of the measurement time of the at least one radar sensor, and thus achieve a particularly unbiased estimate.

[0013] It can be specified that during cornering, at least one velocity vector from each of at least two radar sensors is determined, the velocity vectors from the at least two radar sensors are transmitted to the module, and the longitudinal velocity of the vehicle and the mounting orientation of each of the at least two radar sensors are estimated based on the transmitted velocity vectors from the at least two radar sensors. Instead of two radar sensors, for example, three, four, or more radar sensors can also be used to determine the velocity vectors. Each radar sensor then provides a longitudinal velocity component and a lateral velocity component. Accordingly, the longitudinal velocity of the vehicle can be estimated more accurately, and the mounting orientation of each of the radar sensors can also be estimated.

[0014] Furthermore, it can be specified that the estimation of the vehicle's longitudinal velocity and the mounting orientation of the at least one radar sensor are performed simultaneously. This has the advantage that the two estimates can be provided simultaneously using only a single common estimation method or by a single common module. The vehicle's longitudinal acceleration can also be estimated simultaneously with the vehicle's longitudinal velocity and the mounting orientation of the at least one radar sensor.

[0015] Alternatively, the module may be implemented on at least one radar sensor or a central computing unit of the vehicle. Here, the central computing unit may be part of a radar system that includes the at least one sensor and the module. The module may be stored as a program or algorithm on the at least one radar sensor, on one of a plurality of radar sensors, or on and implemented in the central computing unit.

[0016] Finally, it can also be stipulated that the at least one velocity vector of the at least one radar sensor is determined based on the relative velocity of the at least one radar sensor, the vehicle's float angle, the horizontal incident angle of the at least one radar sensor, and the vertical incident angle of the at least one radar sensor. The relative velocity of the at least one radar sensor is measured as radial velocity. The radial velocity allows for the determination of the longitudinal and lateral velocity components of the velocity vector, provided that the vehicle's float angle, the horizontal incident angle of the at least one radar sensor, and the vertical incident angle of the at least one radar sensor are measured or perceived accordingly. The following equation represents this relationship:

[0017]

[0018] Here, v r It is the relative velocity of the radar sensor, v x It is the longitudinal velocity component along the vehicle's longitudinal axis, v y It is the transverse velocity component orthogonal to the longitudinal velocity component. ε is the horizontal angle of incidence, ε is the vertical angle of incidence, and δ is the float angle. Here, the float angle is formed during the so-called float phase, which describes the vehicle's unstable driving characteristics in terms of direction. The float angle is the angle between the vehicle's motion around its center of gravity and the vehicle's longitudinal axis.

[0019] Alternatively, it can be specified that at least one inaccurate longitudinal speed of the vehicle is determined by the vehicle's odometer sensor, and said inaccurate longitudinal speed is transmitted to the module, wherein at least one scaling factor for the transmitted inaccurate longitudinal speed of the vehicle is estimated simultaneously with the vehicle's longitudinal speed and the mounting orientation of said at least one radar sensor. The longitudinal speed measured or provided by the odometer sensor is inaccurate in this respect because it has inaccuracy or measurement error relative to the actual longitudinal speed. This measurement error can be detected by the method in the form of an estimated scaling factor. The scaling factor can correct for the measurement error. This can be done by multiplying the inaccurate longitudinal speed by the estimated scaling factor.

[0020] Furthermore, it can be specified that at least one inaccurate yaw rate of the vehicle is determined by at least one yaw rate sensor, and the inaccurate yaw rate is transmitted to the module, wherein at least one scaling factor for the inaccurate yaw rate of the vehicle is estimated simultaneously with the vehicle's longitudinal speed and the mounting orientation of the at least one radar sensor. The yaw rate, or yaw speed, measured or provided by the yaw rate sensor is inaccurate in this respect because it has inaccuracy or measurement error relative to the actual yaw rate. This measurement error can be detected by the method in the form of an estimated scaling factor. The scaling factor can correct for the measurement error. This can be done by multiplying the inaccurate yaw rate by the estimated scaling factor.

[0021] Furthermore, it can be specified that the vehicle's yaw rate is estimated simultaneously with the vehicle's longitudinal velocity and the mounting orientation of the at least one radar sensor. Thus, another precise measurement parameter can be obtained using the module based on measurements from the at least one radar sensor. It can also be specified that yaw acceleration is determined simultaneously with other parameters.

[0022] Furthermore, it can be specified that the mounting orientation of the at least one radar sensor is determined by estimating the difference between the parameterized mounting angle and the actual mounting angle.

[0023] Alternatively, the module can be specified as a Kalman filter. It has been shown that this provides a particularly unbiased estimate of the vehicle's longitudinal velocity and the mounting orientation of the at least one radar sensor.

[0024] Furthermore, it can be specified that, within the module, the measurement vectors having the longitudinal and lateral velocity components of the at least one radar sensor are combined with the state vector to be estimated having the vehicle's longitudinal velocity and the mounting orientation of the at least one radar sensor to form a state-measurement equation (Zustands-zu-Messgleichung).

[0025] It can be specified here that a state-measurement matrix (Zustands-zu-Messmatrix) is established in the module by means of the state-measurement matrix, and the module estimates the longitudinal speed of the vehicle and the mounting orientation of the at least one radar sensor by means of the state-measurement matrix.

[0026] According to a second aspect of the invention, the aforementioned task is further solved by a radar system for a vehicle, wherein the radar system has at least one radar sensor and at least one module, wherein the radar system is configured to perform the method according to the first aspect of the invention.

[0027] Accordingly, the radar system may have multiple radar sensors, such as two, three, four or more radar sensors.

[0028] It can be specified here that the at least one radar sensor is connected to the module via a proprietary or open data channel, particularly a CAN bus. Accordingly, the transmission of the at least one velocity vector to the module can be performed via a proprietary or open data channel, particularly a CAN bus.

[0029] According to a third aspect of the invention, the aforementioned task is further accomplished by a vehicle having a radar system according to a second aspect of the invention.

[0030] The invention will now be described in detail with reference to the accompanying drawings. All features derived from the claims, description, or drawings are essential not only individually but also in any different combinations thereof. The drawings schematically illustrate, respectively:

[0031] Figure 1 A vehicle according to an embodiment of the present invention is shown in the form of a vehicle turning.

[0032] Figure 2 This shows the difference between the parameterized mounting angle and the actual mounting angle for the vehicle.

[0033] Figure 3 Showing the use of Figure 1 and Figure 2 A radar system for a vehicle according to one embodiment;

[0034] Figure 4 Show Figure 3Measurements of radar sensors in a radar system;

[0035] Figure 5 The state-measurement matrix is ​​shown;

[0036] Figures 6a to 6d Showing the use of Figure 3 In the case of a radar system, there are graphs showing the estimated and actual parameters, as well as...

[0037] Figures 7a to 7d Showing other uses Figure 3 A graph showing the estimated and actual parameters of the radar system.

[0038] Components with the same function and mode of operation Figure 1 The same reference numerals are used in Figures 7 and 8. If an element exists multiple times, it is numbered consecutively, with the consecutive numbers following the reference numerals and separated from them by a dot.

[0039] Figure 1 A vehicle 1 according to an embodiment of the present invention is shown in the form of a turn. Here, vehicle 1 is shown in a Cartesian xy coordinate system.

[0040] Vehicle 1 has four radar sensors 2.1, 2.2, 2.3, and 2.4. Alternatively, the vehicle may also have only one, two, three, or more than four radar sensors 2. Each of the radar sensors 2.1, 2.2, 2.3, and 2.4 experiences a speed, which can be represented as a vector. (Velocity vector) Each has a longitudinal velocity component v sersor,x lateral velocity component v sensor,y and vertical velocity component v sensor，z The velocity vectors of each radar sensor (2.1, 2.2, 2.3, 2.4) are thus adopted in the form... Among them, the following is by v sensor，z =0 (Starting point)

[0041] The velocity vectors shown in radar sensors 2.1, 2.2, 2.3, and 2.4 Hereinafter labeled 3.1, 3.2, 3.3, and 3.4. Due to the different mounting positions of radar sensors 2.1, 2.2, 2.3, and 2.4 on vehicle 1, when turning, radar sensors 2.1, 2.2, 2.3, and 2.4 pass through different radii relative to the object or reflector 10, and radar sensors 2.1, 2.2, 2.3, and 2.4 perceive the object or reflector as a point target. Accordingly, the velocity vectors 3.1, 3.2, 3.3, and 3.4 are different from each other. Accordingly, the turning radii 4.1, 4.2, 4.3, and 4.4 of radar sensors 2.1, 2.2, 2.3, and 2.4, and the turning radius 5 of vehicle 1 are plotted. Accordingly, in Figure 1 The diagram also shows the floating angle δ of vehicle 1 as the angle between the longitudinal axis of vehicle 1 and the direction of movement of vehicle 1 when turning.

[0042] The method described herein determines the longitudinal velocity v of vehicle 1 by estimating it as accurately as possible. x Meanwhile, the mounting orientation of radar sensors 2.1, 2.2, 2.3, and 2.4 can be determined through precise estimation. The mounting orientation is estimated here as the orientation error relative to the parameterized mounting orientation or mounting position.

[0043] exist Figure 2 In the diagram, the orientation error is shown as the orientation angle α between the velocity vector 3 of radar sensor 2 as measured by radar sensor 2 and the actual velocity vector 6. The actual velocity vector 6 can also be referred to as the ground-based velocity vector.

[0044] Figure 3 Now shown is an embodiment of the invention for... Figure 1 and Figure 2 The radar system 100 of one of the vehicles 1 has one or more radar sensors 2.1...2.N, the velocity vector of the radar sensors... The evaluation is conducted in module 14 (currently in the form of a Kalman filter). Module 14 can be implemented here on a computing unit (not shown) of the radar system 100. The evaluation should then proceed from N=2, i.e., evaluating the two radar sensors 2.1 and 2.2.

[0045] When vehicle 1 is turning, radar sensors 2.1 and 2.2 detect one or more reflectors. Radar sensors 2.1 and 2.2 then independently determine their velocity vectors based on sensor speed determination. To determine the sensor velocity, consider the relative velocity v measured by radar sensors 2.1 and 2.2. rThis is the radial velocity. For each of radar sensors 2.1 and 2.2, it can be determined by the floating angle δ of vehicle 1 and the horizontal incident angle of radar sensors 2.1 and 2.2. The vertical incident angle ε of radar sensors 2.1 and 2.2 is expressed as follows:

[0046]

[0047] Here, as explained above, v x It is the longitudinal velocity component of the corresponding radar sensor in radar sensors 2.1 and 2.2 in the direction of the vehicle's longitudinal axis, and v y These are the lateral velocity components of the corresponding radar sensors in radar sensors 2.1 and 2.2.

[0048] Figure 4 The example illustrates the relative velocity v of one of the two radar sensors 2.1 and 2.2 with respect to different stationary reflectors or targets 10. r The measurement results. To determine the velocity vector. Calculate the normal vector 7 of the plane unfolded from the stationary target 10 and in the coordinate system. and Differentiate in the direction of . The negative reciprocal gradient now represents the estimated longitudinal velocity component v of the corresponding radar sensor in radar sensors 2.1 and 2.2. x and the estimated lateral velocity component v y .

[0049] Here, the velocity vector obtained for each radar sensor in radar sensors 2.1 and 2.2 is... It can be done through relational expressions To model, among which, yes The rotational velocity vector, It is a previously parameterized position vector from the rear axle of the vehicle to the corresponding parameterized positions of radar sensors 2.1 and 2.2, and This is the velocity vector of vehicle 1 measured on the rear axle of vehicle 1. Therefore, the velocity vector of each radar sensor in radar sensors 2.1 and 2.2... and position vector This understanding enables us to determine the yaw rate or yaw rate w of vehicle 1, as well as its longitudinal velocity or longitudinal vehicle speed v. x .

[0050] The velocity vector obtained for each radar sensor in radar sensors 2.1 and 2.2 Here, data can be transmitted to radar sensors 2.1 and 2.2 via inter-sensor communication. However, specifically, these velocity vectors are transmitted to module 14. For this purpose, an open or proprietary data channel, particularly the CAN bus 13, can be used. Furthermore, inaccurate longitudinal velocity v is transmitted from the odometer sensor 11 and the yaw rate sensor 12 via the CAN bus 13. CAN and imprecise yaw speed w CAN The longitudinal and yaw velocities mentioned are inaccurate because they do not correspond to actual ground conditions or are not real, but are subject to measurement errors.

[0051] The measured values ​​transmitted to the module are combined into a measurement vector z in module 14, and the measurement vector can be represented as: Here, the velocity component v x1 v x2 v y1 v y2 These are the lateral velocity component and the longitudinal velocity component of the two radar sensors 2.1 and 2.2, respectively.

[0052] The desired outcome is to estimate the state vector x, which is estimated and output by module 14. The state vector can be represented as... Among them, v x This indicates the longitudinal speed of vehicle 1. The longitudinal acceleration of vehicle 1 is indicated by , and the yaw rate or yaw velocity of vehicle 1 is indicated by . The yaw acceleration of vehicle 1 is indicated by α1 and α2, which represent the error orientation angles of radar sensors 2.1 and 2.2, respectively. β is the error angle for the imprecise longitudinal velocity v. CAN The scaling factor or correction factor, and γ is for the imprecise yaw rate w CAN The scaling factor or correction factor.

[0053] If radar sensors 2.1 and 2.2 are mounted in vehicle 1 with relative rotation due to tolerances, then the velocity vector Similarly, the sensor speed is measured after relative rotation. Then the speed corresponding to the relative rotation of vehicle 1 The state-measurement equation h(x) can be composed as follows:

[0054]

[0055] Accordingly, in Figure 5 The state-measurement matrix shown in the figure has a solution that produces the desired state vector x to be estimated.

[0056] Figures 6a to 6d as well as Figures 7a to 7dThe estimation results, i.e., the estimated values ​​relative to the actual ground conditions, i.e., the true values, are presented for different parameters to be estimated. According to... Figures 6a to 6d It can be seen that the motion state was estimated unbiasedly over the entire duration. According to... Figures 7a to 7d It can be seen that the mechanical system state is estimated unbiasedly after the Kalman filter starts oscillating.

[0057] List of reference numerals

[0058] 1 vehicle

[0059] 2. Radar Sensors

[0060] 3, Velocity vector of radar sensor

[0061] 4. Turning radius of radar sensors

[0062] 5. Vehicle turning radius

[0063] 6. The true velocity vector of the radar sensor

[0064] 7. Normal vector

[0065] 10. Targets, reflectors

[0066] 11 Odometer Sensor

[0067] 12 Yaw rate sensor

[0068] 13 CAN bus

[0069] 14 modules

[0070] 100 Radar System

[0071] α Error Orientation Angle

[0072] δ Floating Angle

[0073] z Measurement vector

[0074] x state vector

Claims

1. A method for determining the longitudinal speed of a vehicle (1) having at least one radar sensor (2) and the mounting orientation of said at least one radar sensor (2) during cornering, wherein, The method comprises the following steps: (a) During the turning motion of the vehicle (1), at least one velocity vector (3) of the at least one radar sensor (2) is determined, the at least one velocity vector (3) comprising the longitudinal velocity component and the lateral velocity component of the at least one radar sensor (2); (b) Transmitting the at least one velocity vector (3) to a module (14) for estimating the longitudinal velocity of the vehicle (1) and the mounting orientation of the at least one radar sensor (2); and (c) Estimate the longitudinal velocity of the vehicle (1) and the mounting orientation of the at least one radar sensor (2) based at least on the at least one velocity vector (3) transmitted to the module (14) and by means of the module (14). The at least one velocity vector (3) of the at least one radar sensor (2) is determined based on the yaw rate of the at least one radar sensor (2), the vehicle's float angle (δ), the horizontal incident angle of the at least one radar sensor (2), and the vertical incident angle of the at least one radar sensor (2). The mounting orientation of the at least one radar sensor (2) is determined by estimating the difference between the parameterized mounting angle and the actual mounting angle.

2. The method according to claim 1, wherein, During cornering, at least one velocity vector (3) of each of the at least two radar sensors (2) is determined, the velocity vectors (3) of the at least two radar sensors (2) are transmitted to the module (14), and the longitudinal velocity of the vehicle (1) and the mounting orientation of each of the at least two radar sensors (2) are estimated based on the transmitted velocity vectors (3) of the at least two radar sensors (2).

3. The method according to claim 1 or 2, wherein, The longitudinal speed of the vehicle (1) and the mounting orientation of the at least one radar sensor (2) are estimated simultaneously.

4. The method according to claim 1 or 2, wherein, The module (14) is executed on at least one radar sensor (2) or central computing unit of the vehicle (1).

5. The method according to claim 1 or 2, wherein, The odometer sensor (11) of the vehicle (1) also determines at least one inaccurate longitudinal speed of the vehicle (1), and transmits the inaccurate longitudinal speed to the module (14). Simultaneously with the longitudinal speed of the vehicle (1) and the mounting orientation of the at least one radar sensor (2), at least one scaling factor for the transmitted inaccurate longitudinal speed of the vehicle (1) is also estimated.

6. The method according to claim 1 or 2, wherein, At least one inaccurate yaw rate of the vehicle (1) is also determined by at least one yaw rate sensor (12) of the vehicle (1), and the inaccurate yaw rate is transmitted to the module (14), and at least one scaling factor for the inaccurate yaw rate of the vehicle (1) is estimated simultaneously with the longitudinal speed of the vehicle (1) and the mounting orientation of the at least one radar sensor (2).

7. The method according to claim 1 or 2, wherein, Simultaneously with the longitudinal speed of the vehicle (1) and the mounting orientation of the at least one radar sensor (2), the yaw speed of the vehicle (1) is also estimated.

8. The method according to claim 1 or 2, wherein, The module (14) is a Kalman filter.

9. The method according to claim 1 or 2, wherein, In the module (14), the measurement vector (z) having the longitudinal and lateral velocity components of the at least one radar sensor (2) is combined with the state vector (x) to be estimated having the longitudinal velocity of the vehicle (1) and the mounting orientation of the at least one radar sensor (2) to form a state-measurement equation.

10. The method according to claim 9, wherein, In the module (14), a state-measurement matrix is ​​established by the state-measurement equation, and the module (14) estimates the longitudinal speed of the vehicle (1) and the mounting orientation of the at least one radar sensor (2) by means of the state-measurement matrix.

11. A radar system (100) for a vehicle (1), wherein, The radar system (100) has at least one radar sensor (2) and at least one module (14), and the radar system (100) is configured to implement the method according to any one of claims 1 to 10.

12. The radar system (100) according to claim 11, wherein, The at least one radar sensor (2) is connected to the module (14) via a dedicated or open data channel.

13. The radar system (100) according to claim 11, wherein, The at least one radar sensor (2) is connected to the module (14) via a CAN bus (13).

14. A vehicle (1) having a radar system (100) according to any one of claims 11 to 13.

Images