Radar system for 3D ego motion estimation

Abstract

The present invention relates to a radar-based system (10) for a vehicle (1) for driver assistance or autonomous driving, the radar-based system (10) comprises a processor configured to receive signals from at least one sensor (3, 3a, 3b, 3c) of the vehicle (1) configured to detect external objects (T1 to T9) of the vehicle (1), wherein the signals include positional information and radial velocity information for at least three objects (T1 to T9) relative to the at least one sensor (3, 3a, 3b, 3c), and to determine the speed of the vehicle (1) based on the received signals.

Description

Technical Field

[0001] The present invention relates to a radar-based system for a vehicle, a vehicle, a computer-implemented method for driver assistance or autonomous driving for a vehicle, a computer program, and a computer-readable recording medium for driver assistance or autonomous driving for a vehicle.

Background Art

[0002] The 3D radar of a motor vehicle does not always correctly provide radar ranging and direction in three dimensions. Therefore, calibration data is important for the 3D radar of the ego vehicle to perform correction of the measured data or uncertainty calculation. In order to provide calibration data for the 3D radar in any driving state of the vehicle, complete 3D motion data including the linear velocity of three axes and the angular velocity of three axes as ground truth is required.

[0003] Advanced inertial measurement units (IMUs), such as ADMA sensors that require high costs, can measure vehicle motion in three axes even during GPS signal loss. They not only provide accurate acceleration, speed, and position data, but also determine dynamic posture and yaw angle.

[0004] However, ADMA sensors are quite expensive. It is difficult to provide continuous and accurate calibration data for the 3D radar in a vehicle not equipped with an ADMA sensor. Furthermore, sometimes, even when an autonomous vehicle already has an ADMA sensor, it is more desirable to have an alternative 3D ego motion sensing method as a validation of the effectiveness of the actual radar calibration performed.

[0005] U.S. Patent Application Publication No. 2021 / 0124033A1 discloses a method for calibrating vehicle sensors of a motor vehicle. The method includes the steps of: verifying sensor data at multiple measurement times using a vehicle sensor, wherein the motor vehicle is moving relative to an object around the motor vehicle; calculating the object position of the object based on the verified sensor data; calculating a Hough transform based on the calculated object position; verifying the alignment of the vehicle sensor with respect to the drive shaft of the motor vehicle based on the calculated Hough transform; and calibrating the vehicle sensor based on the verified alignment of the vehicle sensor with respect to the drive shaft of the motor vehicle.

[0006] U.S. Patent Application Publication 2017 / 0212215A1 discloses a method and apparatus for determining the misalignment of a radar sensor unit mounted on a vehicle, including locating targets on a matching device. The vehicle is located at a predetermined location on a test station at a precisely constant distance from the matching device. The actual locations and distances of the targets from each other, as well as the actual locations and distances of the targets from the vehicle's radar sensor unit at the test station, are known and stored in advance. At least one target is further from the vehicle than it is from the other targets. The targets receive and return radar waves from the radar sensor unit. The radar sensor unit determines the location and distance of the targets and determines the misalignment of the radar sensor unit by comparing it to a given or actual location and distance of the targets. A calibration program automatically calibrates the azimuth and elevation angles to adjust for the misalignment. [Overview of the project]

[0007] The present invention provides a radar-based system for a vehicle for driver assistance or autonomous driving according to claim 1, a vehicle according to claim 5, a computer-based method for driver assistance or autonomous driving for a vehicle according to claim 6, a computer program according to claim 9, and a computer-readable recording medium according to claim 10.

[0008] Furthermore, advantageous embodiments and improvements of the present invention are enumerated in the dependent claims. Advantages of the present invention In a first aspect, the present invention provides a radar-based system for a vehicle for driver assistance or autonomous driving, comprising a processor configured to receive signals from at least one sensor of the vehicle configured to detect external objects, wherein the signals include positional information and radial velocities of at least three objects relative to at least one sensor, and to determine the speed of the vehicle based on the received signals.

[0009] When there are multiple stationary objects, especially when there are two or more stationary objects around the 3D radar of an Ego vehicle traveling in a straight line, it is advantageous for the radar to be able to dynamically estimate its own velocity vector without wasting time tracking the objects. The 3D velocity corresponding to the velocity vector can be calculated from a single measurement that includes the 3D position information and Doppler / radial velocity of the target / object.

[0010] The relative velocity of a stationary target is the additive inverse of the velocity of the moving radar. In a single measurement, this velocity is identical for all stationary targets. Therefore, a system of equations may be used for each measured point corresponding to a detected target / object. This system of equations is linear with respect to the components of the velocity vector, such that the components of the velocity vector can be estimated by linear regression.

[0011] The 3D velocity vector contains three unknown variables, which means that at least three stationary targets must be measured in a single radar shot for the estimation process to work. Linear regression provides the optimal solution (by least mean squares error), so to achieve a more accurate solution, it is advantageous to involve more stationary objects as measured points, taking into account the existence of measurement errors.

[0012] In a preferred embodiment, the processor is further configured to receive signals from each of at least three sensors of the vehicle, which are configured to detect objects outside the vehicle, wherein the signals from each of the at least three sensors include positional information and radial velocities of at least three objects relative to each sensor; and to determine the linear and angular velocities of the vehicle based on the received signals, the relative positions of each of the at least three sensors with respect to the vehicle, and rotation matrices from the vehicle to each of the at least three sensors.

[0013] If a vehicle is not traveling in a straight line but is turning and / or has begun to move uphill / downhill and / or has encountered a set road or construction site where the road is not perfectly flat, the vehicle will have not only linear velocity but also angular velocity, and therefore the measured data may also need to be corrected accordingly.

[0014] In this case, a different set of equations may be used for an ego vehicle having at least three radar sensors (in particular, each sensor capable of detecting three different stationary objects around the vehicle). Thereafter, the components of the linear velocity vector and angular velocity vector can be estimated by linear regression on this set of equations with the help of data on stationary objects measured by at least three synchronous radars.

[0015] In other words, in an environment containing many stationary targets, a vehicle can estimate its complete 3D ego motion state using three mounted radars with the solution provided by the present invention.

[0016] Furthermore, the present invention provides an alternative 3D ego motion sensing method for verifying the effectiveness of actual radar calibration performed by, for example, ADMA sensors in self-driving vehicles. In a further preferred embodiment, the position information includes the azimuth, elevation, and distance of the object relative to each sensor.

[0017] In a further preferred embodiment, the angular velocity of the vehicle includes the yaw angle and / or pitch angle. In most cases, when a vehicle is on a road, if it is not moving straight forward / backward, it will turn left, right, or make turns. In this regard, the yaw angle is measured as one of the vehicle's motion data, and the measured yaw angle must be corrected as necessary. For example, on rural roads, a vehicle may frequently encounter uphill and downhill slopes. In this case, the pitch angle is measured, and the measured pitch angle must be corrected as necessary. In some undeveloped areas outside of cities, such as fields or construction sites, a vehicle may travel with violent up-and-down motion or on uneven roads. In this case, the roll angle is measured, and the measured roll angle must be corrected as necessary. The last case may occur less frequently than both of the cases mentioned above. Therefore, this embodiment of the present invention focuses primarily on calibrating the measured yaw angle and / or pitch angle in order to accelerate the calibration process. However, the calibration of the measured roll angle of a vehicle is also part of the solution of the present invention.

[0018] The present invention further provides a vehicle having a sensor configured to detect objects outside the vehicle, according to a second aspect of the present invention, and a radar-based system according to a first aspect of the present invention.

[0019] The present invention further provides a computerized method for driver assistance or autonomous driving for a vehicle having a radar-based system, the computerized method comprising: receiving a signal from at least one sensor of the vehicle configured to detect external objects, wherein the signal includes positional information and radial velocity of at least three objects relative to the at least one sensor; and determining the speed of the vehicle based on the received signal.

[0020] In a preferred embodiment of the computerized method, the step of receiving a signal from at least one sensor of the vehicle configured to detect an object outside the vehicle includes the step of receiving a signal from each of at least three sensors of the vehicle configured to detect an object outside the vehicle, wherein the signal from each of the at least three sensors includes positional information and radial velocity of at least three objects relative to each sensor, and the step of determining the speed of the vehicle based on the received signals includes the step of determining the linear velocity and angular velocity of the vehicle based on the received signals, the relative positions of each of the at least three sensors to the vehicle, and the rotation matrix from the vehicle to each of the at least three sensors.

[0021] In a further preferred embodiment of the computer implementation, the position information includes the azimuth, elevation, and distance of the object relative to each sensor. The present invention further provides a computer program, according to a fourth aspect, which includes instructions that cause a computer to perform a method according to a third aspect of the present invention when the computer program is executed by the computer.

[0022] The present invention further provides a computer-readable recording medium according to a fifth aspect, provided that the computer-readable recording medium stores a computer program, and when the computer program is executed by a computer, the computer can implement the method according to the third aspect of the present invention.

[0023] Further advantageous details and features can be understood from the following description of some exemplary embodiments of the present invention in relation to the drawings.

Brief Description of the Drawings

[0024] [Figure 1] It is a diagram schematically showing that a vehicle in an embodiment of the present invention includes a radar for measuring three objects / targets. [Figure 2] It is a diagram schematically showing that a vehicle in an embodiment of the present invention includes three radars for measuring nine objects / targets. [Figure 3] It is a diagram schematically showing an ego vehicle including an embodiment of a radar-based system of the present invention. [Figure 4] It is a diagram showing a block diagram of an embodiment of a computer-implemented method of the present invention. [Figure 5] It is a diagram showing a block diagram of an embodiment of a computer program of the present invention, which includes instructions for causing a computer to execute an embodiment of the method of the present invention according to FIG. 4 when the computer program is executed by the computer. [Figure 6] It is a diagram showing a block diagram of an embodiment of a computer-readable recording medium of the present invention, where the computer-readable recording medium stores a computer program, and when the computer program is executed by a computer, the computer can implement an embodiment of the method of the present invention according to FIG. 4.

Modes for Carrying Out the Invention

[0025] It should be understood that the terms used herein are for the purpose of describing individual embodiments and are not intended to limit them. Unless otherwise specified, all technical and scientific terms used herein should be understood to have the meaning equivalent to that of a person skilled in the art in the relevant field of this disclosure, and should be understood neither too broadly nor too narrowly.

[0026] Furthermore, it should be noted that terms such as “a / an,” “one,” “two,” and “three” used in the claims and / or this description should not be understood as numbered words, but rather as non-exclusive numbering indications for scope protection. For example, the term “an ABC” means “at least one ABC.” However, it is further applicable to the described content of the invention that terms such as “a,” “two,” and “three” used in the claims and this description may also be expressed as numbered words, and therefore simultaneously as numbering indications to be considered final.

[0027] Figure 1 schematically shows that in one embodiment of the present invention, a vehicle 1 includes a radar sensor 3 having a detection range 7 and measuring three stationary objects / targets T1 to T3. In a single measurement, the relative velocities V1 to V3 of the stationary targets T1 to T3 with respect to the radar are identical and correspond to the additive inverse of the velocity 2 of the moving radar 3 / vehicle 1.

[0028] When ego vehicle 1 travels in a straight line, vehicle 1 has only a linear velocity vector and no angular velocity vector. The following equation can be used to estimate the ego motion of vehicle 1 based on the measured targets T1~T3.

[0029] dot(Position,-Velocity)=Radial_Velocity*|Position| Formula 1, However, dot() represents the dot product of two vectors. Position is the 3D position vector of each target relative to radar 3, converted from the radar's angle-based measurements of the target, such as the elevation angle, azimuth angle, and distance of the target relative to radar 3. Radial_Velocity is the radial component of the relative velocity (V1~V3) of each target, indicated by the reference codes Rv1~Rv3 in Figure 1.

[0030] According to the algebraic definition of the inner product, assuming that Position is a 3D vector as [a1, a2, a3] of target T1 and Velocity is the velocity of vehicle 1 as [x, y, z], then Equation 1 is a1*x+a2*y+a3*z=Radial_Velocity*|Position| This is the result.

[0031] Since the position and radial velocity of target T1 can be measured directly or calculated by radar 3, the product of radial velocity * | position | is a constant represented by symbol A1, which becomes equation 1 as a1*x + a2*y + a3*z = A1, further transformed.

[0032] In this transformed equation 1, there are three unknown variables x, y, and z. To solve this equation, at least three targets T1-T3 must be measured in a single radar shot. Linear regression provides the optimal solution (by least mean squares error), so to achieve a more accurate solution, it is advantageous to involve more stationary targets as measured points, taking into account possible measurement errors.

[0033] Figure 2 schematically illustrates that in one embodiment of the present invention, vehicle 1 includes three radars 3a-3c mounted on the body of vehicle 1, each measuring three objects / targets T1-T3, T4-T6, and T7-T9 respectively. In the example shown in Figure 2, vehicle 1 is not traveling in a straight line. Therefore, when estimating the complete 3D ego motion state of vehicle 1, its angular velocity vectors (i.e., [roll angle 22, pitch angle 24, yaw angle 26]) must also be taken into consideration.

[0034] In a single measurement, since radars 3a-3c are mounted on vehicle 1 as a rigid body, they have the same linear velocity vector as vehicle 1 (represented by reference point 5), indicated by reference numeral 20 in Figure 2. As in the embodiment shown in Figure 1, the relative velocities of the stationary targets T1-T9 with respect to each of these radars are identical and correspond to the additive inverse of the velocity 20 of the moving radars 3a-3c / vehicle 1.

[0035] The velocities of stationary objects T1 to T9 from the viewpoint of each vehicle-mounted radar can be calculated using the following formula. Velocity=Radar_orientation*(V_hull+M_rolling*(Radar_orientation -1 *Position+Radar_position)) Equation 2, However, V_hull is the 3D velocity vector of vehicle 1 at reference point 5 of vehicle 1. Radar_position is the relative position of radars 3a~3c with respect to the vehicle's reference point 5. Radar_orientation is the 3D rotation matrix from the vehicle's reference point 5 to the radar orientation. M_rolling is a 3x3 matrix containing the three-axis angular velocities of vehicle 1 (i.e., roll angle 22, pitch angle 24, and yaw angle 26).

[0036]

number

[0037] The angular velocities 22, 24, and 26 in the matrix V_hull, which is the 3D velocity vector (Vx, Vy, Vz) of the vehicle, and M_rolling can be estimated by linear regression to Equation 2 using data from the stopping points (T1 to T9) measured by the three synchronized radars 3a to 3c.

[0038] This allows a vehicle to estimate its complete 3D ego motion state in an environment containing many stationary targets using three stationary radars. Linear regression provides the optimal solution (by least mean squares error), so to achieve a more accurate solution, it is advantageous to involve more stationary targets as measured points, taking into account possible measurement errors.

[0039] Figure 3 schematically shows an Ego Vehicle 1 having three vehicle-mounted radar sensors 3a, 3b, and 3c and a radar-based system 10 for driver assistance or autonomous driving according to the present invention. The system 10 has a processor configured to receive signals from each of the at least three sensors 3a, 3b, and 3c of the vehicle 1, which are configured to detect objects outside the vehicle, wherein the signals from each of the at least three sensors 3a, 3b, and 3c include the respective position information and radial velocity of the at least three objects relative to the respective sensor, and to determine the linear velocity and angular velocity of the vehicle 1 based on the received signals, the respective relative positions of the at least three sensors 3a, 3b, and 3c with respect to the vehicle 1, and rotation matrices from the vehicle 1 to each of the at least three sensors 3a, 3b, and 3c.

[0040] An embodiment of the computer implementation of the present invention for driver assistance or autonomous driving for a vehicle having a radar-based system as shown in Figure 4 includes steps S10 and S20. In step S10, a signal is received from at least one sensor of the vehicle configured to detect objects outside the vehicle, the signal including positional information and radial velocity of at least three objects relative to at least one sensor. In step S20, the speed of the vehicle is determined based on the received signal.

[0041] An embodiment of the computer program 200 of the present invention shown in Figure 5 includes an instruction 250 that causes the computer to execute an embodiment of the method of the present invention shown in Figure 4 when the program 200 is executed by the computer.

[0042] The embodiment of the computer-readable recording medium 300 of the present invention shown in Figure 6 stores a computer program 350. When the computer program 350 is executed by a computer, the computer can carry out an embodiment of the method of the present invention shown in Figure 4.

[0043] The present invention is described and illustrated in detail by the preferred embodiments described above. However, the present invention is not limited by the disclosed examples, and other variations may be derived from the disclosed examples, while remaining further within the scope of protection of the present invention.

Claims

1. A radar-based system (10) for a vehicle (1) for driver assistance or autonomous driving, The receiving of a signal from at least one sensor (3, 3a, 3b, 3c) of the vehicle (1) configured to detect external objects (T1 to T9) of the vehicle (1), wherein the received signal includes the position information and radial velocity of at least three objects (T1 to T9) relative to the at least one sensor (3, 3a, 3b, 3c). The speed of the vehicle (1) is determined based on the received signal. A radar-based system (10) having a processor configured to perform the following.

2. The aforementioned processor further, The receiving of signals from each of at least three sensors (3a, 3b, 3c) of the vehicle (1) configured to detect objects outside the vehicle (1), wherein the received signals from each of the at least three sensors (3a, 3b, 3c) include the position information and radial velocity of at least three objects (T1 to T9) relative to each of the sensors (3a, 3b, 3c). Based on the received signal, the relative positions of the at least three sensors (3a, 3b, 3c) with respect to the vehicle (1), and the rotation matrices from the vehicle (1) to each of the at least three sensors (3a, 3b, 3c), the linear velocity and angular velocity of the vehicle (1) are determined. A radar-based system (10) according to claim 1, configured to perform the following:

3. The radar-based system (10) according to claim 1 or 2, wherein the position information includes the azimuth angle, elevation angle, and distance of an object relative to each of the sensors (3, 3a, 3b, 3c).

4. The radar-based system (10) according to any one of claims 1 to 3, wherein the angular velocity of the vehicle (1) includes a yaw angle (26) and / or a pitch angle (24).

5. A vehicle (1) having sensors (3, 3a, 3b, 3c) configured to detect external objects (T1 to T9) of the vehicle (1), and a radar-based system (10) according to any one of claims 1 to 4.

6. A computer implementation method for driver assistance or autonomous driving for a vehicle having a radar-based system, Step (S10) of receiving a signal from at least one sensor of the vehicle configured to detect an object outside the vehicle, wherein the signal includes position information and radial velocity of at least three objects relative to the at least one sensor. Step (S20) of determining the speed of the vehicle based on the received signal and A computer implementation method including

7. The step (S10) of receiving a signal from at least one sensor of the vehicle configured to detect an object outside the vehicle, A step of receiving signals from each of at least three sensors of the vehicle configured to detect objects outside the vehicle, wherein the signals from each of the at least three sensors include position information and radial velocity of at least three objects relative to each sensor, The step (S20) of determining the speed of the vehicle based on the received signal, The computer implementation method according to claim 6, comprising the step of determining the linear velocity and angular velocity of the vehicle based on the received signal, the relative positions of each of the at least three sensors with respect to the vehicle, and the rotation matrix from the vehicle to each of the at least three sensors.

8. The computer implementation method according to claim 6 or 7, wherein the position information includes the azimuth angle, elevation angle, and distance of the object relative to each of the sensors.

9. A computer program (200) that, when executed by a computer, includes instructions (250) that cause the computer to perform the method according to any one of claims 6 to 8.

10. A computer-readable recording medium (300) for storing a computer program (350), wherein when the computer program (350) is executed by a computer, the computer can carry out the method described in any one of claims 6 to 8.

Images