EP-4530574-B1 - METHOD AND DEVICE FOR DETERMINING A LATENCY BETWEEN IMU DATA AND SENSOR DATA OF A SURROUNDING OF A VEHICLE

Inventors

• Marosi, Csaba
• Aranyosy, Adam

Dates

Publication Date

20260513

Application Date

20230927

Claims (8)

1. A method for determining a latency between a first time reference frame of a system under test (4) in a host vehicle, and a second time reference frame of a reference system (5) in the host vehicle, wherein the system under test (4) comprises at least one sensor for acquiring sensor data of a surrounding of the host vehicle, and wherein the reference system (5) comprises an inertial measurement unit, IMU, wherein the method comprises the steps: determining (S1) a time-dependent first velocity of the host vehicle in the first time reference frame, using the system under test (4); determining (S2) a time-dependent second velocity of the host vehicle in the second time reference frame, using the reference system (5); determining a time-dependent acceleration of the host vehicle in the second time reference frame, using the reference system (5); and computing (S3) the latency between the first time reference frame and the second time reference frame, using the first velocity, the second velocity and the acceleration of the host vehicle, which comprises minimizing a difference between the first velocity in the first time reference frame and the second velocity in the second time reference frame shifted by the latency.
2. The method according to claim 1, wherein the difference between the first velocity and the second velocity is minimized using linear regression.
3. The method according to claim 1 or 2, wherein the first velocity is determined using a first sampling rate; wherein the second velocity is determined using a second sampling rate; and wherein the second velocity is resampled to the first sampling rate.
4. The method according to any of the preceding claims, wherein the IMU comprises three orthogonally positioned gyroscopes for measuring rotational motion and/or three accelerometers for measuring linear movements of the host vehicle.
5. The method according to any of the preceding claims, wherein a sensor drift of the reference system (5) is compensated for by using data from a global navigation satellite system, GNSS.
6. The method according to any of the preceding claims, wherein the system under test (4) comprises at least one of a radar sensor, LiDAR sensor, infrared sensor and camera sensor.
7. The method according to any of the preceding claims, further comprising the step of bringing (S4) first data obtained by the system under test (4) and second data obtained by the reference system (5) into a common time reference frame, using the latency.
8. A device (1) for determining a latency between a first time reference frame of a system under test (4) in a host vehicle, and a second time reference frame of a reference system (5) in the host vehicle, wherein the system under test (4) comprises at least one sensor for acquiring sensor data of a surrounding of the host vehicle, and wherein the reference system (5) comprises an inertial measurement unit, IMU, wherein the device comprises: an interface (2) configured to receive a time-dependent first velocity of the host vehicle in the first time reference frame, determined by the system under test (4), configured to receive a time-dependent second velocity of the host vehicle in the second time reference frame, determined by the reference system (5), and configured to receive a time-dependent acceleration of the host vehicle in the second time reference frame, determined by the reference system (5); and a computing unit (3) configured to compute the latency between the first time reference frame and the second time reference frame, using the first velocity, the second velocity and the acceleration of the host vehicle, which comprises minimizing a difference between the first velocity in the first time reference frame and the second velocity in the second time reference frame shifted by the latency.

Description

The present invention relates to a method and a device for determining a latency between a first time reference frame of a system under test in a host vehicle, and a second time reference frame of a reference system in the host vehicle. Prior Art Advanced driver-assistance systems (ADAS) provide functions which help drivers in driving and parking maneuvers. US 11,187,793 B1 relates to the calibration of sensor streams. CN 112129317 A relates to determining an information acquisition time difference. Mair et al., "Spatio-temporal initialization for IMU to camera registration", Robio 2011, pages 557-564 relates to inertia-visual sensors. Perception is the subsystem in ADAS responsible for processing and condensing sensor data into a low-dimensional representation of the environment around the host vehicle. The tracked traffic participants can be selected as targets for a certain function such as automatic emergency braking or adaptive cruise control. The system-level performance can be measured on the function-level. However, top-level performance reflects the superposition of the deficiencies of all lower-level system components, which makes the identification of root causes a lengthy process and decreases the modularity of subsystems. To remedy this problem, some form of perception-evaluation framework may be implemented. Measuring tracking accuracy also requires some form of reference data, which can be provided by double automotive dynamic motion analyzer (ADMA) or differential GPS (dGPS) setups. A further approach relates to advanced reference systems such as lidar auto-labeler pipelines. DE 10 2021 212 632 A1 relates to the testing of environment sensors and environment perception of vehicles, in particular for the purposes of driving assistance systems. Disclosure of the Invention The invention provides a method and a device for determining a latency between a first time reference frame of a system under test in a host vehicle, and a second time reference frame of a reference system in the host vehicle. Preferred embodiments are set out in the dependent claims. According to a first aspect, the invention provides a method according to claim 1. According to a second aspect, the invention provides a device according to claim 8. Advantages of the invention In order to have a proper evaluation it is crucial to ensure that the states of reference data objects obtained by the reference system and the objects obtained by the system under test are compared at the same timestamps. In certain cases, no common time base exists between the reference system and the system under test, which introduces an additional error to the evaluation. This can be interpreted as a latency, which will offset the reference data states from the measured states. The invention allows determining the latency und can therefore contribute to a more reliable perception evaluation. According to the invention, the latency is determined based on a first velocity obtained by the system under test and a second velocity obtained by the reference system. By comparing or matching these velocities, the latency can be determined. According to the method, computing the latency between the first time reference frame and the second time reference frame comprises minimizing a difference between the first velocity in the first time reference frame and the second velocity in the second time reference frame shifted by the latency. If the correct latency is used, the difference becomes minimal because the velocity data from both systems basically coincides. In an embodiment of the method, the difference between the first velocity and the second velocity is minimized using linear regression. This is a computationally efficient way to determine the latency. According to the method, a time-dependent acceleration of the host vehicle in the second time reference frame is determined, using the reference system. The latency between the first time reference frame and the second time reference frame is further computed using the acceleration of the host vehicle. The acceleration can be used in the linear regression method. In an embodiment of the method, the first velocity is determined using a first sampling rate. The second velocity is determined using a second sampling rate. The second velocity is resampled to the first sampling rate. By resampling, the first and second velocities can be compared. In an embodiment of the method, the IMU measures a velocity, an acceleration and a position of the host vehicle. In an embodiment, a Kalman filter is used for measuring a velocity, an acceleration and a position of the host vehicle. In an embodiment of the method, the IMU comprises three orthogonally positioned gyroscopes for measuring rotational motion and/or three accelerometers for measuring linear movements of the host vehicle. In an embodiment of the method, a sensor drift of the reference system is compensated for by using a global navigation satellite system, GN