Method, Computer Program, Storage Medium, Device for Tracking an Object

Abstract

A method is for tracking an object using an environment sensor. The object is represented by an object status. The method includes detecting a sensor value of the environment sensor, predicting a future object status of the object, and updating the object status using a Bayesian filter. The updating includes using an artificial intelligence module (“AI module”). The AI module is trained such that the detected sensor value is associated with the object and the object status of the object is updated based on the predicted future object status of the object and the detected sensor value.

Claims

1. A method for tracking an object using an environment sensor, the object being represented by an object status, the method comprising: detecting a sensor value of the environment sensor; predicting a future object status of the object; and updating the object status using a Bayesian filter, wherein the updating includes using an artificial intelligence module (“AI module”), and wherein the AI module is trained such that the detected sensor value is associated with the object, and the object status of the object is updated based on the predicted future object status of the object and the detected sensor value.

2. The method according to claim 1, wherein the AI module comprises a grid-based artificial neural network.

3. The method according to claim 1, wherein the AI module comprises an artificial neural network based on adaptive lists.

4. The method according to claim 1, wherein the AI module is trained to map the detected sensor value from a sensor measurement space into an object status space.

5. The method according to claim 1, wherein measurement noise is derived based on the AI module.

6. The method according to claim 1, wherein the method is used for longitudinal and/or lateral control of a motor vehicle.

7. The method according to claim 1, wherein a computer program is configured to carry out the method.

8. A non-transitory machine-readable storage medium on which the computer program according to claim 7 is stored.

9. A device configured to carry out the method according to claim 1.

Description

[0057] Embodiments of the present invention are explained in more detail below with reference to drawings.

[0058] In the drawings:

[0059] FIG. 1 is a flow chart of an embodiment of the method according to the present invention;

[0060] FIG. 2 is a block diagram of a device that implements the method according to the present invention.

[0061] FIG. 1 shows a flow chart of an embodiment of the method 100 for tracking an object comprising an AI module according to the present invention.

[0062] In step 101, a sensor value of an environment sensor is detected.

[0063] In step 102, a future object status of the tracked object is predicted.

[0064] In step 103, the object status of the tracked object is updated.

[0065] Steps 102 and 103 take place in the Kalman filter.

[0066] The method of the present invention is based on the knowledge of using, for the hitherto conventional steps of associating the detected sensor values with the tracked object and updating the object status of the tracked object, an AI module which is trained in such a way that the detected sensor value is associated with the tracked object and the object status of the tracked object is updated on the basis of the predicted future object status of the object and the detected and associated sensor value.

[0067] It is clear that the object status cannot be updated if the detected sensor value cannot be associated with the tracked object.

[0068] There may be substantially two reasons for this. Either the detected sensor value has no reference to the tracked object, for example because the sensor value comprises information about a further tracked object (“true negative”). Or the detected sensor value has a reference to the tracked object, but is not associated because the association is too weak and an excessive residual probability remains that the detected sensor value does not have a reference to the tracked object (“false negative”).

[0069] FIG. 2 shows a block diagram of a device that implements the method according to the present invention.

[0070] Block 11 represents the detection of sensor signals by means of an environment sensor (measurement). The stars represent respectively detected signal reflections. An environment sensor is typically understood to mean a sensor or a sensor system that detects waves, in particular electromagnetic waves, and optionally emits corresponding waves. Typical environment sensors are video, radar, ultrasound, infrared and lidar sensors.

[0071] The detected sensor signals are fed to a method for object tracking, shown in block 10.

[0072] If the object tracking is carried out by means of a Kalman filter, the steps of predicting a future object status (prediction), shown in block 12, associating the sensor signals with a tracked object (association), shown in block 13, updating the object status (update), shown in block 14, and compiling the current list of tracked objects (tracked objects) to be output to a further-processing system such as an adaptive cruise control (ACC) system, shown in block 15, are typically performed cyclically.

[0073] The present invention is based on the knowledge that the step of association 13 and part of the step of updating 14 can be implemented by means of an AI module 21 that receives the detected sensor signals, possibly preprocessed, and, as part of the processing, carries out both the association and the required transformation from the sensor measurement space into the object status space 22.

[0074] This characterizing part of the present invention is shown in block 20.