Search

EP-4047392-B1 - POSITION TRACKING WITH MULTIPLE SENSORS

EP4047392B1EP 4047392 B1EP4047392 B1EP 4047392B1EP-4047392-B1

Inventors

  • Vagarappan Ulaganathan, Raghavendran
  • Baheti, Ashutosh
  • JUNGMAIER, REINHARD-WOLFGANG
  • SANTRA, Avik
  • TROTTA, SAVERIO
  • VAISHNAV, Prachi

Dates

Publication Date
20260513
Application Date
20210218

Claims (15)

  1. A method (100, 200) for an electronic device (300) of tracking an object, comprising: predicting (110, 210) a predicted state (Xpred), in a global coordinate space (450), of an object based on a state (X) of the object; said method being characterized by further comprising: determining (120, 220) in local coordinates (410, 420) the predicted state (Xpred); determining (150, 250) a plurality of measurements (M_1, M_2) of the object, in the local coordinates (410, 420), with first and second radar sensors (310, 320) of the electronic device (300) arranged along perpendicular lines (330, 340) which intersect at the origin (350) of the global coordinate space (450); determining (130, 230, 232) a matching of the predicted state (Xpred) and the plurality of measurements, in the local coordinates (410, 420), thereby obtaining a matching result; updating (140, 240, 242) the state (X) of the object based on the matching result.
  2. The method of tracking an object of claim 1, wherein: determining (130, 230, 232) the matching includes: determining (230) a first matching result of the predicted state (Xpred) and a plurality of first measurements (M_1) of the first radar receiver (310); and determining (232) a second matching result of the predicted state (Xpred) and a plurality of second measurements (M_2) of the second radar receiver (320).
  3. The method of tracking an object of any preceding claim, wherein: determining (120, 220) in local coordinates (410, 420) the predicted state (Xpred) includes: coordinate-transforming the predicted state (Xpred) of the object into a first local space (410) of the first radar receiver (310), and coordinate-transforming the predicted state (Xpred) of the object into a second local space (420) of the second radar receiver (320).
  4. The method of claim 3, wherein: the first local space (410) originates at the first radar receiver (310); the second local space (420) originates at the second radar receiver (320); the second local space (420) is rotated at an angle from the global coordinate space (450) about a z-axis which extends away from an x-y plane which includes the radar receivers (310, 320) and the origin (350) of the global coordinate space (450); and the orientation of the second radar receiver (320) is rotated at the angle along z-axis from alignment with the global coordinate space (450).
  5. The method of any of claims 3-4, wherein: determining the matching (130) includes: a first matching (230) of the predicted state and a first plurality of first measurements (M_1) of the first radar sensor (310), and a second matching (232) of the predicted state and a second plurality of second measurements (M_2) of the second radar sensor (320).
  6. The method of claim 5, wherein the updating (140, 240, 242) includes: a first update (240) to the state (X) between the first matching (230) and the second matching (232) and a second update (242) to the state (X) after the second matching (232).
  7. The method of claim 5, wherein: the coordinate-transforming the predicted state (Xpred) into the first and second local spaces (410, 420) is before determining the matching (130) and after the determining (120) in global coordinates (450) the predicted state (Xpred); the determining (130) the matching is based on the first plurality of measurements (M_1) and the second plurality of measurements (M_2); and the updating (140) is after the determining (130) the matching.
  8. The method of any preceding claim, wherein the radar sensors (310, 320) are separated by 3-30 cm, and the origin (350) is at the center of the electronic device (300).
  9. The method of any preceding claim, wherein the first radar sensor (310) is at a first edge (311) of the device (300) and the second radar sensor (320) is at a second edge (322) of the device (300).
  10. The method of any preceding claim, wherein the state (X) includes a position and velocity of the target.
  11. The method of any preceding claim, wherein the predicting (110, 210) is based on an unscented Kalman filter.
  12. The method of any preceding claim, wherein the matching (130, 230, 232) is based on a Hungarian bipartite matching which determines matching of the measurements and an existing track which includes at least one state (X) determined at a time preceding the measurements.
  13. The method of tracking an object of any preceding claim, wherein the determination (150, 250) of the plurality of measurements triggers the determining (130, 230) the matching, and the updating (140, 240, 242) follows the determining (130, 230) the matching.
  14. An electronic device, comprising: a processor, and first and second radar sensors (310, 320) arranged along perpendicular lines (330, 340) which intersect at the origin (350) of a global coordinate space (450); wherein the processor is configured to: predict (110, 210) a predicted state (Xpred), in the global coordinate space (450), of an object based on a state (X) of the object; determine (120, 220) in local coordinates (410, 420) the predicted state (Xpred); determine (150, 250) a plurality of measurements (M_1, M_2) of the object, in the local coordinates (410, 420), with the first and second radar sensors (310, 320); determine (130, 230, 232) a matching of the predicted state (Xpred) and the plurality of measurements, in the local coordinates (410, 420), thereby obtaining a matching result; and update (140, 240, 242) the state (X) of the object based on the matching result.
  15. A computer program having a program code for performing a method of any one of claims 1-13 when the program is executed on the processor of an electronic device according to claim 14.

Description

Field Examples relate to methods and devices for tracking an object. Background Tracking of objects is finding application across many fields for many purposes, such as for tracking a target, motion sensing, and gesture sensing. US 2016/103214 A1 proposes to locate a target and to associate a track with the target in the fusion coordinate system. An estimate/prediction of the target' s velocity is developed within the tracker, as well as T, a vector representing the distance from the fusion center to the target as estimated by the tracker, and Sf, a vector representing the known distance from the fusion center to the sensor. The sensor's range vector, K (the distance from the sensor to the target as predicted by the tracker) is transformed to fusion coordinates. Using the sensor's range vector, normalized to unit length, in fusion coordinates and the estimated target's velocity, an estimate of the target's speed projected in the direction of K is derived. The estimated range-rate is compared per update to the sensor's measured range-rate in the form of an error measurement. The error is then used to correct the track's velocity prediction. Summary Herein are disclosed methods of tracking and devices for tracking that may improve tracking accuracy in challenging conditions such as when measurement signals are weak, dropped, or sometimes spurious. A method for an electronic device of tracking an object is disclosed, including predicting a predicted state, in a global coordinate space, of an object based on a state of the object; determining in local coordinates the predicted state; determining a plurality of measurements of the object, in the local coordinates, with first and second radar sensors of the electronic device; determining a matching of the predicted state and the plurality of measurements, in the local coordinates, for a matching result; updating the state X of the object based on the matching result. The first and second sensors are along perpendicular lines which intersect at the origin of the global coordinate space. Herein is further disclosed a device including first and second radar sensors arranged along perpendicular lines which intersect at the origin of a global coordinate space and a processor that is configured to execute the method. Brief description of the Figures Some examples of apparatuses and/or methods will be described in the following by way of example only, and with reference to the accompanying figures, in which: Fig. 1 illustrates a method of tracking an object;Fig. 2 illustrates a method of tracking 200;Fig. 3 illustrates a device 300;Fig. 4B illustrates a state;Fig. 5 illustrates a block diagram of matching;Fig. 6 illustrates a block diagram of a method of tracking;Fig. 7 illustrates a cycle of a method of tracking an object;Fig. 8 illustrates a cycle of a method of tracking an object;Fig. 9 illustrates mathematical operations and definitions;Fig. 10 illustrates data of an example;Fig. 11 illustrates tracker output of an example; andFig. 12 illustrates a normalized innovation squared metric of an example. Detailed Description Various examples will now be described more fully with reference to the accompanying drawings in which some examples are illustrated. In the figures, the thicknesses of lines, layers and/or regions may be exaggerated for clarity. Same or like numbers refer to like or similar elements throughout the description of the figures, which may be implemented identically or in modified form when compared to one another while providing for the same or a similar functionality. Fig. 1 illustrates a method 100 of tracking an object. The method 100 of tracking an object can include predicting 110 a predicted state (e.g. a subsequent state) in a global coordinate space of an object, based on a state of the object, such as a previous state, of the object. The method 100 can also include determining 120, in local coordinates (e.g. the coordinates used by the sensors), the predicted state (e.g. by transforming coordinates 120 from global coordinates used in predicting the state of the object into local coordinates). The method 100 can include determining 150 a plurality of measurements of the object in the local coordinates with a plurality of radar sensors, including first and second radar sensors. Having a plurality of sensors, particularly when optimized for placement, can aid in reducing blind spots and/or regions where the object is out of the field of view of the tracking device. Alternatively/additionally, multiple sensors may allow for continued data acquisition related to the object's state (e.g. position and/or velocity) when one of the sensors loses signal and/or produces spurious data. Having exactly two sensors is particularly contemplated, such as two radar sensors. Two sensors may allow for redundancy and/or improved field of view coverage while keeping computational costs low. Redundancy, e.g. with multiple sensors, can aid in enabling tracking to cont