Method, fusion filter, and system for fusing sensor signals with different temporal signal output delays into a fusion data set

Patent metadata

Field	Value
Publication number	US-2016209236-A1
Application number	US-201414917063-A
Country	US
Kind code	A1
Filing date	Nov 19, 2014
Priority date	Nov 20, 2013
Publication date	Jul 21, 2016
Grant date	—

Abstract

A method, a system, and a use of the system, for fusing sensor signals with different temporal signal output delays into a fusion data set. Each of at least three different sensor signals describe a measured value of at least three different sensor systems. The signal output delays are unique to the sensor. Each sensor signal includes a timestamp describing a detection time of the measured value. Erroneous measured values of a first sensor system are determined by a comparison with the measured values of other sensor systems. The erroneous values are considered constant at least for the duration between two consecutive comparisons. The erroneous values and/or the measured values are available for a specified time. The erroneous values and/or the measured values of the first sensor system constant are corrected by a correction process, and corrections carried out during the time period are considered when applying additional corrections.

First claim

1 . A method for fusing sensor signals with different temporal signal output delays into a fusion data set, wherein at least three different sensor signals in each case describe a measured value of at least three different sensor systems, wherein the signal output delays are individual to each sensor, wherein the sensor signals in each case comprise a timestamp describing an acquisition time of the measured values, wherein error values of measured values of a first sensor system are defined by a comparison with measured values of further sensor systems, wherein measured values with an identical timestamp are used for the comparison, wherein the error values are assumed as constant at least for the time period between two consecutive comparisons, and wherein the error values and/or the measured values of the first sensor system are retained for a specified time period, wherein the error values and/or the measured values of the first sensor system are continuously corrected by a correction, wherein corrections carried out during the time period are taken into account in an application of further corrections during the time period.

2 . The method as claimed in claim 1 , wherein the first sensor system has a comparatively shortest signal output delay.

3 . The method as claimed in claim 1 , wherein the retained error values and/or the retained measured values are corrected.

4 . The method as claimed in claim 1 , wherein the corrections are offset against one another during the time period.

5 . The method as claimed in claim 1 , wherein the corrections are carried out based on measured values of the further sensor systems.

6 . The method as claimed in claim 1 , wherein the sensor signals are filtered before being used for the fusion.

7 . The method as claimed in claim 1 , wherein the sensor signals are sensor signals of an inertial navigation system, sensor signals of a global satellite navigation system and sensor signals of an odometry navigation system.

8 . The method as claimed in claim 1 , wherein the error values are defined by an error state space filter.

9 . The method as claimed in claim 1 , wherein measured values with a required timestamp are generated by interpolation for defining the error values if no measured values with an identical timestamp are present.

10 . The method as claimed in claim 9 , wherein changes in the measured values are assumed as proportional to time for the interpolation.

11 . The method as claimed in claim 1 , wherein the error values are defined for the measured values of all sensor systems.

12 . A fusion filter for fusing sensor signals with different temporal signal output delays into a fusion data set, wherein the fusion filter acquires at least three sensor signals from at least three different sensor systems in each case describing a different measured value, wherein the signal output delays are individual to each sensor, wherein the sensor signals in each case comprise a timestamp describing an acquisition time of the measured values, wherein the fusion filter defines error values of measured values of a sensor system with the shortest signal output delays by a comparison with measured values from other sensor systems, wherein the fusion filter uses measured values with an identical timestamp for the comparison, wherein the fusion filter assumes the error values as constant at least for the time period between two consecutive comparisons, wherein the fusion filter is designed to retain the error values and/or the measured values of the first sensor system for a specified time period, and wherein the fusion filter is designed to correct the error values and/or the measured values of the first sensor system continuously by means of a correction, wherein the fusion filter takes account of corrections carried out during the time period in an application of further corrections during the time period.

13 . A system for fusing sensor signals with different temporal signal output delays into a fusion data set, comprising at least three different sensor systems which output sensor signals describing measured data with signal output delays individual to each sensor, wherein the system furthermore comprises a fusion filter as claimed in claim 11 .

14 . The system as claimed in claim 13 , wherein the system carries out a method for fusing the sensor signals with the different temporal signal output delays into the fusion data set, wherein at least three different sensor signals in each case describe a measured value of the at least three different sensor systems, wherein the signal output delays are individual to each sensor, wherein the sensor signals in each case comprise a timestamp describing an acquisition time of the measured values, wherein error values of measured values of a first sensor system are defined by a comparison with measured values of further sensor systems, wherein measured values with an identical timestamp are used for the comparison, wherein the error values are assumed as constant at least for the time period between two consecutive comparisons, and wherein the error values and/or the measured values of the first sensor system are retained for a specified time period, wherein the error values and/or the measured values of the first sensor system are continuously corrected by a correction, wherein corrections carried out during the time period are taken into account in an application of further corrections during the time period.

15 . A use of the system as claimed claim 13 in a motor vehicle.

16 . The method as claimed in claim 2 , wherein the retained error values and/or the retained measured values are corrected.

17 . The method as claimed in claim 1 , wherein, the error values are defined by an error state space Kalman filter.

Assignees

• Continental Teves Ag & Co Ohg

Inventors

• Steinhardt Nico

Classifications

• G01C25/005 Primary

  initial alignment, calibration or starting-up of inertial devices · CPC title

• G01C21/165 Primary

  combined with non-inertial navigation instruments · CPC title