US 12,054,179 B2
Kurtosis based pruning for sensor-fusion systems
Syed Asif Imran, Moorpark, CA (US); Jan K. Schiffmann, Newbury Park, CA (US); and Nianxia Cao, Oak Park, CA (US)
Assigned to Aptiv Technologies AG, Schaffhausen (CH)
Filed by Aptiv Technologies AG, Schaffhausen (CH)
Filed on Feb. 22, 2023, as Appl. No. 18/173,045.
Application 18/173,045 is a continuation of application No. 17/129,775, filed on Dec. 21, 2020, granted, now 11,618,480.
Claims priority of provisional application 63/115,142, filed on Nov. 18, 2020.
Prior Publication US 2023/0192146 A1, Jun. 22, 2023
Int. Cl. B60W 60/00 (2020.01); B60W 30/095 (2012.01); B60W 50/02 (2012.01); G01S 13/931 (2020.01); G06T 7/20 (2017.01)
CPC B60W 60/0027 (2020.02) [B60W 30/0956 (2013.01); B60W 50/0205 (2013.01); G01S 13/931 (2013.01); G06T 7/20 (2013.01); B60W 2420/40 (2013.01); B60W 2420/403 (2013.01); B60W 2420/408 (2024.01); B60W 2420/54 (2013.01); B60W 2554/20 (2020.02); B60W 2554/4029 (2020.02); G06T 2207/10028 (2013.01); G06T 2207/10044 (2013.01); G06T 2207/10048 (2013.01); G06T 2207/10132 (2013.01); G06T 2207/30252 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A method comprising:
receiving, by a processor, a set of vision tracks corresponding to one or more first objects detected by a vision-sensor system on a vehicle;
receiving, by the processor, a set of non-vision tracks corresponding to one or more second objects detected by a non-vision-sensor system on the vehicle;
for each one of the vision tracks in the set of vision tracks, determining whether a first object of the one or more first objects corresponding to the one of the vision tracks matches a second object of the one or more second objects corresponding to any of the non-vision tracks by:
determining possible associations between the one of the vision tracks and each of the set of non-vision tracks;
determining a likelihood of a false detection for each one of the possible associations of a match produced between the one of the vision tracks and the non-vision track for the one of the possible associations;
determining weighting evidence for each one of the possible associations based on a function applied to the likelihood of the false detection determined for the one of the possible associations;
determining a distribution of the weighting evidence for each of the possible associations;
determining a Kurtosis value of the distribution for each one of the possible associations to weight the one of the possible associations among the possible associations; and
selecting, based on the Kurtosis value of each of the possible associations, a high weight subset of the possible associations; and
determining whether the first object corresponding to the one of the vision tracks matches the second object corresponding to any of the set of non-vision tracks by matching the one of the vision tracks to the second object corresponding to the non-vision track of at least one of the possible associations selected for the high weight subset.